Digitized by the Internet Archive

in 2022 with funding from

Kahle/Austin Foundation

https://archive.org/details/ohysicstechnologO000brod
 The Physics and
Technology of Tennis

Howard Brody
Rod Cross
Crawford Lindsey

Racquet Tech Publishing

An Imprint of the USRSA
Solana Beach, California USA
Copyright © 2002 USRSA

USRSA
337 S. Cedros Ave., Suite D
Solana Beach, CA 92075
www.racquettech.com

All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the express written
consent of the publisher.

The following chapters by or with Howard Brody incorporated materials that

were reprinted with permission by TennisPro Magazine. TennisPro is the official
publication of the Professional Tennis Registry.

Chapter 3: TennisPro, November/December 1998 .

Chapter 6: TennisPro, March/April 1998; May/June, 1998; July/August, 1998
Chapter 15: TennisPro, May/June 1993
Chapter 17: TennisPro, September/October 1997
Chapter 19: TennisPro, March/April 2001
Chapter 21: TennisPro, March/April 1997
Chapter 22: TennisPro, January 1990; November/December 2000
Chapter 23: TennisPro, May/June 1995
Chapter 24: TennisPro, March/April 1993
Chapter 25: TennisPro, March/April 2000

Library of Congress Control Number: 2002094432

Cover design: R. Christian Anderson, Ph.D.

Chapter head photography: Ron Waite/Photosportacular
Contents photos of ball bounce: courtesy of the University of Sheffield and the
International Tennis Federation
Contents photos of wind tunnels: Courtesy of NASA Ames Research Center
Fluid Mechanics Laboratory and Cislunar Aerospace, Inc.

Printed in the United States of America

ISBN 0-9722759-0-8
Contents
ees
Contents il

Preface xiii

| Choosing A Racquet
Rod Cross

Choosing a Racquet
How Does It Feel?
Grip Size Above and the next 17 photos: Ball
Racquet Weight and Power impact with a head-clamped racquet
Sweetspots at 90 mph with no spin at a 90
Balance Point and Swingweight degree angle (courtesy of the
Headsize University of Sheffield and the

Frame Stiffness International Tennis Federation).

String Type and Diameter

String Tension

2 Racquet Materials
Crawford Lindsey

Graphite
Modulus and Strength
Making High Modulus Carbon
Making a Racquet
Racquet Lay-up
Titanium
Piezoelectric Materials

3 Weight, Balance & Swingweight

Crawford Lindsey & Howard Brody

Force, Mass and Inertia

Weight
Balance Point
Pickup Weight
What Balance Point Means to Tennis Players
Customizing the Balance Point
Swingweight
What Swingweight Means to the Player
Customizing Swingweight
The Relationship of Weight, Balance and Swingweight
Is It the Weight of the Swing, or the Swing of the Weight?
As Weight Goes Down, Balance Goes Up
iv Contents

4 Racquet Stability 37
Crawford Lindsey & Howard Brody

Weight In Past Racquet Developments 37

Stability By Weight Polarization 38
What, How, When, Where and Why of Polarization 39
Twistweight: The Stability of the Racquet Against Twisting 39
Interpreting the Stability Table 4|

5 Measuring Swingweights 45
Rod Cross & Crawford Lindsey

Measuring Swingweight 46
Measuring Twistweight 48
Measuring Recoilweight 50

6 Sweetspots 5|
Howard Brody, Rod Cross, & Crawford Lindsey

The Center of Percussion 52

The Node 55
The Power of the Tennis Racquet: COR and ACOR 58
The Deadspot a7

7 Introduction To Shock & Vibration 61

Howard Brody & Crawford Lindsey

Introduction 61
What Is Vibration? 62
Bending Shapes (Modes) 62
Amplitude of Vibration and the Node 62
Frequency 65
Stiffness 66
Dampening 67
Injury 67
What Is Shock and What Is Its Relation to Vibration? 68
Lessons From Vibrations 69
How Does Shock Absorption Work? 69
Conclusion 7\

8 Racquet Vibration Theory 73

Rod Cross

Frame Bending and Vibrations 73

Vibration Shapes 74
Racquet Motion During Impact 75
Vibration Frequency 76
Vibration Modes of a Rod or Beam 77
Impact with a Beam or Racquet 79
Bounce Speed Off a Vibration Node 79

9 Racquet Feel 81
Rod Cross

Effects of String Tension on Jarring 82

Simple Stiffness and Vibration Experiments 83
Measurements of Shock and Vibration 83
 Contents

Vibration Frequency and Amplitude 86

Vibration Frequency of a Racquet 88
Sweetspot Softness and Feel 88
Sweetspot Size 89
Shock and Feel 90
Recoil Speed of the Handle 9|
Impact Duration, Force, and Feel 92
Effect of Frame and String Stiffness on Frame Vibrations 93
String Vibrations 94
Measurements of the Force on the Hand de,
String Vibrations and Feel 96

10 Force and Momentum 99

Crawford Lindsey

Introduction 9
Newton's First Law of Tennis 100
Newton's Second Law ofTennis 10]
Newton's Third Law ofTennis 10]
Momentum 103
Impulse = Change in Momentum Theorem 103
Law of Conservation of Racquet and Ball Momentum 108
Conclusion 109

11 Energy and Coefficient of Restitution

Crawford Lindsey

Introduction 111
Kinetic Energy [12
Law of Conservation of Racquet and Ball Energy 114
Difference between Racquet Energy and Momentum 115
Energy Return and Coefficient of Restitution (COR) 117
Final Ball and Racquet Velocities — The Hard Way 19
How To Use the Speed Equations 120
Apparent Coefficient of Restitution (ACOR) 121
Final Ball Speed Formulas — The Easy Way 124
Conclusion 126

12 Collisions 127
Rod Cross

Effect of Racquet Mass on Serve Speed 128

Elastic Energy Stored During a Collision 130
Calculation of the COR 133
Rotation 134
Relation Between Angular Speed and Actual Speed 134
Moment of Inertia 136
Angular Momentum 137
Collisions Involving Rotation of the Racquet 138
Effective Racquet Mass 140
Estimating the ACOR from the Effective Racquet Mass 142
Axis of Rotation of a Racquet 143
Notes 144
Appendix 144
Formula for Effective Racquet Mass for Impact Along Center Line 144
Formula for Effective Racquet Mass for Impact Anywhere 145
vi Contents

147
13 Tennis Springs
Rod Cross

Stiff Springs and Soft Springs 147

Springs in Series or Parallel 148
Potential Energy 149
Potential Energy Stored in a Mass 149
Potential Energy Stored in a Spring 150
The Tennis Ball As a Spring 150
Strings 15]
How the Elastic Energy Is Shared 152
Effect of String Tension on Ball Speed 153
Sideways Stiffness of a Single Stretched String 154
Impact of a Ball on a Stretched String 156

14 Racquet Power 159

Rod Cross

Summary 159
Introduction [59
Factors Affecting Power 160
Physics of Power 162
Maximum Possible Racquet Power 163
Estimating Racquet Power 163
Moving Racquet vs Free Racquet at Rest 165
Effects of Frame Stiffness on Racquet Power 166

15 Determining On Court Ball Speed 169

Howard Brody

Measuring Serve Speeds 169

Measuring Groundstroke Speeds 172

16 Customizing Racquets 175

Rod Cross

Introduction 175
Balance Point 176
Swingweight 177
Racquet Power 178
Rotation About the Long Axis 178
Shifting the Sweetspots 179
Summary 180

PART 2: STROKES
EET

17 Racquet Technology and Strokes 183

Howard Brody

Traditional Racquets 183

Modern Racquets 184
Classic Strokes 184
New Strokes 184
 Contents vi

Sweetspots 184
Stability 185
Power 185
Court Surfaces and Strokes 186
“Bring Back the Wood Racquet” 186

18 Impact Location 189

Howard Brody

Where on the Racquet Head Should You Hit the Ball? 189
Transverse Direction 19|
Conclusion 19|

19 Improving Your Serve 193

Howard Brody

Introduction 193
Vertical Angular Acceptance (Window) 194
Ball Speed 195 ee

Impact Height 195

Topspin 196
Higher Ball Toss 197
Location on Court 198
Racquet Parameters 199
Summary 200
The next 3 photos: Wind tunnel
20 Serve Speed 201 results with equivalent ball speed of
Rod Cross & Howard Brody
87 mph and spin of 72 rev/sec (cour-
tesy of NASA Ames Research Center
Serve Speed vs Racquet Speed 201
Fluid Mechanics Laboratory and
Serve Speed Formulas 202
Cislunar Aerospace, Inc.).
Impact Duration 203
Arm Rotation 203
Serve Speed Strategy 204
Case Study |: Second Serve Speed 205
Case Study 2: Serve Speed 206
Summary 207
Backspin
21 Receiving Serve 209
Howard Brody

22 Strokes and Spin 213

Howard Brody

Topspin vs Backspin 213

Watch That Spin 215
Slicing the Slice 2G.
The Approach Shot — Why Keep It Low and Deep? 218

23 The Drop Shot 221

Howard Brody

Hitting a Drop Shot 2 Topspin

Returning a Drop Shot 223
viii Contents

24 Time and Strokes 225

Howard Brody

25 Error Reduction 229

Howard Brody

Lateral Errors 229 ieee

Errors of Depth 231 Noein
Equipment and Error Reduction 234

PART 3:STRINGS
roe ae ye te The next 18 photos: Ball impact with
26 Overview of Strings 239 floor at 90 mph with no spin at a 90
degree angle (courtesy of the
Howard Brody
University of Sheffield and the
Introduction 739 International Tennis Federation).
Air Resistance 239
The String Plane Deformation 240
The Ping of the String 240
Dwell Time 241 |
Power 242 a
The Feel of the Strings 244 |
Elasticity 244
Gauge LAS
Wear and Breakage 245
String Patterns 246
Control 246
Rebound Angle 247
Grommets 248

27 String Chemistry 251

Crawford Lindsey

Polymers 252 |
253 ]
Bonding
Bonding on a Chain 2o4
Bonding Between Polymer Chains 254
String Materials 255
Copolymers Meh
Polymers Are Silly 257
Conclusion 258

28 String Materials 259

Crawford Lindsey

What Materials? 259)

What Is Strength? 260 L
What Is Stiffness? 260
Tension and Stretch 260 /
Strength and Stiffness by Weight 262 ;
Conclusion 263

29 String Construction 265

Crawford Lindsey

Core Principles 265

i Ww :

Natural Gut Construction and Play 266

 Contents ix

Wraps:To Serve and Protect 266

Twists, Angles and Fiber Counts 267
Resin: The Glue that Bonds 268
Coatings: Slip Sliding Away 269
How Are Strings Different? 270
Why the Cost Differences? 270

30 String Properties 271

Rod Cross & Crawford Lindsey

Dynamic Stiffness — It Determines Almost Everything 272

Impact Force and Duration — It’s What You Feel 274
Calculations 276
Do Strings Ever Lose Power or Resilience? 279
Player Perceptions of Power 7Mfe|

31 Power and Strings Part | 283

Crawford Lindsey

String Tension and Power 283

String Length and Power 287
String Gauge and Power 288

32 Power and Strings Part 2 291

Rod Cross

Serve Speed 292

Typical ACOR Values 293
Increasing Power by Increasing Bounce 293
Increasing Bounce Speed by Lowering Tension 294
Stored Energy vs Lost Energy in Strings and Balls 294
What if the Ball and Strings Are Equally Stiff? 295
Effect of Lowering Tension on Stored Energy and Power 296
How Does This Affect the Serve Speed? 296
Energy Calculations 297

33 String Tension 301

Crawford Lindsey

What Is Tension? 302

String Tension Is a Moving Target 303
Where Does It Go and by How Much? 303
Why Does It Go? 303
Tension Loss While Stringing 305
Tension Loss and Prestretching 306
Time: Every Second Counts 307
Consistency and Attention to Detail 308
Pull Angle and Tension 308
Tension Head Location and Final Tension 309
Temperature and Tension 309
Conclusion 309

34 Measuring String Tension 311

Rod Cross

Racquet Diagnostic Machines 311

The Stringmeter 312
X Contents

The ERT700 33
Artificial Tension Calibration 314
Four Reasons for Tension Loss During Stringing 314
String Tension: Can Players Tell the Difference? 315
Determining Actual Stringbed Tension: Method | 316
Determining Actual Stringbed Tension: Method 2 316
Determining Tension in Each String 317
a
ee
Comparing Pull Tension, ERT, Pacific RA, and Calculations 317 <<

Comparing the Stringmeter with Calculations 318

Where To Go From Here? 319
The Stringmeter Revisited 319
Conclusion 320

35 String Gauges 321

Rod Cross

Summary 321
Introduction 321
Controlled Experiment 325
Conclusions 326

PART4: BALLS AND COURTS

_A.0™
GC
‘sisiddidriaibaaias

36 The Tennis Ball 329 SH

ROUTE
Howard Brody \S

Pressurized and Pressureless Balls 330

High Altitude Balls 330
Big Balls 331
Air Resistance and the Big Ball 331
i
37 Ball Testing for Bounce 333
|

Rod Cross

Compression Test |
333 :
yy SF
Slow vs Fast Compression 334
Bounce of Spinning Ball 336
.
|
38 The Bounce of the Ball 339
Howard Brody &

Vertical Velocity 339

Horizontal Velocity 340
Fast Courts — Slow Courts 341

39 Ball Spin and Bounce Off the Court 343

Rod Cross iy
Zo

Summary 343
Introduction 343
Ball Spin 345 Midas
ecmssassssssssessilddldi/

Friction Force 345

Rolling 347
Spaghetti Strings 348
Ball Speed and COR 349
Horizontal Speed of a Court 350
 Contents

Measurements of Court Speed 350

Measurements of Court Speed on Different Surfaces 352
Measurements of Ball Spin 354
Measurements of Vertical Speed (COR) 354
Measurements of Court Speed on Actual Courts 356
Measurements of the Coefficient of Friction S57

40 Ball Bite and Bounce 359

Rod Cross

Introduction 359
Apparatus 359
Friction Force Results 360
Explanation of Friction Force Results 362
Slamming on the Brakes 364
Perception of Court Speed 366

4! Ball Spin and Bounce Off A Racquet 369

Rod Cross

Introduction 369
Bounce Basics 370
Topspin or Backspin Into the Strings? 371
Ball Speed, Spin and Angle Measurements 373
How To Calculate the Bounce Off a Moving Racquet 376
Example of Bounce Calculation 376
Serving with Topspin cya

42 Ball Trajectories 381

Rod Cross

Introduction 381
Examples of Ball Trajectories 382
Drag Force 383
Magnus Force 384
Serve Trajectories 385
Appendix |: Trajectory Equations 387
Appendix 2: Simple Ball Speed Measurement 388

Appendix A: Glossary 391

Appendix B: Units and Conversions 405

Appendix C: Equipment Rules 409

Appendix D: The Mathematics of Ball Bounce 413

Brody’s Bounce Model 413

Case Where Ball Slides Throughout the Bounce 414
Case Where Ball Rolls 414
xii Contents

Appendix E: Tennis Court Dimensions 417

Appendix F: References 419

Aerodynamics 4l9
Ball Bounce 420 J
Biomechanics 420 |J
7Z
General Interest 42\
Measurements 42 7
4
Racquets 42 |
Related Sports 423 |
Statistics 424
Strings 424
Strokes 425
Sweetspots 425
Tennis Courts 426

Index 427

About the Authors 437

 his book fills several niches. It was conceived to be the all-in-one source of enlightenment for all things
scientific and technological about tennis equipment and its affect on a player’s performance. This task was
not undertaken simply for the pure joy of intellectual pursuit. There is a great need and want in the mar-
ketplace and on the tennis courts for information about equipment. Tennis equipment has become so techno-
logically advanced that buying a racquet or choosing a string requires either a Ph.D., in order to make an intel-
ligent choice, or total blissful ignorance, to avoid the stress of making the wrong decision. Other sports have
become the same way — golf in particular.

For years, players have pondered many tennis mysteries:

Is a stiff frame good or bad?

e Where is the sweetspot and what does it do for me?
e What causes shock and vibration? Is it injurious, and does it atfect my shots?
e Does string elasticity make a difference, and how can I influence that elasticity?
e What difference does string tension make?
e Where should I impact the ball on my racquet?
e Should I use a lighter or heavier racquet?
e Can I and should I alter the weight and balance of my racquet?
¢ How much difference do different string materials, constructions and gauges make?
¢ How can I create ball spin, react to spin, and read spin?
e How can I hit with more power?
e And many more ...

All of these questions can be answered with a little understanding of the physical fundamentals underlying the
universe — tennis, of course, being its most important microcosm. In the past, these questions have been
answered with one part science and two parts folklore, and a pinch of myth. Sometimes the answers are cor-
rect, but the explanations are wrong. Sometimes both are correct, but if you measure the physical quantities
involved, they are not enough to make a difference anyway. And sometimes, players’ perceptions of what is
happening at ball impact with the strings, racquet and court are at odds with the measured physical reality. And
we all know that you can’t talk players out of what they feel.

So this book is here to set the record straight. This book is a hybrid of sorts. It is part textbook, part reference,
and part "good read." There is tennis science here for every level of expertise and curiosity. It is "popular sci-
ence," but without leaving out the hard explanations that make for true understanding. Concepts are fully
explained in words. But, where necessary, words are supplemented with formulas, and the formulas, if very
important, interesting, or intrinsically fascinating, are derived. In all cases, the math can be skipped without loss
of the conceptual train of thought, and most of the chapters are formula free anyway. Where present, the for-
mulas help to better understand the relationships that are involved in any equipment associated event on the
tennis court. And, of course, if you want the answer to a specific question like, “How fast will the ball go, or
where will the ball land if I hit it in such and such a way with such and such a racquet,” then you will need to
Xiv Preface

use the equations. The book is also heavily illustrated with diagrams and graphs to complement the text and to
add visual understanding and eye appeal.

There is some duplication of discussions and explanations in the book. This is on purpose. Not every reader is
going to read the book from front to back. Nor is it our intention that they do so. Rather, our intention is to have
each chapter stand by itself as much as possible, so the reader can delve in anywhere one’s need or whim dic-
tates. To do this, some chapters require concepts and explanations from other chapters. We make no apology
for that. Quite the contrary. The end result is that each chapter's discussion can be understood in a larger con-
text, without reading any prerequisite chapters.

Appendices complete the presentation. Here you will find a glossary, handy units and conversions, equipment
rules, court dimensions, references and other more esoteric information.

The authors would like to thank the following for photographs: Ron Waite/Photosportacular for the player pho-
tographs at the beginning of each chapter; The University of Sheffield and the ITF CUnternational Tennis
Federation) for the ball bounce photos in the Contents; and NASA Ames Research Center Fluid Mechanics
Laboratory and Cislunar Aerospace, Inc. for the wind tunnel photos in the Contents. Thanks also to R. Christian
Anderson for his beautiful cover design.

Recognition is due also to Racquet Tech Publishing, which is a division of the United States Racquet Stringers
Association (USRSA). The USRSA is a worldwide organization of over 7000 racquet technicians, teaching pro-
fessionals, racquet sports retailers, manufacturers, sales organizations and tennis (racquet sports) enthusiasts. Its
mission is to educate all its constituencies to better understand, service, perform with, and enjoy the techno-
logical wonders known as racquets, strings, balls, courts, shoes, and stringing machines. For information about
membership, products, or services, go to its website at www.racquettech.com, call 858-481-3545, or write to
USRSA, 337 S. Cedros, Suite D, Solana Beach, CA 92075.
 Why Some Racquets
Are Better Than Others

Pete Sampras, United States

By Rod Cross

difficult question to answer is why some racquets are better than others even though they all look much
the same and they are all made from the same basic ingredients. If you pick ten racquets at random
® and hit a few balls with each, one or two are likely to feel a lot better than the others. It is as if some
racquets have a magic ingredient that makes them perform better — at least better than others for some peo-
ple. Because of differences in abilities and styles of play, there is no universally acknowledged “best racquet.”
The perfect racquet has not yet been invented. If it had, then everyone would be using it. There definitely are,
however, some underlying basic principles or ingredients that make some racquets very good and others just
ordinary.

No one uses wood or aluminum racquets any more. They have been replaced by lighter, stiffer and stronger
graphite versions with larger heads. These factors, then, must be part of the formula that makes one racquet bet-
ter than another. But what is the optimum combination of weight, stiffness, weight distribution and head size
that best suits the individual player? And what of string type, string gauge, string tension, sweetspot size and the
amount of vibration damping built into the racquet? If one mixes all these ingredients in the right proportions,
then it should be possible to come up with a really good racquet.

In theory, it should be possible to come up with a formula for the ideal racquet based on the physical proper-
ties of a racquet and a player’s individual style of play. If we knew what the formula was, it would be relative-
Chapter 1

ly easy to design the ideal racquet for each different type of player. In fact, this
has already happened to some extent. Racquets these days are commonly clas-
sified as either “player’s” racquets or “recreational” racquets. A “player’s” rac-
quet is one designed for professional or elite players, and it typically has a
medium to small size head and is relatively heavy. Recreational racquets are
for everyone else and are generally lighter and have a larger head. Racquets
used by the professionals all look much the same and the lengths are all
between 27.0 and 27.5 inches. Professionals don’t use any of the high-tech lat-
est models and they don’t choose oversize or widebody or extra long frames.
The differences are mainly in the weight and the way that the racquets are cus-
tomized with lead tape in the head and the handle to suit each player.

There are also intangibles in good racquets. For example, many players
(including professionals) cannot play without a string dampener inserted in the
strings. But why? The string dampener stops string vibrations after the ball has
left the strings, but it has no effect on ball speed or ball control or frame vibra-
tions. It just deadens the sound of the strings. So a good racquet must be one
that not only feels good, it also has to sound good.

The best way to choose a good racquet is to take a dozen or so different rac-
quets onto a court and hit a few balls with each racquet. Some tennis shops
allow you to do this but they can be hard to find. In fact, twelve racquets is a
bare mimimum. Ideally, each of the twelve racquets you try should be multi-
plied by three so that you can try each model strung at low, medium and high
tension. Then multiply that by four so you can try gut, nylon, polyester and
kevlar. You will never actually find your ultimate weapon, since no one will
lend you three identical racquets strung at three different tensions, and new
models keep appearing faster than you can try them. You will never reach your
ideal playing standard either. But it is fun trying to get there and a good rac-
quet will make a big difference.

The first racquet I ever bought felt perfect in the shop, but it was also the worst
racquet I have ever owned. I made the mistake of assuming that a racquet that
feels good when you swing it will also feel good when you hit the ball.
Nothing could be further from the truth. The truth is, hitting a ball is vastly dif-
ferent from hitting air. There is no jarring and no vibration and no sensation of
power when you hit the air. It is absolutely essential to hit a ball when choos-
ing a racquet to buy.

The main thing to look for is whether the racquet feels comfortable to swing
and feels good when you hit the ball — not too much shock and vibration, not
too heavy and not too light. Every racquet you try will feel different from the
next. However, if you hit the ball in the middle of the strings then there will
be no vibrations at all in any racquet. Racquets vibrate only when you hit a
few inches away from the center, especially if you hit the ball near the tip or
the throat. So, if one racquet vibrates more than another, it might be because
you didn’t hit the ball in the middle of the strings. Don’t assume that it was the
racquet’s fault. Keep hitting until you can decide if it is you or the racquet that
is causing the vibrations.
 Choosing A Racquet

Some racquets DO vibrate more than others. Stiff racquets always vibrate less
than flexible racquets since they are harder to bend. A racquet that vibrates
does so because it bends rapidly one way and then the other. Consequently,
flexible racquets vibrate more than stiff racquets. Also, racquets strung at high
tension vibrate more than racquets strung at low tension. It is perhaps surpris-
ing that a racquet vibrates less when the frame is stiffer but it vibrates more
when the strings are stiffer. The reason is that the ball sits on soft strings for a
longer time, which helps to dampen both the string and the frame vibrations.
The least amount of vibration therefore occurs in stiff racquets strung at low
tension. Try hitting with an old wood racquet if you can find one. They are
very flexible compared with a modern graphite frame, and they vibrate MUCH
more than a modern racquet.

An important factor in choosing a racquet is to select the correct grip size. The
best racquet in the world will feel terrible if the grip is too large or too small
for your hand. Good racquet shops will have a selection of handle sizes that
you can try and they should also have a chart based on hand size to select the
correct grip. Grips are cheap and easy to change. Some people prefer a tacky
feel, and others like a soft feel since it helps to cushion the blow when you
mishit a ball. If you decide that the handle size is too small, the problem can
be fixed by an overgrip or by building up the handle size with plastic heat-
shrink tubing under the grip.

Racquet Weight and Powe

Should you choose a light or a heavy racquet? This is largely a matter of per-
sonal taste, but light racquets are easier to swing and you can swing them
faster. Light racquets are more maneuverable so you can move them from point
A to point B quicker and you can rotate them faster around the long axis.
However, the ball can also rotate a light racquet faster and further when the
ball pushes on the strings. If the ball is hit in the middle of the strings, the
impact normally causes the wrist to rotate backwards. With a “wristy” shot, the
impact slows the rate of wrist rotation even if it does not result in a backwards
rotation of the wrist. If the ball impacts towards the top or bottom of the frame,
the racquet will rotate about its long axis. Light racquets rotate further, result-
ing in a larger error in the rebound angle off the strings. Ball control suffers if
you mishit the ball. Light racquets are less forgiving in this respect.

If you want raw power, then the most powerful racquet possible would weigh
about 1000 gm — about the same as a baseball bat. However, baseball players
swing their bat only a few times in a whole game; they don’t have to run with
it to get to the ball; they don’t have to serve the ball; and they can play only
one half-decent shot — the double-handed forehand. Obviously, raw power is
not the only factor in choosing a racquet. In fact, power tends to be overrated
as a selling point. Ball speed off the strings is easy to increase simply by hit-
ting the ball harder. Players will do so when it is appropriate, but maximum
possible ball speed is not necessary or desirable on every shot.

Tennis players can get enough power (and much better maneuverability) with
a racquet weight between 250 and 400 gm. The speed of the ball off the strings
depends on three factors. The first is the speed of the incoming ball. Apart from
the serve, the incoming ball speed is determined by your opponent, so there
Chapter 1

are really only two factors to consider. The most important of these is the speed
of the racquet. If you swing the racquet faster, then the speed of the ball off
the strings will also be faster. The other factor is how well the ball bounces off
the strings. The main factor here is the weight of the racquet. String tension
also plays a small role since it affects the softness of the impact and that influ-
ences the energy loss in the ball and the energy lost in vibrations of the frame.
There are no other factors involved, unless you want to include tiny effects
such as wind speed or the type of string used. Many players and string manu-
facturers will tell you that certain strings give more power than others, but no
one has ever measured this, and I am fairly certain that it is psychological,
either something to do with the sound made by the strings or the way the
impact duration influences the feel of the racquet.

The speed of the ball for a serve is given by the simple formula, v = (1 + eV,
where v is the serve speed, V is the speed of the racquet at the impact point
and e, is a number known as the apparent coefficient of restitution (ACOR)
typically about 0.4 that describes how well the ball bounces off the strings. For
example, if V = 72 mph and e, = 0.4 then v = 1.4 x 72 = 100 mph.
Consequently, 72% of the ball speed is determined by the speed of the racquet.
The value of e, is determined by (a) the relative weight of the ball and the rac-
quet, (b) the energy loss in the ball and the racquet, and (c) the actual impact
point on the strings. For an impact in the middle of the strings, e, can be as
high as 0.45 for a heavy racquet, but heavy racquets cannot be swung as fast
as light racquets. Even if they could, and if e, increases from 0.4 to 0.45, the
serve speed increases by only 4% to 1.45 x 72 = 104 mph. But if V decreases
by 4% when e, increases to 0.45, there is no change in serve speed at all. Tests
have shown that players can serve almost as fast with an old wood racquet as
with a modern graphite racquet.

A tennis ball bounces faster off the strings than a baseball since a tennis ball is
lighter, but it bounces slower than a superball or a golf ball. A tennis ball will
bounce better off a heavy racquet since a heavy racquet is not pushed back-
wards by the ball as far as a light racquet. For example, if you throw a tennis
ball at a brick wall it will bounce back a lot faster than off the strings of a rac-
quet held in the hand. But you can’t swing a brick wall as fast as a tennis rac-
quet. When these two factors (swing speed and bounce) are considered togeth-
er, it turns out that there is very little difference between a light racquet swung
fast and a heavy racquet swung a bit slower. The ball speed will be noticeably
smaller only if the racquet weight is less than 250 gm, or more than 1000 gm.
That’s why it is largely a matter of preference. Similarly, string tension is large-
ly a matter of preference. Lower string tension will give you more power, but
its only about 1% more. You can get 1% more power by swinging the racquet
1% faster.

There is an interesting lesson here. Whenever you read an article or whenev-

er someone tells you that a certain racquet or technique gives “more” power
or “more” control or “more” of anything else, then make sure you find out
exactly how big “more” is. If it is only 1% more, then it’s usually not worth wor-
rying about. If “more” is not backed up by a number that has actually been
measured or calculated, then the person giving you the information is giving
you an opinion rather than a statement of fact. In fact, very few performance
measurements of racquets have ever been made, and those made by the man-
ufacturers are closely guarded trade secrets.
 Choosing A Racquet

A common selling point for a racquet is that it has a large sweetspot. This
needs to be taken with a grain of salt. There are two sweetspots on the strings
of a racquet, but they are so close together that it is impossible to distinguish
between them with anything other than sensitive vibration equipment. The two
spots are near the center of the strings, only a few centimeters apart, so the
ball can land on both of them at once.

If a racquet hits a ball, then three things happen to the racquet. It starts
to vibrate, it suddenly slows down, and it suddenly starts to rotate about
a different axis to the one you used to swing the racquet. This all hap-
pens in a split second, and you would have to concentrate hard or take
high speed film to see that the racquet actually slows down. It is easier
to see this if you hold the racquet at rest in your hand and someone
throws a ball at the strings, or if you drop a ball onto the strings. The
ball bounces off in one direction, the racquet recoils in the other direc-
tion, and the racquet forces your wrist to rotate backwards since the
whole racquet rotates about an axis through your wrist. The same things
happen when you swing the racquet towards the ball, but you would
need a small video camera strapped to your arm to capture the action.

1. The node. When a racquet vibrates, most of the vibration occurs at

the tip, the throat and at the butt end of the handle. There is a spot near
the middle of the strings that doesn’t vibrate at all (Figure 1.1). It is actu-
ally not a spot but a cufved line extending from the middle of the strings
to points at about 2 and 10 o’clock on the frame. The spot that most peo-
ple talk about is the intersection of this curved line with the long axis of
the racquet. It turns out that if a ball hits this spot (or anywhere else
along the curved line) then the racquet will not vibrate at all, even at the
tip or the butt end. This spot is therefore a sweetspot, technically known
as a vibration node. If the ball hits the strings say one inch away from
this spot, then the racquet will vibrate a small amount. The further away
from the spot, the more the racquet vibrates. If the ball impacts at a point
on the node line near the 2 or 10 o’clock position, then the frame will
not vibrate, but it will twist in your hand. Consequently, the sweetest
point on the cuved node line is the one that intersects the long axis. Figure 1.1
Time exposure of a vibrating rac-
Hence, we can describe different parts of the string bed as feeling per- quet — actually a stiff cardboard
fect, pretty good, or terrible in terms of the amount of vibration trans- model with one elastic main and
mitted to the handle. A stiff racquet will have a bigger “pretty good” area one cross string. A real racquet
would not vibrate as much as this.
than a flexible racquet since the overall level of vibrations is reduced in
The node points in the frame (at 2
a stiff racquet.
and 10 o'clock) and the strings
(near but not at the intersection of
2. The COP. The second sweetspot is a few centimeters closer to the the two strings) are clearly visible.
throat and it is known technically as the center of percussion or COP. A The handle is connected toa
ball impacting at the COP will not cause any sudden motion of the han- vibration generator (a metal rod
dle either forward (towards the net) or backward, but the whole racquet connected to a speaker cone) in
and the wrist will rotate regardless of where the ball impacts on the order to vibrate the handle back
strings. Sudden motion of the handle towards the net occurs if the ball and forth 1 mm at 25 Hz. Real
impacts near the tip of the racquet, in which case the handle jerks for- racquets are stiffer and vibrate at
ward and tugs on the hand and arm. Sudden motion backward results frequencies from about 120 Hz up
to about 200 Hz.
when the racquet strikes the ball in the throat area. An impact at the COP
8 Chapter 1

Figure 1.2
Series of photos taken from video film showing the
location of the COP. A racquet was suspended by
its handle using a long length of string, and a ball
was thrown by hand at low speed to impact at
either the tip, throat or COP on the strings. Dotted
white line is starting position of string and racquet.

Impact at the tip: The handle swings to the right.

The ball bounces at low speed and drops vertically
downwards. Racquet tip moves rapidly away from
the ball at about the same speed as the incident
ball. Time between photos = 0.04 s.

Impact at throat: The handle swings slightly to the

left. Ball bounces best here. Racquet tip moves
slowly away from the ball. Time between photos =
0.04 s. ;

Impact at COP: The COP for this racquet was clos-

er to the throat than to the middle of the strings.
Racquet rotates, but string and butt end of handle
don't move. Ball bounces at slightly lower speed
than off throat. Time between photos = 0.08 s.

generates a small amount of vibration but no

shock, while an impact at the first sweetspot
generates a small shock but no vibration. An
impact at the tip or throat results in both shock
and vibration (Figure 1.2).

3. Maximum power. Many people (especially

the manufacturers) like to include a_ third
sweetspot — the point at which the ball comes
off the strings with maximum speed. However, I
don’t like the label “sweet” for this third spot
since it doesn’t necessarily feel sweet. If you
hold a racquet in your hand and drop a ball on
the strings, the ball bounces best off the throat
and worst off the tip. In other words, the ball
comes off the strings with maximum speed
when the ball impacts in the throat region. This
is an impact region that causes the whole rac-
quet to vibrate and should not be identified as a
sweetspot.

The ball comes off the strings with maximum

speed in the throat area only when the racquet
is initially at rest or moving at relatively low
speed. If the racquet is swung at a higher speed,
then the point at which the ball comes off the
strings with maximum speed shifts towards the
center of the strings since the center is moving faster than the throat. At the
opposite extreme where the ball is initially at rest (as in a serve), then the ball
comes off the strings fastest at a point roughly half way between the center and
the tip. The maximum power point can be shifted even closer to the tip by
 Choosing A Racquet

adding small lead weights to the tip. Even though the racquet travels fastest
right at the tip, the maximum power point is not right at the tip. As described
above, the speed off the strings is given by (1 + e,)V, where V is the speed of
the racquet at the impact point. Since e, is smallest right at the tip, it turns out
that the serve speed from the tip is about 10% lower than the serve speed from
the middle of the strings.

Manufacturers like to specify the size of the sweetspot in terms of the area of
the string bed that generates about 50% or more of maximum power. This is a
larger area than the size of the comfortable hitting zone, and therefore appears
more impressive. This is another example where the advertising brochures can
be misleading and where the actual figures are usually left off the graphs.

Racquets are classified as either head-heavy, head-light or evenly balanced

depending on whether the center of mass is respectively nearer the tip, near-
er the handle, or in the middle of the racquet. The center of mass is also called
the balance point since a racquet can be balanced on one finger, or on the
edge of some other object, if the point of support is directly underneath the
center of mass. Most players have no idea where the balance point of their rac-
quet is located, or why it might be important.

The balance point is closely related to swingweight but it is not the same thing.
Swingweight depends on the balance point, but it also depends on the actual
weight of the racquet. Both the balance point and the swingweight depend on
how the weight is distributed along the racquet, but the balance point doesn’t
depend on how much actual weight there is, whereas the swingweight does.
The situation is complicated by the fact that two racquets of the same length
and the same weight and having the same balance point can have different val-
ues of the swingweight, depending on how the weight is distributed. The
important thing to know is that the swingweight determines the resistance felt
in the hand and arm when you swing a racquet through the air. It is greater
for heavy racquets, it is greater when the racquet is head-heavy and it also
increases if the racquet is longer. For that reason, extra long racquets are gen-
erally very light so that the swingweight will be about the same as shorter rac-
quets.

You can get a good feel for balance point and swingweight by screwing a nut
onto a bolt through the strings at the tip so that it is attached firmly to the
strings. A couple of washers will be needed. A small 10 or 20 gram bolt will
make a big difference to the position of the balance point and to the swing-
weight, even though the total weight of the racquet is increased by only 10 or
20 gm. If you tape the same bolt to the handle, there is no change in the total
weight, but the balance point will shift about 30 mm closer to the handle, and
the swingweight is a lot less than when the bolt is located at the tip. Extra
weight at the tip will slow down the head speed and it increases racquet power
a small amount, but it doesn’t slow down your arm speed by much. Extra
weight in the handle has no effect on racquet head speed relative to the wrist
and no effect on racquet power, but it will slow down your arm a tiny bit.
However, the handle and the hand is already relatively heavy, so an extra 10
gm in the handle is unlikely to achieve much. Some players say it helps, prob-
ably because extra weight in the handle increases the moment of inertia about
the racquet’s center of mass, and hence, it will reduce shock caused by sud-
10 Chapter 1

den rotation of the handle. The same result could be achieved by adding mass
at the tip, but that has a much bigger effect on the swingweight about an axis
through the wrist. Depending on your style of play, and the amount of wrist
action, an extra 10 gm of lead tape at the tip could make a significant differ-
ence to the way you hit the ball. It is easy to add lead tape to a racquet to see
if it helps. If it doesn’t, peel it off and throw it away or try another spot.

In theory, the head size of a racquet needs to be only slightly larger than the
ball. The head size of a golf club or a baseball bat is not much bigger than the
ball, so what is different about tennis? In golf, the ball sits on the ground wait-
ing to be hit. The idea is to hit the ball in the middle of the club head, so the
head can’t be a lot bigger than the ball, otherwise it might hit the ground.
Nevertheless, the head size of a modern golf club has grown to about twice
the size of older club heads. In baseball, it is the batter who remains at rest
waiting to hit the ball. Even so, the batter often misses the ball completely. That
would be embarrasing in tennis, which is why the head is considerably larger
than the ball.

Professional or elite players hit the ball in the middle of the strings most of the
time, so they don’t need to use a racquet with a really large head. If they did,
it would be too cumbersome and not very maneuverable at the speeds at
which they play. On the other hand, recreational players benefit from a large
head since they don’t always hit the ball in the middle of the strings, and they
don’t swing the racquet as fast. A larger head has a larger polar moment of iner-
tia since the mass of the frame is further from the rotation axis. As a result, a
hit off-center (towards the top or bottom of the frame in a groundstroke) does
not cause the racquet to twist in the hand as much.

In addition, a racquet with a larger head has a larger sweetspot area over which
the shot still feels pretty good, partly because the sweetspot is closer to the
center of the strings. If the head size is reduced, the sweetspot tends to shift
towards the throat. The sweetspot remains at the same distance from the tip,
but if the head size is reduced and the length of the racquet is kept the same,
then the center of the strings shifts towards the tip. This can be countered by
locating extra mass in the tip in racquets with a small head, which also helps
to increase racquet power.

Fram i

Modern graphite racquets are much stiffer than older wood or aluminum rac-
quets. They are also stronger, which means that they can be made lighter and
still not break. Part of the reason is that graphite itself is stiffer and stronger
than wood or aluminum, but an equally important factor is that graphite rac-
quets are hollow whereas old wood racquets were solid. For any given mass
of material, a hollow section is stiffer than a solid section. For example, a solid
cylindrical aluminum rod weighing about 300 gm is much easier to bend than
a 300 gm aluminum tube of the same length, provided it has a reasonable wall
thickness, since it will have a larger overall diameter. That is why the metal legs
of a chair are made from hollow tubing rather than solid material.

If you take a wood ruler and bend it, it is much easier to bend across the nar-
row part than edgewise. In other words, a length of material that is only 1 mm
 Choosing A Racquet 11

thick and 20 mm wide is much easier to bend than material that is 20 mm thick
and 1 mm wide. For this reason, tennis racquet frames have a hollow cross sec-
tion that is longer in the bending direction (at right angles to the string plane)
and narrower in the other direction. One cannot make the frame very wide and
narrow like a ruler since it would buckle as soon as it was strung. The cross-
section of the frame is therefore elliptical rather than circular or flat. Widebody
racquets are quite a bit longer in the bending direction, which makes them very
stiff.

An interesting question is whether a stiff frame will feel stiffer than a more flex-
ible frame. The answer is that stiff racquets reduce the impact force on the
hand. The opposite situation prevails with the strings. A stiff string plane acts
to increase the impact force on the hand and therefore feels stiffer than a soft
string plane. If you hit the ball properly in the middle of the strings, then the
racquet will not vibrate at all and it will not slam into your hand, regardless of
the frame stiffness. Frame stiffness is a factor only if you mishit the ball or if
you deliberately hit the ball at a point nearer the tip or throat of the racquet.

Stiff racquets bend less when they vibrate compared to flexible racquets, but
they vibrate faster. The vibration frequency of most racquets is around 120 or
130 Hz, but very stiff racquets vibrate at frequencies up to about 200 Hz. By
contrast, old wood racquets vibrate at about 90 Hz. If you hold the handle near
your ear and tap the frame, you can hear how stiff the racquet is. It’s the same
as hitting the strings so that you can hear how tight they are. However, in order
to hear the frame vibrate, you need to hold the handle lightly between two fin-
gers at a point about half way along the handle, otherwise your hand will
damp the vibrations.

There are four basic string types you can choose from. In increasing order of
stiffness they are natural gut, nylon, polyester and kevlar. Each type is avail-
able in a range of different diameters from about 17G (1.20 mm) to 15G (1.40
mm). Natural gut is the most expensive and is the choice of many profession-
al players. Thicker strings are more durable since they take longer to break.
Most commonly it is one of the central main strings that breaks first since they
move along the cross strings every time you hit the ball. This cuts a groove in
the main strings and weakens them.

Further details on the general behaviour and properties of particular strings are
discussed in the chapters devoted specifically to this topic.

The best tension at which to string a racquet is a matter of personal choice.

Most people prefer around 60 lb (27 kg), but some people like 70 lb (32 kg),
and others like 50 Ib (23 kg). Lower tensions give slightly more power and
higher tensions give slightly better control. Recent tests show that most people
cannot pick a 10 lb difference in tension either in terms of the feel or per-
formance of the racquet. This result is surprising since many players say they
can definitely pick a difference, even though when tested blind, many of them
can’t. However, just as many players admit that they can’t pick a difference and
they don’t have any preferred tension. If you can notice a difference then
12 Chapter 1

maybe you are in the minority who can, in which case you should get your
racquet strung at whatever tension you prefer.

Even though a 10 lb difference in string tension seems a lot, the actual effect
on racquet feel and performance is unlikely to be large. At 60 lb tension the
string plane is moderately stiff and the string plane will stretch about 20 mm
in a fast serve. At a tension of 50 Ib the string plane is softer so it will stretch
further, with the result that the increase in tension during impact is larger, and
the maximum tension during impact is about the same as that for strings at 60
lb. The peak force acting on the ball will be about the same and the speed of
the ball off the strings will also be about the same.
 Graphite, Titanium
and Piezoelectric

im Cli . Belgium

By Crawford Lindsey

APHITE
arketing campaigns brag a lot about the materials with which racquets are made. In days of yore, it
was wood. Then came aluminum and fiberglass. But all the recent advancements in racquet technol-
, ogy are due to one thing — graphite composite construction. Without graphite we wouldn't have over-
size, superoversize, widebody, or superlight racquets. And graphite itself comes in many flavors. Manufacturers
also brag about racquets being made of “high,” “ultra-high” or “hyper-high” modulus graphite. What is graphite
anyway? What is modulus? How is graphite used in the construction of racquets? Why does higher modulus
graphite cost so much? How do you achieve stiffness and strength with these materials?

The basic racquet graphite recipe is to combine carbon fibers with a polymer resin to form a composite mate-
rial. The composite material is then layered, shaped, and molded into a racquet with various properties includ-
14 Chapter 2

ing weight, strength and stiffness. But all along the way, manufacturers are able
to alter different variables to change a material's final contribution to the prop-
erties of the product. Each point of difference is eligible to end up as a mar-
keting point of difference somewhere, someday.

Modulus, as it is most commonly used, describes the

Modulus Defined
relationship of stress to strain on a object when a pull of
Before After a given force is applied along its length (Figure Pal):
Area =0.5"x 0.5" Area = 0.4" x 0.4" More specifically,

force/cross section area

Modulus = stress/strain = - —-
elongation/original length

elongation = 0.1" Under normal conditions, stress and strain always

change in the same proportion, so the modulus (stress
divided by the strain) is always the same, and that is
why “modulus” is considered a property of a material. If
Stress = force/area = 100 Ib/0.16 in.? = 625 Ib/in.?
you double the stress, the strain will also double.
Strain = elongation/original length = 0.1 in./1 in. = 0.1
Modulus = Stress/strain = 625 Ib/in.2/0.1 = 6250 Ib/in.?
Manufacturers have typically advertised the graphite in
Figure 2.1 their racquets as either intermediate or high modulus
graphite. Then, “ultra-high” modulus started to creep
Although racquets are not stretched, the material mod-
into promotional literature. And then, Wilson recently
ulus most often quoted is the tensile modulus, not the
came out with a graphite called Hyper Carbon. This
bending modulus.
caused quite a stir because their aggressive marketing
campaign really forced tennis players to become edu-
cated about the different grades of graphite, if they were
to make heads or tails of all the competing claims.

The first thing to understand is that “hyper” is not a new

Carbon Fiber Products category above ultra-high. The materials industry simply
1000 T uses four modulus categories:
| © Buff-ite 1 Halucin-ite ©®
|
900 |- | Standard = 30-39 Msi (million pounds per square inch)
| Intermediate = 40-49 Msi
!
| High = 50-58 Msi
I Ultra-high = > 58 Msi.

Strength
)
(ksi “Hyper” is a brand name for a particular ultra-high prod-
600 | uct from a specific racquet manufacturer. The same is
true for Prince’s GraphitExtreme or any of the other
brand’s naming of the material. Hyper Carbon has a
Ir Limp-ite 8 Stiff-ite e
i modulus of 63.3 million pounds per square inch (Msi)
20 40 60 80 100 and a breaking strength of 612 thousand pounds per
Modulus (Msi) square inch (ksi). So, “Hyper” is the marketing tag,
Figure 2.2 “ultra-high” is the modulus classification. Figure 2.2
shows a plotting of some generic carbon fiber products
Is higher modulus good or bad? It depends on how you
on today’s market that are used in racquets.
use it. As is discussed later in the text, as modulus goes
beyond 40 Msi, strength tends to go down. Moving along You will notice that each category is a range. So, if you
the diagonal line gives equal increases in strength and really wanted to know which racquets have the highest
stiffness. As you can see, that is difficult to achieve. That modulus graphite, you would need to know the actual
is why we have playfully named that quadrant “Halucin-
specs of the raw material. Numbering of the modulus, of
ite.”
course, gives consumers a better comparative guideline
 Racquet Materials 15

than category names do, especially if marketing brand names are then added
to each category. But even if you know the exact modulus specs of the mate-
rial in a racquet, how do you know whether that is good or bad for the desired
product and whether it makes a difference to the product’s performance? That
is more complicated. To answer that, we must first look at how graphite is
made and how it is used in racquets.

_Making High Modulus Carbon

Phase 1 — The Precursor Material: Carbon fibers are made from one of
three carbon-rich starting materials — rayon and polyacrylonitrile (PAN), both
polymer processed fibers, and pitch, a residue of coal and oil refining. All three
are heated into liquids and extruded through a disk about the size of a nickel
with thousands of microscopic holes that shape the liquid into filaments. The
filaments are cooled, congealed and bundled into fiber-yarns containing thou-
sands of filaments. The fibers are then rolled onto large bobbins. Pitch-based
fiber is the precursor for very high modulus fibers over 75 million pounds per
square inch (Msi). Rayon and PAN are used in clothing or further processed
into carbon fiber. PAN is the precursor for most racquet carbon.

Phase 2 — The Carbon Fiber: The PAN fibers (literally, “acrylic sweater
yarn”) are pulled through two ovens of varying high temperature and oxygen-
free atmosphere to burn off all non-carbon elements (carbonization). The fibers
are tensioned throughout the process to align the carbon molecules in hexag-
onal crystal sheets along the fiber. At this stage, modulus is affected by tem-
perature, time at temperature, and the amount of
orientation achieved in pulling. Temperature expo- What Is Resin?
sure varies from 1-3 hours. Higher temperatures
produce purer carbon fiber. The grades of carbon ee esin binds the fibers together. In fact, a composite is
are typically 90-99 percent carbon. Though the often referred to as “fiber reinforced matrix,’ which
terms “carbon” and “graphite” have been used implies the matrix has top billing in this relationship.
interchangeably over the years, graphite is defined Behind every fiber there’s a matrix. Resins are polymers (long
as greater than 95% pure carbon or as having a chains of molecules) and are of two kinds — thermoplastic
and thermosetting. Resins, too, have their own strength and
modulus greater than 50 Msi. The higher the car-
elasticity properties.
bon content, the higher the modulus.
Thermoplastic resins soften with temperature. These resins
Higher moduli are created by higher temperatures. are elastic because the individual molecules are not bonded
Temperatures for good carbon fiber might be to their neighbors and can slip past each other when being
1,200-1,400 degrees centigrade and for good stretched. Consequently, thermoplastic resins are good can-
graphite between 2,000 and 3,000 degrees centi- didates for string.
grade. The energy costs and quality control for
attaining higher modulus are considerable. Also, Thermosetting resins are used in racquets. These are two
higher modulus fibers are more brittle and difficult types — polyester and epoxy — of which the epoxy predomi-
to handle, so they have to be processed slower, nates in racquets. Thermosets harden, or are cured, under
high temperature. The distinguishing feature of thermosets is
and that also costs more.
that molecules link to each other, forming a cross-linked sup-
porting structure. This results in additional strength and stiff-
Emerging from the ovens, the fibers go through a
ness. Because the molecules in the matrix are not aligned in
cleaning bath and are coated with a one-percent any one direction, the strength and stiffness properties are
epoxy solution known as a “sizing.” This facilitates the same in all directions — they are isotropic. Fibers, on the
handling. Without sizing, the fibers are like wispy other hand, are anisotropic, meaning the properties change
microscopic hairs floating in the air. The fibers are depending on the direction of the stress to the long axis of the
gathered in either 1K, 3K, 6K or 12K bundles fiber. In many stress situations, the matrix does more than the
(tows) composing a carbon fiber yarn that is then fiber. But it is not strong enough to do it all, so fibers are
rolled onto spools. Racquet manufacturers typically placed at many angles to handle all possible load conditions.
16 Chapter 2

use 12K bundles because they cost less and are easier to handle in the next
step — prepregging. Fiber strength and modulus specs refer to these fibers
before they are combined This is important to note, because,
into a composite.
as we will see, the properties of the composite will depend on both the
strength and modulus of the composing graphite and resin, and also on the rel-
ative proportion of each material in the composite.

Phase 3 — Prepreg: This is where a “composite”

Resin: The Weakest Link is created, Hundreds, or even thousands, of spools
of bundled fiber are placed on one end of the
f the matrix is the weakest link, why isn’t that the strength prepreg line, and the ends are pulled over a comb
of the composite? Epoxy resin has a strength index up to to spread and align all the fibers into “sheets” 12”
about 20,000 Ib/in? (20 ksi) and a modulus of 600,000 Ib/in* to 48” wide. The sheet is pulled through an epoxy
(0.6 Msi). In the case of Hyper Carbon, the strength is 612 ksi resin bath. The excess resin is scraped off with a
and modulus is 63.3 Msi. So why doesn’t the composite break squeegee device and the sheet is rolled onto a
when the stress exceeds that of the matrix? The answer is drum with a removable paper backing. Because all
that the resin doesn’t share the load equally with the fibers the graphite fibers are arranged parallel to each
because it is more elastic. Elasticity is a measure of how far
other, this is known as unidirectional graphite
you can stretch a material before the bonds between atoms
prepreg. The term “prepreg” refers to the fibers
and molecules build up enough force to balance and counter-
act the stress — as when tensioning a string at 50 pounds
being impregnated with resin before the actual
and it stretches until it works up 50 pounds of resistance. molding process. The combination of the resin and
the fiber is referred to as a composite. The resin
The relationship between the matrix and the fiber is much like also contains a curing or hardening agent (catalyst)
gluing a Kevlar and nylon string together. If you tension the and a coupling agent that together causes the resin
combination in a stringing machine, each fiber will elongate molecules to link together in a supporting cross-
the same amount (because they are glued together), but that linked network and to adhere to the carbon fiber.
amount of stretch will not build up very much stress in the
This occurs under heat and pressure in the racquet
nylon. Instead, most of the stress will be transferred to the
molding process.
Kevlar, which will bear most of the load. If the Kevlar is 10
times as stiff as the nylon, it will bear 10 times the load. That’s
exactly how fiber and matrix work together. The epoxy resin is also known as the matrix.
Neither the fiber nor the matrix is much good with-
out the other. The loose fibers by themselves can’t
transmit loads to each other, and the matrix is not
nearly as stiff or strong as the fibers. So the matrix serves three purposes: to
bind fibers together, transmit loads to each fiber, and to protect the fibers. The
strength and stiffness of the composite falls somewhere in-between those of
the fiber and the matrix. In other words, the strength and modulus of the com-
posite is less than that of the fiber specs by themselves.

The strength and modulus properties of the composite depend on the volume
ratio of fiber to matrix. In fact, the modulus of the composite as a whole equals
the modulus of the fiber times it’s volume percentage in the composite plus the
same for the matrix. The fiber/matrix ratio typically varies between 50/50 to
60/40. But there is a theoretical limit to how high the ratio can be and still
allow the epoxy resin to serve its purpose. That limit is around 90%, but it’s
not approachable in any practical manufacturing application. Middle 70s might
be doable in some high-end space applications, but the generally accepted
practical limit is around 60/40 fiber-to-matrix. The theoretical limit is deter-
mined by the geometry of the fibers and by how much space is left over when
they are packed as close together as possible. If the packing is too tight, the
heated resin can’t flow into the spaces between fibers during molding. You end
up with weakness-causing voids as a result.

The sheets of prepreg are then either stored or first cut into shapes for later
lay-up, catalogued by shape or product, and then stored. Prepreg is
usually
 Racquet Materials 17

Thin Prepreg

trength and stiffness can be increased by using thin prepregs — also known as low FAW (fiber-average-weight)
prepregs. This involves an additional step in prepreg manufacture. Before the fibers are resin-coated and rolled onto
the paperbacked drums, they go through a series of rollers that spread the fibers as much as possible. The goal is to
minimize the number of fibers per inch and yet not have any space between fibers. In other words, as little overlapping as
possible. Thinner prepregs have lower weight per square meter of material. The thinnest FAW available is between 60-70
grams per square meter. Just ten years ago, the lightest prepreg was 115 grams per square meter.

Thin prepreg offers a couple advantages. Thinner layers allow a manufacturer to use more layers while maintaining the same
weight and thickness. More layers allow more combinations of fiber angles, allowing greater customization and optimization
of strength and stitfness properties. This facilitates using less material overall, resulting in thinner frame walls and a lower
weight racquet.

Low FAW prepreg costs more because it requires more machine time and equipment. Many technologies have been devel-
oped, but there is a cost associated with each one. If you want a lightweight racquet, you start with more expensive materi-
al, processing takes more time, and quality standards are higher. All these things build in cost.

refrigerated to stabilize any room temperature curing (hardening), though it

can last several months without refrigeration. One prepreg company even
claims one year.

Making a Racquet
Now we want to make a racquet. This is where it becomes fun. There are a
number of considerations:

1. Fibers are very stiff along the length-wise axis, but not so at any fiber angle
off that axis. If the load is applied just 15 degrees off axis, the modulus may
decrease to 60% of maximum. At 45 degrees it could be as low as 20% and so
on down to 0 at 90 degrees (Figure 2.3).

2. As modulus goes up, strength goes down (see Figure

2.2). Why? It has to do with the number of imperfections Angle, Stiffness, Strength
on the fiber surface per unit of length. Imperfections are
0 30° 60° 90°
both uncovered and created during the heating process. It ‘ A r
has been proven that microscopically short fiber lengths of y
different moduli have approximately the same strength. yy
However, a one-inch piece of high modulus fiber will have
more imperfections than one inch of a lower modulus, high
strength fiber.

3. As modulus goes up, fiber becomes more brittle. When

the atoms and molecules of a material are so completely
oriented that there is no room left for movement, there is
no elongation or deformation prior to breaking. There is
I
Stiffness (modulus) drops off very steeply
just instantaneous failure — especially if impacted at an when the load (direction of arrows) is applied
angle off axis to the strongest direction of the fiber. just alittle off the fiber direction.As the angle
Brittleness is an impact phenomenon. increases, the pull is less on the fibers and
more on the matrix.

4. As modulus goes up, the fiber’s density goes up because

the lighter elements have all been burned off. Higher mod- Figure 2.3
18 Chapter 2

ulus material is heavier per foot or volume.

5. Higher modulus costs more. There are higher energy, quality control, and
handling costs.

6. The lighter the racquet, the more strength and stiffness is needed to make
up for less material. But stiffness and strength move in opposite directions.

These factors limit how much of any material you can use and what you can
use it for. Every racquet must combine a blend of different modulus material to
balance strength, durability, stiffness, and economics. This is accomplished in
a racquet’s lay-up.

Most racquets are made of multiple layers aminations or plies) of unidirec-

tional prepreg. One layer won't do it. So no matter what the fiber’s or com-
posites’ specs, racquet stiffness and strength depend on prepreg. The art of lay-
up lies in determining the angles at which the different layers of unidirection-
al prepreg — each perhaps with different modulus and strength properties —
will be oriented to each other.

Five to ten layers are common for tennis racquets, so several angle combina-
tions are possible. If we consider the length of the racquet as the reference axis,
a zero degree angle is parallel to that. Zero degree high modulus fiber makes
the racquet stiff to bending. Ninety degree prepreg is used to stiffen the hoop
in the string pull-through direction, and every angle in-between will increase
torsional stiffness. Manufacturers typically think in terms of 7.5- or 15-degree
increments. The trick is to stack all the layers so the strength and stiffness are
optimized to match the racquet’s target player (see thin prepreg sidebar).

Modulus classifications are not good or bad in themselves. High modulus is not
better than intermediate or low modulus carbon fiber. It depends on how they
are used and to what end purpose. The key is to use the material in a design
that will optimize the material’s properties. Without the design, the material’s
properties may be wasted. Without the material, the design may be impossible.

Titanium is the ninth most abundant element in the earth’s surface and is the
fourth most used metal in industrial applications — behind aluminum, iron and
magnesium. It is the strongest and lightest metal. It has very good torsional
stiffness, and like most metals, is isotropic, meaning it is equally strong in all
directions. Graphite, on the other hand, is anisotropic. Its strength and stiffness
is more oriented in the direction of the fibers and not, for example, at right
angles to them. However, titanium is about two and one-half times heavier and
three times less stiff than high modulus graphite (see Table 2.1). As the stiff-
ness/weight column shows, it theoretically takes 3.15 times (5.235/1.66) the
titanium to equal the graphite’s stiffness.

The question then arises, why are racquets containing a small amount of tita-
nium so light? Manufacturers tell us that it is how titanium and graphite are
combined that makes the difference. Though its specific properties don’t equal
 Racquet Materials 18

graphite in weight and _ stiffness, Table 2.1

their use together allows their prop-
erties to complement each other — Mechanical Properties Comparison
sort of the whole is greater than the
sum of the parts argument. The the- Tensile
ory is thatin a graphite racquet, ’ Ptrengih Benelly modulus bese
separate layers of graphite fibers sb al) GBM) eae) Ubn:) mot
eon ied an ane meen) 1401000 0.09548 10,100,000 1.057
tions to gain strength. But by
adding isotropic titanium fibers toa titanium —-130,000 0.16257 27,000,000 1.660
single layer of unidirectional
graphite, less right-angle graphite Kevlar 203,052 0.052745 11,020,000 2.089
has to be added. It then becomes a
complicated mathematical/mechan- — Graphite 217,000 0.057803 26,251,830 4.541
ical balancing act between titani-
um’s isotropic strength benefits and HM Graphite 638,166 0.06647 34,800,000 5.235

torsional — stability, but heftier

Source: Mark’s Standard Handbook for Mechanical Engineers and Composite
weight, and the weight savings of
Materials Handbook.
eliminating some of the high modu-
lus graphite.

Furthermore, the titanium is only

used in limited strategic strength
areas like the throat and shoulder
area where there are increased tor-
sional and bending stresses. It is
also typically used in small percent-
ages. Does it work? It does in golf.

Titanium is used differently in golf

and tennis, however. In golf heads,
metals are used as the predominant
structural element. Titanium, rather
than graphite, is used here for three Figure 2.4
main reasons. It is much more
Titanium fibers interwoven with graphite in
abrasion and impact resistant than
Head’s “graphite/titanium matrix.”
graphite; it is more resilient upon
impact, thus adding to ball velocity;
and its isotropic properties are ideal
for bearing the golf ball’s impact loads. Arranging enough light-weight graphite
fibers in the directions necessary to resist golf ball impact might make the
graphite head almost as heavy as the titanium head, without the other advan-
tages of titanium. So sometimes light weight can be heavy (golf) and some-
times heavy can be light weight (tennis).

For these reasons, titanium is used as a predominant material in golf but plays
more of a supporting role in tennis. Racquets became superlight just as titani-
um was being marketed. Titanium is not the reason racquets are light. So, in
spite of the marketing hype, titanium plays only a supporting role in racquet
construction. Graphite is the workhorse material that makes everything possi-
ble.

But manufacturers are constantly pushing the materials envelope. Head was
the first racquet company to put “smart” materials into their racquets.
20 Chapter 2

and stiffening system. What the heck is that?

In a nutshell, piezoelectric material generates an electric charge when a force

deforms it. The charge produced is proportional to the force — more force,
more charge. The opposite is also true. Piezo material is deformed when it is
subjected to an electrical charge. In this case, deformation is proportional to
the amount of charge. They are reverse processes to each other. One will
exactly undo the other.

Mechanical energy is turned into electrical energy and electrical energy is

turned into mechanical. This all happens by the miracle of positive and nega-
tive charged particles pushing each other around as materials are pressed, bent
and electrified. If you get the right pushing and pulling, you get an electric cur-
rent as your reward (Figure 2.5).

This creates two benefits. First, energy turned into electrical energy does not
go into vibrating the racquet. That's the damping effect. Second, there's a self-
induced restorative force that acts to stiffen the racquet the more it is bent.
Racquets with piezo material are intelligent to the extent that they react vari-
ably and adaptively to their environment. They sense the force and dissipate it
in electrical energy in proportion to the size of the force. At the same time, they
produce some internal stiffening that is also proportional to the force. But as
smart as these racquets are, they are still passive.

Head also developed a racquet that is an active participant in its universe. It

possesses a higher intelligence in the form of an electronic brain that is resi-
dent in the handle and connected to its motor system by an electronic nervous
system. The brain takes the electric signal, measures it, times it, processes it,
and redirects it back to whence it came to do work. The return signal tells the
piezo material (‘intellifibers”) how much to contract to counter the impact
force and when to do it so as to create a canceling vibration to the impact
vibration.

So we have stiffness and damping as calculated by the brain. Upon impact, the
piezo material acts as a sensor (“generator” in piezo lingo) device, sends the
data to a control device, and then actively performs a motor (“actuator") func-
tion. In this case, that would be the “active counter-force" that acts within one
millisecond to limit the amplitude of the impact-induced bend. Then this
“counter-bend," if you will, creates vibrations that will help cancel the residual
impact vibrations because it is timed to be “out of phase" with the latter. This
residual cancellation is just that — residual — because mechanical vibrations trav-
el through the racquet at the speed of sound, whereas the electrical signals
travel at the speed of light.

Piezo materials are one of a class of materials called “smart materials.” Besides
piezoelectric materials, there are shape memory alloys, magnetostrictive mate-
tials and polyelectrolytic gels. Basically these all match a change in shape with
a change in one of the following variables: electric charge, magnetic field, or
temperature. And it goes both ways—a change in shape (like bending) due to
a force will result in an electric charge, magnetic field, or change in tempera-
ture. And within certain ranges of behavior, these changes are very precise,
 Racquet Materials 21

consistent, proportional, and can be calculated in nanometers. (So exactly, in

fact, that piezo applications are copiously abundant in such high tech areas as
positioning scanner heads, precision tooling, and extremely accurate measuring
devices.) As such, a change in shape can be used to measure a change in charge
or temperature, just as a change in charge can be used to measure a change in
shape. Or, thinking less passively, a change in shape can be used to produce
an electric charge, which in turn can produce a change in shape. In other words, Figure 2.5
you can use these changes to do things, as the Intelligence racquets do. How piezoelectricity and com-
puter chips work in a racquet

1. Make ceramic molecules with charged ends

Atoms are electrically neutral. They have an equal number of positive
protons in the nucleus and negative electrons flying about the nucleus.

Atoms combine into ceramics (the material used in the Intelligence

Series is lead zirconium titantate [PZT]) by giving and receiving
electrons. Atoms try to get 8 electrons in their outer shell by combining
with other atoms that have the number needed to complete the shell.

When this happens, one atom becomes negatively charged and the
other positive. They then electrically attract each other (opposites
attract) and form a bond uniting them into a molecule of the new
material.

One end of the combination is negative and the other positive. This is
called a dipole.

Billions and billions of molecules of the material come together to

form the material. (Actually a powder is mixed, milled, shaped and
heated.) At this stage, the dipoles are facing every which way, and
the material as a whole is neutral as dipoles arrange themselves so
they are all happy in electrical equilibrium.

3. Polarize the material--make it piezoelectric

Negative charges attract positive ends

The key to the unique piezoelectric effect is the ability to orient and
freeze the dipoles in an aligned manner.

To make the material piezoelectric, it is polarized by exposing it to

an electric field. The charged ends of each dipole align themselves
toward the opposite charge electrodes that are exerting the electric
field. The field and heat are then removed and the dipoles "freeze"
permanently into alignment.
Direction
Polarization

The result is an electrically neutral material on the inside (all the

charges balance each other out), but the top surface has a positive
charge and the bottom surface a negative charge. These charged
surfaces attract opposite charges from the atmosphere around
them to balance the charge so the surface is neutral also, but the
stage is set for intellifibers to do their thing.

Figure 2.5 continued on next page

22 Chapter 2

Figure 2.5 (continued)

4. Create fibers and attach electrodes

The dipoles within each fiber are aligned from top to
bottom. Before impact, neither the electrodes or the fibers
have a charge. Everything is neutral.

Electrodes
ME \
SA.

2
[pases]

Electrodes are metal conductors with "loose" electrons that

jump from atom to atom away from an electric field of
negative charge... or towards a field of postive charge. As
they do so, they repel other electrons in their path creating a
domino effect called electric current.

5. Stretch fibers to create charge

Racquet bends at impact, stretching intellifibers
When piezo material is stretched or compressed, the polarized dipoles are also stretched
(in this case) and displaced from their neutral positions and the dipoles become even more
polarized because the charges separate. This chases away like charges on the surface
electrodes, and attracts opposites, thus creating a current.
Electrons flee down wire to chip

Electrode 2: positive dipole end

attracts electrons, creating negative charge
(i.e., a surplus of electrons).

Electrode 1: negative dipole end repels 2 eS OC S

electrons leaving positive charged electrode ©
(i.e., scarcity of electrons).

The current produced instep 5 travels to the chip, turns it on, and then
is sent back to the fiber electrodes at a timed interval that will contract
the fibers in the opposite frequency at which they were stretched,
thus minimizing vibration by cancellation. The current from the chip
sends electrons into the positive electrode, reversing its charge to
repel the negative end of the dipole. Meanwhile, electrons leave
electrode 2 toward the now positive chip.

Repeat the entire process back to step 5, and so on until damping is

complete.

2) Current from chip negatively charges electrode

a Electrode 2: Electrons move toward chip balancing

shortage left by chip sending electrons to electrode 1.
This leaves positive charge on electrode, repulsing and
shortening the positive end of the dipole.

Electrode 1: As electrons flood electrode and = OC 2)

repel the negative end of the dipole causing O C
it to shorten.
 Distribution of Weight
and a Player’s Swing

Bjorn Borg, Sweden

By Crawford Lindsey and Howard Brody

' hy do I feel one kind of weight when I pick up a racquet, and another kind when I swing it? It seems
that weight takes on different guises. First, there is weight, which is a force — the force of gravity on
* an object. Then there is mass — an object’s resistance to changes in its motion — its inertia. Mass and
peient are different physical quantities, but we will neglect that (see Note 1). The racquet weighs ten and a half
ounces and has a mass of 300 grams. That’s easy. But why does it feel like it weighs more depending on where
I pick it up — at the handle, head, or somewhere in the middle? This is where distribution of mass and balance
point come into play. And last, a racquet may have the same weight and balance point as another, yet still be
more difficult to swing. This is where distribution of mass combines with racquet acceleration in swingweight.
This chapter will explain the concepts of mass, weight, balance, and swingweight. Other chapters will deal
specifically with their role in power, maneuverability, feel, shock, and vibration.

Mass and Inertia

Nothing changes the motion of an object without a force — tennis balls included. Without force, all objects tend
to maintain their speed, direction, and rotation. All objects have mass or inertia, which is the resistance to change
in motion. Force, therefore, doesn’t make the universe go (it goes by itself), but it does change how it goes. And
to change an object’s motion, you must change its speed or direction. In other words, you must accelerate it.
Thus, we have the most famous law in physics and tennis for making things happen: F=ma (force equals mass
24 Chapter 3

times acceleration). This says that you need force applied to a mass to get the
object moving to the desired velocity. So, mass is the primary restraint in what-
ever you want to do. And the effect of mass will depend on its amount and
also its location.

The implications for racquet sports are obvious. You have to get your racquet
in motion, but once you do, it takes little force to keep it going, just enough to
overcome the force of air resistance. But then you have to stop it, change its
direction, and start all over — all of which require additional force. Sometimes
inertia helps you, sometimes it hurts you. Your racquet resists everything you
want to do with it, but once it is doing it, it doesn’t want to stop. So every atom
in your racquet is a little mass of resistance. Sounds somewhat daunting, does-
n't it? But before we try to swing these rascals, let’s see what mass has to do
with the feel of the racquet.

You know very well what the weight of a racquet means. You put your racquet
on a scale and read how much it weighs — 10.5 ounces, 300 grams, etc. If you
hold the racquet at the balance point, the weight is the amount of force your
hand must apply to keep the racquet at rest and the force your hand is apply-
ing is equal and opposite to the force of gravity on the racquet. Most racquets
feel remarkably light when held at the balance point instead of the grip, but
you don’t hold the racquet at the balance point when you play tennis. For the
first half of the twentieth century, most racquets were made of wood and
weighed 14 ounces (400 grams or 0.4 kg), or more. With the introduction of
graphite composite frames, engineers were able to produce racquets with
weights as low as 7 or 8 ounces (200 — 230 grams). This is a great technologi-
cal feat, but may not be the best racquet weight to play tennis with. The Rules
of Tennis do not specify a maximum or minimum racquet weight.

You probably know what the balance point of a tennis racquet means, but do
you know its significance? The balance point is the place near the center of the
racquet where you can balance the frame on one finger (that’s why it is called
the balance point). Sometimes the balance point is called the center of mass,
or the center of gravity. You normally measure the balance point as a distance
from the butt end of the racquet. Technically, the distance from the butt end to
the balance point is the sum of each little bit of mass in the frame multiplied
by that bit's distance from the butt end of the racquet, all divided by the total
mass of the racquet. If that resulting distance from the butt is more than half
the racquet’s length, you say the racquet is head-heavy. If the balance point is
closer to the butt end than the geometric center of the frame, you say the rac-
quet is head-light. Often the distance from the balance point to the geometric
center of the racquet is specified as 1/2-inch head-light or 1-inch head-heavy,
Sikey

Though balance seems straightforward, it is sometimes confusing. For example,

since racquets now come in different lengths, two evenly balanced racquets
(balance point halfway between the tip and the butt) may have their balance
points at different distances from the butt end. What the racquet manufacturers
and racquet testers should tell you is the actual distance of the balance point
from the butt end of the racquet.
 Weight, Balance & Swingweight 25

There’s another common misconception. Many

(a) Meter Stick
people think balance point means there is an equal
amount of weight on each side of that point. This ee ee
can be true, but it is not necessarily so. For exam- 0 10 20 30 40 50 60 70 80 90 £4100
ple, if you balance a meter stick (a ruler 100 cen-
timers long — Figure 3.1a), it balances at 50 cen-
timeters (Figure 3.1b). Now, if you place a 10-gram (b) Balance Point = 50 cm

weight at each end of the meter stick, it still bal-

ances (Figure 3.1c). But if you place either weight a 0 10 20 30 40 60 60 70 80 90 100
different distance from its nearest end than the
other weight, the meter stick no longer balances at
the same point (Figure 3.1d). To make it balance (c)
again at that point, you have to add more weight to 10g 10g
md = 10 x 50 = 500 g*cm md = 10 x 50 = 500 g*cm
the “high” side of the stick. The amount you need
to add depends on where you add it. Thus, in this
case, to balance the meter stick at 50 centimeters,
you need to have a different amount of weight on
each side of the balance point (Figure 3.1e). If you
have 10 grams at one end (50 centimeters from the
balance point) and 10 grams 20 centimeters from
the other end (30 centimeters from the desired bal-
ance point), then to keep the balance at 50 cen-
timeters, you must make the weight times the dis-
tance from the original balance point the same on
each side. So if one side is 10 grams times 50 cen-
md = 10 x 50 = 500 g*cm
timeters equals 500 gram centimeters, the other side
must be a total of 16.7 grams times 30 centimeters
(in other words, we must add 6.7 grams to the exist- (e)
ing 10 located at 30 centimeters from the desired 10g

balance). Or, we could have added any amount of

weight anywhere along that side of the meter stick
such that each weight times its distance from the 0 10 20 30 40 60 60 70 80 90 100
balance together added to 500 (Figure 3.1f). So md = 10 x 50 = 500 g*cm md = 16.7
x 30 = 500 g*cm

adding 10 more grams 20 centimeters from the bal-

ance would have also worked. That would result in
20 grams on one side of the balance point and 10
grams on the other. Yet, the balance point is the
same (50 centimeters) as when we had 10 grams on
each end of the meter stick.

A weight feels heavier or lighter depending on its 0) 20) SO} AN 60 60 70 80 90 100

distance from its support point. To feel this, pick up md = 10 x 50 = 500 g*cm md = (10 x 20) +(10 x 30) = 500 g*cm
the meter stick (or a yardstick) from one end and
place a weight at the middle at 50 centimeters. Now Figure 3.1
move the weight to the end (100 centimeters away from your hand). It should
Weight acts differently
take about twice as much effort to hold up the end of the stick in this case.
depending on where it is
The same weight seems heavier. In fact, the weight being twice as far feels located.
twice as heavy. So the balance point is determined by the combination of the
mass and location of each little piece of the racquet. The actual formula for the
balance point of a racquet is to add each infinitely small racquet piece times
its distance from the end of the racquet and divide by the total mass of the rac-
quet:

(m,d, + m,d, + ... m,d_) + total mass = balance point (3.1)

26 Chapter 3

This formula is not balancing just mass or weight — it is bal-

Balance Point And Feel ancing weight multiplied by distance. That, of course, is why
two racquets can have the same weight but different balance
points or different weights and the same balance point.

13" = Balance Point |~

veight
300 gm = Weight | The balance point is important for another somewhat magical
reason. The average of all these individual weights and dis-
13" tances come together in this spot such that the entire racquet
behaves as if all its mass were concentrated right there. That is
why the balance point is often called the center of mass. When
you pick up the racquet from the handle (hence the name “pick
up weight”) and hold it horizontal to the floor, the weight you
feel is exactly what it would feel like to hang the racquet’s total
Figure 3.2
mass at the balance point. So, for a 300 gram racquet with bal-
If you pick up a racquet by the handle, ance point of 13 inches, you are feeling the same thing as if you
Brent wouldeel, exactly the: same it;the were holding onto a weightless rod with 300 grams hanging off
whole mass was concentrated at the
its end at 13 inches (Figure 3.2).
balance point at the end of a weightless
rod witha a weightless handle
This makes the concept experiential, but how do you compare
a racquet of 300 grams with a balance at 13 inches (33.02 cm)
with one of 290 and a balance at 13.5 inches (34.29 cm)? Which one will actu-
ally feel heavier due to a larger pickup weight? Easy — just multiply the weight
times the balance point (convert to centimeters first to measure in the same sys-
tem) and divide by 100 to make the number manageable. In this case, we get
99.06 gm-cm for the 300 gram racquet vs. 99.44 for the lighter racquet. So the
lighter racquet feels just a bit heavier (or more head-heavy) even though it’s
lighter. Or vice versa, the heavier racquet feels slightly lighter.

This pickup weight is also known as the “first moment” or the “static moment.”
“Moment” is a physics term related to length of the lever arm from the force
(the weight) to the point of application (your hand). That is also the definition
of a torque — force times distance Cength of the lever arm). A torque tries to
make an object spin faster or slower, depending on which way it is applied.
So, the weight of a racquet torques or twists your hand by the amount of all
the racquet’s weight acting as if it were at the center of mass times the distance
to the end of the racquet. Your hand must apply an equal but opposite torque
(so the sum of the torques on the racquet are zero) to keep the racquet from
twisting.

o Tennis Players
What does this mean to a tennis player? The player has to supply a counter
force and torque to the racquet. When you hold a racquet in the normal man-
ner (with your hand on the grip), you must apply a torque with your fingers
just to keep the head up (remember, a torque is a force times a distance — it
is a twist). You are pulling up with your fingers and pushing down with the
heel of your palm. The amount of torque you must apply to the handle to keep
the racquet in equilibrium is the same as the racquet is applying to your hand
— the weight of the racquet multiplied by the distance from your hand to the
balance point. This torque that you apply to the racquet is the feeling you
relate to whether you think a racquet feels heavy, or feels light, when you are
just holding, not swinging it. You can always make a racquet feel lighter in
 Weight, Balance & Swingweight 27

your hand by choking up on it just a bit. This

reduces the distance from your hand to the
Balance Point And Forces on a Racquet
balance point, therefore it reduces the amount
of torque your hand must apply to keep the
racquet in equilibrium Gust holding it steady).
-——R' = Balance ———_1
However, there is a second way to make the
Palm
racquet feel lighter in your hand. Spread your
fingers out.

If you examine the situation carefully, you will

discover that your index finger is pushing up
on the racquet handle while the butt of your
palm is actually pushing down on the butt end i
of the racquet. The force your index finger
must exert on the handle to keep the racquet Fingers W = Weight
head up is the weight of the racquet (W) times
the distance from the butt of your palm to the
racquet balance point (R) divided by the dis- Figure 3.3
tance from the butt to your index finger (H)
(Figure 3.3). As a formula, this reads: The upward and downward forces due to your hand produce a
counter-clockwise torque on the racquet which balances out
the clockwise torque produced by the force of the racquet’s
Force = Wx R/H weight around your hand.

This ratio of distances multiplied by the rac-

quet weight is about four or five (the distance
to the balance point is about 13.5 inches and the upward hand force is applied
about 3 to 4 inches from the butt end), so the racquet will “feel” four or five
times as heavy when you hold it by the handle as when you hold it at the bal-
ance point.

By spreading your fingers, you can increase H from 4 inches to almost 5 inch-
es. With your fingers tight, a 12-1/2 ounce racquet will require 40.6 ounces of
upward force to keep the head up, while with the fingers spread, it will require
only 32.5 ounces of force. It is clearly advantageous to spread your fingers.

All the tennis books, all instruction, and all the teaching professionals tell you
to spread your fingers when you grip the racquet handle, and not to hold it
with a tight fist like a baseball bat grip. This gives more control, power, etc.,
but there is one other major benefit that spreading your fingers produces that
never seems to be mentioned in any of these instructions .. . it makes the rac-
quet feel lighter to hold.

You can also reduce the force required to keep the head up by moving your
entire hand up the handle toward the head, but this is not as effective as
spreading your fingers. Moving your entire hand does not increase H, and you
have only reduced R by a small fraction of its original length. To get the same
reduction as the spread fingers provide, you would have to choke up about
2.6 inches, which is a considerable distance.

If two racquets have the same weight and the same length, the head-heavy one
will require you to apply a greater torque to keep its head up compared to the
head-light frame, and you will think the head-heavy frame feels heavier, even
though it has the same weight. (To keep the racquet head up, you will have
to apply a larger force with your fingers.) If this were the case, why would any-
28 Chapter 3

one want a head-heavy racquet? The bang you get out of a frame (the “power”)
depends on where the weight is located. As a general rule, a head-heavy frame
will give you more ball speed for the same swing speed, compared to a head-
light racquet of the same weight. Another correlated general rule says that the
power you get out of a racquet falls as the ball impact point to balance point
distance increases if you are translating the racquet or moving it in a straight
line (as you might in a punch volley as opposed to a groundstroke). In such a
case, if you hit the ball closer to the balance point, you will get more power
for a given racquet weight and head speed at the impact point. A head-heavy
racquet has its balance point closer to the center of the strung area, which is
where many people like to hit the ball, so it will produce more power.
However, racquets are usually swung about an axis, not translated in a straight
line, in which case the maximum power point moves away from the balance
point, as we will discuss in the swingweight section.

When you buy a racquet, the balance point is set. But a knowledgeable rac-
quet technician can apply lead tape to adjust the weight and balance of a rac-
quet to alter the racquet’s feel and peformance. Table 3.1 shows how much the
balance point will change by adding specific amounts of weight at specific dis-
tances from the existing balance point. The heavier the racquet, the less effect
per gram of added weight, since it is a smaller percentage of the total weight
of the racquet.

If the weight is added on the handle side of the balance point, the change is
negative (towards the handle). The table shows increments up to 5 grams. For
increments greater than five grams, simply add gram increments together. For
example, 10 grams (2 x 5 grams) at 12 inches for a 300 gram racquet adds 0.4
inches to the balance point (0.20 + 0.20). However, these calculations assume
all the weight will be added at the same spot, but lead tape will cover a few
inches, so you may have to do some averaging.

So far, all we have done is picked up the racquet and held it steady.
Racquet Speed It felt heavier or lighter, depending on where we held it. But weight
feels differently when we put it in motion with a swing. We’ve just
established that the effect of weight depends on how far it is from
the handle. The same holds true for the velocity of any point on the
racquet. Velocity varies with distance from a pivot point because a
racquet swing travels in a circular motion. (Swingweight is usually
calculated at 4 inches from the butt end. This is primarily because
the only commercial machines available to measure swingweight
swing the racquet from that location. See Chapter 5 for more on
measuring swingweight.) All points on an object that is translating
Figure 3.4
or traveling in a straight line are moving at the same speed. But if
The racquet tip travels farther (d,) than the object is rotating or traveling on a curved path, like a swinging
the throat (d,) in the same time. Speed racquet, then the mass points of the object far from the pivot are
depends on distance from pivot point (r,, moving faster than those closer to the pivot. Figure 3.4 makes this
r,). But mass at greater distances from
clear.
the pivot point is harder to accelerate.
Why? Because it travels a greater dis-
The tip of the racquet travels a greater distance than the throat or
tance in the same time, so it accelerates
fasterand’thé force necdeto pe larder the handle, but they all travel their respective distances in the same
 Weight, Balance & Swingweight 29

Table 3.1
Balance Customization
Amount balance will change in inches for racquet of specific weight by adding 1-5 gram weights in 1-inch increments from the balance point. For
increments greater than 5 grams, add results of separate increments together. Changes are negative if weight is added on handle side of balance.
Added Inches Racquet Weight in Grams
From
Grams 280 290 300
Balance
00
01
77

From Balance

a"

From Balance

3"

From Balance

4"

From Balance

5"

From Balance

6"

From Balance ~

via

From Balance ~

8"

From Balance

9"

From Balance ~

10"
From Balance |

1’
From Balance ©

ee
From Balance

13°
From Balance |

14"
From Balance

15"

From Balance |

16"
From Balance
ATAARWONATDAWON]HAARWNHHAMAWNAUMTAWNHAAAWONHTAAWNHAAWNHATARWONHATAWNHAAWNHARWNHATAHRWONHA
ARONA
30 Chapter 3

time. So, racquet speed depends on where on the racquet you are talking
about, and that velocity increases with the distance from the butt end of the
racquet. All points on the racquet obviously travel the same angle in the same
time. So all points on a racquet have the same angular velocity, no matter their
distance from the handle; whereas every point has a different linear velocity
depending on that same distance.

The feel of a swinging racquet is twice dependent on the average distances of

masses on the racquet from the handle — once for the effect of the mass with-
out motion (balance point) and once for the velocity of the mass in motion.
This is distance times distance, or distance squared. Swingweight is an objects
resistance to change in its rotation. Mass in circular motion is what swingweight
is about, so we end up with:

Swingweight = mass x distance x distance = md? (3.2)

This is swingweight for a point of mass at a given distance from its point of
rotation — we’re not up to a racquet yet. To equate this to an entire racquet,
let’s do the same exercise we did with balance point — consider the racquet
as an accumulation of small bits of mass. Each bit of mass has a swingweight
of md’, and the sum of all these individual swingweights is the total swing-
weight of the racquet. The units of swingweight, which are usually left out in
the popular tennis literature in favor of unitless numbers, are kilogram times
centimeters squared (Figure 3.5) or kilogram times meters squared in the
Figure 3.5 physics literature.

Depending on the distribution of mass, a racquet’s swing-

weight will be the sum of each bit of mass times the dis-
How Swingweight Is Determined tance squared to the pivot point at 4 inches on the han-
(Example: 324 gm, 27", even balance) dle. (Figure 3.5 is just an approximation because it is pre-
Note: Distance is determined from 4-inch (10.16 cm) mark
tending all the mass of each bit exists only at inch inter-
_ Weight per vals along the racquet and not spread out between each
betes, ae ole a inch.) Swingweight is ultimately the sum of every atom’s
{ 0.70 mass times its distance squared from the pivot point.

34
5
8Z ia
O79 The swin gweightight of of a racquet is a measure of f Ihow much
9 1.94 torque (twist) you must apply to the racquet handle to
i as get the racquet to swing. (You do not apply a force to the
42 4.95 handle of a racquet to cause it to swing, you apply a
1s b20 torque.) Swingweight can only be measured by an exper-
15 9137 iment where the racquet is moving. You cannot distin-
16 11.15 guish one swingweight from another by just holding a
us ee racquet in your hand. To determine a racquet’s swing-
19 17.42 weight you must swing the frame or put it into a machine
a AaCe that moves it. A racquet with a large swingweight will be
ae 35.08 difficult to swing and a racquet with a small swingweight
23 27.95 will swing easily. As a general rule, for a given racquet
BE Be weight, a head-heavy racquet will have a larger swing-
36 37 47 weight than a head-light frame of the same total length
27 40.95 and weight.
Swingweight = 335.84 kg-cm?
 Weight, Balance & Swingweight 31

While most people have a “feel” for weight and balance, few people have a
good feel for swingweight. In tests run with non-tennis players, the participants
could only distinguish racquets having one swingweight from racquets having
another swingweight when the difference was about 25 percent. When the
same tests were run with college level varsity players (who practice or play
about 2 hours every day) and teaching pros, they could distinguish frames that
differed by as little as 2 percent in swingweight.

A racquet with too large a swingweight will be sluggish, hard to swing, and
not feel maneuverable. When racquet manufacturers were contemplating pro-
ducing longer (28-, 29-inch length) racquets they realized that if all they did
was to add the extra inches to the butt end of their existing frames, the swing-
weights would have come out too high. Players would not have immediately
switched over to the longer racquets. A complete redesign of the frame was
needed to provide the extra length, yet have a feel and swing similar to the
racquet players were used to and liked playing with. That is one of the reasons
why sales of the extended length racquets did so well initially, and are still
doing well. The Rules of Tennis do not specify a minimum racquet length, but
do specify 29 inches as the maximum length.

Why would anyone not want a racquet with the smallest swingweight if it were
easier to swing and maneuver? Racquet manufacturers have tried very hard to
keep swingweights down, since they will swing much easier. However, there
is often a correlation between swingweight and the speed of the ball as it
leaves your racquet. In general, for a given swing speed, the racquet with the
larger swingweight may give you a bit more ball speed on your shots. Of
course, if the swingweight is too high, you may not be able to achieve that
given swing speed.

If you find that your present racquet seems to be sluggish when you swing it
and you cannot generate the head speed you want, the problem may be too
large a swingweight. Just because a racquet has been made lighter, that does
not guarantee that the swingweight has been reduced. It is swingweight, not
weight, that determines your swing speed. To find out if this is your problem,
try choking up on the handle just a bit when you swing. Since choking up
reduces the effective swingweight markedly, see if this solves your problem.

Just as you can alter the balance of a racquet, you can alter its swingweight by
adding lead tape. Table 3.2 demonstrates the effects on swingweight of adding
weights at different locations along the racquet. (The table is in units of kilo-
grams x centimeters squared. These are the units racquet technicians are most
familiar with and the units that the automated measuring machines use.) The
swingweight of a racquet depends very critically on where the racquet’s weight
is located and how far the weight is from your hand. Adding a few grams of
lead tape to the tip of your racquet will increase the swingweight much more
than adding the same weight at the throat of your frame. Adding extra weight
to the handle region will hardly change the swingweight at all, but it will
change the racquet’s total weight and balance point.

The real value of the chart is for customizing. Each number in each column sig-
nifies the amount the swingweight will be changed if you add that much
weight at that distance.
32 Chapter 3

Table 3.2

Swingweight Customization
Columns are the amount swingweight changes (in kg*cm*) by adding the specified
Inches weight at the top of each column at distances from the butt equal to the far left
From column. Add columns together for larger weights (measured 4" from the butt).
Butt to Cm
Add From Amount of Weight Added (Grams)

us ae 1 3 4 5 6 7 8 9 10
0 0.00 0.10 0.21 0.31 0.41 0.52 0.62 0.72 0.83 0.93 1.03
1 2.54 0.06 0.12 0.17 0.23 0.29 0.35 0.41 0.46 0.52 0.58
2 5.08 0.03 0.05 0.08 0.10 0.13 0.15 0.18 0.21 0.23 0.26
3 7.62 0.01 0.01 0.02 0.03 0.03 0.04 0.05 0.05 0.06 0.06
4 10.16 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
5 12.70 0.01 0.01 0.02 0.03 0.03 0.04 0.05 0.05 0.06 0.06
6 15.24 0.03 0.05 0.08 0.10 0.13 0.15 0.18 0.21 0.23 0.26
U 17.78 0.06 0.12 0.17 0.23 0.29 0.35 0.41 0.46 0.52 0.58
8 20.32 0.10 0.21 0.31 0.41 0.52 0.62 0.72 0.83 0.93 1.03
9 22.86 0.16 0.32 0.48 0.65 0.81 0.97 1.13 1.29 1.45 1.61
10 25.40 0.23 0.46 0.70 0.93 1.16 1.39 1.63 1.86 2.09 2.32
11 27.94 0.32 0.63 0.95 1.26 1.58 1.90 2.21 2.53 2.85 3.16
12 30.48 0.41 0.83 1.24 1.65 2.06 2.48 2.89 3.30 3.72 4.13
13 33.02 0.52 1.05 1£57/ 2.09 2.61 3.14 3.66 4.18 4.70 5.23
14 35.56 0.65 1.29 1.94 2.58 3.23 3.87 4.52 5.16 5.81 6.45
is 38.10 0.78 1.56 2.34 3.12 3.90 4.68 5.46 6.25 7.03 7.81
16 40.64 0.93 1.86 2.79 3.72 4.65 Sidi 6.50 7.43 8.36 9:29
17 43.18 1.09 2.18 3.27 4.36 5.45 6.54 7.63 8.72 9.81 10.90
18 45.72 1.26 2.53 3.79 5.06 6.32 7.59 8.85 10.12 11.38 12.65
19 48.26 1.45 2.90 4.35 5.81 7.26 8.71 10.16 11.61 13.06 14.52
20 50.80 1.65 3.30 4.95 6.61 8.26 9.91 11.56 13.21 14.86 16.52
21 53.34 1.86 3.73 5.59 7.46 9.32 11.19 13.05 14.92 16.78 18.65
22 55.88 2.09 4.18 6.27 8.36 10.45 12.54 14.63 16.72 18.81 20.90
23 58.42 2.33 4.66 6.99 9:32 11.65 13.97 16.30 18.63 20.96 23.29
24 60.96 258 5.16 7.74 10.32 12.90 15.48 18.06 20.65 23.23 25.81
25 63.50 2.85 5.69 8.54 11.38 14.23 17.07 19.92 22.76 25.61 28.45
26 66.04 3.12 6.25 9.37 12.49 15.61 18.74 21.86 24.98 28.10 31.23
27 68.58 3.41 6.83 10.24 13.65 17.06 20.48 23.89 27.30 30.72 34.13
28 Talet2 3.72 7.43 nee 14.86 18.58 22.30 26.01 29.73 33.45 37.16
29 73.66 4.03 8.06 12.10 16.13 20.16 24.19 28.23 32.26 36.29 40.32
30 76.20 4.36 8.72 13.08 17.45 21.81 26.17 30.53 34.89 39.25 43.61

For example, if you add 10 grams (= 0.010 kilograms) at 23 inches from the
butt (23 inches from the butt is 23-4 = 19 inches = 48.26 centimeters from the
pivot point), you will add approximately 23.29 swingweight units to the exist-
ing swingweight (0.010 x 48.26 x 48.26 = 23.29). At 27 inches, it would be
34.13. And if you add 15 grams at 27 inches, the additional swingweight is
51.19 (the 5 and 10 gram columns added together, or the 8 and 7, etc.). Also,
notice the effect of adding weight vs. adding distance. Doubling the added
weight doubles the increase in the swingweight. Doubling the distance from
the pivot point (subtract 4 inches from whatever row you are in and double
the result to locate the point twice the distance from the pivot point) while
keeping the weight the same will add four times the additional swingweight.
Where you add weight makes much more difference than how much you add.
And notice too, there is no such thing as lowering swingweight, even though
you can lower the balance point.

Figure 3.6 is a weight, balance, and swingweight map. It plots all performa
nce
racquets of a recent year (2000) by weight and swingweight with balance
num-
 Weight, Balance & Swingweight 33

Figure 3.6

Weight, Balance & Swingweight Map

Lightweight Weight (gm) Heavyweight
200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390

Set aS eee
nNo a

Swingweight
Low yUBiambuims
mo7q

Swingweight
(kg*cm2)

Swingweight
High ybiyq
wbiembuims
4 Center of Swingweight/Weight Universe (includes all performance frames — i.e., over $100 suggested retail price — in year 2000.
Approximately half the racquets are to each side and half above and below the center axis cross.

Balance point (% of length): 1 Super Head-light < 46% 2 Head-light = 47-49% 3 Even = 50%
4 Head-heavy = 51-53% 5 Super-Head-heavy = 54%

ber coded as follows: 1 = super head-light (balance at less than or equai to 46%
of length); 2 = head-light (balance equal to 47-49% of length), 3 = even (bal-
ance at 50% of length); 4 = head-heavy (balance equal to 51-53% of length); 5
= super head-heavy (balance greater than or equal to 54% of length). The bal-
ance is figured as the percentage of racquet length from the handle. That
allows comparing head-light and head-heavy racquets of all lengths on the
same map.

The map reveals slightly different nuances to the relationship between these
three components of weight distribution. Some are obvious, some surprising,
and all are interesting. The map is designed to gain a better overall under-
standing of how these variables interact and affect racquet performance.

What immediately stands out is the range of racquets represented in each row
and column. For example, in the weight column of 290 gram racquets, there
34 Chapter 3

are swingweights ranging from 290 to over 370. Similarly, in

Swingweight and
the 315 swingweight row, weights vary from 230 to 370 grams.
Customizing Length? That’s a 140 gram difference, yet they both swing equally as
easy. How can that be?
ou have been playing with a standard length
Y (27-inch) racquet for your entire life, and you
wonder what it would be like to move up to a I's not. Swingweight measures maneuverability along the arc
longer (28- or 29-inch) frame. You have read all the of your swing and weight measures maneuverability if you
ads, read the reviews, and talked to (and played were to push at the balance point, which you never do. Every
against) people who have the longer frames. You swing is a combination of the swing weight (second moment)
have even hit once for five minutes with a longer and the weight times balance (first moment) though rotation-
racquet. However, you are reasonably happy with al maneuverability is probably the most noticeable. So in a
your present racquet. It cost a bundle 3 years ago, given row with equal swingweights, racquets to the left have
and it still plays great. Could another inch or two less mass or inertia but tend to be head-heavy. (The sum of
take your tennis up to the next level? You would like swingweight plus weight is less for these racquets.) Likewise,
to know. Do you really want to go out and buy all
all racquets in a column have the same weight, but those to
new frames? So when you see an ad telling you of
the top offer less resistance in a circular swing and are thus
a company that will add 2 inches to your present
racquet for a reasonable fee, what should you do?
more maneuverable overall. If two racquets have the same
DON’T DO IT! weight, the one with the higher swingweight will feel heavier.
If two racquets have the same swingweight, the one that
You will usually end up with a racquet that you will
weighs more will still feel heavier.
have difficulty playing with, and as a result, you
won't like it. The modification could turn the careful-
ly balanced, well-tuned frame that you presently It’s also interesting what you can’t say about racquets by look-
play with, into a club that maybe only a caveman ing at the chart. For instance, you can’t say, “For a group of
would enjoy playing with. racquets, as weight goes down, swingweight goes down (or
up).” Likewise, as swingweight goes up and down, you can’t
There is a parameter called swingweight that deter-
mines how difficult it is to swing a racquet and
say anything definite about weights — they still stretch across
maneuver it from one location to another. The larg- the entire row. (Note: this does not apply to an individual rac-
er the swingweight, the more torque you must apply quet. If you add weight to the head, swingweight will go up.)
to get the racquet to do what you want it to do. It appears that a racquet can be made to fit just about any box
Swingweight can be measured using a Babolat on the map.
Racquet Diagnostic Center (RDC), and almost all
modern racquets have a swingweight in the 300 to
360 range (in units of kilogram x centimeters
squared). If you take a typical 27-inch racquet that
has a swingweight of 330 and a weight of 320 grams
In general, for this universe of racquets, as weight goes down,
(11.3 ounces), adding 2 inches to the handle will balance goes up, and as weight goes up, balance goes down.
increase the swingweight to at least 425. This is Notice how the number codings drift from super head-light on
greater than any other modern racquet on the mar- the right to super-head-heavy on the left.
ket, except the 32-inch long Gamma Big Bubba. The
added 2 inches will increase the swingweight by The head of a racquet is under tremendous stress because of
about 30%, which means it will be 30% more slug- the tension in the strings and the constant pounding due to
gish than the racquet you started with. The hitting the ball. With less weight (material), you have to use
increased benefits of the longer frame may be can-
what’s available in the head to keep it strong. Still, it’s tempt-
celled by the extra swingweight you will be forced to
drag around on your shots.
ing to think that if two racquets have the same swingweight,
a heavier racquet will have at least as much mass in the head
However, if you buy a typical 28- or 29-inch frame as the much lighter head heavy racquets. For example, look
that has been designed from scratch to handle the
at the 315 swingweight row. There’s almost a 140 gram dif-
extra length, you probably will be pleased with the
ference between the racquet on the left and the racquet on the
way it feels and the way it plays. The manufacturers
right. You’d think that all the extra weight would spread out
have cleverly redesigned their longer racquets and
so the heavier racquet would have more material everywhere,
redistributed the weight so that the swingweights
come out very similar to the values found in the even though the balance is toward the handle, not the head.
average 27-inch frame. This will give you all the But if they both had the same mass located at the same dis-
advantages of a longer racquet (since it is longer), tance in the head, plus the heavier racquet had more mass
yet when you swing it, the racquet will feel like the everywhere else, then it would have a higher swingweight
standard 27-inch frame that you are used to. than the lighter but higher balanced racquet. It follows that
 Weight, Balance & Swingweight 35

the lighter racquet has more absolute (not just relative) mass in the head. This
makes perfect sense, it’s just difficult to imagine. (Note: you can think of
bizarre combinations of weight distribution that violate this statement, but at
the end of the day, we still have to have a racquet that does its job and remains
intact.)

Notes

1. Mass and weight are different things but we often treat them in everyday lan-
guage as being the same. For example if we hang a 60 Ib weight on the end
of a string then we say the tension in the string is 60 lb. In doing so we have
made several technical errors. Suppose we take our 60 lb weight to the moon.
Will the tension in the string still be 60 Ib? And what if we take a racquet strung
at 60 lb to the moon? Will the strings be loose and floppy? Technically, the 60
Ib object we hung on the string has a mass of 60 lb, not a weight of 60 lb.
Technically, weight is a measure of force. When we take our 60 Ib object to
the moon it still has a mass of 60 Ib but the moon doesn't pull on it as strong-
ly as the earth so the tension in the string is less than it is on earth and the
weight of the object is less. Put the object on a set of scales and it will read
less on the moon than on earth. The scales measure the weight of the object
or the force on the scales rather than the mass on the scales.

Back on earth, we often say that mass and weight are the same, but then we
get into trouble when doing calculations since mass and force are different
things. For that reason, physics calculations are usually done in the SI system
of units where mass is measured in kilogram and force or weight is measured
in Newtons. Then the weight of an object of mass M (in kg) is Mg (in Newtons)
where g on Earth is 9.8 but g on the Moon is only 1.67 m/ s’. The tension in a
string is a measure of force on the string and is measured in Newton. The string
tension won't change when you take your pre-strung racquet to the moon, and
neither will the mass of your racquet, but the racquet will feel a lot lighter. The
rules of tennis will need to be changed drastically when we start living and
playing on the moon.
=_ —~ a fan aol! |
yah Os i
a Male mild
me ee co
a4 us
wen De Le ug
* vet Piro ina bs
a

- ne
=

ae ran ad. 77=

aeate
a) _
7 ain
oer
ae

of | 1ALa Tae 5Si]

, ?.

eo)| < | ¥
ae a*am

y dant 7
= (@ Me « } ain < a
- 7 Vibe
a _
 Twistweight
and Recoilweight

Jelena Dokic, Yugoslavia

By Crawford Lindsey and Howard Brody

aking a lightweight, stable, maneuverable, and powerful racquet is not easy. Some of these features
benefit from more weight, some from less. So, adding, subtracting, and moving weight around is what
’ ™ racquet design is all about. The history of major racquet innovations is, to a large extent, a story of
weight distribution and doing more with less.

The introduction of metals, and then graphite, were billed as materials stories. But what made these materials
so important (especially graphite) was the ability to eliminate and relocate weight because of these materials’
higher stiffness and strength-to-weight ratios. You could now build the same structure with greater stiffness and
strength and use less material weight to do it.

The first oversize racquet (produced commercially by Prince in the 1970s, but handmade experiments had been
around for 50 years) was also indirectly about weight. It was marketed as a “headsize” story, but not only did
the larger head’s longer strings mean more power in the middle of the sweetspot, but a wider head also meant
more power on off-center hits due to the stabilizing influence of weight at greater distances from the center line
of the racquet.
38 Chapter 4

Wilson’s Hammer system was also about weight. It was the most radical elim-
ination and redistribution of weight ever, removing it from the handle and
resulting in radical head heaviness.

Even the widebody (racquet frames with a thick cross-section perpendicular to

the stringbed) revolution could be interpreted as weight related because it was
graphite’s strength-to-weight and stiffness-to-weight ratios that made the radi-
cal shape of widebody construction structurally possible.

And, of course, the era of super lightweight frames of the present is, by defi-
nition, about weight.

lity ByWeight Polarizati

In the last chapter, we discussed how the amount and location of weight affects
maneuverability and the feel of the racquet through the air. Amount and loca-
tion of weight also becomes important at impact. Here we discuss how weight
affects stability. Recently, Prince started polarizing the weight at 2 and 10
o'clock and in the butt of the handle. This strategy stabilizes the racquet in two
directions — along the width and length of the racquet (Figure 4.1). True, the
story sounds familiar — perimeter weighting and butt weighted racquets are

Three Kinds of Swingweight and Their Relation To Stability

Swingweight Twistweight Recoilweight

(handle axis = swing axis) (length axis = twist axis) (balance point axis = recoil axis)

Figure 4.1
Swinging and impacting a ball torques the racquet in three ways and each torque rotates the racquet around a different
axis. The resistance to rotation around any axis is the “swingweight” around that axis.

Swingweight about the handle axis (swing axis) is called just that — “swingweight.” This measures the distribution of weight
along the length of the racquet, which in turn determines maneuverability and stability around that axis when you swing.
Swingweight around ou axis drawn from butt to head (twist axis) can be called “twistweight,” and it measures
weight dis-
tribution from side to side and its effect on maneuverability and Stability to twisting by you or the ball.

Swingweight around an axis parallel to the racquet face and through the balance point (recoil axis) can
be coined “recoil-
weight.” This measures weight distribution from tip to butt and the resulting resistance to recoil.

Parallel Axis Theorem: If you know the moment of inertia about any given axis, then you can calculate
it about any other
parallel axis by using the parallel axis theorem (see Note 1)
 Racquet Stability 39

not new concepts, but they have never been integrated into, nor marketed as,
a philosophy of racquet design. The emphasis on head to butt polarized
weighting is a new twist.

What, How, W en, Where and Why ofPolar

1. Why stabilize? Lightweight racquets require extra stabilization because they
have less inertia, get pushed around more, and incur more shock. Impact caus-
es a racquet’s motion to change primarily in two directions. First, there is back-
wards recoil, like a door on a hinge (on lengthwise off-center hits). Second,
the racquet twists like a top in your hand (on widthwise off-center hits). The
amount of recoil and twist represents wasted energy.

2. Weight stabilizes. More weight stabilizes more, and less weight, less. But
the greater the distance the weight is from its center of rotation, the more it sta-
bilizes. So, if you double the weight at a certain location, you double its effect
on stability. If instead, the weight’s distance from the racquet’s swing pivot
point is doubled, the effect on stability is quadrupled. Likewise, adding one-
fourth the weight at twice the distance has the same effect as all the weight at
the original location. In short, weight stabilizes and polarized weight stabilizes
more with less.

3. What does stabilization do? Stabilization reduces recoil and twist and their
effects. Maximum stability against both kinds of torque is achieved by placing
weight at 3 and 9 o’clock for twist resistance, and 12 o’clock and the butt for
recoil resistance. These are the greatest polarized distances on the racquet. This
is why bigger head sizes and longer racquets are more stable, all else being
equal.

Twistweight: The Stability of a Racquet Against Twisting

Howard Brody

it a ball off-center, and you know what happens. The racquet tries to twist in your hand. The ball comes off at the wrong
“ angle, and your shot does not have the speed that you hoped it would have. Chalk up another unforced error on your
stat sheet.

You know that if you can only get the ball to hit closer to the axis of the racquet, the frame will not twist as much. Then your
shot will land in the court, usually closer to the location where you want it to go. Making sure that the ball hits near to the cen-
tral axis of the frame requires good hand-eye coordination, concentration, and practice — lots of practice. Is there anything else
you can do to keep the racquet from twisting so much on miss-hits? Yes, and here science comes to your rescue.

You can use a racquet with a wider head. The width of the head determines something called the polar or roll moment of iner-
tia (physics for twistweight). Wider racquets (as distinguished from “widebody” racquets which have a thicker frame) have a larg-
er polar moment and provide more stability against twisting. One great virtue of the original Prince oversize racquet was that it
had almost a 40% greater stability against twisting than the old small head racquet that was in use at that time. To measure a
racquet’s stability against twisting (the moment of inertia about its long axis) is a rather technical exercise that can be performed
with something called a torsion pendulum or by turning the frame into a pendulum with the handle horizontal (see Chapter 5).
The definition of the moment of inertia is the sum of every little mass piece in the racquet multiplied by that mass piece’s dis-
tance from the axis of rotation squared. However, the racquet weight multiplied by its head width (or diameter) squared will give
you a very good approximation of the stability factor. This is shown in Figure 4.2 for a number of racquets of different mass and
width. As a specific example, a 0.3 kg (300 gram) frame that has a 25 cm inches wide head will have a stability factor of 187.5
(weight x width x width = 0.3 x 25 x 25 = 187.5). Selecting a racquet with the same weight, but a head just three centimeters
wider (28 cm instead of 25 cm), will produce a stability factor of 235.2 (0.3 x 28 x 28 = 235.2), which is an increase of 25%.

There are corrections to this due to the exact weight and shape of the head. A square head (sides parallel) will have slightly
more stability than an oval or egg shaped head, which in turn will have somewhat more stability than a triangular (“yield sign”
40 Chapter 4

racquet of
shaped) head, for the same total head width. A head-heavy frame will be a bit more stable than an evenly balanced
exactly the same dimensions and total weight.

in rebound ball
A racquet with an increased stability factor will have a more uniform response across its face (less variation
speed for slight miss-hits) as well as less twisting on the miss-hits. This may be thought of as having an enlarged sweetspot
and being a more forgiving racquet.

The reduction in the twisting may have a very beneficial effect in reducing the probability of developing the dreaded tennis
elbow. When you hit a ball well off-axis, the racquet will twist regardless of how tightly you are gripping the handle. But the rac-
quet rarely twists out of your hand. Your hand and your forearm twist along with the racquet, but your upper arm does not rotate
that way. That is what causes the problem in the elbow where your forearm and upper arm meet — one rotates, the other does
not. A racquet that is more stable and twists less on off-axis impacts will be kinder to your arm and elbow.

Measured Polar Moment vs From these arguments, it seems that everyone

Mass x Width? should play with a racquet that is constructed at
the very limit of width that the ITF allows (31.75
cm). Why don’t they? There is the matter of
maneuverability. More stability usually leads to
less maneuverability. However, this is a particu-
lar type of twisting maneuverability, and it may
not affect all players the same way. If you hit
very flat shots, or if you chop and slice a great
deal, the larger stability factor about the long
axis of your racquet may not be a major detri-
i sa ment. If, on the other hand, you use a great deal
r of arm pronation (rotating about the long axis of
42 [ your arm) on your serve and on your ground-
(kg-cm?)
Moment
Measured
[ strokes, this extra frame stability will slow down
AL ail | eal pol ea ed oe el ee your racquet head and consequently take some
10 - ho Lit of the snap out of your shots.
0.02 0.022 0.024 0.026 0.028 0.03
Mass x Width? (kg-m?) You have to make the decision on your racquet
depending on whether you have an aggressive,
arm pronating style of play, or a more conser-
Figure 4.2 Measured Polar Moment of Inertia ver-
vative, less error prone stroking. If your arm is very strong, and
sus Mass x Width x Width.
the extra stability factor does not slow down your stroke, then
The measured polar or roll moment of inertia for a the answer is simple. For the rest of us, a decision (or compro-
number of tennis racquets with different masses and mise) must be made.
widths is plotted against the mass of each racquet
multiplied by that racquets head width squared. Remember: A racquet with a wider head will be more stable
Note that the data point for the different racquets against twisting, give a more uniform response across the head,
seems to lie along a smooth curve. and provide more control. It also will be somewhat less maneu-
verable on groundstrokes and it will not be as user-friendly on
your serve and volley game as a narrow head racquet.

4. Stability adds power. Recoil and twist steal energy and reduce power.

5. Stability adds comfort. Energy siphoned off by twist and recoil goes into
your hand and arm, not the ball.

6. Stability adds control. Twisting and recoil reduce control because the rac-
quet face rebounds at uncontrollable angles from which the ball also rebounds
in undesired directions.

7. Stability expands the power zone. Sweetspot size is an indication of the

degree of stabilization. The most widely used definition of sweetspot is the area
of the stringbed that produces an outgoing ball velocity that is at least 45 per-
cent (40-50 percent, depending on whom you ask) of the incoming velocity.
 Racquet Stability 41

8. Torque robs energy. The further off center you hit, the more the racquet
will twist, recoil or both. The more the racquet is stabilized, the further off-cen-
ter you can hit with less energy loss and, therefore, more power, comfort, and
control. The sweetspot is therefore larger. Unless you hit dead center every
time, the size and location of the sweetspot area is more important than the
maximum “sweetness” at the center.

9. Stability raises the sweetspot. If stability follows weight and sweetspot

follows stability, then sweetspot follows weight also. Weight toward the rac-
quet tip stabilizes and drags the sweetspot to the upper half of the stringbed
where most balls are hit.

10. Its more difficult to polarize weight and stabilize a light racquet.
That’s because there is less weight to distribute, and in the extreme case, all
the weight goes into the bare-bones structure just to keep the racquet from
breaking. There is little leftover to move around for additional benefits. The
lighter the racquet, the more it needs polarization, but the more difficult it is
to accomplish and the less you have to do it with. Thus, at a certain point, a
lighter racquet can never compete with a much heavier one in terms of stabil-
ity.

The lightweight race (culminating, so far, with the lightest racquets in the
neighborhood of 200 grams strung) has taught manufacturers how to remove
weight where it isn’t needed in order to build the lightest possible foundation
structure. Then, some of that weight can be added back to produce added ben-
efits besides simple structural integrity. Getting the most out of every fraction
of a gram is important.

The idea of weight polarization is that recoil does not occur around the han-
dle axis. Ball impact causes the racquet to rotate around the balance point, not
the handle. This is illustrated by looking at two extremes. If a ball hits an
unheld racquet, it flies off rotating around its balance point. If a steel rod were
inserted through the handle and cemented into the floor, the racquet would
instead rotate around the rod (as in Figure 4.1). You may be tempted to say a
player holding a racquet would be more like the steel rod case, but you’d be
wrong. Studies have shown that a person is not strong enough to stop rotation
around the balance point during the 5 ms of impact. The player’s hand itself
recoils at the wrist because, it too, is rotating around the balance point. The
player slows the rotation about the balance point, but does not relocate it to
the handle like a secure steel rod would until after the impact is over.

Thus, to maximize resistance against recoil, weight must be polarized around

the balance point. So, putting weight in the butt does not make it substantial-
ly more difficult to swing around the handle axis because it is located close to
the axis. But that weight does significantly increase the swingweight for swing-
ing around the balance point (recoilweight) because it is at a much greater dis-
tance to the balance point.

Swingweight around any axis (whether it be the swing, twist, or recoil axis)
depends on the amount and distance of the weight from the axis. Thus, all
things being equal, a longer, wider or a heavier racquet will be more stable.
For these reasons, racquets being compared must be the same length and with-
42 Chapter 4

in a certain weight range. That’s the only way to make an “apples-to-apples”

comparison. To that end, the stability table compares 28-inch racquets below
300 grams.

The table ranks recoil stabiilty in two ways. The first is “polarization index.”
This measures the racquet’s degree of weight polarization. It is calculated as
the recoilweight divided by the racquet weight. This tells you “pound-for-
pound” how polarized the weight is. It is an indication of what is done with
the weight available. The top half of the table ranks racquets by polarization.

The bottom half of the table ranks the same racquets in order of recoilweight
(the higher the number, the more stable to recoil). For a given weight range,
the recoilweight depends on the degree of polarization. The recoilweight is the
feature the player is looking for (at a certain weight or swingweight). The
polarization index helps explain how the racquet’s weight and distribution
combine to determine the recoilweight rankings.

Notes

1. Parallel axis theorem: If you know the moment of inertia about any given
axis, then you can calculate it about any other parallel axis by using the par-
allel axis theorem. For example, if you know the moment of inertia about the
balance point, to determine the moment of inertia about the butt, use the for-
mula Toe = De funds
= 2
.*. where dour = distance from butt to cm, m = mass of
== ; =

racquet, and I.cm. = moment of inertia about balance point (recoilweight). If you
know the swingweight about an axis 4 inches from the butt and want to know
ieee
the recoilweight ‘
the formula is I,,, == I,,, - md,,,”,2 where d.,, == distance
Ak :
from
4-inch axis to cm, m = mass of racquet, I. = moment of inertia about balance
point, and I,,, = moment of inertia about the 4-inch axis. If you know the
swingweight about the 4-inch axis and want to know the swingweight about
the butt, the formula is Da
. v2 2 2: aes .
ec es 0 wine d.,, = distance from
4-inch axis to cm, dbutt = distance from butt to cm, m = mass of racquet, I,
= moment of inertia about 4-inch axis, and Thue = Moment of inertia about the
butt end.
 Racquet Stability 43
Table 4.1

Stability 1998: 28” Racquets Under 300 Grams

Polarization Index Ranking = Recoilweight/Weight (measures the degree of weight polarization). If two racquets are close in
weight, the one with the highest polarization will be more stable. If two racquets have the same, or very close polarization index-
es, the heavier one will be more stable. (For example, see the four racquets with a 0.51 index.)

Recoil-
Polarization Weight Weight Swingwt Headsize Bal. COP*
Mfg. Racquet Index (kg:cm?) (gm) (kg-cm?) (in.?) (cm) (in.)
Prince Ring 0.56 150 267 339 125 36.75 20.50
Prince Rip 0.55 146 268 332 ks 36.50 20.25
Pro Kennex Kinetic Pro Ti 20G Reach ORS5 156 279 336 110 CIS) TAS) 20H\S
Prince ThunderStorm LongBody OS 0.54 142 263 328 120 3615 20.25
Dunlop Max Superlong +1.00 OS 0.51 149 291 347 105 36.25 19.83
Gamma Diamond Fiber C-3.0 OS 0.51 WeZ 220 292 116 38.75 20.44
Head ii S7 0.51 124 243 316 124 38.25 20.32
Prince ThunderLite LongBody MP 0.51 149 292 326 95 34.75 19.48
Prince Thunder Extreme Titanium OS 0.51 147 289 322 as 34.75 19.45
Dunlop Revelation Lite +1.00 0.50 129 260 S13 100 36.75 19.79
Prince Thunder Extreme Titanium MP 0.50 143 287 310 100 34.25 19.23
Wilson Hammer Titanium 3.0 Stretch 0.50 130 259 Sei, 118 SMfoi/S 20.09
Gamma Diamond Fiber C-3.0 MP 0.49 110 222 288 105 38.50 20.21
Wilson Hyper Sledge Hammer 2.0 St. 115 0.48 WiZ 236 312 WS 39.25 20.23
Yonex Super RQ Ti-600 Long 0.48 136 286 331 110 36.25 19.45
Prince Thunder Superlite Titanium OS 0.47 101 213 266 11S 38.00 19.87
Prince ThunderLite LongBody OS 0.47 139 296 325 110 S525 19.11
Wilson Hammer 6.2 95 Stretch 0.47 13a 278 346 95 38.00 19.83
Yonex Super RQ Ti-800 Long 0.47 111 238 316 110 39.50 20.20
Gosen Twin Pro 0.46 122 264 342 120 39.00 20.04
Wilson Sledge Hammer 3.8 Stretch OS PH 0.46 125 270 350 112 39.00 20.04
Wilson Hyper Hammer 2.3 Stretch SOS 0.46 iz 243 318 120 39.25 20.09
Wilson Hyper Sledge Hammer 2.0 St. 125 0.45 115 254 326 25 39.00 ORO
Wilson Hammer 6.2 110 Stretch 0.45 128 285 349 110 38.00 19.62
Yonex Super RQ Ti-900 Long 0.45 ls 255 Soo 120 39.50 20.07
Yonex Super RQ Ti-450 Long 0.45 128 283 332 110 37.00 19.38

Recoil Stability Ranking = Recoilweight (swingweight taken at the balance point). Within a given weight range, high polariza-
tion can make a racquet more stable than heavier less polarized ones. However, after a certain point, light racquets (polarized
weighting or not) are less stable than heavier ones.

Pro Kennex Kinetic Pro Ti 20G Reach 0.55 153 279 336 110 SORT 20.13
Prince Ring 0.56 150 267 339 125 36.75 20.50
Dunlop Max Superlong +1.00 OS 0.51 149 291 347 105 36.25 19.83
Prince ThunderLite LongBody MP 0.51 149 292 326 95 34.75 19.48
Prince Thunder Extreme Titanium OS 0.51 147 289 322 als) 34.75 19.45
Prince Rip 0.55 146 268 332 US: 36.5 20:25
Prince Thunder Extreme Titanium MP 0.50 143 287 310 100 34.25 19.23
Prince ThunderStorm LongBody OS 0.54 142 263 328 120 36.75 20.25
Prince ThunderLite LongBody OS 0.47 139 296 325 110 SOrZo 19.11
Yonex Super RQ Ti-600 Long 0.48 136 286 331 110 36.25 19.45
Wilson Hammer 6.2 95 Stretch 0.47 131 278 346 95 38.00 19.83
Wilson Hammer Titanium 3.0 Stretch 0.50 130 259 327 118 SW ATAS) 20.09
Dunlop Revelation Lite +1.00 0.50 129 260 313 100 36.75 19.79
Wilson Hammer 6.2 110 Stretch 0.45 128 285 349 110 38.00 19.62
Yonex Super RQ Ti-450 Long 0.45 128 283 332 110 37.00 19.38
Wilson Sledge Hammer 3.8 Stretch OS PH 0.46 25 270 350 We 39.00 20.04
Head TLS 0.51 124 243 316 124 38.25 20.32
Gosen Twin Pro 0.46 122 264 342 120 39.00 20.04
Wilson Hyper Sledge Hammer 2.0 Stretch 125 0.45 WS 254 326 125 39.00 19.91
Yonex Super RQ Ti-900 Long 0.45 115 255 335 120 39.50 20.07
Gamma Diamond Fiber C-3.0 OS 0.51 112 220 292 116 38.75 20.44
Wilson Hyper Hammer 2.3 Stretch SOS 0.46 2 243 318 120 39.25 20.09
Wilson Hyper Sledge Hammer 2.0 Stretch 115 0.48 112 236 312 Ws 39.25 20.23
Yonex Super RQ Ti-800 Long 0.47 da 238 316 110 39.50 20.20
Gamma Diamond Fiber C-3.0 MP 0.49 110 222 288 105 38.50 20.21
Prince Thunder Superlite Titanium OS 0.47 101 213 266 Wis 38.00 19.87

*COP is distance from butt end. It is the impact location that causes a “no shock” point at the butt end of the handle during the
 &»-

Daas

Lea pe
o>
a a
q 7 x)

ng >
=
>
 Do-It-Yourself Swingweight,
Twistweight and Recoilweight

By Rod Cross and Crawford Lindsey

he weight of a racquet can be measured by placing it on a set of scales. The swingweight is different
and can be measured only by swinging the racquet. This is typically done using a commercial swing-
™ weight machine like the Babolat RDC or Alpha Accuswing. But not everyone wants to invest in these.
Hence, this chapter offers do-it-yourself techniques for determining the different kinds of swingweights (swing-
weight, twistweight, and recoilweight). A swingweight is different from weight as we usually know it. The resist-
ance of an object to motion in a straight line is called its inertia and is measured by its mass (or its weight). The
resistance to rotation is called rotational inertia and is measured by its moment of inertia with respect to a cer-
tain axis of rotation. Depending on which axis is chosen (perpendicular to the handle or balance point, or along
the racquet’s entire length), we call the moment of inertia the “swingweight,” “recoilweight” or “twistweight”
respectivly.

Ideally, one should swing the racquet about the actual axis of rotation of interest (near the end of the handle),
but it is not easy to mount a racquet in this way. It is much easier to mount the racquet as a pendulum so that
it rotates about an axis in the string plane. The swingweight about a parallel axis in the handle can then be cal-
culated using the formula on the next page. (Note: The industry standard is to publish swingweights measured
4 inches from the butt end on the handle. This has become standard since that is where the commercial
machines calculate the swingweight. However, using the parallel axis formula, you can calculate the swingweight
about any parallel axis.)
46 Chapter 5

As shown in Figure 5.1, a racquet can be supported vertically by resting

one of the cross strings on two horizontal prongs about 6 inches apart.
A suitable arrangement is to use two rods (about 1/4-inch diameter)
placed on a table with weights on top so that each rod protrudes about
2 inches over the edge of the table. Pushing the handle gently allows the
racquet to swing back and forth smoothly as a pendulum and it will keep
swinging for several minutes if the rods do not not rub against any of
the main strings.

If the racquet is supported on the bottom cross string, it will take about
3 seconds to complete one cycle of oscillation. This is quite a slow oscil-
lation and the time can be measured accurately with a stopwatch. The
racquet swings faster if it is supported on the top cross string or on a
string near the middle of the stringbed. In that case, it is best to meas-
ure the time for 10 oscillations and then divide by 10 to get the time (T)
for each oscillation. An interesting property of a pendulum is that the
time for one oscillation does not depend on the size of the swing, pro-
vided the size is not too large. In practice, this means that you should
Figure 5.1 keep the swing angle less than about 15 degrees each side of the verti-
cal.

Swingweight Setup The formula for the swingweight is a bit complicated since the axis of
rotation used to measure the swing time is not the same as the axis nor-
meeteralsts NS I} 9)SHUTING, mally used to swing the racquet (Figure 5.2 and 5.3). The formula is:

Swingweight = ry cL Lee (5.1)

P 40.28 1000 1000

in kg-cm* where,
AS Balance
M = the mass of the racquet in grams.

T = time in seconds for one oscillation.

¥.
BP P = distance in cm from the balance point to the rotation axis (the bot-
tom edge of the cross string used to support the racquet).

Y = BP - 10.16 = the distance in cm from the balance point to a point 4

inches (10.16 cm) from the butt end of the handle.
Figure 5.2
BP = the distance in cm from the balance point to the
butt end of the handle.
Swingweight Axis
The above formula gives the swingweight in units of
kg-cm? for a racquet that is swung about an axis 4
inches from the butt end of the handle. This is the
standard way of specifying the swingweight, or of
comparing one racquet with another, even though the
actual axis of rotation in normal use is beyond the end
of the handle. A distance of 4 inches is used since rac-
quets are normally positioned in swingweight
Figure 5.3 machines so that the actual axis of rotation is 4 inch-
es from the butt end.
 Measuring Swingweights 47

Example

Measurements were made using a racquet with M = 291 gm, BP = 35.7 cm and
Y = 35.7 - 10.16 = 25.54 cm. The time for ten complete cycles of oscillation
back and forth on the bottom cross string was 32.72 sec for the first measure-
ment and then 32.90, 32.72 and 32.78 sec for the next three measurements. The
average time T for one cycle of oscillation was T = 3.278 sec. The distance P
from the balance point to the point of support (the bottom edge of the bottom
string) was 1.3 cm. These figures give

Swingweight = 291 x (1.3 x 3.278 x 3.278/40.28 - 1.3 x 1.3/1000 + 25.54 x 25.54/1000)

= 291 x (0.3468 - 0.0017 + 0.6523) = 290.2 kg-cm’.

In order to check if the answer is the same by swinging the racquet on a dif-
ferent string, the racquet was supported with the top cross string (nearest the
tip) resting on the two prongs. The bottom edge of the top string was a dis-
tance P = 30.2 cm from the balance point. This time, the average period of
oscillation was T = 1.323 sec. The formula then gives

Swingweight = 291 x (30.2 x 1.323 x 1.323/40.28 - 30.2 x 30.2/1000 + 25.54 x

25.54/1000) = 306.3 kg-cm’.

This is a bit higher than the previous value, but if we take an average, we get
298.1 plus or minus 7.9. This amounts to an error of plus or minus 2.6%. The
two values should be exactly the same in theory, but small errors in measur-
ing T and in measuring the distances contributed to the two different answers.
A small error in measuring P has a relatively large effect when P is small. For
example, if P was actually 1.4 cm in the first case rather than 1.3 cm, then the
swingweight is 298 which is much closer to the second value. An error of 0.1
cm in P in the second case is not so important. For example, if P was 30.3 cm
instead of 30.2 cm, then swingweight = 305.8 kg-cm’.

A better result can be obtained if the racquet is supported on one of the cross
strings near the middle of the string bed rather than the bottom cross string.
The values here were T = 1.285 s with P = 16.1 cm. This gives swingweight =
308.4 kg-cm? which is very close to the value 306.3 kg-cm* obtained using the
top string. These two values agree within plus or minus 0.3%. The lesson here
is that the top string, or a string near the middle, gives a more accurate meas-
urement of swingweight than the bottom string. These calculations can all be
done online at www.racquettech.com (Figure 5.4).

Figure 5.4
Online calculations at www.racquettech.com

ANSWER
WILL
APPEAR
48 Chapter 5

Twistweight Axis
The above method is easily extended to
include a measurement of the polar moment of
inertia. This is just a technical name describing
the resistance or inertia of a racquet for rota-
tion around its long axis (Figure 5.5). If a rac-
quet strikes a ball towards the top or bottom
edge of the frame, then the racquet will tend
to twist in the hand and the ball will come off
the strings at the wrong angle. The amount of
Figure 5.5 twist will be small if the racquet has a large
polar moment of inertia. By analogy with
swingweight, the polar moment can be called
the “twistweight.” Extra mass at the 3 and 9
o'clock positions on the frame is often used to
increase the twistweight.

In theory, all that needs to be done to measure

the twistweight is to rotate the racquet through
90 degrees, replace it on the prongs resting on
Twistweight with Normal Pattern a main string, and measure the time for 10
Support racquet on
oscillations. In order to get the racquet to bal-
Balance
furthest string from axis ance on the prongs in a horizontal position,
axis |
that stradles the both prongs must be located on opposite sides
balance point and can of the balance point (Figure 5.6). It is not nec-
fit a rod for swinging.
essary that the prongs be an equal distance on
either side, but the balance point must be
somewhere in the throat area of the strings and
not outside the stringbed. If the main strings
are all parallel, then the same string must rest
Measure P from the middle on both prongs. If necessary, the main string
of the two center mains to used to support the racquet can be moved
the bottom of the support slightly to ensure that the rotation axis is par-
string.
allel to the long axis of the racquet. If the main
strings are not parallel, but fan out, then the
Figure 5.6 only way to get the racquet to rotate about an
axis parallel to the long axis is to balance the
racquet on two different mains as shown in
Figure 5.7. The best way to do this is to adjust
the spacing between the prongs (located on a
horizontal table) until the racquet itself is hor-
izontal.

A slight push on the bottom edge of the frame

sets the racquet into oscillation about an axis
that is parallel to the long axis. If T is the time
in seconds for one oscillation, and P is the dis-
tance (in cm) from the long axis to the bottom
edge of the strings used to balance the racquet,
then twistweight is calculated as:

anes
Twistweight = M( PT? - —“—
—— p?
Figure 5.7 (40.28 1000 ) G2)
 Measuring Swingweights 49

The best string to use is the one that is furthest from the axis but still allows
the prongs to be inserted on each side of the balance point. The Wilson Triad
3.0 uses a fan pattern and can be balanced on the 6th and 7th main strings
from the middle, at P = 5.90 cm. In this case, T = 0.767 s, giving twistweight =
15.0 kg-cm*. On the Sth and 6th strings, P = 4.95 cm, T = 0.791 s and twist-
weight = 15.0 kg-cm’, exactly the same as the first measurement. A sawn-off
ruler is helpful in getting a good measurement of P. If the ruler is too long it
extends over the frame and it doesn't sit nicely on
the string plane.

On some racquets, the balance point lies outside

the string plane. In that case, a solution is to sus-
pend the racquet horizontally using two lengths of
string tied to a horizontal support frame as shown
in Figure 5.8. The formula for the twistweight is the
same as before, but P is the distance from the long
axis of the racquet to the actual axis of rotation near
the top end of the string.

The twistweight of a racquet is a lot smaller than its

swingweight. The twistweight can be increased by
adding lead tape of mass M at 3 o'clock and addi-
tional tape of mass M at 9 o'clock. If the tape is Figure 5.8
located a distance R (in cm) from the long axis
(passing through the middle of the handle) then the
twistweight will increase by MR*/500 (in kg-cm’)
where M is in gm. For example, if M = 10 gm and
R = 14 cm then the twistweight increases by 3.9
Z 0 2 :
kg-cm”. This amounts to about a 30% increase for
most racquets.

An alternative method of measuring the twistweight is described by Howard

Brody in The Physics Teacher (April 1985). The racquet can be suspended ver-
tically by a length of stiff wire attached to the handle so that the racquet twists
about its long axis as a torsional pendulum. It is necessary to calibrate the sys-
tem using a metal rod or bar with a known moment of inertia.

Twistweight can also be calculated online at www.RacquetTECH.com (Figure

Do):

ANSWER
Twist Weight Online WILL
APPEAR
Enter Data and Calculate Twist Weight

\
cont _ 8.28 ed
Calculate Twist Weight} Reset|

Figure 5.9
50 Chapter 5

Recoilweight Axis

Recoilweight is the swingweight around the

balance point or center of mass (Figure
5.10). Once you have the moment of inertia
about any axis (in this case we already have
the swingweight about the 4-inch point on
the handle), you can determine the moment
of inertia around any other axis parallel to
the swingweight axis. The formula is the par-
allel axis formula and its general form is as
follows:

l=, + MY" (5.3)

where:

Figure 5.10 I = moment of inertia (swingweight) about

axis parallel to the axis of the balance point.

I, = recoilweight (swingweight about the

|) ANSWER
center of mass or balance point).
Recoil Weight Online A WILL
APPEAR
M = mass of racquet in grams.

Y = distance in cm from swingweight axis

i oe
(10.16 cm from butt end) to the balance
Calcula point.

So, since the recoilweight is the moment of

Figure 5.11 inertia about the cm, we solve for I.

Recoilweight = Swingweight - MY” (5.4)

The units are kg-cm’.

The recoilweight can also be calculated at www.racquettech.com (Figure 5.11).

 The COP, Node,
Power Point,
and Dead Spot

ne of the terms often used in the tennis world is “sweetspot,” particularly by racquet manufacturers in
advertising their products. What is a sweetspot? Where on the racquet is it located? Can one location
be sweeter than another? Are some sweetspots bigger than other sweetspots? How can you find the
sweetspot on your own racquet?

Ask a tennis player what or where a sweetspot is, and the answer you will probably get is, “The sweetspot is
the ball impact location where it feels good to me,” or possibly, “It’s the location where the ball seems to fly off
of the strings with little effort by me.” The problem then is to turn these “touchy-feely” definitions into meas-
urable quantities that can be tested in the lab or on the court.

When the problem is analyzed, it turns out that there are actually three separate sweetspots on a tennis racquet
(Figure 6.1). They are the impact location of the ball on the strings where the shock or jar your hand feels is
minimized (called the center of percussion), the ball impact location that produces the least vibration of the rac-
quet (called the node), and the impact location that gives you the most return ball speed or “power” (where the
coefficient of restitution is a maximum). It is clear that each of these sweetspots, in one way or another, fits or
agrees with the player response quoted above. If the ball impacts at a point well away from the sweetspots, the
impact jars. The maximum power point is not a jar-free location, but it delivers the sweet feel of “oumph.”
52 Chapter 6

Spots on a Racquet Sometimes, when you hit the ball, there is a sharp
shock, or jar, to your hand and wrist. Sometimes
Tip you feel almost no sensation at all due to the ball
impact on the racquet’s strings. What is it that
determines whether you will have pain or pleas-
Dead spot
ure?

Sweet- Node Suppose that the racquet is initially at rest. If a ball

spots COP strikes a racquet at the racquet balance point (also
known as the center of mass, CM, or center of
Best bounce
gravity) and then rebounds, the entire racquet will
recoil in the direction of the original ball motion in
Point order to conserve linear momentum. All parts of
the racquet will recoil at the same speed V. The
faster the ball is moving, the faster will be the
recoil. Since the CM, or balance point, is usually in
Handle the throat of the racquet, you will probably not
want to hit the ball there. If the ball strikes the rac-
quet well above the balance point, near the center
Butt of the head (a more likely location to hit a ball),
the racquet will still recoil to conserve linear
momentum, but now it will also twist or rotate
Figure 6.1
about the CM or balance point in order to conserve
The ball bounces best near the throat only if the racquet is
angular momentum (Figure 6.2).
initially at rest.

This rotation, will cause the head of the racquet to

recoil backwards in the original ball direction, and
the handle to rotate forward. The actual motion of
the racquet will be the sum of both of these effects.
Since the motion of the entire racquet due to linear
momentum conservation is backward, and the
motion of the racquet handle or grip due to angu-
lar momentum conservation is forward, they are in
opposite directions. The whole racquet then moves
How a Racquet Recoils away from the ball with a speed V, due to the
recoil, but the handle simultaneously moves
te aS towards the ball with speed V, due to rotation of
ied |
the racquet. If there is any point in the handle
| |
_— COP where V, = V,, then that point will remain station-
| |
| l ary and the rest of the racquet will rotate about that
|p le}CM + = point as shown in Figure 6.2.
3
| | Therefore, it is possible for a particular ball impact
| | location on the strings to have the two motions
| | Motions
exactly cancel at the point where the hand holds
= (eG;ancel
the racquet. If this were to happen, and the two
motions cancelled, then the shock or jar that the
Translation + Rotation = Rotation about
about CM end of handle
hand experiences due to the ball impact on the
strings would be minimized.

Figure 6.2 The particular ball impact location on the strings

where the two motions cancel at the hand is called
 Sweetspots 53

the center of percussion (COP), and its location depends upon how the weight
is distributed along the racquet as well as where the hand is located on the
handle. The axis of rotation is called the conjugate point with respect to the
impact point, and the impact point is called the center of percussion (COP) for
that particular axis of rotation. The axis and the COP form a pair of conjugate
points. For an impact near the tip of the racquet, the axis of rotation is about
half way between the end of the handle and the CM. For an impact near the
throat of the racquet, the axis of rotation is beyond the end of the handle.

If the ball hits the strings above the COP point (closer to the tip of the racquet),
the racquet handle will try to come out of your hand by causing a force on
your fingers. If the ball hits the strings between the COP and the throat, there
will be a sharp force pushing on the palm of your hand. It is clear that if you
hit the ball at or very close to the COP point that corresponds to where you
are holding the racquet, you will feel less shock or jar from the ball impact.
You will have hit that sweetspot (Figure 6.3).

Related Pairs: COPs and Pivot Points

Every hitting location A has a related pivot point B. And vice versa—every gripping point B is a pivot point for some
hitting point A. If you hold the racquet about 4" from the end, there is a point in the racquet face that when hit, will
cause no rotation into or out of the hand. That hitting point is the center of percussion (COP) for that gripping
location. Because the handle does not slam into or pull out of the hand here, the COP is called a sweetspot.

CASE 1 CASE 2 CASE 3 CASE 4

Hit at CM Hit near CM Hit middle of face Hit near tip

Rotatio Se

CE "4

B B A
Pivot Hit Pivot Hit

No COP A=COP
of B A=COP
of B A=COP of B
Hitting at CM causes only Hitting at A, close to CM, Hitting at A, near the center of Hitting at A, near the tip,
backward motion. There is no causes the racquet (at your the strings, causes the racquet causes the racquet (at your
off-center torque in this case. gripping location) to move (at your gripping location) to gripping location) to rotate
The racquet goes straight backward faster than it rotates. translate backwards and rotate faster than it translates. The
back into the hand. The result is a pivot at B forwards at the same speed. result is a pivot at B closer to
beyond the end of the handle. The result is a pivot at B in the the CM. The handle will try to
You'll feel some pressure of grip right near the base of your pull out of your hand.
the handle into your hand. index finger. The handle will be
motionless at this spot.

Notes: (1) The amount and speed of translation is the same regardless of the impact point. It is only the amount of rotation that varies since
the torque increases from case 1 to case 4. (2) In Case 3, if the impact point A moves closer and closer to the CM, the pivot point will move
to different points under the hand and ultimately to the end of the racquet. Besides the base of the index finger, the end of the racquet is
most often seen as the most desireable spot for the pivot to occur. Regardless of the impact point, the racquet will rotate and tend to bend
the wrist backward. Soon after the initial impact, the hand takes over from the ball as the primary force on the racquet, with the result that
the axis of rotation shifts to the wrist immediately after the ball leaves the strings.

Figure 6.3

It is important to understand that this COP pivot is momentary. For a brief

moment during impact, the racquet is rotating or pivoting around a point on
the handle and not around the CM. Pivoting implies a stationary position while
the rest of the racquet moves around it. But this pivoting about the handle is
54 Chapter 6

just a snapshot in time. But it lasts just long enough that if you’re holding the
racquet there, the impact will involve a minimum jolt to your hand. After a few
milliseconds, the racquet would be rotating about the CM again, if it weren't
for your hand restraining it and relocating the axis to your wrist after impact.

It seems that no matter where the ball strikes the strings, there may well be a
point under the hand where the force is zero, but there will always be other
points where the force is not zero. In fact, that is exactly what is measured. The
impact causes the racquet head to recoil, so the whole racquet rotates in the
hand, exerting a force on the upper part of the hand and a force in the oppo-
site direction on the lower part of the hand. The torque causes the hand to
rotate about an axis through the wrist. However, if the impact is at the COP for
rotation about the end of the handle, then the force on the upper part of the
hand is equal and opposite the force on the lower part of the hand, so there
is no net force on the hand or the forearm. Therefore, the forearm will not
receive a sudden jolt if the ball impacts at this COP (Figure 6.4).

The location of the COP on the head of the

Figure 6.4. racquet does not depend on how hard you hit
F, is the force on the upper part of the hand, F, is the force on the ball or the type of stroke you are using. It
the lower part of the hand, and F, = F, + F, is the net force on is a property of the racquet frame itself and
the forearm. where you are holding the racquet. If you
move your hand up and grip the racquet high-
Tip er on the handle (choke up), the correspon-
ding COP point will also move upwards on
the head.
<() // Throat : ”
| You can determine the location of the COP of
| your own racquet by turning the racquet into
| a pendulum, with the pivot at the location on
the grip where the hand is located. The peri-
+ od (the time for a complete swing) of the rac-
quet-pendulum is obtained by allowing the
é F3 racquet to swing through a full cycle. The dis-
{ tance between the pivot point and the COP
location (in inches) is given by
|
COP = 9.78 x T? (6.1)
where T is the period of the racquet-pendulum (in seconds). If it takes the rac-
quet-pendulum 1.25 seconds to swing through a full cycle (T = 1.25 sec), then
the distance from the pivot point to the conjugate COP point where you want
to hit the ball is 9.78 x 1.25 x 1.25 = 15.3 inches.

There's another quick method to determine the COP location. It is to use the
following formula:

x =I, / (Mb) (6.2)

where

I, = the moment of inertia about the CM with units of kg-cm’,

b = the distance in centimeters from the impact point to the CM,

and
 Sweetspots 55

M = mass in kilograms of racquet,

x = distance from conjugate point to CM. The conjugate point is usually taken
at the butt end of the handle or at some point underneath the hand.

Another way to approximately locate the COP is by holding the end of the han-
dle between your finger and thumb and throwing a ball onto the strings. If the
handle jumps out of your hand, then you missed the COP. It is usually located
about 5 cm away from the center of the strings, as shown in Figure 6.1. A more
accurate measurement of the COP can be made with a piezo disk (extracted
from a piezo buzzer) between the thumb and the handle to measure the force
acting on the thumb.

Assuming the base of your index finger is about 3 to 4 inches from the butt
end of the racquet, the COP for a typical, old, classical, wooden racquet came
out near the throat of the frame, and not near the center of the head, where
players tend to hit the ball. With the introduction of the oversize head, the COP
point came out close to the center of the head. This was not because the COP
was a different distance from the hand, but because the center of the enlarged
head has been moved down toward the racquet handle. It is also possible to
raise the COP location higher in the head by adding weight to the tip of the
racquet, or by removing weight from the throat area.

A tennis racquet, even when made with the most advanced composite materi-
al, is not a completely rigid body — it can flex. When the racquet hits the ball,
the frame deforms, then snaps back, and it vibrates. The principal mode of
vibration of a hand-held racquet is shown schematically in Figure 6.5. Note that
in the fundamental mode the racquet vibrates so that it has a large amplitude
(big oscillation) in the middle and at each end. Since you hold the racquet at
the end, you will feel those vibrations. The other dotted lines are the next high-
est mode of vibration. This mode is not stimulated by a tennis ball, but might
be if you hit your racquet on a hard court or tap the frame with a hammer.
Looking at Figure 6.5, you can see that there are also two locations along the
racquet where the racquet vibrating in the fundamental mode has no ampli-
tude. Those locations are called nodes.

If the racquet is not struck at one of the two

nodes, the frame will vibrate. The amount of Vibration of a Uniform Beam
vibration depends how hard the ball is hit,
how flexible the racquet is, and how far from
the node the ball strikes. When a racquet Fundamental Mode 2nd Mode
vibrates, most of the vibration occurs at the
tip, the throat, and at the butt end of the han-
dle. There is a spot near the middle of the
strings that doesn't vibrate at all. It is actually
not a spot, but a curved line extending from
the middle of the strings to points at about 2
and 10 o'clock on the frame (Figure 6.6). The
spot that most people talk about is the inter-
section of this curved line with the long axis
Node Node
of the racquet. It turns out that if a ball hits
this spot (or anywhere else along the curved
Figure 6.5
56 Chapter 6

Vibrations at Different Locations on Racquet

TOce voltage signals from the piezo are shown in Figure 6.7a, for impacts at the tip, middie, and throat of (a) an old
Slazenger wood racquet, (b) a Wilson Triad 3.0 racquet, and (c) a Prince Viper racquet. Each of the racquets was sus-
pended freely from a horizontal beam by two lengths of string so that it was free to rotate and vibrate. A tennis ball was
also suspended from the beam, as a pendulum, so that it could swing towards the stringbed and impact at the same speed on
every racquet and at the exact spot we wanted.

The results in Figure 6.7a are typical of all racquets in that (a) the largest vibra-
tions of the handle occur for impacts at the tip and throat of a racquet, and (b)
the smallest vibrations occur at the sweetspot near the center of the stringbed.
50
a aes
, The initial motion of the handle is either towards the incoming ball for an impact
Ny 4 | near the tip, or away from the incoming ball for an impact near the throat. This
Tip Sweetspot_ Throat can be seen from the piezo traces where the first part of the signal is positive for
30 v 5 i v =| an impact near the tip and negative for an impact near the throat.

20 metas \ ! =| There is another sweetspot near the first, called the center of percussion (COP).
If a ball impacts at the COP, then the racquet rotates about an axis through the
amplitude
Vibration
Rollers’ butt end of the handle. In that case, the butt end does not move towards or away
LON :
from the incoming ball. The node sweetspot results in minimum vibration of the
Viper
“J handle. The COP sweetspot results in minimum sudden motion of the handle and
6G © & 10 13 BO 2s 20 25 hence minimum shock. The first spike in waveforms is the intial acceleration of
Impact distance from tip (cm) the handle due to handle rotation and translation. The rest of the spikes are the
6.7a vibrations at the handle. In these examples, we missed hitting exactly at the node
or COP. That’s because the impact locations were preset to be the same for each
racquet.

pierces cratonge 000d) hear Ee The four racquets shown in Figure 6.7a were chosen because they each illustrate
some interesting differences between racquets (Figure 6.7b-6.7d). Old wood rac-
quets vibrate a lot more than modern graphite racquets and they vibrate at a
lower frequency. Both of these effects are due to the fact that old wood racquets
are heavier and more flexible than graphite racquets. The Triad is interesting
because the vibrations are strongly damped by a rubber Iso-Zorb strip glued
Middle J
lr between the head and the handle. Additional strong damping occurs in all rac-
signal
Piezo
ie quets when the handle is held by hand. Damping doesn’t mean that the first vibra-
THroat tion cycle is significantly reduced in amplitude. Damping refers to the fact that the
DAVAVAVAVAVAVACA subsequent vibrations get smaller quicker, and it may also act to reduce the
L | amplitude of the first vibration to some extent.

: oy =e Bf ee The Viper had the smallest vibrations of all the racquets tested. It was also one
of the stiffest and it vibrated at the highest frequency (202 Hz with the wood block
6.7b
attached or 227 Hz without the wood block). Frame vibrations are almost com-
pletely absent in this racquet for an
; impact near the throat. The single
Prince TT Viper (202 Hz) Wilson Triad 3.0 (157 Hz) : : 4 :
Lo ae Se SS large spike in the piezo waveform is
due to the sudden acceleration of the
Tip handle when the ball strikes the
strings. That cannot be avoided and it
therefore occurs in all racquets. It
Middle
results in shock, but no vibration in the
Viper (or in any other racquet that
Piezo
signal signal
Piezo
| vibrates at 200 Hz or above). With
Throat Throat other racquets, the effects of rotation
| and vibration merge together right
| 4 from the beginnining of the impact,
!
roto ty tii increasing the overall acceleration of
0 20 40 60 80
eS the handle and the shock felt by the
6.7c 6.7d hand.
 Sweetspots 57

line), then the racquet will not vibrate at all, even at the tip
or the butt end. This spot is therefore a sweetspot, techni-
cally known as a vibration node. If the ball hits the strings,
say one inch away from this spot, then the racquet will
vibrate a small amount. The further away from the spot, the
more the racquet vibrates. If the ball impacts at a point on
the node line near the 2 or 10 o'clock position, then the
frame will not vibrate but it will twist in your hand.
Consequently, the sweetest point on the cuved node line is - 4 =
the one that intersects the long
ma eeaxis. \
ee Lo
GEeeeeas,
Baae® i
es sESmEg
However, when a ball strikes the racquet at the node, that m | hill : eee!
mode of oscillation does not occur, and the hit feels sweet. \ poo -
The vibrations of a hand-held racquet frame have a fre-
quency in the range of 100 to 200 hertz (vibrations/second),
depending on the frame stiffness, frame length, and the rac-
quet weight. Stiffer frames and lighter racquets tend to
vibrate at a higher frequency (more oscillations per second).
You can determine the location of the node of your own
racquet without fancy electronic equipment by holding the
racquet between two fingers at the approximate location of
Figure 6.6
the node in the handle (the top of the grip) and striking the
strings at various places with a ball held in your other hand. | Curved node line. An impact anywhere on the line
You will be able to feel which impact locations cause vibra- Will not result in vibration.
tions and which location on the strings does not cause the
frame to vibrate. That point will be the node.

Much design effort has gone into both reducing vibrations and damping out
vibrations. A stiffer racquet will have a smaller amplitude of vibration for the
same off-node ball impact. Various methods have been tried to damp out those
vibrations, which do occur, and the best one seems to be the human hand. In
fact, the tighter the racquet handle is gripped, the quicker the vibrations damp
out (with more of the energy of the racquet vibration going into the hand and
arm).

The small, soft objects that are placed in the stringbed close to the throat with
the hope of reducing or eliminating vibrations, do essentially nothing to pre-
vent, reduce, or damp out vibrations of the frame of the racquet. However, they
do quickly kill off the 500 to 600 hertz string vibrations. (They turn the “ping”
that you hear when the ball hits into a “thud” sound.) The energy contained in
string vibrations is quite small compared to the energy in frame vibrations,
because the frame has about twenty times the mass of the strings.

It is obviously quite desirable to hit the ball at the node of the frame, or as
close to it as possible, to eliminate or minimize the subsequent annoying vibra-
tions of the frame. With the old, small headed, wooden racquets, the node
location tended to be below the center of the head, closer to the throat. When
racquet heads were enlarged by extending them down toward the handle, the
center of the head was moved closer to the hand. On modern, oversized rac-
quets, the node is now located at or above the center of the strung area of the
head. It is also possible to move the location of the node slightly by changing
the stiffness or mass distribution of the frame or the shape of the head.

While vibrations are annoying and bothersome, there is no clinical evidence

that vibrations of either the frame or the strings can cause any type of hand,
58 Chapter 6

wrist, and arm or elbow damage. There is some experimental evidence (wear
pattern on the strings, high speed photography, etc.) that on groundstrokes,
good tennis players tend to hit the ball at the location of the node of their rac-
quet, rather than at one of the other two sweetspot locations.

The Power ofthe Tennis Racquet: COR an

The term power, as used in tennis, refers to the speed of the outgoing ball rel-
ative to the effort the player has exerted in swinging the racquet. A racquet
which provides high ball speeds for small effort on the part of the player is said
to be powerful. This is to be distinguished from the definition of power used
in physics and engineering where the term means the rate at which work is
done or energy transformed.

As important as the maximum power of a racquet is, the variation of the power
with respect to where the ball hits on the strings may be more important. If the
ball strikes the strings at a small distance from the desired location and the
result is a very different ball rebound speed, the shot will probably fall short
and a point will be lost. When this happens, the power sweetspot is consid-
ered to be quite small and the racquet will be difficult to play with. If, on the
other hand, some miss-hits do not lead to an appreciable difference in rebound
ball speeds and they result in acceptable ball trajectories (the shot still goes in),
the racquet is said to have a large power sweetspot, be “forgiving”, and enjoy-
able to play with.

A heavier racquet will be more forgiving and it will also be more powerful, if
you swing it at the same speed as a lighter racquet. A lighter racquet will give
you more power if you can swing it at a higher speed. When you try to swing
a racquet at a higher speed, you run the risk of a loss of control. Most players
do not swing their racquets as fast as is physically possible when _ hitting
groundstrokes, so as to avoid a loss of control. On the other hand, on a serve,
players often swing the racquet as fast as they can.

To measure the effective power of tennis racquets, a manufacturer will fire

balls at various locations on the head of a freely suspended racquet and meas-
ure the rebound ball speed. The ratio of ball speed out (rebound) to ball speed
in (incident) is called the apparent coefficient of restitution (ACOR). The term
ACOR is used instead of the commonly used term, coefficient of restitution
(COR), because in this case, the recoil velocity of the racquet is neglected. The
larger this ACOR ratio, the more powerful the racquet. A map of the ACOR
value across the strung portion of the racquet will show where it is lively and
where it is dead — where you will get power and where you will not.

As a general rule, the ACOR is a maximum near the balance point of the rac-
quet and it falls off as you approach the tip or move to the sides of the head.
The fall-off to the sides can be reduced somewhat by making the racquet wider
(as Howard Head did when he introduced the oversize) or adding weight to
the head at 3 and 9 o'clock. Some of the loss of power as the impact point
moves toward the tip is overcome by the fact that a racquet is swung in an arc,
not translated or pushed in a straight line. This means that the tip is moving
faster than the throat, which tends to move the maximum power point up into
the head and away from the throat. By enlarging the head on the modern rac-
quet, the center of the strung area has been brought closer to the maximum
power point.
 Sweetspots 59

However, for groundstrokes, unless you have a very wristy swing, for many
racquets, the maximum power point is often still well below the center of the
head and below the other two sweetspots. To raise the power spot higher into
the head, racquet designers have done a number of things. They have moved
the balance point up (by removing weight from the handle or adding it to the
head), made the head stiffer (wide body construction), or made the head get
wider toward the tip, with the strings fanning out as well. The resulting rac-
quets allow players to hit the ball at the node or at the center of the head and
not pay a penalty of a loss of power.

For a typical serve (where the incoming ball speed is zero), the modern, stiff,
graphite composite racquet has its maximum power point well above the cen-
ter of the head. This has the added advantage of allowing you to hit the ball
when it is higher, thus increasing the chances of the serve going in (greater
height gives you a bigger acceptance window). The position of the maximum
power point is illustrated in Figure. 16.1 for a typical serve and in Figure 106.2
for a typical groundstroke. The position is different in each case and it depends
on how the weight is distributed in the head. For a groundstroke, it is possi-
ble for the maximum power point to coincide with one or other of the other
two sweetspots, or at least be reasonably close, in which case all three sweet
spots will contribute to the fact that the shot feels good. For a serve, the max-
imum power point is closer to the tip. Since locations in the upper half of the
stringbed are further away from the COP and the node, there will be more
shock and vibration. In that sense, the feel will not be so “sweet” but the out-
come should be a slightly faster serve.

Clamp the end of the handle on a table, using your hand to press on the han-
dle, so the rest of the racquet hangs over the edge of the table. Then drop a
ball onto the strings at various points. The ball will bounce best near the throat.
There is a spot near the tin where the ball doesn't bounce at all. That's the
deadspot. At the deadspot, all of the energy otf the ball is given to the racquet,
and the racquet does not give any energy back to the ball. The reason is that
the effective mass of the racquet at that point is equal to the mass of the ball.
The effective mass is the ratio of the force at that point to the acceleration at
that point (F = ma, so m = F/a). In a completely elastic collision where no
kinetic energy is lost, if a ball of mass m collides head-on with another ball of
mass m at rest, then the incident ball stops dead and gives all its energy to the
other ball.

If a moving racquet strikes a stationary ball at the deadspot, then all the rota-
tional energy of the racquet is given to the ball. One can serve a ball at high
speed when serving from the deadspot. However, when returning a fast serve,
the deadspot is the worst place to hit the ball. The best spot is nearer the throat
of the racquet since that's where the ball bounces best.
 =
S
=
=
fe)
(om

lo Shock
& Vibration
What It Is and
How It Happens

By Howard Brody and Crawford Lindsey

Introduction _ ee

here are two different types of vibrations that occur in a tennis racquet — vibrations of the frame of the
racquet and vibrations of the strings. The vibrations of the strings are discussed in Chapter 26, so it will
not be dealt with here. We will restrict this article to vibrations of the frame of the racquet.

The frame of a racquet seems like a rigid body with no elastic properties, yet on the time scale of a ball impact-
ing and with the forces involved in such an impact, the racquet frame is flexible. When the strings are hit by a
ball, the local tension around the impact region can increase to several hundred pounds. This causes the rac-
quet to deform and then the frame snaps back, overshoots its equilibrium configuration, and oscillates for a peri-
od of time, depending on how it is dampened. The frequency at which this happens depends on the stiffness
of the racquet frame and its moment of inertia or mass distribution. The racquet can vibrate in a number of dif-
ferent modes of oscillation — bending, torsional, etc. We will restrict our discussion to the lowest bending mode
62 Chapter 7

of oscillation because this is the mode with the largest amplitude and it
involves the most energy.

Before the ball hits your racquet, it is in equilibrium. At impact it starts bend-
ing, shaking, and twisting in all manner of being. In short, it goes into vibra-
tion. Vibration is everywhere in our lives, but what is it?

It’s nothing but movement. In the simplest case, it’s back and forth movement.
When the racquet bends, a restorative force attempts to return to equilibrium.
This force is nothing but the attractive forces between atoms which gets
stronger in proportion to how far the atoms get pulled apart (up to a point).
This is how it works. Imagine the racquet composed of little billiard ball parti-
cles (mass) all connected together by rubber bands (stiffness). When the ball
hits, it knocks the first particle backward stretching the rubber band. This in
turn pulls the next particle, and so on right down to the other end of the rac-
quet. This occurs as a force that begins at the point of impact and radiates out
at a speed depending on the type of material. After a certain amount of stretch,
the rubber bands get-stiffer and pull the particles back (restoring force). But in
heading back to its starting position, each particle gains momentum and can’t
put on the brakes. By the time it reaches its equilibrium position (no bend), it
has gained maximum velocity and momentum and it flies past until the rubber
bands stretch and reverse its direction again — and so on. This motion finally
dies out when energy is transferred and “lost” to the air by fighting drag, to
heat by internal structures rubbing against each other and getting “tired” from
frictional resistance, and to your hand by expending energy compressing flesh.
The impact force sets vibration in motion, the restoring force keeps it going,
and damping interactions dissipate it.

The modes of oscillation of a racquet (or any other object) depend on how the
racquet is held or clamped. If the frame is clamped near the butt end, the low-
est bending mode of oscillation is about 25 to 40 hertz. If the racquet is freely
suspended, the frequency of its lowest mode of oscillation is between 100 and
250 hertz. These are schematically shown in Figure 7.la. The question then
arises: when in play, does a hand-held racquet act like a clamped frame or as
a free frame?

When the vibration frequency of the lowest mode of a hand-held racquet is

measured, it comes out very close to, but slightly lower than the frequency of
a free racquet. Looking at oscilloscope traces of the vibration signal of a hand-
held racquet, there is essentially no sign of the low frequency pattern one
would get with a clamped racquet. This clearly shows that a hand-held racquet
behaves as if it were a free racquet. We conclude that the hand is not strong
enough to act as a rigid clamp, and the racquet handle can still shake back and
forth even when the hand is gripping as tightly as possible.

The amplitude of the oscillation of the frame depends very strongly on where
on the head the ball impacts. There is one impact location on the head where
the amplitude of the oscillation is reduced to zero — no measurable amplitude.
 Introduction to Shock & Vibration 63

Figure 7.1 a. Bending Shapes During Vibration

Clamped Handle ———— Free or Hand-held Racquets

Equilibrium Diving Board Mode First Bending Mode Second Bending Mode

Frequency is number of times per 25-40 Hz 125-200 Hz 250-450 Hz

second the racquet oscillates
between one extreme of the AG AVG WWW
mode’s shape and back again. VW

@ The Diving Board Model is not appropriate because the hand cannot hold the racquet rigid.
@ The first or fundamental mode is the predominant shape for racquet vibration.
# The second mode is a higher vibration mode. It has so little energy and occurs so quickly that the mass of
the ball dampens virtually all this vibration before it leaves the racquet. However, it will appear if you hit
the racquet with a golf ball or a hammer. In these cases, the impact occurs so quickly that the higher
frequency vibrations will show up. Still, higher frequencies die out much faster, anyway.

b. Nodes

Nodes

@ Nodes are locations where there is

no movement, no vibration.

If you are holding at a node, you will

feel no vibration at that point. @ When impact is quick enough to cause a
If you impact at the node in the higher frequency, the racquet actually
head, the racquet will not vibrate vibrates in a composite wave form that is
anywhere. the sum of the two separate frequencies.
@ Each point on the racquet travels a
different "back-and-forth" distance.
64 Chapter 7

Figure 7.2 Anatomy of a Vibration

Handle Vibration at Different Impact Locations

# What is an oscilloscope tracing? 250 impact distance

The tracings are the result of piezo from tip

accelerometer located on the 4 cm

handle that measures the
200
acceleration of the handle at
impact and after. It shows the rate
and direction of velocity change,
not the magnitude of displacement 150
in space. That can be derived,
however.

@ Shock: The first peak is shock — 100

acceleration of the racquet
backward by translation and
rotation. The ball has left the
50
racquet by the second peak —5 ms.

¢ Vibration: All the peaks and

troughs after the first are racquet (mV)
signal
Piezo
bending of vibration.

e Frequency: One cycle of back-and-

forth vibration is represented by
-50
going from peak to peak. The
number of full cycles that occur in
one second is the frequency. One
cycle per second is known as a -100
Hertz (Hz).

@ Amplitude: The height of the crest

above or below the horizontal -150
starting axis.

e Impact: Tracings are the vibration at the handle when the impact is at the designated distance from the racquet tip.

Sd Frequency signature: Frequency is constant and independent of impact location.

You can see this because the crests
line up right above each other at each location.

Sd What you feel: Amplitude of the vibration at your hand depends on the impact location. Shock and vibration felt at the
hand are greatest when impact is at the tip or throat. Vibration is least when impact is at the node. Shock is least
when impact is at the COP.

@ COP: The COP is located somewhere between 16 and 20 cm here. This is evident because the first peak (the
backward motion of the racquet due to impact) has changed direction from positive to negative. That means the
racquet handle is being pushed in the opposite direction than impact above the COP. At the COP, the first peak is
nonexistent, and there is very little vibration.

Node: At the node, the first peak is small (still some backward movement), but the vibration component is flatlined.

¢ Damping: Each successive crest has a smaller amplitude. This is due to damping — energy dissipation due to pushing
air molecules around, internal friction between atoms, and with a hand-held racquet, fighting against your hand.

¢ Note: (1) Neither the COP or node was exactly hit in this series of impacts. (2) Each signal shows the voltage
generated by a piezo disk located on the handle when the racquet was freely suspended. The ball impacted at the
same speed in each case. The tracings have been artifically moved under each other, so the signal readings on the
left axis opposite each tracing do not show the actual voltage of each respective trace.
 Introduction to Shock & Vibration 65

This location is called the node of the first bending mode, and it is usually
located near the center of the strung area of the head (Figure 7.1b). There is a
second node of this mode of oscillation and it is located in the handle, slight-
ly above the grip. Since the ball rarely strikes this node, we will neglect it. The
magnitude or amplitude of the oscillation of a racquet due to the ball impact
is directly related to the distance between the impact location and the node. If
the ball hits on the head far from the node, the oscillations will be relatively
large in amplitude and can produce an uncomfortable feeling. Figure 7.2 shows
the vibrations of the frame of a racquet for different ball impact locations on
the head. Note that for a ball impact close to the center of the head (where the
node is located), the resulting vibrations have a very small amplitude. The
amplitude of vibration increases as the impact location moves away from the
node.

The amplitude of the vibration depends on three factors:

1. impact location,
2. racquet stiffness, and
3. how hard the ball is hit (the relative ball-racquet velocity).

It is obvious that the harder the ball is hit, the more the racquet frame will
deform, and the bigger the amplitude of the oscillation will be. It is also obvi-
ous that a stiffer racquet will deform less upon impact, so a stiffer frame will
have a smaller amplitude of oscillation, if all else is the same. A flexible frame
will deform or distort more on impact, so its amplitude of oscillation will be
bigger. Because the energy in the oscillation goes as the stiffness times the
square of the amplitude, a more flexible frame, having a larger amplitude of
oscillation, will have more energy in its vibration than a stiff racquet will have.

The frequency of vibration of the lowest bending mode does not depend on
where the ball impacts the head (assuming it misses the node), or how hard
the ball is struck. The frequency depends on:

1. racquet stiffness, and

2. racquet mass (actually the moment of inertia).

To first order, the frequency of this mode is proportional to the square root of
the stiffness divided by the moment of inertia about the center of mass. When
the frequency of the oscillation of the first bending moment is plotted against
the square root of the racquet stiffness divided by the racquet’s moment, the
result is a fairly smooth curve, with little scatter of the data over many differ-
ent racquets. Because a racquet frame can be stiff in some areas (like the han-
dle) and flexible in other regions (such as the throat or tip), the value one gets
from a stiffness measurement depends on how it was measured.

If the racquet is held fixed at one end (like a cantilever), and a force is applied
to the tip, you may get a different value of stiffness than if the racquet is sus-
pended at both ends and a force applied in the middle. In addition, there is
the problem of the softness of the grip material deforming, which can give an
erroneous stiffness reading. An alternate method of specifying racquet stiffness
would therefore be to specify the frequency at which the racquet vibrates. This
66 Chapter 7

method would tend to over emphasize the stiffness

Higher Frequency Vibrations — What Are They?
of the frame near the center (throat area) and not
Ty: fact that an object can vibrate with multiple frequencies count the stiffness of the frame near the nodes as
at once is easily demonstrated by two examples. First, much.
consider what happens if you strike middle C on a piano.
Middle C has a frequency of 256 Hz. However all notes with fre- You will often hear about higher frequency modes
quencies of 2, 3, 4, 5, etc., times 256 Hz also begin to vibrate than the first bending (or fundamental) mode. A
and no others. The C note created air and piano frame vibrations racquet actually vibrates simultaneously in a theo-
at all those frequencies that then set the strings into vibration retically infinite number of modes. However, we
that are tuned to each frequency. don’t concern ourselves much with these since
they have so little energy that the mass of the ball
The second demonstration requires a little setup (see Figure
virtually dampens them all out before it leaves the
7.3). Loop a rubber band over the head of an unstrung racquet.
racquet. All that remains of any significant ampli-
Hold the racquet horizontally in front of a turned on computer
screen (this acts like a strobe light with different horizontal and
tude (thus, energy) after the impact is the first
vertical blinking rates). If you pluck one string of this band (while mode. If you hit the racquet with a light striking
holding the other off to the side), it appears to vibrate up and hammer or a golf ball, these higher modes will still
down as a whole. If you turn it perpendicular, you see the vibra- be visible after impact. That’s because these
tions occurring in many waves along the rubber band. This is how impacts occur so rapidly (because they are stiffer
the racquet also behaves under the proper conditions. and have less mass), the higher vibrations are not
dampened by the very impact that caused them.

Demonstration
Multiple Frequencies and Shapes
Stiffness
The stiffness of a racquet frame (hence the ampli-
tude and frequency of the oscillation) depends on:

1. frame construction,
2. material used in the frame, and
Rubber Band 3. frame thickness.
Figure 7.3
If all else is held constant, the stiffness of a beam
goes as the cube of its thickness. Therefore,
increasing the thickness of the racquet frame from 19 mm to 38 mm could give
you a frame that is twice as thick, and since two cubed is eight, the frame will
be eight times as stiff. Even as little as a ten percent increase in thickness can
provide as much as a thirty percent increase in stiffness. As an experiment, try
bending a ruler or a yardstick, first in one direction, then in the other direction.
It will usually bend quite easily when bent against the thin dimension, but it
will be very stiff when you try to bend it against the thicker dimension.

What else can influence the frequency of vibration? A strung racquet will be a
little more flexible than an unstrung racquet, and in addition, there are 15
grams of string added to the head. Both of these effects will tend to lower the
racquet’s vibration frequency a little.

Adding weights to the frame at the tip (12 o’clock), the throat (6 o’clock), or
the butt end will lower the frequency of oscillation and possibly move the
node location a little. Adding weight to the head near the node (3 and 9
o'clock) will not appreciably change the frequency of the first bending moment
or the location of the node.
 Introduction to Shock & Vibration 67

Once the racquet is vibrating, it will continue to oscillate until

the energy involved is dissipated. For a free, undamped
frame, it can take a second or so until the oscillations die out.
There are materials that can be built into the frame or added
to the frame that will dissipate energy faster than the free rac-
quet alone can. Such devices include hollow spaces with
small lead shot or other energy dissipating beads in them and
energy absorbing layers of material fabricated into the
graphite/epoxy frame. The various vibration dampeners that
are placed in the string bed of a racquet do essentially noth-
ing to reduce or diminish frame vibrations. Figure 7.4 shows
the vibrations of the racquet with and without a string vibra- Figure 7.4
tion dampener in place. There is no apparent difference Oscillations of a free racquet with and without a
between these two oscilloscope traces. vibration dampener in the strings.

Stiffer frames have less amplitude,

energy and dampen faster.

Oscillations in stiffer racquets tend to damp out faster than oscillations in more
flexible frames for several reasons. The energy involved in a vibrating frame
depends on the stiffness of that frame, with stiffer frames storing less energy.
Since the frequency of a stiffer frame is higher than the frequency of a flexible
frame, within a given time period, the stiffer frame will go through more cycles
of oscillation, with each cycle dissipating some of the vibrational energy.

The hand is the best dampener.

What does a marvelous job of dissipating the energy associ-

ated with the vibration of a racquet is the human hand, when
the racquet is gripped by the handle. The tighter the racquet
is gripped, the quicker the vibrations are damped out. Most
of the previously mentioned mechanical methods for damp-
ing out vibrations work well for a free racquet, but are not
that significant compared to the effect of a tight grip. The
hand actually dampens most of the vibrations before the
other methods have much of a chance to take effect. Figure
7.5 shows the oscilloscope traces of the vibrations of a freely
suspended racquet and one that is gripped by a player. It is
clear from the figure that the hand does a very good job of
damping out vibrations quickly. Wouldn’t that be an argu-
ment for a very tight grip since it helps to quickly damp out Figure 7.5
vibrations? Not necessarily, since all the energy quickly taken
Oscillations of a free and a hand-held racquet.
out of the frame’s vibrations end up in the hand and arm, Note how quickly the oscillations damp out when
rather than the racquet’s frame. If the racquet is gripped fur- the racquet is held in the hand.
ther up the handle toward the second node (choked up),
then the dampening effect of the hand on the vibrations of
the frame is diminished.

At present, there is no clinical evidence that vibrations of the racquet frame (or
the strings) are a cause of any sort of arm, elbow, or shoulder injury. Vibrations
68 Chapter 7

may be irritating, annoying, and uncomfortable, and you may want to elimi-
nate them, but they may not be harmful. Since the node’s location is usually
fairly close to the center of percussion location, a ball impact that misses the
node will usually miss the COP as well. This will produce a shock or jar to the
hand and arm, which to many players will be indistinguishable from vibration.

Some tennis articles tell the player to grip the racquet very tightly so that the
mass of the hand and arm is added to the mass of the racquet, giving more ball
speed (or power). As was mentioned earlier, a hand-held racquet has a vibra-
tion frequency slightly lower than a free racquet. As an experiment, a racquet
handle was gripped as firmly as possible by this author’s hand, and the result-
ant frequency (for off node impacts) was measured. The frequencies of a free
racquet with various masses added on to the butt end were then measured. It
was discovered that the frequency shift observed with a very firm grip could
be duplicated with 20 grams attached to the butt end of the same racquet, but
freely suspended. This means that gripping the racquet as firmly as possible
will add less than an ounce to the racquet’s weight — and that weight is at the
butt end where it does not appreciably change the racquet’s hitting power.
Another tennis myth is shown to be wrong. There is no need to grip a racquet
any tighter than is necessary to get the racquet to its correct location at the cor-
rect time and with the correct velocity. It is claimed by some that an exces-
sively tight grip can lead to arm (elbow) injury. This is one tennis myth that
may be true.

IsShoc k and What Is Its Relation to Vibratio

Upon impact, the racquet experiences a force. A stiff racquet will experience a
high force for a short time and a more flexible racquet will feel a lesser force
for a longer time. In either case, the force times the time for each are equal so
that the same energy is involved in each case. This force causes three things to
happen:

1. The racquet recoils.

2. ine racquet rotates:
3. The racquet bends.

This is a complicated array of movements. All three motions contribute to

shock. The recoil and bending will slap your hand and the rotation will twist
it. This doesn’t happen instantly. The racquet head “knows” there has been a
collision before the handle does. Granted, it seems instantaneous to us, but the
sequence of events is important to the engineering of racquets.

The ball is on the strings for a total time of about 5 ms (0.005 seconds). The
bending wave travels up and down the racquet in 8 ms (0.008 seconds) if the
racquet’s vibration frequency is 120 Hz, or 5 ms if the vibration frequency is
200 Hz (very stiff racquet, but typical of wide body racquets). So, let’s take an
average racquet somewhere in the middle of this range, say 150 Hz. If the ball
strikes the racquet at 10 cm (4 inches) from the tip of a 68.58 cm (27 inches)
racquet, the vibration must travel 58.58 cm to the end of the handle and 58.58
cm back to the impact point (total of about 118 cm) if it is to influence the
speed of the ball. If the racquet frequency is 150 Hz, that means the wave does
one complete cycle (.e., back and forth, or once up and down the racquet
which is about 137 cm) every 0.00067 seconds or 20.5 cm/ms = 205 m/s = 459
 Introduction to Shock & Vibration 69

mph. If the wave must travel 118 cm back to the impact point
How Does Shock Absorption Work?
at that speed, it will take about 5.8 ms to do so. The ball is
already gone. Even for the very stiffest racquets, the wave will T= is a difference between shock absorption
be arriving back at the impact point just as the ball is leaving, and shock isolation. Also, anything that is
so it has no, or almost no, effect on the ball. advertised as a “shock absorber” is also a
vibration dampener. Shock absorption involves turn-
So, even with the stiffest racquets, it takes 2.5 ms for the ing energy into heat whereas shock isolation
bending wave to get to the hand. But what of the rotation and changes the way the energy acts on a system.
the recoil? How long do they take to get to the handle? The Strictly speaking, most shock absorption systems in
answer is they all arrive at about the same time. At impact, racquets work as shock isolation systems, the
exceptions being systems with any kind of movable
the racquet head rotates with respect to the handle. How
mass. In general, any resilient material can be used
does the handle know to move also? In an extended object
as an isolator. “Resilience” means that the material
like a racquet, the message that the head has been hit cannot will store and return energy and recover to its origi-
get to the handle until the bending wave is transmitted to the nal shape without losing any of its properties so that
handle. The bending wave travels quite a bit slower than the it will be ready to perform again. Depending on the
speed of sound in a material. Umpact also causes a compres- application, the isolator may be a spring, a metal,
sion wave that travels at the speed of sound, but the direction closed or open-celled foam, or in most cases, an
of that force is along the length of the racquet and therefore elastomer. An elastomer is another name for any
of no consequence to your hand or the ball.) The speed of type of natural or synthetic rubber. Each elastomer
the bending wave is the “signaling speed” of the racquet and has different strain, strength, and temperature prop-
is about 205 meters per second in a racquet with a frequen- erties, so choice of material is critical depending on
the application.
cy of 150 Hz. So, for the handle, news of the impact and
instructions to recoil, rotate, and bend all arrive at the same
The key to an elastomer’s success is that it be of the
time on the wings of the wave. proper stiffness to both isolate the shock and not
interfere with the performance of the equipment.
Because the racquet bends, it is not easy to see that the rac- Shock isolators work by storing the input energy and
quet actually rotates — it rotates at one end but not the other! giving it back at a much slower rate than the input.
Imagine throwing a bit of jelly that wobbles as it rotates. How So the energy is transferred to the handle over a
do you tell if it is wobbling, or translating, or rotating? You longer period of time than it entered the material.
can’t tell until you watch it wobble a few cycles. And it will The longer time period means the peak force expe-
change during the first few cycles anyway. In a racquet, the rienced by the handle will be less. The energy does
initial bend in the head travels up to the handle, reflects up not magically disappear, it is almost all still transmit-
ted, but it is done so in a kinder, gentler fashion.
and down the racquet a few times, and eventually sorts itself
out into a definite amount of rotation, translation and vibra-
Elastomers work differently depending on the direc-
tion. But it is impossible to see this in the beginning.
tion of the force to their cross-section. Cushion han-
dles or grips are compressed perpendicular to the
The first spike in a tracing in Figure 7.2 represents this com- material. Materials in the frame usually are sheared
bination translation/rotation/vibration acceleration. That is (like putting a deck of cards between your hands as
why this spike is the biggest. The hand resists and stops the if you were praying and then pushing up with one
translation and rotation, but the vibration continues. All the hand and pulling down with the other). The secret to
peaks and valleys after the first spike are racquet accelera- most elastomers is that they act 2 or 3 times thicker
tions due to vibration only. in the shear direction than their actual dimensional
thickness. When you compress an elastomer that is
too thin for the force, it will start to react by getting
stiffer — exactly what you don’t want it to do. In the
shear direction, the elastomer can distort 2-3 times
These arguments raise some interesting questions: further without getting stiffer. This makes material
selection very critical when choosing an elastomer.
1. Does the whole racquet influence the hit of the ball or You don’t want a material that will actually increase
does just part of it? shock at a certain force level. All elastomers do get
2. If only part of the racquet is involved, what parts and what stiffer above a certain load and many elastomers
role does each play? have different stiffness properties in different direc-
tions of force application. For these reasons, elas-
3. If only part of the racquet is involved, does stiffness along
tomeric material does not have to have the same
the entire length matter?
elastic properties in all directions (isoelasticity).
4, How much of the racquet’s weight is involved in the hit?
70 Chapter 7

As the questions imply, only part of the racquet is involved during impact. We
can refer to a playful experiment by Howard Brody to get an intuitive feel for
this answer. Several years ago, Dr. Brody made a racquet with a hinged head,
such that the head literally swung freely from halfway down the handle. The
surprising results of hitting the ball with this racquet were, first, that you could
hit it, and second, that the ball had almost the same velocity as with a real rac-
quet.

The lesson is that it is predominately “local” head stiffness and mass that hits
the ball, not the entire racquet. As far as the actual hit and resulting power is
concerned, the handle serves primarily to add reach and to accelerate the head
to impact speed. Figure 7.6 shows how this can be so.

BENDING TIME LINE

(Who knows what, how much, and when)

HEAD Max Deflection Ball Leaves

0 5 1 1.5 2 2.5 3 3.5 4 4.5 bY

Milliseconds

HANDLE Max Deflection

Figure 7.6

It shows that at any time during impact, each part of the racquet has a differ-
ent “knowledge” of the impact, ranging from none to everything.
Consequently, the so-called “effective hitting mass” and “effective hitting stiff-
ness” are defined only for as much of the racquet to which the impact message
can get delivered before the strings start to send the ball back. For example,
the amount a racquet will bend (and energy it will store) during the initial
backward impact before it reverses its oscillation (2.5 ms) depends on only the
stiffness of the length of the racquet over which the bending wave has trav-
elled in those 2.5 ms. That is the only part of the racquet that has been hit in
that time. It is the only part of the racquet that knows that it has been hit before
the ball changes direction and the strings begin the ball’s rebound. The amount
of bend also depends on how much of the wave has arrived at a location. If
only the front of the wave has arrived, that location thinks it has only been hit
by a tiny impact. If the peak of the wave has arrived, that location knows the
whole story.

Two objects hitting each other will always satisfy conservation of momentum.
If there are no energy losses, they will also satisfy conservation of energy. All
 Introduction to Shock & Vibration 71

this remains true, even if only one end of a racquet is involved in the collision.
In that case, and if the ball begins to rebound before the handle starts to move,
then the ball will behave as if it hit the relatively small mass of just the racquet
head. Similarly, the racquet head will rebound as if the handle wasn’t even
there. However, that sets up some significant bending between the head and
the handle, so the energy in the head starts to get redistributed along the whole
racquet, in the form of rotation, translation, and vibration. After the bending
wave has traveled up and down the racquet a few times, the energy gets dis-
tributed in exactly the way that you would predict from the conservation equa-
tions.

So, that’s the theory and the principles behind shock and vibration. But all that
said, three things need mentioning. First, the hand still does most of the shock
absorbing and continuous jolts to the hand can potentially cause arm problems.
Consequently, anything that lessens the hand’s task is welcome. Second, the
easiest method of shock absorption is lowering string tension (along with using
a softer string). However, the significant tension lowering required could have
greater negative consequences on performance than positive ones on shock
relief. And last, some shock is what “feel” is all about. Some players like a crisp
feel and others a softer feel, but everyone wants to feel something.
 The Why
of Vibrations

Amelie Mauresmo, France

By Rod Cross

hen a tennis player hits a ball firmly with a racquet, the strings stretch and exert a large force on the
frame. Even though you won't be able to see it by eye, the ball squashes approximately in half and
the frame bends out of shape during the impact. As the ball springs back to its original shape, it pro-
pels itself off the strings, assisted by the fact that the strings and the frame also spring back towards their orig-
inal shape during the latter half of the impact. Ideally, the frame would recover its original shape just as the ball
leaves the strings, giving all its stored elastic energy back to the ball. In that case, the frame would not vibrate
at all after the impact, and no energy would be wasted in frame vibrations. This, indeed, happens at one spot
along the axis of the racquet, near the middle of the strings. It also happens at any point on a curved line join-
ing this spot to the 2 or 10 o'clock positions on the frame, 12 o'clock being at the tip of the racquet. At any
other impact point on the strings, the frame remains bent at the instant the ball loses contact with the strings,
and the frame vibrates after the collision is over. Frame vibrations are largest when the ball impacts near the tip
or the throat of a racquet, but they can be reduced by making the racquet stiffer.
74 Chapter 8

Most tennis racquets vibrate at a frequency between 120 and 150 Hz (cycles
per second) taking only 7 or 8 ms (0.008 s) to complete one vibration cycle,
depending on the mass and stiffness of the frame. The frame bends backwards
and then springs back towards its initial position in about 4 ms, which is a typ-
ical time that the ball spends on the strings. One might therefore expect that
the racquet should always give back most of its elastic energy, especially if the
impact duration coincides exactly with one half of the vibration period.
However, this is not generally the case.

When a racquet is used to hit a ball, it bends into a curved arc and tends to
straighten out just as the ball leaves the strings. It can straighten out in sever-
al different ways, only one of which prevents the racquet vibrating. For exam-
ple, the racquet may straighten out at sufficient speed that it overshoots the
straight position and bends back the other way, with the opposite curvature. If
that happens, the racquet will continue to vibrate after the ball has left the
strings. The racquet will also vibrate if the head section straightens out while
the handle remains bent. Normally, one would consider a modern racquet to
be very stiff and it should hardly bend at all during a collision. But it is not infi-
nitely stiff, so it does bend and it does vibrate. A consequence is that the head
can bend while the handle remains straight, and then, while the head is
straightening out, the handle may start to bend. The latter effect is due to the
fact that a bending wave travels down the racquet. The wave reflects off the
handle end and arrives back at the impact point just as the ball is leaving the
strings. The wave also reflects off the tip of the racquet and arrives back at the
impact point well before the ball leaves the strings. The speed at which a ball
bounces off the strings is determined partly by the mass of the racquet, partly
by amplitude of the bending wave, and partly by the time taken by the reflect-
ed waves from each end to get back to the impact point.

Even if the whole racquet vibrates as a result of hitting a

Figure 8.1 ball, there is a point near the middle of the strings that
The impact of a ball with a racquet doesn't vibrate at all. This point is called a vibration
causes it to bend when the ball hits node. The node is a special point and it has the proper-
near the throat (a), or near the tip (b), ty that when a ball impacts at the node it does not gen-
but not in the middle of the strings (c). erate any vibrations of the racquet frame, regardless of
the string tension or the impact duration. When a ball
impacts at the node point, there is very littlhe bending
Bending at Different Locations during the impact. Furthermore, the racquet straightens
out just as the ball leaves the strings, in such a way that
it does not overshoot the straight position and it remains
straight along its whole length after the ball leaves the
strings.

The effect of an impact at various positions along a rac-

quet is shown qualitatively in Figure 8.1. An impact in
the middle of the racquet causes the racquet to bend in
the middle with the racquet curved towards the ball. An
impact near the tip causes the racquet head to bend with
the racquet curved away from the ball. An impact
between the tip and the middle results in almost no
bending at all.
 Racquet Vibration Theory 75

These effects are consistent with detailed Impact Near Tip

bending calculations for impacts at various
points on a racquet, as shown in Figure 8.2.
Each curve shows the shape and location of
the racquet at 0.8 ms time intervals after the
ball first contacts the strings. The results were
calculated by modeling the racquet as a uni-
form beam of length 71 cm, assuming that the
racquet was initially at rest and suspended in
midair just before the impact without any sup-
port at either end. In practice, it is impossible
to levitate a racquet this way, but one could
support a racquet at one or both ends with a
long length of string to give the same effect.
The ball was incident at right angles to the -20 0 20 40 60 80
racquet, at a speed of 1.0 m/s, and it impact- x (cm)
ed at a point either near one end, or near the
vibration node 15 cm from the end, or at a
point 25 cm from the end. An incident speed impact at x = 12.5 cm
of 1.0 m/s is, of course, much lower than nor-
ae!
mal ball speeds in tennis, but the results can
simply be proportioned. For example, if the
ball is incident at 10 m/s then the y scale
should be multiplied by 10. The time intervals
are unaffected. The values of e, represent the
ratio of the rebound speed of the ball to the
incident speed.
Handle
Racqa ee odor Qube inpect
The whole racquet rotates about its center of
mass as a result of the impact, and it is also
pushed backwards, and it vibrates. The com-
bined effect of rotation about the center of
mass (in the middle of the racquet) and trans-
lation of the whole racquet is that the actual
axis of rotation is shifted towards the handle. Impact at Throat
For an impact near the throat, the axis of 8
rotation is actually beyond the end of the 15.6 ms
handle. In this case, the axis is not a well
: : nae : 6
defined point (as it is for an impact at the
vibration node) since the axis wobbles due to
ne
&
: iS
Figure 8.2 1 Handle
al
Impact of a tennis ball on a freely suspended
uniform beam of mass, 360 gm, and length, 71
cm, showing the position of the beam at 0.8 ms 0 - 0.4 ms
intervals. The ball is incident at speed v,, = 1.0
m/s and bounces off the beam at a speed v,,, =
@,V,,- The impact duration is 5.4 ms near the tip, -2
5.5 ms at x = 12.5 cm and 5.9 ms near the -20 0 20 40 60 80
throat.
 76 Chapter 8

the frame vibrations. There is one particular

Impact Near Vibration Node
impact point, known as the center of percus-
sion, where the axis of rotation is right at the
end of the handle near the wrist. For such an
impact, the hand rotates about an axis through
the wrist but there is no sudden motion of the
forearm either forwards or backwards. At
other impact points, the hand still rotates
backwards about an axis through the wrist,
but the whole forearm will jerk forwards or
(mm)
y backwards depending on the impact point.

There is a small residual vibration for an

impact near the vibration node, but there is
negligible vibration for an impact right at the
A Ball (e, = 0.40) node. The racquet does bend slightly when
the ball impacts near the node, but the rac-
-20 0 20 40 60 80 quet quickly recovers. Figure 8.3 shows
x (cm) details of the impact near (but not at) the
Figure 8.3 node point during the impact period itself.
Detail of impact near the vibra-
tion node during the first 4.4 ms.

A racquet is not a simple geometric shape,

having an approximately elliptical head and
an approximately cylindrical handle. Nevertheless, the vibrations of a racquet
frame are qualitatively similar to those for a simple rod, or beam, or indeed,
for any other relatively stiff object, even a knife, or spoon, or a baseball bat.
Whatever the shape, a bending wave can travel along the length of the object,
reflect off each end, and set up standing waves at certain well-defined fre-
quencies. A standing wave represents back-and-forth vibrations where one or
more sections of the object vibrates “back” while other sections vibrate “forth.”
More than one vibration frequency can be excited by an impact, like the over-
tones of a stretched string, but (a) the frequencies for a rod or beam are not
integer multiples of each other, and (b) the only vibration mode of any signif-
icance in a tennis racquet is the lowest frequency mode called the fundamen-
tal mode. The impact duration between the ball and the strings of a racquet,
around 5 ms, is too long to excite any of the higher frequency modes. The
higher frequency modes can be observed only if the frame is tapped with a
sharp object like a metal rod.

For a tennis racquet, the bending wave takes about 8 ms (0.008 s) to travel the
70 cm length of a racquet and back again. The wave travels at a speed of about
140 cm/0.008s = 175 m/s (about three times faster than a very fast serve). As a
result, the racquet vibrates once every 8 ms or about 120 times a second. The
velocity of the bending wave increases with the racquet stiffness and it decreas-
es as the racquet mass increases. As a result, the vibration frequency can be as
high as 200 Hz for a very stiff and light racquet. This is similar to the behav-
iour of a transverse wave on a stretched string where the wave velocity increas-
es with string tension and decreases as the mass of the string increases. For that
reason, strings at high tension vibrate faster, or at a higher frequency, than
strings at low tension, while thick, heavy strings of the same length, vibrate at
a lower frequency than thin, lightweight strings.
 Racquet Vibration Theory

The difference between a racquet and a string is that a racquet has its own
built-in stiffness and does not need to be tensioned to support a wave. For a
stretched string, the wave speed does not depend on wave frequency. By con-
trast, high frequency transverse waves propagating along a beam travel faster
than low frequency waves. This result can be understood in a qualitative sense
from the fact that it is easier to bend a long beam than a short beam. The beam
is effectively stiffer for a high frequency wave since high frequency waves have
a short wavelength, so the beam is bent over a relatively short distance.

In the case of a stretched string, the string is anchored at both ends and the
vibration frequency of the fundamental mode is determined by the fact the
string is exactly half a wavelength long, with a vibration node at each end of
the string. The speed of a transverse wave on the string depends only on the
string tension and the mass of the string. The vibration frequency of a stretched
string is given by the formula

wave speed = frequency x wavelength. (8.1)

This formula is derived from the fact that speed = distance/time. A wave always
travels a distance of one wavelength during one vibration cycle. The time for
one vibration cycle (the vibration “period”) is given by time = 1/frequency.

The same formula holds for any type of wave, including a wave traveling along
a beam. The node points in a beam are located at each end only if the beam
is anchored down at each end. If each end of a uniform beam is free to vibrate,
then the node points of the fundamental mode are located at x/L = 0.22 and
0.78 where x is the distance along the beam and L is the beam length. Even
though the speed of a transverse wave on a beam is a function of frequency,
there is only one frequency that has the correct wavelength with nodes at these
two spots.

The vibration of a rod or beam is a subject that is not treated in elementary

physics text books since it is a bit too complicated and specialized. We will
simply quote the results here without detailed explanation. The velocity of a
bending wave along a rod or beam is given by v = cf” where f is the wave
frequency (in Hz), and C is a constant which depends on the mass and stiff-
ness of the rod or beam. C is given by

C = [4n7EW(da)] (8.2)

where E is Young's modulus of the material, I is the area moment, d is the den-
sity, and A is the cross-sectional area of the material in the beam. The value of
I depends only on the cross-sectional dimensions of the rod or beam and
whether it is solid or hollow. For a solid rectangular beam of width b and
thickness a (in the bending direction), A = ab and I = ba°/12. For a circular rod
of radius R, I = nR“/4. For an aluminum beam or rod, E = 7 x 10” N/m? and
d = 2700 kg/m*. Graphite/epoxy composite materials used in modern racquets
have a larger value of E (about 10 x 10'° N/m? ) and a lower of d (about 1650
kg/m? ), meaning that they are stiffer and lighter.
78 Chapter 8

The vibration frequencies of a uniform rod or

beam of length L are given by

f = [GC/(2nL)]? (8.3)

First Three Modes of a Cantilever where the value of G depends on how the beam
is supported. The constant C can be regarded as
a measure of the stiffness to weight ratio. The
vibration frequency increases as C increases (stiff
ness increases or weight decreases) and as the
beam length decreases.

Cantilever beam

If a beam is rigidly clamped at one end and com-

pletely free at the other end, as in a cantilever,
then G = 1.875, 4.694, 7.853, 10.996 for the four
Figure 8.4 lowest frequency modes. The first three of these
modes is shown in Figure 8.4. The lowest fre-
First three vibration modes of a cantilever beam.
Each curve represents a picture of the beam at
quency mode is called the diving board mode.
one instant of time. Half a cycle later, the beam This is the mode that gets excited when you
bends in the opposite direction. The diving board clamp a ruler on the edge of a table and flick the
mode is the one that doesn't cross the y = 0 axis free end, or when a diver jumps off a diving
(dashed line). The highest frequency mode cross- board.
es the axis twice.
A simple experiment with a ruler clamped at one
end shows that the vibration frequency increases
as the overhang length decreases, and it varies as
ivAte Many different modes will be excited at
once if you flick the free end of the ruler, but the
most obvious one is the diving board mode. It is
much more difficult, if not impossible, to excite
this mode by holding one end of the ruler rather
than clamping it, especially if it is a standard 30
cm ruler. However, one can excite the diving
board mode by holding one end of a 1 m-long
ruler in two hands and giving the ruler a sharp
First Three Modes of a Free Beam flick of the wrist so that the ruler swings quickly
and then comes to a sudden stop.

Free beam

If the beam is free at both ends (or one end sup-

ported lightly by a length of string) then G =
Third Second First 4.730, 7.853 and 10.996 for the first three modes.
These modes are shown in Figure 8.5. The fre-
quencies are very similar to the previous case,
but the diving board mode is missing. The lowest
frequency mode with G = 4.730 (the fundamen-
Figure 8.5
tal mode) is the one that gets excited in a racquet
First three vibration modes of a free-free beam when you hit a tennis ball. The handle end is not
(both ends free). The fundamental mode has completely free, but neither is the hand rigid
nodes at x = 0.22 and x = 0.78
enough to act as a clamp. As shown in Figure 8.5,
 Racquet Vibration Theory 79

the fundamental mode has nodes located at x = 0.22 and 0.78 (where x is the
fractional distance along the beam). These are the points in the standing wave
that don't vibrate at all. Furthermore, if a ball impacts at these points, then the
fundamental mode does not get excited (see Figure 8.1 — no bending takes
place). Other modes can get excited, but the ball sits on the strings for too long
and effectively damps all the high frequency modes. In a tennis racquet, the
nodes are about 15 cm from each end of the racquet.

If the fundamental vibration frequency is 150 Hz, then from the formula
f=[GC/2nL]’, the next mode has a frequency of 150 x (7.853/4.730) = 150 x
2.76 = 413 Hz.

mpact with a Beam or

As shown by the graphs in Figures 8.2 and 8.3, the handle begins to rotate
before the ball leaves the strings. One might therefore expect that a hand-held
racquet would behave quite differently since motion of the handle would be
resisted by the hand. In fact, the ball bounces off the strings at the same speed,
regardless of whether the handle is free, or held in the hand, or clamped in a
vice. The hand itself is not as rigid as a vice, and offers very little resistance to
the motion of the handle during the time the ball is on the strings. One can see
this effect by dropping a ball on a hand-held racquet. No matter how firmly
the racquet is gripped, the wrist will always jerk backwards when the ball lands
on the strings. The wrist and the racquet eventually stop moving after the
impact is over (within about 0.05 sec), but the force of the hand on the han-
dle is much smaller than the sudden large force exerted by the handle on the
hand when the ball impacts the strings.

A vice will resist the handle motion completely and will prevent it moving at
all. The racquet is therefore unable to rotate or to translate, but it can still bend
and vibrate. Nevertheless, the ball bounces off the strings at the same speed as
when the racquet is completely free. This effect is very surprising, but there is
a simple explanation. When the ball impacts the strings, it causes the racquet
head to bend locally near the impact point. As a result, a bending wave starts
to propagate towards the tip of the racquet and also in the opposite direction
down to the handle. The wave takes about 8 ms for a round trip, so it takes
about 5 ms to travel down to the handle and back to the impact point. But the
ball leaves the strings after 4 or 5 ms, and doesn't receive any information from
the handle until it after it bounces off the strings. As far as the ball is con-
cerned, the handle plays no role in the bounce process and it may as well be
6 feet away. It is for this reason that the grip on the handle makes no differ-
ence to the bounce speed. However, it does make a difference if the tip of the
racquet is clamped in a vice. It takes only about 1 or 2 ms for a wave to reflect
off the tip and get back to the impact point. As a result, the ball bounces much
better off the tip if the tip is held in a vice.

When a ball impacts at the vibration node near the middle of the strings, the
frame recovers its original shape just as the ball leaves the strings, giving all its
stored elastic energy back to the ball. One might expect that the ball should
then come off the strings with the maximum possible speed. However, this is
80 Chapter 8

only one of several ways that energy is shared between the racquet and the
ball. One also needs to consider rotation and translation of the racquet to find
the impact point where the ball speed is maximized. In general, the maximum
power point is indeed near the middle of the strings, but the speed off the tip
or throat is only slightly smaller. There is no dramatic increase in power when
the ball impacts at the vibration node. The maximum power point shifts
towards the tip of the racquet if the incident ball speed is smaller than the rac-
quet head speed, and it shifts towards the throat if the incident ball speed is
larger than the racquet head speed. In the latter case, the speed off the throat
is a lot larger than the speed off the tip, as one can easily verify by dropping
a ball on a stationary hand-held racquet.

References

R. Cross, Impact of a ball with a bat or racquet, American Journal of Physics,

67, 692-702, (1999).

R. Cross, Flexible beam analysis of the effects of string tension and frame stiff-
ness on racket performance, Sports Engineering 3, 111-122 (2000).

A. Nathan, Dynamics of the baseball-bat collision, American Journal of Physics,

68, 979-990 (2000).
 What Makes
One Racquet Feel
Different From Another?

Leyton Hewitt, Aus

By Rod Cross

espite the fact that most racquets these days look much the same and are made from the same basic
@ materials, every racquet feels different from the next. The reason is that no two racquets are the same in
™ every respect. The most obvious differences are in the weight and headsize. Other differences that con-
tribute to racquet feel include the swingweight, the twistweight, trame stiffness, stringbed stiffness, head shape,
handle grip, sweetspot size, and shock absorbing properties. The most relevant property, in terms of feel, is
probably the stiffness of the frame since it determines the amount of shock and vibration generated on impact
with the ball. The amount of shock and vibration depends also on where the ball makes contact. If the ball con-
tacts the middle of the strings, then there is almost no vibration in any racquet and only minimal shock, regard-
less of the frame stiffness. In that case, the feel of the racquet will depend mainly on its weight, swingweight,
and string tension. These factors determine how easy it is to swing the racquet and how far and how fast the
wrist rotates backwards when the racquet strikes the ball.

A common view is that a racquet with a stiff frame will feel stiff and will generate more shock at impact. In fact,
measurements show that stiff racquets vibrate less than flexible racquets and generate less shock, not more. A
flexible frame bends further than a stiff frame on impact with the ball. As the ball leaves the strings, the frame
springs rapidly back to its original shape, but it may overshoot and then bend in the opposite direction. The
frame will then vibrate, bending back and forth rapidly through a large distance if the frame is flexible, but only
82 Chapter 9

a small distance if the frame is stiff. The high level of vibration in flexible rac-
quets adds to the total shock or jarring sensation experienced during impact.
The only exception would be an extremely flexible racquet made from some-
thing like rubber. A rubber racquet would have very little power, but it would
feel really soft since the handle itself would be softer than the hand, and it
would bend just beyond the end of the hand. The handle in a modern racquet
is a lot stiffer than the hand and it does not bend in this way. Consequently,
any sudden motion of the handle arising from an impact tends to jar.

Even though stiff frames reduce jarring, stiff strings normally increase jarring
since they cause the handle to vibrate more and to slam into the hand at a
higher speed if the player misses the sweetspot. The strings are not nearly as
stiff as the frame and have very little effect on frame stiffness. The function of
a soft stringbed is to reduce the impact force on the racquet, in the same way
that jumping on a trampoline feels nicer than jumping up and down on con-
crete. You could hit a ball at about the same speed with a solid wood paddle
or bat, but the strings are there to reduce the jarring effect and to reduce the
drag force through the air when you swing the racquet (see Note 1).

As string plane stiffness decreases, the impact force on the frame decreases,
and hence, the racquet bends less. The force on the ball is also reduced, but
it acts over a longer time with the result that the ball speed is a tiny fraction
larger than the speed off stiff strings. The fact that the ball sits on soft strings
for a longer time also helps reduce frame vibrations. In effect, the ball helps to
damp both the string and the frame vibrations. This effect is quite significant
and is explained in more detail below. The most comfortable racquet is, there-
fore, one with a stiff frame strung at low tension. In fact, one can eliminate
frame vibrations completely if the frame is stiff enough, and light enough, and
if the string tension is low enough. As described below, one of the racquets
tested was almost ideal in this respect. It vibrated slightly for impacts near the
tip, but frame vibrations were almost undetectable for impacts in the throat
area. This is not necessarily a good thing. It depends on individual taste. Some
players like feeling the vibrations since it gives them some feedback on where
they are striking the ball and whether they hit a good shot. But too much vibra-
tion gives “bad vibes” to most players.

Professional tennis players normally string their racquets at high tension rather
than low tension, despite the fact that an impact away from the middle of the
strings will feel harsh. However, professionals are generally more consistent
than the average player at hitting the ball in the middle of the strings. For the
professional player, shock and vibration is not such an important issue. In fact,
if a player hits the ball in the middle of the strings, high string tension will
reduce the shock caused by sudden rotation of the wrist.

An impact in the middle of the strings produces essentially no vibrations and

no sudden movement of the handle in one direction or the other. What you
feel is due almost entirely to rotation of the racquet. The whole racquet rotates
about an axis through your wrist when you hit a ball in the middle of the
strings, even if you “lock” your wrist. Some players grip the handle loosely and
use a lot of wrist rotation when hitting a ball. Others grip the handle more firm-
ly to prevent the wrist rotating when swinging the racquet. Wrist rotation is
used by squash players and can also be used by tennis players to good effect.
 Racquet Feel 83

Wrist rotation just before impact can be controlled by the player since racquets
are reasonably light. However, wrist rotation during the impact itself cannot be
controlled since the impact force is so large. The force of the ball on the strings
is typically around 100 lb (45 kg or 450N) and can be as large as 300 lb in a
very fast serve. Try lifting a 5 lb weight on the end of your racquet using wrist
action alone. If you have a strong wrist you may be able to lift a 10 Ib weight,
but no one can lift a 20 lb weight, let alone 100 or 300 lb. The impact force
causes the wrist to rotate backwards, but not very far since the force acts for
less than one hundreth of a second. Heavy racquets, or racquets with a large
swingweight, will resist this rotation better than light racquets. Racquets strung
at high tension will also rotate less on impact, despite the larger impact force,
since the dwell time on the strings is shorter (see Note 2 for an explanation).
This should also give a player better control on the angle of the ball coming
off the strings.

le Stiffness and Vibration Experiments ©

The effect of stiffness on vibration can be demonstrated by some simple exper-
iments that you can do yourself. Take a ruler and hold, or clamp, one end firm-
ly on the edge of a table so that the other end extends past the edge of the
table. If you pluck the far end, the ruler will vibrate back and forth at a fre-
quency of about 10 cycles per second. This mode of vibration is called the div-
ing board mode, since one end vibrates like a diving board while the end does-
n't move at all. If you decrease the length of the ruler extending past the edge
of the table, then the ruler becomes stiffer (meaning harder to bend), the vibra-
tion frequency increases (you hear a higher pitch tone), and the vibration
amplitude decreases (the tip moves back and forth through a smaller distance).

Now hold one end of the ruler in one hand and pluck the other end. If your
hand is stiffer than the ruler, the ruler will vibrate just like it did when you
clamped it on the table. However, most rulers are stiffer than the hand and it
is usually impossible to get them to vibrate in this way. Racquets are even
stiffer than rulers, but they can still vibrate when they are used to hit a tennis
ball. If the handle of a tennis racquet is clamped firmly in a vice (but not too
firmly or it will break) then it will vibrate like a diving board at low frequen-
cy, around 20 cycles per second. In that case, the handle doesn't vibrate at all.
When the handle is held by hand and the racquet is used to strike a ball, the
racquet vibrates at a much higher frequency, around 100 to 200 cycles per sec-
ond: That is a different mode of vibration, where both ends of the racquet
vibrate strongly, even though the handle is held firmly in the hand. The hand
is not stiff enough to act like a rigid clamp, and both ends of the racquet vibrate
simultaneously. The hand and the forearm also shake back and forth when the
racquet is vibrating in this manner. The mode of vibration of a hand-held rac-
quet can be demonstrated by holding a long, thin rod, or a length of stiff wire,
or a narrow strip of cardboard in the middle and shaking it back and forth.
These objects are very flexible and each end will vibrate back and forth at only
one or two cycles per second. In fact, they are flexible enough to support both
the low frequency diving board mode (when held at one end) and the higher
frequency tennis racquet mode.

surements of Shock and Vibration —

The language used to describe feel is not very precise. Every racquet “jars”
every time it is used to hit a ball, but exactly what happens depends on the
84 Chapter 9

Wilson Rollers 2.6

Wilson Triad 3.0 Voc toatl
ea en epee) (a Vi
250
|
ac ae Treat i (ig ae ae
250 Impact distance Impact
distance
from tip from tip
al
Toalents 4 cm
4cm
200 200

150

Ss 100 qd
& E 16cm |
S 100
ow zy £
=
D ie 20cm — G
in 50 Pe NNN SI en ee ee {sa
Boy
D0)
2 c
a UNs
N [ 24cm 4
oO
Oo Or Es N
o
a

-50
-100 =
= WO tak f= 137 kz

eee |e epee 4 ett Son 2 (ore i L ee | a

-150 -100
-20 0 20 40 60 80 100 -20 0) 20 40 60 80 100
t (ms) t (ms)
Figure 9.1 Figure 9. 3

Slazenger Challenge (Wood) Prince TT Viper

250
Impact Impact distance from tip
distance 6cm
from tip
4 cm 100.0

200

L 8 cm

= aa = 18cm
Ss
=
Si
L 12 cm
zD 22cm
42) no
no fo}
(e) 9 0.0
a
fall - 26 cm

r 20 cm
30 cm 4

50 L 22 cm -50.0

34 cm ]

[f = 87 Hz Center of strings 14 cm from tip

f = 202 Hz
0 SA
ah es areal en rere ola -100.0 a | a ee ee| re |

-20 0 20 40 60 80 100 0 20 40 60 80

t (ms) t (ms)
Figure 9.2 Figure 9.4
 Racquet Feel 85

Table 9.1

Racquet Mass Length Frame stiffness Stringbed stiffness

Wilson Triad 3.0 291 gm 699 mm 70-76 DA 62 DA

Prince TT Viper 261 gm 697 mm 81 DA 60 DA
Slazenger Challenge 374 gm 685 mm <0 DA 45 DA

circumstances. It could be a sudden motion of the handle in one direction or

a sudden vibration back and forth or a combination of both and it might last
for a short time or it might last for a longer time. To help sort this out, shock
and vibration measurements were made on a number of different racquets
using a piezo sensor attached to the handle near the butt end. The piezo gen-
erated a voltage signal proportional to the acceleration of the handle. The rac-
quets were not held in the hand but were suspended at rest from a horizontal
beam using two lengths of string. A tennis ball was also suspended from the
beam as a pendulum so that the ball hit a selected spot on the strings at the
same speed on every spot and on every racquet. The method is described in
more detail in the side box “How To Measure Vibrations.” Results for four of
the racquets are shown in Figures 9.1-9.4 and the properties of each racquet
are indicated in Table 9.1. The Slazenger racquet was an old wood racquet
known from its feel to vibrate a lot more than the other three modern graphite
racquets. The frame and stringbed stiffness was measured on a Pacific diag-
nostic machine (higher reading = stiffer). The wood racquet was so flexible that
the frame stiffness reading dropped below zero.

Results of measuring handle motion of the Triad racquet are shown in Figure
9.1 for eight different impact points on the strings, all lying on the long axis
through the handle. These results are typical of those for all racquets in that:

1. The level of vibration in the handle is largest for impacts near the tip or the
throat of a racquet.

2. The vibration level in the handle is zero for one particular impact spot near
the middle of the strings. This spot is called a vibration node.

3. The impact of the ball on the strings lasts about 5 ms, after which the ball
bounces off the strings. During this time the racquet head accelerates back-
wards, but the handle itself can accelerate forwards or backwards depending
on the impact point. If the ball hits the strings near the tip, then the handle
accelerates forwards and the voltage signal from the piezo is large and posi-
tive. If the ball hits in the throat area, the handle accelerates backwards, and
the signal from the piezo is large and negative. This can be seen by the rever-
sal in direction of the first part of the piezo signal. There is an impact point
between the tip and the throat where the handle doesn't accelerate forwards
or backwards at all. This point is called the center of percussion (COP) and it
is located close to the vibration node. However, every point in the handle has
a different COP on the strings. For the results in Figures 9.1-9.4, the piezo was
located 50 mm from the butt end of the handle and the COP is then only a few
mm from the vibration node.
86 Chapter 9

4. The region extending from the node to the COP is the sweetspot region
since the total shock and vibration is minimized. An impact at the node results
in zero vibration, but there is a tiny amount of shock since the handle has a
small acceleration forwards. An impact at the COP generates zero shock, but
there is a small level of vibration generated in the handle.

The sweetspot is located about 16 cm from the tip of all racquets. The
sweetspot was located near the middle of the strings for the three graphite rac-
quets, but it was closer to the throat in the wood racquet since the headsize
was smaller. In all cases the sweetspot could be located within 5 mm. The
sweetspot did not fill up the whole string plane, or even half of it, or even 10%
of it. The level of vibration remained small over a distance about 1 cm either
side of the sweetspot in the wood racquet and over a larger distance in the
graphite racquets. In fact, the Prince Viper was very surprising, with almost no
vibration anywhere past the middle of the strings, and with no vibration at all
in the throat. The explanation of this provides the key to understanding rac-
quet vibrations. Racquet manufacturers have tried all sorts of tricks to reduce
frame vibrations, and the Viper demonstrates how it can actually be done.

One obvious difference between the wood and the graphite racquets is the
vibration frequency. The graphs show that one complete cycle of oscillation
takes 11.4 ms for the wood racquet, but it takes only 6.3 ms for the Triad and
5.0 ms for the Viper. The wood racquet vibrated one cycle in 0.0114 seconds.
This is equivalent to 1/0.0114 = 87 cycles per second, commonly designated as
87 Hz (Hz is short for Hertz). The Triad vibrated at 157 Hz and the Viper at 202
Hz. The vibration frequencies were measured with wood blocks on the handle
to hold the piezo and to simulate the weight of the hand. If the wood blocks
are removed, the racquets vibrate at slightly higher frequencies, but not a lot
higher. The different vibration frequencies
Figure 9.5 makes a big difference to the feel of each rac-
quet. The handle vibrates back and forth in the
hand, reversing direction 157 times a second
50 with the Triad, but only 87 times a second with
the wood racquet. But the biggest difference in
feel is due to the size of the vibrations.
40
The piezo signals shown in Figures 9.1-9.4 rep-
Throat resent the acceleration of the handle as a func-
tion of time. The actual distance traveled back
oe) S)
and forth by the handle during each cycle of
oscillation is shown in Figure 9.5 for each of the
four racquets. The distance (technically called
Ne) (=) Triad the vibration amplitude) is given by the accel-
eration divided by the vibration frequency
Vibration
amplitude
Rollers squared. The first positive or negative peak in
ay oO
a
the piezo signals was ignored when calculating
the vibration amplitude since the first peak is
Viper potas eene due mainly to sudden rotation of the handle
rather than vibration. The accelerations for all
four racquets are similar in magnitude, but the
15) 0 5 me) ToreZ0" 25 3310) 35
vibration amplitudes are a lot different. The
Impact distance from tip (cm) wood racquet vibrated with much _ greater
 Racquet Feel 87

How To Measure Vibrations

he arrangement used to measure motion of the handle is shown in Figure 9.6. A racquet was supported by two lengths
of string tied to a metal frame at the top end and to two points on the racquet at the bottom end, one at the 3 o'clock
position and one near the butt end of the handle. This is obviously not the same as swinging a racquet on the court, but
it still gives useful and relevant information. One problem with swinging a racquet is that the handle must be held in the hand,
just where an instrument needs to be located to measure the vibrations in the handle. Another problem is that any instrument
on the handle needs to be connected with wires to a measuring system, and that interferes with the ability to swing the racquet.
The way around both of these problems is to suspend the racquet freely, with a force gauge or an accelerometer located on the
handle, and then allow a ball to impact on the strings. The racquet swings around and vibrates, but that is just what we need to
measure. The sudden motion of the handle generates the shock that you feel as the handle suddenly slams into, or tries to pull
out of, your hand. The vibration of the handle is what you feel after the initial shock is over.

The simplest and cheapest force gauge is a piezoelectric disk. This type of disk is used in piezo buzzers and in musical greet-
ing cards and it costs about $2. Normally, one applies a voltage to the disk which causes it to vibrate and to generate a buzzing
noise. A piezo disk can also be used the other way around. If the piezo is made to move or vibrate, it will generate a voltage
that can be measured on an oscilloscope. One can attach a piezo disk directly to the handle of a racquet to measure move-
ment or vibration of the handle, but the size of the voltage signal generated depends on how well the disk is attached to the
handle. It also depends strongly on whether the disk remains flat or whether it bends slightly when it is attached to the handle.
A piezo disk is typically about 0.3 mm thick, about 1/2 in. diameter, and is made from ceramic. Since racquet handles are not
perfectly flat, and since every handle is different from the next, the best approach is to attach a flat surface firmly to the han-
dle and then attach the flat disk to the flat surface. That way, the result does not depend on the exact shape of the handle in
each racquet or on how firmly the disk is attached to the grip.

The arrangement used to attach a piezo to the handle is shown in Figure 9.7. Two identical wood blocks were shaped to fit over
the handle and two steel bolts were used to clamp the blocks firmly onto thin rubber pads on the handle to obtain good con-
tact. A 16 mm diameter piezo disk was taped firmly onto one of the wood blocks. The total weight of the blocks and the bolts
was 170 gm, about the same as a human hand. In this situation, the piezo works as an accelerometer. If the handle acceler-
ates, a voltage is generated by the piezo. If the handle is accelerated from rest to a constant speed, then the voltage drops to
zero while the handle is moving at constant speed.

The arrangement used to impact a ball on the strings is also shown in Figure 9.6. The ball was mounted as a pendulum so that
it could impact precisely at any selected spot and at exactly the same speed every time, regardless of which spot or which rac-
quet was chosen. To control the impact speed, the pendulum was moved back off the strings by the same amount each time,
using a mechanical stop to control the distance off the strings. When the ball was released, it swung towards the racquet and
hit the strings at a fixed speed. The ball hit the strings at a low speed, much lower than the normal impact speed in tennis, but
the location of the sweetspot has nothing to do with impact speed. The vibration level is larger at high racquet or ball speeds
but the issue of interest is how the vibration level changes when the impact point is varied, and not how it varies when the ball
speed changes.

Figure 9.6 Figure 9.7

88 Chapter 9

amplitude than the graphite racquets. The Triad vibrated much less than the
wood racquet, but more than the Viper, at least at the beginning. The Triad has
a vibration dampening rubber strip glued between the head and the V-section
at the top end of the handle. As a result, the frame vibrations are damped much
faster than in the other racquets.

It is possible to measure the amount of vibration in a racquet, how it varies

with the impact point on the strings, and how it varies from one racquet to the
next. This provides a better indication of the actual properties of a racquet than
the information commonly provided by the manufacturers. The manufacturers
usually claim that their sweetspot is bigger and that there is less shock and
vibration, but it would be much more useful if the manufacturers provided
actual numbers on shock and vibration, in addition to the usual information
provided on the weight, length and headsize of each racquet. In fact, only one
number is needed, as we will see below. The number that is missing from the
advertising brochures and magazines is the vibration frequency of the racquet.

The measurements show very clearly that as racquet stiffness increases, the
vibration frequency of the racquet increases, and the vibration amplitude
decreases. The amplitude drops to zero when the vibration frequency of the
racquet is about 200 Hz or higher, depending on the impact duration and
hence on the string tension and the impact point. A surprising result is that it
is not necessary for a racquet to be infinitely stiff to stop all vibrations. It is nec-
essary only that f > 1/T, where f is the vibration frequency, and T is the time
that the ball is in contact with the strings before it bounces off. This time is
known as the dwell time, or the impact duration. For example, if T = 0.005 s
then f must be greater than 1/0.005 = 200 Hz. Most racquets vibrate at around
140 Hz. Widebody racquets vibrate at around 170 Hz, but 200 Hz racquets are
rare.

Most tennis players like to hit the strings of their racquet with their hand, or
with another racquet, to check the string tension. Provided there is no string
dampener, a nice ping tells them if the strings are tight or loose. Hardly any-
one listens to the racquet itself. If you tap the frame or the strings with the end
of your finger and hold the handle near your ear, you will hear the frame
vibrating. It works best if you hold the handle lightly about 6 inches from the
butt end. Holding it anywhere else will damp out the frame vibrations. A low
pitch means the frame is heavy and flexible and a high pitch means the frame
is stiff and light. More than anything else (apart from hitting the ball) listening
to pitch will tell you how much vibration you can expect from the racquet. A
medium pitch could mean that (a) the frame is heavy and stiff, or (b) it is light
and flexible, or (c) it is medium in both weight and stiffness, but it is easy to
figure out which is which just knowing the weight or the “flex” (stiffness) of
the frame.

_Sweetspot Softness and Fe«

Racquet feel before impact is related primarily to the weight and swingweight
of the racquet. The feel during and after impact is related to the sudden motion
and vibration of the handle arising when a player hits the ball. When you hit
a ball, what you feel is not the ball pushing on the strings, but the handle push-
ing against your hand. The impact of the ball on the strings lasts for less than
 Racquet Feel

one hundreth of a second. This results in a sudden change in the motion of

the whole racquet. The racquet head slows down more than the handle, with
the result that your hand rotates backwards about an axis through the wrist. It
takes about 50 ms after the impact for the racquet to stop vibrating and it also
takes about 50 ms for your hand to stop rotating backwards. The initial sud-
den motion of the handle at impact generates a shock that depends primarily
on where on the strings the ball is struck. Maximum shock occurs when the
ball impacts near the tip or throat of the racquet. Minimum shock occurs at a
sweetspot near the center of the strings called the center of percussion. Part of
the initial shock is due to rotation of the racquet and part is due to the fact that
the racquet also starts vibrating. Maximum vibration occurs for an impact near
the tip of the racquet and zero vibration occurs at the vibration sweetspot. The
two sweetspots are quite close together and are almost impossible to distin-
guish in terms of feel. They can, however, be distinguished in the laboratory
by careful measurements of handle motion. In practice, an impact at the vibra-
tion sweetspot will result in a barely perceptible amount of shock and an
impact at the center of percussion generates only a tiny amount of vibration.

If the handle slams into your hand at impact, the shot does not feel as nice as
when the handle barely moves. Excessive vibration of the handle does not feel
nice either. Minimum motion of the handle due to shock and vibration occurs
when the ball strikes the sweetspot near the middle of the strings. Even then,
not all racquets feel the same. Some sweetspots seem to feel different than oth-
ers. The difference is mainly in the stiffness or the softness of the sweetspot.
Several factors contribute to sweetspot softness. If the racquet has a soft
sweetspot, the ball stays on the strings a bit longer, and the force on the hand
will be smaller. This is partly due to differences in the string tension or gauge
or type of string or headsize, and partly due to the stiffness of the racquet itself
or the way the stiffness varies along the length of the racquet. A racquet can
be stiff at one end and soft at the other, or it might be stiff at both ends and
soft in the middle.

Comfort by itself is nice, but good players prefer to win points regardless of
the comfort. If that means using a bone-rattler, then some players might like
the way it feels simply because that's what works best for them. Alternatively,
a player might like a racquet because it feels solid or powerful rather than
because there is very little shock and vibration. Very stiff and light racquets give
the least shock and vibration, but they are not necessarily as powerful as a
heavier racquet.

Another difference between racquets could be the size of the sweetspot. The
sweetspot of a tennis racquet is not actually a tiny spot, but neither is it as big
as some manufacturers would have you believe. Some manufacturers show
diagrams where the sweetspot on their racquet is almost as big as the whole
stringbed. To most of us, the sweetspot is a small region near the middle of the
strings where the impact with the ball feels best. In this region, there is almost
no shock or vibration transmitted to the hand or the arm. It feels almost as if
you didn't actually hit the ball and you just swung at the air. It is not really that
good, but it feels much better than a miss-hit where your arm and elbow and
hand get tugged and shaken. If you hit the ball near the sweetspot, it still feels
pretty good, but the further away from the sweetspot, the worse it feels, and
the closer to the sweetspot, the better it feels.
90 Chapter 9

Suppose you rank the spots on your racquet on a scale from 0 to 10 where 0
feels terrible and a 10 is perfect. A 9 is almost perfect and an 8 is pretty good.
By definition, a 10 is a single spot. It is the only spot that feels perfect. But for
all practical purposes, you would probably not be able to pick the difference
between say a 9.5 and a 10. In that case, the spot is slightly bigger than a sin-
gle point. In fact, the area that feels really good is probably about the size of
a tennis ball. All points inside this area would be around a 9.5 or higher.

It seems that some manufacturers include all points that are about a 3 or above,
since their sweetspot is so large. The only way to find out for sure is to meas-
ure the size of the sweetspot. One way to do that would be to mark a ball with
colored chalk so you can tell where on the strings you hit the ball. That would
tell you how big your sweetspot is for that racquet, but a shot that feels nice
to you might feel less than nice to someone else. A better method is to meas-
ure the actual shock and vibration in the handle when the ball impacts the
strings at various spots. :

For any given impact point on the strings, the acceleration of the handle was
about the same for the three graphite racquets, since the ball speed was the
same, and since the weight and the swingweight of each racquet was about
the same. The wood racquet accelerated more slowly since it was heavier and
the swingweight was larger. The sudden motion of the handle due to the ball
striking the strings results in a sudden shock that is felt by the hand and arm.
The nature of this sudden motion is best illustrated by the results in Figure 9.4
for the Viper racquet. The vibration amplitude is much reduced for this rac-
quet, but the impact shock occuring in the first 5 ms is still clearly visible. For
the other racquets the effects of impact shock and vibration tend to merge
together and probably feel about the same, resulting in a jarring feel that is part
shock and part vibration.

The amount of shock depends on the impact point on the strings and it increas-
es with the ball speed, but it also depends on the mass and stiffness of the rac-
quet. Light racquets accelerate faster and will slam into the hand faster, but
since they are light, the shock experienced by the hand is likely to be smaller
than that felt by a heavier but slower racquet. To exaggerate the point, the
shock felt by a heavy truck hitting you at low speed would be much greater
than the shock felt by a small insect hitting you at high speed.

The shock of an impact depends on the following factors:

1. A ball hitting a light racquet will usually bounce at a lower speed than a ball
bouncing off a heavy racquet if swing speed is the same or the racquet is sta-
tionary. As a result, the momentum transferred to a light racquet is less than
the momentum transferred to a heavy racquet. Your hand absorbs this momen-
tum by bringing the racquet to rest, and it will therefore absorb a greater
momentum from a heavy racquet.

2. The shock depends on how big a force is exerted on your hand, and that
depends on how long the force acts. For example, suppose you catch a base-
ball in your bare hands. You can lessen the shock by allowing your hands to
move backwards as you catch the ball. That way you can stop the ball by exert-
ing a smaller force for a longer time. If you don't move your hands backwards
 Racquet Feel 91

then the ball will slam into your hands with a bigger force and the shock will
be greater even though your hands absorb exactly the same momentum in
both cases.

3. The shock felt through the handle of a racquet might depend as much on
the softness and thickness of the grip as on the weight or stiffness of the rac-
quet itself because it affects the duration of the force acting on your hand.

4. Soft strings in a stiff frame tend to give a soft feel, while stiff strings in a flex-
ible frame tend to give a stiff feel. Experienced players can usually pick
whether it is the frame or the strings that are stiff. Stiff strings will give a short
impact duration, while a stiff frame will vibrate less and at a higher frequency
than a flexible frame.

5. Some racquets are stiff in the head and soft in the handle, others are the
other way around and some are soft in the middle. Imagine hitting a ball with
a racquet made from stiff graphite at one end and soft rubber at the other. If
the racquet has a stiff head and a flexible handle, the shock transmitted to the
hand will be quite small.

The speed at which the handle slams into the

hand at impact can be calculated from the Handle Recoil Speed
results shown in Figures 9.1-9.4. The piezo sig-
nals represent the acceleration of the handle 150
during and after the impact, starting with the
handle at rest. The racquet was not swung
towards the ball, but the other way around in Rollers
this experiment. The racquet was suspended 100
horizontally from a beam with two lengths of
string and the ball was swung onto the strings
as a pendulum. This way, the impact speed of
Handle moves
the ball could be kept constant for every impact on j=)
towards ball
point, and the impact point itself could be accu-
rately controlled. In effect, this is what an insect
would see if it was sitting on the handle of a
racquet that is swung towards the ball. The =)

insect thinks that it is at rest and sees the ball

coming in and bouncing off, but the insect itself
jerks forwards or backwards depending on speed
recoil
Handle
where the ball hits the strings.
i]
o1 S Handle moves
away from ball
Speed is acceleration multiplied by time, pro-
vided the acceleration remains constant. In
Figures 9.1-9.4 the acceleration is not constant, -100
but the handle speed can be found by calculat-
ing the area between the acceleration curves Throat
and the time axis. This was done on a comput- 150
er and the results are shown in Figure 9.8. The 0 5 10
19 20 PRS) 30 35
results plotted in Figure 9.8 give the maximum
handle speed for each impact point on the Impact distance from tip (cm)
strings. The handle speed increases to a maxi-
mum just as the ball leaves the strings, and then
Figure 9.8
92 Chapter 9

decreases and increases rapidly with time due to vibration of the handle.
Vibration of the handle acts to increase the speed of the handle just as the ball
leaves the strings. At this instant (or shortly afterwards) the handle has its max-
imum rotation speed and it also has its maximum vibration speed, both caus-
ing the handle to move in the same direction at the maximum possible speed.
It would be nice if the handle rotated one way and vibrated the other way, but
this would just delay the time at which the handle reached its maximum speed.
If rotation and vibration are in opposite directions during the first half cycle of
the vibration, then they are in the same direction during the next half cycle.

For an impact near the tip of the racquet, the handle swings around towards
the incoming ball and the handle speed is taken to be positive. For an impact
near the throat, the handle recoils away from the incoming ball and the han-
dle speed is taken to be negative. The impact point where the handle speed
remains zero during the impact is the center of percussion. Figure 9.8 shows
that:

(a) the handle speed is greatest for impacts near the tip and throat of the rac-
quet,

(b) the handle speed is about the same for all four racquets, and

(c) the center of percussion (where the handle speed is zero) is located about
16 cm from the tip for all four racquets, but is a bit closer to the tip for the
Viper and a bit closer to the throat for the Triad.

The handle of the heavy wood racquet actually rotates at the slowest speed of
all four racquets, but it also vibrates the most. As a result, the combined speed
of the handle due to rotation, translation, and vibration is just as large, if not
larger, than the other racquets. The Triad was modified before these measure-
ments by adding an extra 7 gm grip to the handle. As a result, the center of
percussion shifted slightly towards the throat.

When a tennis ball bounces on a tennis court it spends only about 4 ms in con-
tact with the surface. The time depends on the stiffness of the ball, but it does-
n't depend much on ball speed. A high speed ball might spend 3.5 ms in con-
tact with the surface, compared with about 4.5 ms for a very slow ball. The
decrease in time for a faster ball is due to the fact that a ball gets a bit stiffer
when it is compressed a long way. A soft ball will remain in contact with the
surface a bit longer. Similarly, the contact time on the strings of a racquet is
extended slightly, to about 5 or 6 ms, due to the softness of the strings. The
time is about 5 ms on a stiff stringbed and about 6 ms on a soft stringbed. As
a result, the force acting on a stiff stringbed is larger than on a soft stringbed.
The racquet head is pushed backwards by the impact and the whole racquet
rotates about an axis through the wrist. The rate of rotation is relatively slow
for an impact on a soft stringbed, but the racquet will rotate further during the
impact.

Suppose that a ball bounces at the same speed off stiff strings as it does off
soft strings. That is not quite correct, but the difference is only a few percent.
If the bounce speed is the same, then the change in momentum of the ball is
the same, and the momentum given to the racquet is the same. If the racquet
 Racquet Feel 93

mass is also the same then the speed of the racquet at the end of the impact
will also be the same since momentum = mass x speed. On soft strings, the
force on the ball or the racquet is relatively small, but it acts for a long time.
The soft-string racquet reaches the same speed as the stiff-string racquet, but it
took longer to get there so it traveled a greater distance. The hand absorbs the
momentum of the racquet in much the same way as it does when you catch a
ball. The shock of the impact will therefore be less using a racquet with soft
strings.

The results for the Viper racquet are especially interesting since there are almost
no vibrations at all for an impact in the throat region, and since the vibration
amplitude is relatively small for an impact everywhere on the stringbed. It is
easy to understand that there are no vibrations at a vibration node in any rac-
quet, regardless of how stiff or flexible it is, but the Viper has a small vibration
level everywhere. Vibrations result from the fact that a racquet bends suddenly
at the impact point, especially if the ball impacts near the tip or throat of the
racquet. The only exception is an infinitely stiff racquet which will not bend at
all, and hence, there will be no vibrations. The Viper seems to be almost infi-
nitely stiff in the throat region and less stiff near the tip. It is true that this rac-
quet is the stiffest of the four tested, but it is only slightly stiffer than the Rollers.
So why should the vibrations disappear almost completely for an impact at the
throat?

The answer is related to the vibration frequency of the racquet and the dura-
tion of the impact. If a golf ball is dropped on the strings of the Viper in the
throat region, then the racquet frame vibrates just like any other racquet. The
vibration amplitude is smaller in the Viper than when a golf ball is dropped on
the other racquets, but at least there is a relatively large vibration, and it is
almost as big as the initial shock caused by sudden motion of the handle. The
difference between a golf ball and a tennis ball is that a golf ball is lighter and
stiffer. As a result, a golf ball bounces off the strings sooner than a tennis ball.
Furthermore, it bounces off the strings before it has any significant effect on
damping the vibrations. A tennis ball sits on the strings long enough to act as
a vibration dampener.

A racquet vibrates at a particular frequency because it takes a certain time for

the initial bending wave to travel along the frame, bounce off the ends and
return back to the starting point. The speed of the wave depends on the stiff-
ness and lightness of the frame. If the wave travels at a speed of 140 m/s along
a 70 cm long racquet it will bounce up and down the racquet 100 times per
second and the racquet will vibrate at 100 Hz. If the wave bounces up and
down 200 times per second, then the racquet vibrates at 200 Hz. If the ball is
pushing on the strings, then it acts in the same way as putting your finger on
a vibrating guitar string. It damps the vibrations. The longer your finger is on
the string, the more it damps the vibrations. A wave that travels up and down
the racquet 100 times per second takes 1/100th of a second = 10 ms for the
round trip or 5 ms from one end to the other. If the ball bounces off the strings
after 5 ms, it doesn't get in the way of the wave, so the wave keeps propagat-
ing up and down the racquet and the racquet keeps vibrating. But it takes only
2.5 ms for a one-way trip when the racquet vibrates at 200 Hz and a tennis ball
DOES get in the way. It damps the frame vibrations strongly. A golf ball sits on
the strings for a shorter time and it doesn't damp the vibrations as strongly. A
94 Chapter 9

similar effect occurs with a guitar string or a piano string. A string vibrates
strongly if it is plucked or given a quick blow, but it won't vibrate strongly if
you push it sideways and release your finger slowly. The finger or impact ham-
mer must be removed from the string very quickly, otherwise it damps the
vibrations.

The vibration frequency of a racquet frame is typically about 140 Hz, but it is
higher in stiff racquets and lower in flexible racquets. The impact between a
ball and a racquet lasts for about 5 ms (0.005 seconds). The impact duration is
longer at low string tension, shorter at high string tension, and increases as the
impact point moves from the tip to the throat of the racquet. If F = vibration
frequency and T = impact duration, then the impact will cause the racquet to
vibrate if F is less than 1/T, but the racquet will not vibrate if F is larger than
1/T. Most racquets do vibrate since 1/T is typically about 1/0.005 = 200 Hz. It
is not necessary for a racquet to be infinitely stiff for the vibrations to disap-
pear. All racquets have a sweetspot near the middle of the strings where no
vibrations are excited, but frame vibrations are excited at all other impact
points, especially near the tip and throat of a racquet.

The above explanation accounts for the fact that there are no vibrations in the
Viper for a throat impact, but what about an impact at the tip? The answer is
that the impact duration at the tip is shorter than the impact duration at the
throat. In effect, the tip is lighter than the throat since the tip accelerates away
from the ball faster than the throat due to the greater rate of rotation of a rac-
quet for an impact at the tip. As a result, the ball sits on the strings for a short-
er time and the damping effect is weaker. A similar effect occurs if the string
tension is increased. A ball will sit on the strings for a shorter time if the ten-
sion is increased, with the result that the racquet will vibrate more for any
impact away from the sweetspot.

The above explanation also accounts for the fact that higher frequency modes
of a racquet are never seen when a racquet strikes a tennis ball. Like a guitar
string, a racquet can vibrate at many different frequencies, but in practice the
only mode of any importance is the fundamental Cowest frequency) mode at
around 100 to 200 Hz. In order to observe the high frequency modes, it is nec-
essary to impact the strings with a small steel ball so that the impact duration
is much smaller than the vibration period. Even higher frequency modes, up
to 10 kHz or higher, can be seen if the frame itself is struck with a metal object
so that the impact duration can be reduced to 100 microseconds or less.
Detailed calculations of impact duration and vibration amplitude for a tennis
racquet can be found in my article in Sports Engineering (2000), 3, 111-122.

There is one small point that appears to be inconsistent with the arguments
presented above. If an impact of duration 5 ms can excite vibrations only up
to 200 Hz, then how come the strings can be heard to vibrate at around 500
Hz? The ping from the strings arises from the fact that the strings vibrate like
guitar strings, at a frequency of about 400 Hz at low string tension and up to
600 Hz at high string tension. If the ball can damp the frame vibrations then it
should do an even better job at damping the string vibrations. In fact, the ball
does help to damp the string vibrations but it doesn't sit on all the strings so
some of them are not well damped. A string dampener located in the throat
also damps the vibrations but it takes about 100 ms to damp them complete-
 Racquet Feel 95

ly. The ball sits on the strings for only about 5 ms. However the main point is
that while most of the energy in a 5 ms impact is concentrated in exciting vibra-
tions at frequencies below 200 Hz, there is a tiny amount of energy available
to excite vibrations at frequencies above 200 Hz. The strings are so much
lighter than the frame that it takes only a tiny amount of energy to start them
vibrating. They sound louder than the frame not because they have more vibra-
tion energy but because (a) they push slightly more air around and generate
sound waves more efficiently than the racquet frame
and (b) the ear is about 20 times more sensitive to
vibrations at frequencies around 500 Hz than at fre-
quencies around 100 Hz. It is impossible to hear any-
thing at frequencies below about 20 Hz.
Force On First Finger — Wood
g NWS
tS ©

The force acting on the hand for each of the four rac-
quets was measured using a small piezo disk located
on the handle under the base of the first finger. In this
case, the piezo signal is proportional to the force on
the piezo, not its acceleration. A piezo will generate
a large voltage even if it is at rest on a table, provid-
ed a force is applied to both sides of the disk. The
disk was only 9 mm in diameter and 0.3 mm thick
(mV)
Signal
Piezo
and it was pressed onto the handle using a firm grip.
The two thin wire leads soldered to the disk were
anchored with tape and the voltage generated when
a ball was dropped on the strings was measured with -100
an oscilloscope and also by a data acquisition system
connected to a PC. The results are shown in Figures
9.9, 9.10 and 9.11. Some variation in the size of each
signal was due to variations in the flatness of the han- Figure 9.9
dle, but the variation was no more than 30%, and it

Force On First Finger — Rollers 2.6 Force On First Finger — Triad 3.0
150 1

Tip
100

(mV)
Signal
Piezo (mV)
Signal
Piezo

Figure 9. 10
96 Chapter 9

had no effect on the shape of each signal. The disk was removed and replaced
many times to check this. The signals were obtained by dropping a tennis ball
from a height of 40 cm onto the tip, middle or throat region of the strings. This
is a very low speed impact, but the signals had the same shape at higher
speeds — they were just bigger.

Vibration of the handle is minimized for an impact in the middle of the strings
for the three graphite racquets. For an impact near the tip or throat, the vibra-
tion level is much larger, especially for the wood racquet. However, there is
very little difference between the two graphite racquets. There are minor dif-
ferences, but there is no particular feature of the force signals that one could
identify as being responsible for a difference in the feel of the two racquets.
One would need to measure lots more racquets to identify which of the wrig-
gles in the signals felt nice and which caused jarring.

One of the differences between the Rollers and the Triad is that they sound a
lot different. The Rollers has two throat sections. The main strings pass through
a rubber insert in the top throat section. As a result, the strings are strongly
damped, even without a string dampener. The
addition of a string dampener stops the strings
String Vibrations vibrating quickly, as shown in Figure 9.12. For
comparison, vibrations of the Triad strings with
the same dampener inserted is shown in Figure
Rollers (f = 490 Hz) 9.12. An impact on the Rollers generates a dull
thud, while an impact on the Triad generates a
short ping. The string vibrations can also be felt
through the handle, but only when the string
dampeners are removed. Vibrations due to the
(Piezo in middle of strings) strings are much weaker than the frame vibra-
tions but they last for a longer time when the
(V)
Piezo dampeners are removed.

Triad (f = 470 Hz)

20 30 40 50

Figure 9.12

Notes

1. The COR is 0.75 for a wood paddle and 0.88 for a strung racquet. If the bat
and the racquet have the same mass, then from Chapter 12, we getie c= 0156
for the racquet and e , = 0.27 for the bat. The serve speed formula shows that
the serve speed increases by only 7% when the strings are added.

2. The speed of a ball off the strings does not depend very much on the string
tension. The ball speed depends mainly on how fast the racquet is swung. The
 Racquet Feel 97

increase in momentum of the ball is accompanied by an equal and opposite

decrease in momentum of the racquet. The change in momentum = (force on
the strings) x Gmpact duration) and change in angular momentum = torque x
Gmpact duration). The force is large and the impact duration is small with stiff
strings and vice-versa with soft strings. The change in momentum and the
decrease in the racquet head speed is essentially the same regardless of
whether the strings are stiff or soft. However, the racquet head speed drops
faster if the strings are stiff, and the racquet rotates through a relatively small
angle during the impact. Conversely, if the strings are soft, the force on the rac-
quet is small and it takes a longer time to reduce its speed. As a result, the rac-
quet rotates through a relatively large angle during the impact. In each case the
final and average speed of the racquet is the same, so the distance traveled dur-
ing the impact is proportional to the impact duration.
 2SS
a
=
6e
¢

An Introduction to
Racquet-Ball Collisions

Gustavo Kuerten, Brazil

By Crawford Lindsey

Introduction

ennis has become a power game. Everyone wants to know how fast they are hitting the ball. This chap-
ter and the next are first steps in determining just that. There is just one caveat, however. We will assume
that all collisions described in Chapters 10 and 11 are occurring on a line through the center of mass of
the racquet and ball. In other words, we are pretending the ball is hitting the racquet at the balance point where
the mass involved in the collision is the same as the racquet’s mass. This makes the explanations simpler. Chapter
12 will then examine how to apply these concepts to a hit anywhere on the racquet. In that chapter we will
define “effective mass.” It is the effective mass that would then be used in all the formulas in this chapter.
What happens when a ball hits a racquet? Some, or all, of the following events, among others, may occur,
depending on where you hit the ball on the racquet:
e strings stretch
e ball deforms
¢ racquet bends longitudinally
e racquet head bends in and out
e racquet hoop twists like wringing a towel
¢ racquet shaft twists like a screw
100 Chapter 10

e strings vibrate
e racquet vibrates
e arm vibrates
e ball vibrates
e wrist bends backward
e string tension increases
* racquet slows down
° racquet rotates longitudinally backward head over heels
* racquet rotates around long axis towards 3 or 9 o'clock
e ball decelerates
e ball stops
ball changes direction
e ball accelerates
ball leaves

These events, which occur in a few milliseconds (4-8 ms depending on sever-

al factors), represent a flow of energy and forces — thosé mysterious and mag-
ical “somethings” that pervade the universe and that are inextricably bound up
in hitting a tennis ball. This single event, which determines the course of a ten-
nis match, is variously described in terms of force, momentum, energy, and
coefficient of restitution. These are the big four concepts that explain the
dynamics involved in purposefully colliding a racquet and a ball into each
other in a tennis shot. Of course, Sir Isaac Newton glimpsed all of this in his
falling apple that led to his laws of motion.

A tennis ball in motion tends to stay in motion;

a tennis ball at rest tends to stay at rest.

Let’s get one thing clear from the outset — the ball is not a force by itself. The
ball is not traveling “with a force” of 30 mph or of anything else. Once in flight,
no force is causing the ball to move along its journey. In fact no forces are act-
ing on the ball at all, except air resistance and gravity, which we will ignore
since they have nothing to do with the ball/racquet collision, which is the
essence of tennis. If you don't think the world is magical, just consider that no
forces at all are necessary to keep this ball speeding toward its fateful collision.
It just goes. Galileo and Isaac Newton figured this out in the 1600s. Newton
formulated three laws of motion, which, for the most part, describe virtually
every physical event we witness, but more importantly, every tennis event.
Newton's First Law is that any object will keep on doing what it is doing until
something stops it from doing so. That “something” is a force. You've proba-
bly heard it paraphrased as “an object in motion tends to stay in motion, and
an object at rest tends to stay at rest.” This “tendency” is also known as inertia
and it is measured by its mass. The more massive an object is, the harder it is
to change what it is doing or not doing. There is no privileged status for an
object at rest or in motion. They are both equivalent natural states of being,
requiring nothing to maintain their states. To emphasize this equivalency, we
sometimes refer to a resting object as having inertial mass and to an object
moving force-free as having inertial motion. In each case, it is the same inertia
(mass) that a force must overcome to change the ball’s velocity.

By itself, a racquet isn't a force either. It, like the ball, is just a mass of stuff.
There are stationary masses and there are moving masses. And whenever two
 Force and Momentum 101

masses come in contact with each other, they create a force. Each mass exerts
a force on the other. There is no force until there is an interaction. The two
colliding objects “decide” what the force between them is going to be. Two
separate sets of two objects traveling at the same speed with the same masses
will create different size forces depending on the properties of the colliding
objects such as stiffness. This goes for two different racquets with equal mass
hitting balls also. The force created at impact with a ball will depend on the
properties of the racquet like string tension and racquet flex, among others.

pesca acme ]
The force required to change a tennis ball’s speed or direction
is equal to the ball’s mass times its acceleration.

Strange but true — a force can never be seen. Only the effects of the force can
be seen. A force shows itself by the changes in the state of motion of the ten-
nis ball. A force has acted if the ball changes its direction or speed. Speed in
any given direction is known as velocity, so a force has acted if the ball
changes its velocity (direction or magnitude). Changing a ball's velocity is the
entire life's purpose of a tennis racquet. Changing velocity from 30 mph to 50
mph or from 50 mph to 30 mph means that you have accelerated (or deceler-
ated) the ball. Newton realized that a force becomes manifest as the accelera-
tion of a mass. His famous Second Law of Tennis thus became

F=ma (10.1)
where,
F = the force,
m = mass of the object, and
a = acceleration.

Notice that if the mass of the racquet or ball is doubled, the force necessary to
achieve the same acceleration is doubled. Or said differently, if the mass is
doubled and the force remains constant, the acceleration will be halved. Since
all balls have about the same mass, changing the mass is only relevant to the
racquet.

Remember, this force isn't seen, only inferred. The force acting on the ball was
equal to the ball's mass multiplied by how much it was accelerated. The force
acting on the racquet was the racquet's mass multiplied by its acceleration. As
you can see, this force, the mover of all things, is hard to pin down. At this
point you might ask, “How are we to measure this force? And what about
acceleration? How are we to determine our final ball velocity from this equa-
tion?” We will see below that we don’t actually have to measure forces to fig-
ure out our ball velocities. But we do have to understand the steps in between
which build on each other to attain our goal of determining these velocities.
Newton's Third Law of Tennis provides us with the next step.

d Law ofTennis
The force of the ball on the racquet is equal
and opposite to the force of the racquet on the bail.

The Third Law states that whenever two objects like a racquet and ball inter-
102 Chapter 10

act and create a force between them, the force exerted by the ball on the rac-
quet is equal and opposite to the force exerted by the racquet on the ball. Both
the ball and the racquet experience the same magnitude force but in opposite
directions. But we usually say the racquet hits the ball, and not vice versa. This
makes sense and it is obvious that it is the ball's behavior that we are trying to
influence. But sometimes our language used in such descriptions gets in the
way of our understanding. Our usual descriptions of collisions are shaped by
three things — which of the objects is moving before the collision, which
object's behavior appears to be most changed, and what is our intent for per-
forming the collision. When a ball is thrown against a wall, we say, “the ball
hit the wall,” rather than “the wall hit the ball.” This is true, even though both
the wall and the ball experience the same force. It is also true even though it
is the motion of the ball that is changed, not the wall's, even though the wall
was the recipient of the “hit.” The reason our language describes it this way is
because the ball was “doing” the action,

When a racquet is swung at a tennis ball, we say, “The racquet hit the ball,”
even if both ball and racquet are moving. In this case, the object being “hit” —
the ball — changes its speed and direction the most. Also, we are hitting the
ball with a purpose, so we say, “I hit the ball.”

These scenarios show how language can corrupt and influence our under-
standing of an event. If you hit the ball, the force is not in the racquet. If the
ball hits the wall, the force is not in the ball. The force is generated between
the ball and racquet at impact and they both experience the same force. This
is the essence of Newton’s Third Law. When a racquet and a ball collide (or
any two objects), the force of the racquet on the ball is the same but in the
opposite direction to the force of the ball on the racquet.

But if they both experience the same force, why is it only the ball that goes
flying? The answer lies in Newton's Second Law — F = ma. The force on the
ball and the racquet are the same, so the mass of the racquet multiplied by its
acceleration due to the force will be equal to the ball's mass multiplied by its
acceleration.

F = force by the First Law

F = ma by the Second Law
F.=F, by the Third Law (10.2a)
m,a, = m,a, by the Second and Third Law (10.2b)

where the subscript r is the racquet, and the subscript b is the ball.

The same force on the lesser mass ball will cause more acceleration of the ball
than the racquet. Basically, the ball's mass is always the same (about 57 grams),
but the mass of the racquet is totally up to you. So depending on the mass of
your racquet, the force of the collision will accelerate the racquet different
amounts, and the player will feel the results.

It's obvious when you analyze it, but it is also very wondrous, strange and
beautiful. The bottom line is this: whenever two objects interact, be it a rac-
quet and a ball, strings and the frame, or the handle and your hand, a force is
created and all parties to the event experience the same force. The visible con-
sequences of that interaction will depend on the relative mass and stiffness of
the interacting objects.
 Force and Momentum 103

Again, I hear you saying, “Interesting, but I still can’t determine how fast I’m
hitting.” True, but this very important fact that the force is the same on both
the racquet and ball allows us to set up an equality — Fr = Fb — that is a key
step in our quest. The next step comes from defining and calculating momen-
tum, to which we now turn.

During an impact, the force on the ball and the racquet are the same. Nature
is apparently very fair. It dispenses its resources with great equity. There's
another example of this equanimity that lies behind Newton's laws. Whenever
the racquet and ball collide, the change in the momentum of the ball is equal
and opposite to the change in the momentum of the racquet. That’s odd. We
know for a fact that the ball’s behavior doesn’t look equal to the racquet’s
behavior in any way. They may experience the same force, but they don’t act
the same way. So, what is this mysterious momentum?

We've seen that mass is defined as the amount of matter there is. It is harder
(requires more force) to change the motion (accelerate) of a more massive
object because there is more matter to move. Once the inertia of the mass is
overcome, the object becomes a moving mass. There are two variables here:
the mass and its velocity. Moving mass is mathematically represented as mv
(mass multiplied by velocity) and momentum is represented as p. So momen-
tum equals mass times velocity, or, p = mv.

Momentum represents how much stuff is moving and how fast. So, given two
racquets traveling at the same speed, but with different masses, the heavier rac-
quet will pack a bigger punch since it has more momentum. Likewise, given
two racquets of the same mass but traveling at different velocities, the faster
racquet will pack a greater wallop.

The easiest way to conceptualize momentum is to consider what it takes to

stop a moving object like a tennis ball. We learned from Newton’s First Law
that changing an object’s velocity requires a force. It takes a force acting for a
period of time to bring an object to a stop. It takes more force acting for the
same amount of time to stop a tennis ball traveling at 27 m/s (approximately
60 mph) than one travelling at 13.5 m/s (approximately 30 mph). A tennis ball
has a mass of approximately 57 grams or .057 kilograms. The momentum
(quantity of motion) is mv. This calculates as follows: 27 m/s x .057 kg equals
1.35 kg-m/s for the faster ball, and 13.5 m/s times .057 kg equals .77 kg-m/s,
for the slower ball. The faster ball has twice the momentum and requires twice
as much force acting for the same amount of time to stop it. Or, as we will see,
it requires the same force acting for twice as much time.

This is intuitively obvious, but it also follows directly Newton’s Second Law.
We can rewrite this law as follows:

By definition, acceleration is the change in velocity over time:

a=(v,-v,)/t

where v, = final velocity, and

104 Chapter 10

v, = initial velocity.

So, the following rewriting of Newton's Second Law brings us straight to the
Impulse Momentum Theorem. We start with the Second Law:

F=ma

Substitute acceleration, a, with the equivalent expression (v, — v,)/ t

Pin vt

Multiply each side by t to get rid of the fraction:

Ft = m(v, — V,) (10.3a)

Scientists use the symbol A for changes in anything between an initial and final
state. So the above equation can be written in the following shorthand:

Ft = mAv (10.3b)

where Av means “change in velocity.” (We will go back and forth between
these equivalent equations depending on our purpose.)

This says that a force acting for a given amount of time is necessary to change
the momentum of an object. Or similarly, a force applied to a mass will accel-
erate it by changing its velocity over a certain amount of time.

This is an incredibly handy relationship Gf only for the fact that velocity has
finally shown up in our calculations, and we now are getting closer to deter-
mining how fast we are hitting the ball). The Ft part of the equation is known
as the impulse and the mv is the momentum. So, in a collision of a ball and
racquet, each object experiences an impulse equal in magnitude to the change
in its momentum:

F.t, = m,Av,
Impulse on the ball = Ball’s change in momentum

Ft. = m,Av,
Impulse on the racquet = racquet’s change in momentum

Notice what else this relationship tells us. It says to change the ball's momen-
tum, any combination of Ft that equals mAv will do. A small force for a longer
time can change the velocity the same amount as a larger force for a shorter
time, provided the amount Ft is the same in each case. You can immediately
see the consequences of dwell time (time of ball on strings) on force here. If
the ball stays on the strings of the racquet for a longer time (by using softer
string or lowering tension), then the force produced by the collision that is nec-
essary to accelerate the ball a given amount will be less; it just will last longer.
And as we will see, the force on the player will also be less. (Note: during a
collision the force starts at zero, rises to a maximum, and goes back to zero.
Therefore, we are assuming F = average force during the collision.)

We also see the following relationships in the impulse equation:

 Force and Momentum 105

1. Force and time are inversely related. To produce the same change in the
ball's velocity, if the force is increased, the time that force acts must decrease.
If the time is increased, the force must decrease.
2. If we divide each side by m to get everything that influences a change in
velocity grouped together, we get

Ft/m = Av

This says that for a given force, the change in velocity is inversely related to
mass. The same force acting for the same time on a heavier ball will change
the velocity less. But if a lighter ball is used, the same impulse (Ft) will cause
a greater change in velocity.

In a collision, according to Newton's Third Law, the racquet and ball exert
equal and opposite forces on each other. The observable consequences of
those forces are quite different, however. The force on the smaller mass ball
sends it across the net, while the same force on the much more massive rac-
quet has little visually obvious effect. However, these consequences provide
the basis for very powerful predictions, as we will see.

Now we are ready to derive our next conceptual step. If the force on the ball
and racquet are the same but in opposite directions (represented by the minus
sign below), then

And if the ball and the racquet collide, then the ball touches the racquet the
same amount of time the racquet touches the ball, so,

tees

Therefore, the quantity Ft, the impulse, must be the same but opposite for the
ball and the racquet, since t is equal for each and F the same but opposite:

Ft, = - Ft,

And if the impulse is the same but opposite, then the change in momentum
must be the same but opposite for each (Equation 10.3a and 10.3b), since the
impulse felt by an object equals the change in momentum of the object. So,

m,Av, = - m,Av.. (10.4a)

Change in ball’s momentum = Opposite of change in racquet’s momentum

Expanding Av, and Av, to (v,, - v,,) and (v,, - v,,), the equation looks like
this:
M,Vp2 ~ MVpy = - (M,V,5 - M,V,1)

and multiplying by the -1 on the right side, we get,

MbYp2 ~ MEVp1 = M,_V,, - MV. (10.4b)

During an impact, the momentum gained by the ball equals the momentum
lost by the racquet. Let's look at an example.
106 Chapter 10

Example: Calculating final velocities with given information. Let's say

that we are going to hit a 57 gm tennis ball and launch it at 27 m/s (about 60
mph). It is coming at us at 13.5 m/s (about 30 mph). The racquet mass is 300
gm and is swinging at 9 m/s (about 20 mph). The time of impact is .005 s. The
change in ball velocity will be (v,, — v,,) = 27 - 13.5) = 40.5 m/s. In other
words, the ball’s velocity actually goes from minus 13.5 m/s to 0 and then
back up to 27 m/s. So the total change in velocity will be 40.5 m/s. The
racquet's velocity does not change direction. The racquet slows down but
continues in the same direction. Therefore, the change in velocity of the
racquet will be (v,, — v,,). From this information, we want to know six
things:

Question 1. What is the final velocity of the racquet?

Question 2.. What is the change in velocity of the racquet?
Question 3. What is the final momentum of the racquet?
Question 4. What is the change in the momentum of the racquet?
Question 5. What is the impulse during collision to change the racquet’s
momentum by the amount in 4 above?
Question 6. What is the force required to change the racquet’s momentum by
the amount determined in 4 above?

Figure 10.1 summarizes what we know starting out.

BEFORE collision DURING collision AFTER collision

Vr =9 m/s Vp1 = -13.5 m/s ; 7 =? = Vb2 = 27 m/s

m = 300 gm m=57 gm ¢ Equal and opposite forces on each. m=57 gm

mv;-1 = 2.7 kg m/s MVp1 = -0.770 kg m/s] * The force starts at 0, climbs to a maximum AVb = 40.5 m/s
just before the ball changes direction at about MVp2 = 1.54 kg m/s
t = .0025 s, and goes back to 0. For simplicity, AMVp = 2.31 kg m/s
we use the average force.
¢ Racquet slows but doesn't change direction.
* Collision lasts for t = .005 s.
¢ 5. Impulse = Ft =? ANS e

Figure 10.1

What we know and the six unknowns that we will determine.

Notice that the ball starts out with a negative velocity and momentum. This
simply indicates that it is moving in the opposite direction than that of the rac-
quet.

From equation 10.4b, we know that the change in the ball’s momentum equals
the opposite of the change in the racquet’s momentum,

T,Vp2 a Mh Vb = (mV,5 7 m,V,) is MV 4 . MV 9

 Force and Momentum 107

and, from above, we know everything except v,,. So let’s plug them into the
equation:

1.54 kg-m/s + 0.77 kg-m/s = 2.7 kg-m/s - 0.3v_,

Now we solve for v_,

0.3v,, = 2.7 kg-m/s -1.54 kg-m/s - 0.77 kg-m/s

0.3v,, = 0.39 kg-m/s

Answer 1: Final racquet velocity = v,, = 1.3 m/s.

Answer 2: The change in the racquet’s velocity is thus v_, - v,, = 1.3 - 9 = - 7.7
m/s.

Answer 3: The racquet’s final momentum is m,v,, = .3 x 1.3 = .39 kg-m/s.

Answer 4: The change in the racquet’s momentum is m,v,, -m,v,, = .39 - 2.7
= -2.31 kg-m/s. That is equal but opposite to the ball’s change in momentum,
just like the law of conservation of momentum tells us.

Answer 5: We actually knew the impulse (Ft) before finding v,,. Because Ft =
mAv, and since we knew m,Av,, we knew Ft. And because m,Av, = m,Av,, it
isthe same tor bots: Ft =m Ay. = 3x -7.7 = =2.31 ke-m/s.

Answer 6: Again, we could have figured this with the ball data before we
calculated the racquet’s final velocity. However, using the racquet data, we
get Ft = -2.31/.005 = -462 N. The force of 462 Newtons is equal to about 104
pounds (1 N = .2248 Ib).

The results of our calculations are summarized in Figure 10.2.

Figure 10.2

BEFORE collision DURING collision AFTER collision

Vr4 =9M/S Vpy =-13.5 m/s 6. F, = -462 N 1. Vio = 1.3 m/s Vp2 = 27 m/s

(On

m = 300 gm m=57gm ¢ Equal and opposite forces on each. m = 300 gm m=57 gm

Mv; =2.7 kg m/s MVpz =-0.770 kg m/s} °¢ The force starts at 0, climbs to a maximum 2. Av, = -7.7 m/s AVp = 40.5 m/s
just before the ball changes direction at about 3. MV;2 = .39 kg m/s MVp2 = 1.54 kg m/s
t = .0025 s, and goes back to 0. For simplicity, 4. Amv, = -2.31 kg m/s) AMVp = 2.31 kg m/s
we use the average force.
¢ Racquet slows but doesn't change direction.
* Collision lasts for t = .005 s.
¢ 5.F,t=-2.31 kg m/s Fpt = 2.31 kg m/s
108 Chapter 10

It is interesting that the impulse can be any combination of force multiplied by

time that, in this case, equals -2.31 for the racquet and 2.31 for the ball. If the
time were .004 seconds, instead of .005 s we used above, the force would be
577.5 Newtons, and if the time were .006 seconds, the force would be 385
Newtons. If the time were .008 seconds, the force would be 288.8 Newtons.
You can see that small changes in time make a big difference in the magnitude
of the force. Impact time will be increased by more flexible racquets, softer
strings, and lower string tensions.

We arrived at final velocities for the ball and racquet here, but only because
we presented one of them, the ball, as a given at the beginning of the exam-
ple, just to see how the relationships work. We still have yet to figure out the
after collisions velocity based only on what we know before the collision. We
now turn to The Law of Conservation of Momentum which sets up our equa-
tion to determine after collision velocities.

Law of Conservatio n of Racquet and Ball Momentu

In a collision, one object loses momentum and the other gains momentum.
And since the change in the momentum of each is the same, the amount of
momentum lost by one is the amount gained by the other. The momentum lost
by the racquet was gained by the ball, and the momentum gained by the ball
was lost by the racquet. In the above example, the change in the momentum
of the ball was 2.31 kg-m/s, and the change in the racquet was -2.31 kg-m/s.
The ball changed its momentum 124 percent and changed direction while the
racquet changed its momentum 71 percent by slowing down but continuing in
the same direction. Even though the ball and racquet changed momentum by
equal amounts, the ball’s momentum change is larger relative to its original
momentum. That is why the change was so much more obvious on the ball.

Since the momentum gained by the ball is the same as that lost by the racquet,
the total combined momentum of the racquet and ball is the same before and
after the collision. This is true of every collision and is known as the Law of
Conservation of Momentum. The conservation equation was lurking in equa-
tions 10.4a and 10.4b. Let’s pick up where we left off in equation 10.4b with
the change in the ball’s momentum equalling the opposite of the change in the
racquet’s momentum:

MV p2 NY pI = mV 4 7 MV:

If we simply rearrange this so that initial states of the ball and racquet are on
one side of the equation, and the final states on the other side, we get the Law
of Conservation of Momentum for balls and racquets:

M,V., + MV,, = M,V,. + M,V,, (10.5)

Total Racquet and Ball Momentum Before = Total Momentum After Collision.

where,
m, = mass of the racquet,
m,, = mass of the ball,
V,, = initial velocity of the racquet,
V1 = initial velocity of the ball,
V2 = final velocity of the racquet, and
Vi = final velocity of the ball.
 Force and Momentum 109

The beauty of analyzing collisions in terms of conservation of momentum is

twofold: you don’t have to know the forces or the time of impact in order to
describe changes in motion, and momentum is conserved in all collisions
(unlike energy). You just set the sum of the momentums before the collision
equal to the sum of the momentums after the collision and solve for your
unknown velocities. For example, if we want to know the ball’s velocity after
impact, we solve equation 10.5 for v,,:

A a mV bi = ~ LL)

™M,Vp2 _ Mm,V,4 MVp1 is Mm,V,9

M,V,_ + M,Vp) - MV.

Vino
™

There we have a formula for the final ball speed. But even if we use that, we
have a problem. The problem, of course, is that we have two unknown veloc-
ities, v., and v,,, so we can’t solve for one without knowing the value of the
other. So, as close as we are to calculating our final velocities, we will need a
second independent equation to solve for one of the unknowns, and then plug
it back into our momentum equation. We will find that equation and more in
our discussion of energy in Chapter 11. There we will demonstrate two differ-
ent ways to determine ball velocities — one using conservation of of momen-
tum combined with the “coefficient of restitution,” and another using only the
“apparent coefficient of restitution.” The latter solution will be our “physics
Rosetta Stone,” leading to what we call “the most important formula in the
book”

Nene )V.

which was introduced in Chapter 1 as the “serve speed formula.”

~

» &
a oe _ -

: =e wees

=o * <eiishgts was
i. a a af a

aoe +m, al
"=. @ oe ,
s we AmnA
fh -
=
<a we

———

»
* =
7 - a7 -

=~

_ .

— o >

7 ee —

ad
~S

- => ©

~ **

—— = -
>= 7

7 =
7 —
=

@ -
 Energy and
Energy Return
in Tennis Collisions
Anna Smashnova, Israel

By Crawford Lindsey

escribing a racquet and ball impact in terms of momentum, as we did in Chapter 10, is one way to look
at a tennis hit. But there is another way — describing the interaction in terms of energy, as we will do
in this chapter.

How many times have you heard players talk about the ball’s energy, the energy return of the racquet, the rac-
quet or ball losing energy, or that a light racquet can swing faster and has more energy to hit the ball faster? We
use the term “energy” everyday, but like force, it is one of those invisible creations of the mind that are every-
where and nowhere at the same time.

So what is energy? Fortunately, it is an outgrowth of everything we discussed in Chapter 10. To recap, we said
that a ball will continue its state of motion (whether it is at rest or moving) unless a force acts upon it to change
its velocity. To change the direction or speed is to accelerate an object. This statement is summarized by F =
ma. Since acceleration is change in velocity over time, we were able to rework the formula into the Impulse-
Change in Momentum formula which is stated as Ft = mAv. In words, that translates to “impulse equals the
change in momentum.” Momentum is something every moving object has and is the quantity mv, and the total
momentum of a colliding racquet and ball is always the same before and after the collision (Equation 10.5). And
112 Chapter 11

we saw that this equality could be very useful to make predictions of what the
outcome of a racquet and ball collision would be, based only on knowledge
of their masses and velocities before impact. There, at the end of Chapter 10,
we hit an impasse. Since we have two unknowns (both the final ball and rac-
quet velocities), we need two independent equations, so we can solve for one
of the unknowns in one and plug the results into the other. And this is where
we find out that every moving object has another quantity — +mv? — known
as kinetic energy. Thus, by exploring the kinetic energy and energy return of
racquets and balls, we will discover our second relationship needed to com-
bine with the conservation of momentum equation. The resulting formula will
give us the power to predict the outcomes of racquet and ball collisions. Just
as we derived momentum conceptually, we will do so for kinetic energy.

The notion of energy also grows out of Newton’s Second Tennis Law: F = ma.
In Chapter 10, by transforming that formula, we arrived at further insights into
the laws and relationships of the tennis universe like the nature of momentum,
the Impulse-Change in Momentum Theorem, and the Law of Conservation of
Momentum. Further transformation will lead us right to energy.

To continue, consider what it feels like when you apply a force to your rac-
quet. If you swing it (push or pull), you move it while you are pushing. The
force is applied over a distance. If it is a very heavy racquet, you sweat, huff
and puff, and then wipe your brow saying, “Wow, that was a lot of work.” I
guess that’s what the first tennis playing physicists must have said anyway,
because a force applied over a distance is known as “work” and is represent-
ed as:

W =Fd (11.1)
But where did that come from? How does that follow from Newton’s laws?

Once again it is all in that simple formula F = ma, which describes a force F
that is applied to accelerate a mass m along a straight line from an initial speed
v, to a final speed v,. In Chapter 10, through a simple transformation, we saw
how F = ma is equivalent to Ft = m(v, - v,). We just have to keep playing with
the terms a little to come up with energy. Let’s continue our transformations of
that formula:

Fiesty, =v,) = OV te,

From here, we just make some substitutions. We know that time (t) equals dis-
tance divided by the average velocity: t = d/(v, + v,)/2 = 2d/(v, + v,). So we
substitute that for t in the impulse-momentum equation;

2Fd/(v, + v,) = mv, - mv,

Then multiply both sides by (v, + v)) to isolate 2Fd:

2d = Gnv, — mv), + Vv)

Divide each side by 2 to isolate Fd:

 Energy and COR 113

Fd = (mv, = nv), Vv)

Multiply out the right side:

Fd = 3 mv,?- 3 mv,? (11.2)

Work = Change in kinetic energy

Stated this way, the above formula is known as the Work-Energy Theorem. The
equation highlights the following fact: a force changes the motion of a racquet
or ball, and energy is a measure of that change. Just like force, energy isn’t seen
by itself, but only inferred from observing changes in velocities.

With equation 11.2, our term for kinetic energy Cimv?’) has turned up. So, we
see that the formulas for our concepts of force, momentum, energy, accelera-
tion, and velocity are all lurking in the simple formula F = ma. This is an impor-
tant observation if you like feeling that everything is connected to everything
else and that everything is derived both conceptually and mathematically from
some common starting point. It seems that everything that happens to our rac-
quets, strings and balls will be derived from this one simple equation that says
that the world is made of things and to change the state of motion of those
things, you have to apply a force.

We’re not quite there yet, however. I don’t know about you, but I’m still not
seeing what Christian Huygens (1629-1695) must have seen in this formula that
made him exclaim one sleepless night, “Aha, there’s energy — it’s }mv7!” Yes,
I see the terms that we call kinetic energy, but we have the benefit of hind-
sight and know what we are looking for. What if you were just playing with
your formulas one day, and you came across this transmutation? How would
you interpret it into a meaningful concept such as energy?

Huygens wasn’t happy with the idea, up to that time, that momentum was con-
sidered the essence of a tennis hit. Although he appreciated how it could help
determine final velocities, and therefore was very useful, he found it troubling
that if a ball and racquet approached each other with equal momentums, the
total momentum was thus zero both before and after the hit, even though the
collision broke a string, hurt his hand, went out of the court and broke his
neighbor’s window. That zero momentum system sure had a lot of non-zero
consequences! Likewise, it concerned him that the total momentum of a rac-
quet and ball before the collision could be less than the momentum of either
one of them separately. Huygens realized that by squaring the racquet’s and
ball’s velocities, he got rid of the directional and cancelling property of
momentum equations. Even though the total momentum of the racquet and
ball system might be zero, there was still an inherent capacity to do work in
both objects. Here, in this energy that was independent of direction, was the
real fuel behind flying tennis balls.

So, back to formula 11.2, we notice that if the initial velocity were 0 (as in serv-
ing a ball), then the second term goes away and we are left with Fd = smv,’.
So if you apply a force over a distance to a stationary mass, it will be associ-
ated with this mysterious +mv?. And this quantity remains after the force is no
longer applied. It is a descriptive quantity that is now a “possession” of this
mass. Applying a force to the ball over a certain distance adds energy to the
ball in the amount of Fd. The ball has an amount of kinetic energy of ;mv,’
114 Chapter 11

before the collision and a kinetic energy of +mv,” after the collision. You can
determine how much work was done and how much energy was added to the
ball by using the Work-Energy equation:

Work = Change in Energy = Fd = mv,” — +mv?2

Nie

The unit of work and energy is the Joule (J), which is a N-m or kg-m*/s?.

LawofConservation of Racq
Huygens also noticed that in many experiments with billiard balls that the sum
of all balls’ smv? was equal before and after the collisions. Whatever this term
represented, it was conserved every time. Whenever there is some quantity in
a system of objects that is the same before and after innumerable transforma-
tions to that system, you have to suspect that that “quantity of something” is
very fundamental. In this case it was, and it came to be known as energy, and
more specifically, kinetic energy when moving objects are involved — energy
of motion.

The net result is the Law of Conservation of Energy, which states that the total
energy before the collision will be equal to that after the collision. The word
“total” indicates that energy comes in many forms (mechanical, chemical, elec-
trical) and sub-forms (kinetic, potential, elastic, heat), and it can change from
one form into another. So, in a totally elastic collision G.e., where there is no
energy loss due to transformation of form) between the ball and racquet, the
conservation of energy equation looks like this:

1
il
aMVi4° aero
+ 5MLVpyy ear!
= 5M,V,o° pare + >MVpo 2 (11.3)
Total kinetic energy before collision = Total kinetic energy after collision

This too, like the conservation of momentum, is a handy equation, because it

does not involve any difficult-to-measure terms like force or time of impact, but
we are obviously going to have to find a way around the assumption of total
elasticity, since a racquet-ball collision certainly is not totally elastic.

But first, we still need to determine WHAT is this thing called “energy”? We
know what a racquet’s mass is; we know its position, velocity and acceleration.
We know and see its motion and momentum. These are all things that we see,
touch, feel, conceptualize and understand in our world. These are everyday
entities. But energy? Energy is a total abstraction. In the case of momentum
(mv), we see a racquet moving at a particular velocity. For kinetic energy
G mv) we don’t see a half of a racquet moving at its velocity multiplied by its
velocity. We just see the whole racquet moving at a certain speed. Up to this
point in our exploration of the physics of tennis collisions, everything has
evolved nicely from both a reworking of F = ma and from ordinary perceptu-
al observations of racquets hitting balls. With energy, we seem to leave our
touch-and-feel world and enter a world of make-believe. Yet, energy is the
Holy Grail of physics and tennis. So, let’s examine the difference between
momentum and energy in order to become completely familiar with this thing
called energy.
 Energy and COR 115

Kinetic energy and momentum look similar in that both their equations con-
tain mass and velocity. They are not, however. Here’s what we know:

¢ Momentum = mv; energy = }my?.

¢ Momentum calculations depend on the relative directions of the ball
and racquet before and after the collision, energy does not.
e¢ Momentum is directly proportional to velocity; energy is directly pro-
portional to the square of velocity. If you double the racquet’s veloci-
ty, you will double its momentum, but you will quadruple its energy
(since (2v)? equals 4v?).
¢ Both momentum and energy are directly proportional to mass.
e Momentum is ALWAYS conserved in every racquet-ball collision;
energy is always conserved, but it often changes form. Kinetic energy
is usually not conserved, and that is what we are interested in. (We will
solve this in the next section.)

Given these differences, is it more important to have more momentum or more

energy when you hit the ball? Let’s look at an example. Figure 11.1 shows the

Momentum and Energy in Elastic Collisions Figure 11.1

These are collisions with no
BEFORE collision AFTER collision energy loss using racquets
with the same momentum but
v=12.5m/s v=-15 m/s v = 5.64 m/s v = 33.14 m/s
different masses, velocities

=O oO = and energies. The racquet

with the greatest amount of
energy hits the ball the
m = 400 gm m=57 gm m = 400 gm m=57 gm fastest. It is also the racquet
mv = 5 kg m/s mv = -.855 kg m/s mv = 2.26 kg m/s mv = 1.9 kg m/s with the greatest change in
5mv2 = 31.25 J 5mv2 = 6.4 J 5mv2 = 6.4 J 5mv2 = 31.3 J velocity, momentum, and
total momentum = 4.15 kg m/s total momentum = 4.15 kg m/s energy. In other words, there
total energy = 37.7 J total energy = 37.7 J was a price to pay for the
extra ball velocity — shock.
v=16.7 m/s v=-15 m/s v = 38.27 m/s

+O
m = 300 gm m=57 gm
mv =5 kg m/s mv = -.855 kg m/s mv = 1.97 kg m/s mv = 2.18 kg m/s
mv? = 42 J 5mv2 = 6.4 J 5mv2 = 6.5 J 5mv2 = 41.8 J
total momentum = 4.15 kg m/s total momentum = 4.15 kg m/s
total energy = 48.3 J total energy = 48.3 J

Oo—>
v=25 m/s v=-15m/s V=7.26 m/s v = 47.26 m/s

m = 200 gm m=57 gm m = 200 gm m=57 gm

mv =5 kg m/s mv = -.855 kg m/s mv = 1.45 kg m/s mv = 2.7 kg m/s
5mv2 = 63 J 5mv2 = 6.4 J 5mv2 = 5.3 J Smv2 = 63.7 J
total momentum = 4.15 kg m/s total momentum = 4.15 kg m/s
total energy = 68.9 J total energy = 68.9 J
116 Chapter 11

final velocities of balls launched by racquets having equal momentums but dif-
ferent velocities, masses, and energies. We are assuming there is no energy lost
in the collision. (In order to illustrate the difference between momentum and
energy, we have cheated by using the formulas we are trying to discover so
we could calculate the final ball and racquet velocities for our illustration.) It
is clear that the more energy in the racquet, the greater the final ball velocity.
Does that mean you should purchase the lightest racquet possible and swing
it as fast as you can? The answer is no for two reasons. First, given today’s rac-
quets, you can swing as fast as you want anyway. Second, most players have
a swing speed they are comfortable with in every situation. The speed of the
swing is really a given according to what feels best to the player, not how fast
he is capable of swinging a particular racquet. That being the case, the prob-
lem becomes, given a swing speed, what racquet will produce more ball
speed. Framed that way, the answer is the heaviest racquet you feel comfort-
able swinging at that speed. If the speed of the swing stays constant, both the
momentum and the energy of the racquet will increase with an increase in rac-
quet mass (Figure 11.2). ;

Momentum and Energy with Constant Swing Speed

BEFORE collision AFTER collision
Figure 11.2
v=12.5m/s v=-15m/s v = 5.64 m/s v = 33.14 m/s
Most players have a “grooved”
swing speed, so lowering the
weight of the racquet does not +O CO)——>
necessarily mean that the swing
speed will increase nor that it is m = 400 gm m=57 gm m = 400 gm m=57 gm
desireable to do so. The colli- mv = 5 kg m/s mv = -.855 kg m/s mv = 2.26 kg m/s mv = 1.9 kg m/s
sions here illustrate what hap- 5mv2 = 31.25 J mv? = 6.4 J mv? = 6.4 J Smv2 = 31.3 J
pens to the racquet and ball total momentum = 4.15 kg m/s total momentum = 4.15 kg m/s
given racquets with the same total energy = 37.7 J total energy = 37.7 J
starting velocity but different
masses. As weight goes down, v=12.5m/s v=-15m/s v=3.72 m/s v = 31.22 m/s

+O
ball speed goes down. And look
at the racquet — it is almost
brought to a complete stop with
the 200 gm racquet. There is
thus more shock and less veloci- m = 300 gm m=57 gm m = 300gm m=57 gm
ty with a lighter racquet, given mv = 3.75 kg m/s mv = -.855 kg m/s mv
= 1.12 kg m/s mv
= 1.78 kg m/s
the assumption that swing speed bmv2 = 23.4 J 5mv2 = 6.4 J mv = 2.1 J .5mv2 = 27.8) J
is constant. total momentum = 2.9 kg m/s total momentum = 2.9 kg m/s
total energy = 29.9 J total energy = 29.9 J
To keep the final ball speed at
33.1 m/s going from the 400 gm

O—>
v=12.5m/s v=-15 m/s v=0.3 m/s Vv = 27.8 m/s

—Y
racquet to the 300 gm racquet,
swing speed would have to
increase from 12.5 m/s to 13.6
m/s (a 9% increase). The trade-
off to get a 9% increase in swing m = 200 gm m=57 gm m = 200 gm m=57 gm
speed just to maintain ball mv = 2.5 kg m/s mv = -.855 kg m/s mv = 0.06 kg m/s mv = 1.58 kg m/s
speed may be less control. Smv2 = 15.6 J 5mv2 = 6.4 J
Smv2 = 0.01 J mv? = 22.03 J
total momentum = 1.65 kg m/s total momentum = 1.65 kg m/s
total energy = 22.04 J total energy = 22.04 J
 Energy and COR 117

In the examples above, we assumed that no kinetic energy of the racquet or

ball was lost or converted during the collision (and remember, also, that impact
is at the balance point). This is an unrealistic assumption. But it is a necessary
assumption in order to use the conservation of energy relationship without
having to also track and calculate all the energy transformations during impact.
So even though energy is the most important component in a collision, it is the
most difficult to keep track of.

Energy Return and Coefficient of Restitution (COR)

There is another way, however, that combines the energy and energy loss con-
cept into a simple, easy to use formula. In Figure 11.1, in each case, the veloc-
ity of the racquet minus the velocity of the ball is the same, but opposite,
before and after the collision. In other words, the relative velocities are the
same, but opposite, before and after the impact. This is true in all completely
elastic collisions where no energy is lost. To illustrate let’s look at the first col-
lision in Figure 11.1. The relative velocity before the collision is 12.5 m/s + 15
m/s = 27.5 m/s. After the collision the relative velocity is 33.14 m/s - 5.64 m/s
= 2/5 10/S.

But what about the real life where energy is lost in the racquet and ball? By
“lost” we don’t mean that it disappears. It means that it has been converted to
another form like heat or been used doing things that are not useful in pro-
pelling the ball, like bending the racquet. Given these losses, as it turns out,
whenever any two objects collide, the amount of energy lost depends on the
nature of the two colliding objects, where they hit each other, and their rela-
tive speeds. The amount of energy lost can be derived from one very impor-
tant relationship — the relative velocities before and after a collision for any
two given objects, will always be proportional to some constant. This constant
of proportionality is known as the coefficient of restitution, also referred to as
COR, or as e. COR is defined as the negative of the ratio of the relative racquet
and ball speeds after collision divided by the relative speeds before:

COR = e:= == (11.4a)

And if we rearrange the terms, we get:

(Vi ~ V2 )=-e (V4 ~ Vo4 ) (11.4b)

This shows that the relative speed of the racquet and ball after the collision is
equal to the relative speed before, multiplied by the constant of proportional-
ity — the COR.
The minus sign in (11.4a) is used to make sure that e is a positive number.
Defined this way, it turns out that e always lies between 0 and 1. If the numer-
ator in (11.4a) is positive then the denominator will be negative and hence e
is positive. Keeping track of the signs can be a bit tricky. We first need to
decide which direction is positive. The usual convention is that an object trav-
eling left to right has a positive velocity and an object traveling right to left has
a negative velocity. Since we are interested in the head-on collision between a
ball and a racquet, one will have a positive velocity before the collision and
the other will have a negative velocity. If the racquet travels left to right before
and after the collision, then the ball will have a negative velocity before the
118 Chapter 11

COR and Bounce Height collision and a positive velocity after the collision. In that case, the
Vv, term in (11.4a) is negative but all the other terms are positive.
ow did we determine that COR was
bo equal to the square root of the bounce It is often easier to keep track of the signs if we change our defi-
height? It is as follows:
nition of the v terms so that they represent the speed of the ball
and the racquet rather than the velocity. Speed values are always
Starting conditions: for
positive, regardless of the direction. In that case, the formula
hy = drop height
h, = bounce height
the COR for a head-on collision can be written as
a = g = 9.8 m/s® = acceleration due to gravity (Vi - V2 (11.5)
COR=e=
Vg = Speed at impact of dropped ball
(Vix + Yi4)
Vv, = speed of ball after the bounce

You will often see COR written both ways — as Equation 11.4a or
We know from equation 11.2 that Fd = 1/2mv?,
and we know that F=ma. We also know, that d =
11.5. But if you use Equation 11.5, you must plug in the speeds of
h, and for free falling bodies, we use a = g. the ball and racquet, independent of their direction (i.e., you don’t
Combining we get, plug in a negative ball speed). The tradeoff is that though this
equation is easier to work with because it gets rid of all the nega-
mgh = mv?/2. tive signs, it also hides the original relationships of relative veloci-
ties that define COR, so it is more difficult to remember why you
Solve for v?: are using the equation.
mv? = 2mgh
v2 = 2mgh/m
In the examples in Figure 11.1, e = 1.0. COR is always between 0
v? = 2gh
and 1.0 — 1.0 being a completely elastic collision with no energy
Free fall speed formula = v = V(2gh)
loss, and 0 would be a completely inelastic collision where the
So the speed at impact = v, = v(2gh,). objects stick together after the collision. COR is about 0.88 for a
tennis ball hitting the center of the strings. However, the COR of a
The bounce speed = v, = v(2gh,). v, is actually tennis ball dropping onto a concrete floor is 0.75. It is different on
in the opposite direction and thus negative, but clay, grass, and hard courts, and even at different locations on the
we are considering only the speeds here and strings. It is not proper to talk about the COR of the ball without
plugging them into Equation 11.5, as explained referencing what it is colliding with, and in the case of the racquet,
in the main text.
where it is impacting. The COR will be different in each case. It
also depends on which ball you are using.
The COR from Equation 11.5 is

COR is easily determined for a ball and racquet collision. Simply

place a racquet on the floor, step on its handle, and drop a ball on
the center of the strings. In this case, racquet velocity before and
but v,. and v,, are 0 in this case of dropping the after the impact is zero, so equation 11.5 becomes
ball on a “clamped” racquet, so we have,

e=
Vp Cr V2 /Viyy-

Vg
Substituting from above, we get
How are we to measure these speeds? There’s a trick — COR can
also be obtained by measuring the bounce height of the ball from
Myo v(2gh,) . vh, a J” the racquet, given any specific drop height (see sidebar: COR and
v vi2gh,) —-vh, hy Bounce Height):

And there you have it. The COR is equal to the @ = Vip / Vp, = V (h,/h,) (11.6)
square root of the ratio of the bounce height to
the drop height. where,
h,, = height ball bounces,
h, = drop height of ball,
V2 = Speed of ball leaving the strings
V1 = Speed of ball at impact with strings
V.2 and v,, are 0.
 Energy and COR 119

So, if you drop a ball from 100 inches onto the center of the strings of a rac-
quet secured by your foot on the floor, it will bounce about 77 inches. COR is
then calculated as follows:

e = v(h, /h,) = V(77/100) = 0.88

The COR is 0.88, and the ball will bounce with 77% of its initial energy. This
is important and often misunderstood. The ratio of the bounce height to drop
height is the ratio of the bounce and impact energy of the ball. Remember, mgh
smv? = the potential energy of an object at a certain height. So, mgh,/mgh,
= h,/h, = ratio of bounce energy to impact energy. The COR is not a ratio of
energy. It is a ratio of relative speeds which must be squared to relate back to
energy.

Final Ball and Racqt

We now have everything we need to derive a general formula that determines
ball and racquet velocities after a collision. We have two unknown variables,
V,, and v,, and two independent equations — the conservation of momentum
(Equation 10.4b) and coefficient of restitution (Equation 11.4a). You will
remember that we use the COR equation (11.4a) instead of the conservation of
kinetic energy equation (11.3) for two reasons. First, because the conservation
equation only works if we assume a totally elastic collision with no energy loss-
es. Second, the COR equation incorporates the energy equation and energy
losses in the sense that if energy is conserved, then COR = 1, but if energy is
lost, then COR is less than 1. The derivation goes like this (we are starting with
Equation 11.4a here so we are using velocity, not speed, in the derivation):

Now we substitute the equation for final racquet velocity into the conservation
of momentum equation as follows:

Mm V4 rs ™MVp1 = MVi2 # MV b2

substitute for v,,

Mm,V 4 ¢ sy m,(Vp2 Bae 1 <s evi) i Ay)

mV + MV51 - M,Vp2 a em V4 a em,Vb1 = TY 2

mV e ™Vb1 ~ Vb2 (m, ze m,,) > MV ¥ |

m.Vi4 a Mp1 a em.Vi4 = em.v),4 = V2 Gn, = m,,)

My if MV51 3 em, V4 ; em,V,,))

Eg = (m, + m,,)
120 Chapter 11

Final Ball Velocity Formula:

to em,(V,4 > Vor) + MV, + MAVpy (11.7)

b2 m, +m,

The final racquet velocity is derived the same way and is:

Final Racquet Velocity Formula:

M,V,1 + M,Vp1 ~ EM(V,4 > Vor)

Vio = (1 1 .8)
m,+ mM,

quations
As you can see, we have arrived at general formulas for determining ball and
racquet velocities after a collision by using only pre-collision given data. If you
plug in the data from Figure 11.1 into Equations 11.7 and 11.8, and if you set
e = 1.0 (i.e., perfectly elastic collision with no energy loss), then you get the
values on the after collision side of Figure 11.1. For instance, if we plug in the
numbers into final ball speed formula into the collision shown in Figure 11.3
(same as the first collision in Figure 11.1) with e = 1.0, we get the following:

1(.400)[12.5 - (-15)] + (.400)C2.5) + (.057)C-15) 15.145

: a = 33.14 m/s
Vb2 -
. (.400 + .057) 457

BEFORE collision AFTER collision

v=12.5m/s v=-15m/s v = 5.64 m/s v = 33.14 m/s

a—_U) CO)——_>
m=57 gm
Figure 11.3

You can then plug the numbers into Equation 11.8 to get final racquet veloci-
ty. However, with v,,, now in hand, it would be easier just to plug v,,, into
Equation 11.5 and solve for v_.:

(Vy - Vio )
COR =—e=
(Vinq + Ver)

Via = Vo ~ (Vpy +V,,) = 33.14 - 1(27.5) = 5.64 m/s

If we substitute all the before collision numbers from Figure 11.1 into
Equations 11.7 an 11.8, but use e = 0.88, then we get the results in Figure 11.4,
Both Figures 11.1 and 11.4 have the same masses and initial velocities, but they
have different outcomes because Figure 11.4 has energy losses represented by
the 0.88 COR.

But working with those big ugly formulas is a lot of work. There must be an
easier way. There is, using the apparent coefficient of restitution, also known
as ACOR or e,, to which we now turn our attention,
 Energy and COR 121

Figure 11.4
Momentum and Energy in Inelastic Collisions (COR = e = .88)
These are the same
BEFORE collision AFTER collision collisions as in Figure
11.1 except that the
v=12.5m/s v=-15m/s v = 6.05 m/s v = 30.25 m/s COR = .88 here and

nO)
COR =1 in Figure 11.1.
C)——_> In all three collisions,
compared to Figure
11.1, the racquet’s final
m = 400 gm m=57gm m = 400 gm m=57gm
velocity and energy are
mv = 5 kg m/s mv = -.855 kg m/s mv = 2.42 kg m/s mv = 1.72 kg m/s greater and the ball’s
bmv2 = 31.25 J 5mv2 = 6.4 J 5mv2 = 7.32 J 5mv2 = 26.1 J
final velocity and ener-
total momentum = 4.15 kg m/s total momentum = 4.15 kg m/s
gy are less.
total energy = 37.7 J total energy = 33.4 J

v=16.7 m/s v=-15m/s v=7.19 m/s v = 35.08 m/s

+O) OO
m = 300 gm m=57 gm m = 300 gm m=57 gm
mv =5 kg m/s mv = -.855 kg m/s mv
= 2.16 kg m/s mv =2kgm/s
5mv2 = 42 J 5mv2 = 6.4 J mv2 = 7.74 J 5mv2 = 35.07 J
total momentum = 4.15 kg m/s total momentum = 4.15 kg m/s
total energy = 48.3 J total energy = 42.8 J

v=-15 m/s Vv = 8.32 m/s v = 43.52 m/s

()———_»
m = 200 gm m=57 gm m = 200 gm m=57gm
mv =5 kg m/s mv = -.855 kg m/s mv = 1.66 kg m/s mv = 2.48 kg m/s
5mv2 = 63 J 5mv2 = 6.4 J 5mv2 = 6.92 J 5mv2 = 54 J
total momentum = 4.15 kg m/s total momentum = 4.15 kg m/s
total energy = 68.9 J total energy = 60.9 J

Even though measurements of the COR allow us to work out the ball speed
for any given racquet and racquet speed, the COR is normally measured only
for the relatively simple case where the racquet is clamped to a heavy frame.
This measurement is simplified by the fact that the racquet speed is zero before
and after the collision, and also by the fact that the racquet can't vibrate. If we
unclamp the racquet and hold it or swing it by hand, then we would need to
measure the racquet speeds before and after the collision, as well as the ball
speeds before and after the collision, in order to measure the COR. The COR
measured this way would be the same as the COR on a clamped racquet only
if the racquet strikes the ball in the middle of the strings, since that is the only
spot where there is no energy loss due to vibrations of the racquet.

A much simpler way to characterize the performance of a racquet is to hold

the racquet at rest by hand and fire a ball at it, at right angles to the string
plane. If the ball is incident at speed v,, and bounces at speed v,,,, then we
define the ratio of these speeds by the formula

ACOR = e, = V,,/ V,, (11.9)

122 Chapter 11

The value of e, is typically about 0.1 near the tip of

ACOR vs Distance from Tip of Racquet
the racquet, about 0.4 in the middle of the strings, and
Tip Middle of strings Throat about 0.5 near the throat. These values increase slight-
0.5 ly with heavy racquets and decrease slightly with light
racquets. A typical graph of ACOR vs impact position
0.4 4 on the strings is shown in Figure 11.5. All racquets and
all strings give a similar result to this in that the ACOR
£03 is largest near the throat and smallest near the tip. If
5 | you think about this for a second, it means that rac-
i. 0.2 quet power is largest near the throat and smallest near
| the tip. If the racquet is at rest when the ball hits it,
] then the power is a maximum near the throat.
os However, if you swing the racquet to hit the ball, then
the point of maximum power shifts away from the
: throat to a point near the middle of the strings. That’s
2 Ww ie because the middle of the strings is traveling at a high-
Distance from Tip (inch) er speed than the throat so the ball then comes off
Figure 11.5 faster from the middle of the strings.

The above method of determining ACOR by firing balls at a stationary hand-

held racquet, requires you to be able to measure the speed of the ball both on
the way in and on the way out. That requires instrumentation. You can also
determine ACOR with a bounce procedure like we used to determine COR. We
found the COR of a racquet by putting it on the floor and stepping on the han-
dle to secure it. Then we dropped a ball from 100 inches onto the middle of
the strings. This made the before and after speed of the racquet zero, so
Equation 11.5 became e = v,,/ v,,, or written differently, v,, = ev,).

We can do the same thing for a hand-held racquet (Figure 11.6). Everything is
the same except we are holding the racquet instead of clamping it down. We
again drop the ball from 100 inches (though any height will do) in the middle
of the strings to prevent frame vibrations (though we don’t have to). This time
the ball bounces to about 16 inches. If we use the bounce height rule that we
used to determine COR, we get

e, = V(h,/hy) = V16/100) = 0.40.

Ball Bounce Heights — Three Scenarios

Q
Q
78 in.
55 in.
16 in.

2 in.
Figure 11.6

Typical bounce heights for a SS S|

Q
ee ea

tennis ball dropped from a

height of 100 inches. Concrete (Head clamped)
 Energy and COR 123

That is much less than COR. The reason is that with a hand-held racquet, the
racquet will recoil at impact. How much and how fast it will recoil depends
primarily on its weight and swingweight. You can drop the ball onto any posi-
tion on the stringbed and measure the ACOR for that location. The bounce
height will reflect the differing amount of translation, recoil and vibration that
accompanies each impact location.

The ACOR is relatively easy to measure compared to the COR, since it does not
involve any simultaneous measurement of the racquet speed. Furthermore, as
we will see, a measurement of the ACOR tells us everything we need to know
to work out the ball speed, for any given racquet speed at any given impact
point (not just at the center of mass), since the ACOR value is measured under
realistic conditions where the racquet can recoil and vibrate as much as it
wants to. Knowing the ACOR, we can work out the ball speed without using
conservation of momentum, without using conservation of energy, and with-
out even knowing the COR. We don’t even need to know the mass of the rac-
quet. The ball speed is the result of the interplay of all these things. They are
necessary to explain why the ball speed is what it is, or to theoretically calcu-
late it. But if we measure the ball speed off a hand-held racquet, then we
already know what the ball speed is. There is no need to calculate something
if we have already measured it! And because that ball speed is the result of the
interplay between the different energy sharings, effective masses, translations,
recoils, and vibrations that occur at different impact points, the formulas
derived from the ACOR relation are predictive for all impacts at all locations at
all impact point velocities. With ACOR, then, we can relieve our restrictions
that impacts are at the center of mass. (Chapter 12 will show us how to relieve
them for all our other calculations, as well.)

Using ACOR for calculating final velocities greatly simplifies the math. It's not
as easy as you're thinking, however. You can't just take any collision and mul-
tiply e, times v,, and get the final ball velocity, unless the racquet has no ini-
tial velocity, such as our hand-held case. ACOR itself is only defined for a
bounce off a racquet at rest, but from that data, we will be able to derive for-
mulas for final ball speed given any initial ball and racquet speeds.

I can hear some readers complaining that a measurement of ball speed off a
stationary racquet has nothing to do with the bali speed off a racquet that is
swung vigorously towards the ball. What has a measurement off a racquet at
rest got to do with say a high speed serve? Surely these two evenis are com-
pletely different so the ball speed will be completely different. True, the ball
speeds ARE completely different, but the events are NOT completely different.
In fact, the events and the ball speeds are exactly the same if we observe them
in the correct frame of reference.

Changing the frame of reference is something that physicists often do to save

the hassle of working things out the hard way. Suppose you are in a train trav-
elling at 100 mph and you drop a ball vertically onto the floor. What you see
is that the ball drops in a straight line at right angles to the floor. Someone out-
side the train looking in sees the ball travelling horizontally at 100 mph and
curving downwards towards the floor. Both of you see exactly the same event,
but the ball speed and the trajectory are totally different depending on whether
you are inside or outside the train. Exactly the same sort of argument can be
used to calculate the speed of a ball served from a high speed racquet if we
already know the bounce speed off a stationary racquet. So, let’s do it.
124 Chapter 11

Fina
Key
se Cin. 8 COO ii ltt, «—._K—-—WN~W”“ twitLams Whe

1. Serve Speed Formula.

In order to calculate the serve speed from the ACOR data, it is necessary to fig-
ure out what happens when things are the other way around. The ACOR is
measured when the ball is incident on a racquet at rest. What we are really
interested in is what happens when the ball starts off at rest and the racquet
slams into it. That's what happens when you serve the ball. The ball is not
quite at rest when you serve it, but it’s almost at rest. It won't make any sig-
nificant difference to the calculation if we assume that the ball is actually at
rest, suspended in mid-air, when the racquet slams into it.

There is a very simple way to figure this out. First, we

BEFORE collision AFTER collision go back to the ACOR measurement. Suppose that the
racquet is at rest and you fire a ball to hit the racquet
at speed V. The ball will bounce off at speed v.
Suppose also that the ball approaches at the right and
it bounces off towards the right, as shown in Figure
11.7. In this case, ACOR = v/V.

Racquet at rest Since ACOR is only defined for a bounce off a station-
ary racquet, we must turn Figure 11.7 into a serve the
Figure 11.7 same way the train turned the vertically dropping ball
into a horizontally flying ball in the example above. We
can do this by changing the frame of reference by
adding a fixed velocity to all the speeds shown in
BEFORE collision AFTER collision Figure 11.7. If we add velocity V to each element of
Figure 11.7, we get the results shown in Figure 11.8
V
Y
eo, The racquet is now moving at speed V to the right, the
ou initial ball velcoity is 0 (adding V in the opposite direc-
tion cancelled its previous velocity), and the ball is
Ball at rest
moving speed v + V to the right. We have added V to
the racquet after the collision too, but we didn’t know
its velocity before the collision in Figure 11.7, so all we
know is that V will be enough to move the racquet to
Figure 11.8
the right.

From the definition e, = v/V, the bounce speed v in Figure 11.7 is given by v
= €, V. From Figure 11.8, we see that the serve speed v, = v + V. Consequently,
V,= & V+ V = (1 + €,)V, so we get THE MOST IMPORTANT BOQUATION IN
THE BOOK):

Serve Speed Formula

v, =(1+e,)V (11.10)

Nothing could be simpler. To determine the serve speed, all you have to know
is ACOR and the racquet speed at the impact point.

Let's look at a specific example as shown in Figure 11.9 and Froure Tir 1o:
Suppose that the racquet is at rest and you fire a ball to hit the racquet at 80
mph. If ACOR = 0.4, then the ball will bounce off at 0.4 x 80 = 32 mph.
Suppose also that the ball approaches at 80 mph from the right and it bounces
 Energy and COR 125

BEFORE collision AFTER collision BEFORE collision AFTER collision

|80 mph

Ball at rest
Racquet at rest e, = 32/80 = 0.4

Figure 11.9 Figure 11.10

off at 32 mph towards the right. Now add 80 mph to the right to ALL these fig-
ures, as shown in Figure 11.10. Then the racquet is moving at 80 mph to the
right, the ball is initially at rest, and the ball bounces off at 32 + 80 = 112 mph
to the right. If the initial conditions were plugged into the serve speed formu-
la, we get

ve Gi 040)80 = 112 mph

2. General Ball Speed Formula.

The general formula follows the same logic. By BEFORE collision AFTER collision
“general” we mean that it encompasses any value
of the initial ball speed and racquet speed, includ-
ing the serve, though the serve is a special case
because the ball is initally at rest and has its own
simpler formula. To begin, refer to Figure 11.11
This looks similar to Figure 11.7, but we have
labeled the velocities differently because we are
going to have more terms to keep track of. Figure 11.11

Again, we change the frame of reference, from the

laboratory frame, where the racquet starts off at
BEFORE collision AFTER collision
rest, to the court frame, where the racquet is swung
towards the ball. To do that, we just add V to ali
the velocities shown in Figure 11.11 , where V is
the racquet speed in the court frame of reference
Crigure. £1702).

Again we do a little math:

Figure 11.12
Givens:
e, = V,/v, (measured in the lab frame)
Vi, = V, - V = the incident ball speed in the court frame
Von = + V Grom Figure 11.12)

Calculation:
Vere; (from above)
Vaomi
1
Ph. (from above)
$0), ie Vn V) OV Sov eV
126 Chapter 11

And there we have our general ball speed formula — THE SECOND MOST
IMPORTANT FORMULA IN THE BOOK:

General Ball Speed Formula

Vout = CaVin + (1 + &,)V (11.11)

Equation (11.11) is actually more general than (11.10) since we recover (11.10)
as a special case when v,, = 0. In that respect, (11.11) is more important than
(11.10), but we do not use it as often in this book since we have focussed more
attention on the serve than on groundstrokes. The serve is the power shot in
tennis, the biggest weapon in a modern player's game, and the stroke that
stretches our understanding of the underlying physics (and the strings) the fur-
thest.

Equations 11.10 and 11.11 look a lot friendlier than Equations 11.7 and 11.8. It
all begins with measuring e, with a simple bounce test, and then plug into the
equation any values you want in order to determine the results of any hypo-
thetical collision.

We have come a long way to arrive at some easy to use formulas to determine
after collision racquet and ball velocities, but in the process, we have learned
why the formulas are what they are. Understanding is often the most impor-
tant part of application. But our journey isn’t quite finished. Many of our expla-
nations of how and why things happen as they do have, so far, been based on
two assumptions — that collisions are on a line through the center of masses
of the ball and racquet, and that the racquet is traveling in a straight line with
all its parts traveling at the same speed. Though ACOR released us from these
restraints for making predictions, there are still are two missing ingredients to
explain why ACOR is what it is. Since you usually don’t hit the center of mass
and the racquet is typically swinging in an arc around some axis, different
impact points of the racquet act as if they weigh more or less than other points,
and different impact points travel at different speeds that others. So, what is the
mass and what is the velocity of the impact point?

Chapter 12 will deal with this question. There we will learn how to determine
“effective mass” and how to calculate the velocity of the impact point. It will
then show us how to incorporate these into our formulas. Once you have
learned these concepts, you will always use effective mass and impact veloci-
ity in all your equations.
 isions
Effective Mass
and Velocity at
the Impact Point

Andres Gomez, Ecuador

By Rod Cross

_n this chapter we will examine several aspects of the collision between a racquet and a ball that were not
considered in the previous two chapters. In particular we will show that:

e If you swing a racquet at a certain speed and keep increasing the mass of the racquet, then your serve speed
will keep increasing. That’s not surprising. The surprising part is that the ball speed can never be greater than
twice the racquet speed, even if the racquet is as heavy as a truck and even if all energy losses are eliminated
from the ball. There is a limit at which any further increase in racquet mass will actually decrease the ball speed
since you will not be able to swing a truck as fast as a racquet.

e The function of the racquet strings is to store elastic energy during a collision with the ball and then give
almost all of it back to the ball. When a ball is dropped onto concrete, all of the kinetic energy in the ball is
converted to elastic energy in the ball since the ball comes to a complete stop on the concrete before it bounces
upwards. When a racquet is swung at a ball, only a small part of the kinetic energy of the racquet is converted
into elastic energy in the strings and the ball, since the racquet does NOT come to a complete stop. As a result,
energy losses in the ball and the strings have only a small effect on the speed of the ball.
128 Chapter 12

e When a ball strikes a stationary racquet anywhere on the strings, the racquet
recoils and it also rotates. The combined recoil and rotation speed at the point
of impact is greater than that for an impact at the racquet’s center of mass. As
a result, the racquet behaves at the impact point as if its mass is less than its
actual mass. The effective mass of a racquet near the tip is about the same as
the mass of a tennis ball. In the center of the strings, the effective racquet mass
is about half the actual mass.
BEFORE collision AFTER collision

At rest
Suppose that a racquet of mass M traveling at speed
Om
V, strikes a ball of mass m at rest, as shown
Figure 12.1. That is the usual situation when a ball
in

is being served. After the collision, the racquet con-

tinues in the same:straight line at speed V, and the
ball is served at speed v,. The only additional infor-
Figure 12.1 mation needed to calculate the serve speed is the
value of the COR for this collision, but that can easily be measured as described
in Chapter 11. There is no need to know anything about the force on the ball,
or the force exerted by the player on the racquet, or the time the ball spends
on the strings or the diameter or stiffness of the ball, or the stiffness of the rac-
quet. All those things are important in themselves but they are not needed to
calculate the serve speed. All that is needed is the equation describing conser-
vation of momentum (as described in Chapter 10) and the value of the COR
(described in Chapter 11). We also need to know the correct value of the rac-
quet mass, since it depends on the impact point on the strings. The correct
value is known as the “effective mass”. It is typically about half the actual mass
of the racquet but it can be calculated as described at the end of this chapter.

The total momentum after the collision is equal to the momentum of the rac-
quet before the collision, so

MV, = MV, + mv, (12.1)

If there is any energy loss during the collision, the relative speed after the col-
lision will be less than the relative speed before the collision. The ratio of the
relative speeds is given by

e=(v,-V,)/V, (12.2)
where e is the coefficient of restitution (COR), typically about 0.88 for a rac-
quet and ball collision. The speed of the racquet after the collision is therefore
given from Equation 12.2 by

V. = Vp - eV, (12.3)

If we substitute for V, in Equation 12.1 and solve for Vv, we get

MV, = MW, - eV). + mv,
MV, = Mv, -eMV, + mv,
MV, = v(M> m) =eMvV,
MV CMV = 7m)
My Cle) = v.(M + m)
 Collisions 129

SO
(1 +e) MV,
oe (M +m) (12.4)
That gives us the speed of the ball. If we now substitute for v, in Equation 12.3,
we find the racquet speed after the collision:

V, Ver ENG

MVC +e)
A Pa eee eV,
5 M+m

V(M +m) = MV -eMy, -eMy,—emv,

Ve a) = V,(M - em)

SO,

(M - em)V,
2" (M+ m) (12.5)
Equations 12.4 and 12.5 are the same as Equations 11.7 and 11.8 for the case
when the ball is initially at rest. There is a lot of interesting information here.
The first thing to notice from Equation 12.4 is that v, is proportional to V,. If
you double the racquet speed, then the ball speed is doubled. If you could
increase the racquet speed 100 times, the ball speed would increase 100 times.

The second thing to notice is that the ball speed increases when e increases,
but it is not zero when e is zero. The maximum possible value of e is 1.0,
which corresponds to a hypothetical situation where there is no energy loss in
the ball or the racquet. The minimum possible value of e is zero which corre-
sponds to a hypothetical situation where all the elastic energy stored during the
collision is lost. Elastic energy is stored when the ball compresses and when
the racquet strings stretch. If all that energy is lost, then the ball will still be
served at relatively high speed, but it is only half the speed that would result
if none of the elastic energy was lost. When e = 0 we see from Equation 12.3
that V, = v,, which means that the ball sticks to the racquet and both move at
the same speed after the collision.

Now consider what happens when M is increased. When M = 0 we see from

Equation 12.4 that v, = 0. Obviously, a racquet of zero mass packs no punch
at all. Now suppose that M = m and e = 1. Then, again using Equation 12.4,
we find that v, = V, and V, = 0 (from Equation 12.5). In this case, the racquet
stops dead and the ball takes off with the same speed that the racquet had
before it hit the ball. This is what happens when one billiard ball strikes anoth-
er head-on and when the second ball is initially at rest.

Now let M = 2m and e = 1. Then v, = AN 1/3 so the ball comes off the racquet
faster than the initial speed of the racquet. That’s what every tennis player likes
to see, SO let's try a racquet with M = 10m, Then vy, = 20V,/11 = 1.82V, if ¢ =
1. That’s nearly twice as fast as the initial racquet speed. That’s even better.
How about M = infinity? Then we can forget about m in the bottom of Equation
12.4, and we get v, = 2V, if e = 1. That’s as fast as we can possibly get. How
come? One could reasonably expect that if M = infinity then the racquet has an
130 Chapter 12

infinite amount of energy and momentum so the

ball should be served with infinite speed (or
maybe at the speed of light).

BEFORE collision AFTER collision The easiest way to see what’s happening is to
change the frame of reference (as was explained
At rest At rest in Chapter 11), as shown in Figure 12.2. A ball
that strikes a brick wall at speed v, will bounce
off the wall with speed v, if e = 1. A bug flying at
speed v, beside the ball sees something else. The
bug’s view is obtained by adding v, to the speed
of the wall and the ball. The bug sees the brick
wall approaching the ball at speed v,. It sees the
ball is initially at rest, and it sees the ball being

: ©
At rest
served at speed 2v,. In the reference frame where
the wall is at rest, all that happens is that the ball
bounces off the wall with the same speed as it hit
What the bug sees
the wall. In the frame where the ball is initially at
What the bug sees
rest, the wall collides with the ball and squashes
Figure 12.2
it. The ball acquires the same speed as the wall,
but in addition, it propels itself off the wall at v,,
Serving the ball with a brick wall.
so it is served at twice the speed of the advanc-
ing wall.

In the previous section, in Equation 12.4, we saw that the ball is served at
speed

ve Cre My OL mm)

when it is struck by a racquet of mass M moving at speed V,. An interesting

question is how much elastic energy gets stored in the racquet and the ball dur-
ing the collision. When racquet and string manufacturers claim that their rac-
quet or their string is more powerful, what they are really saying is that they
can somehow reduce the elastic energy that gets lost during the collision. If
none of the energy is lost, then e = 1, and the ball will be served at speed v,
= 2MV, / (M + m). If all of the elastic energy is lost, then e = 0, and the ball
will be served at speed v, = MV, / (M + m). There is a factor of two difference
here, but that is not really relevant because the rules of tennis require that the
ball must lose about 45% of its stored elastic energy when it has a collision. So
the manufacturers can increase the ball speed not by a factor of two but by
something that is a lot less than a factor of two.

There are two ways to figure out how much elastic energy is stored during the
collision. One way is to assume that e = 0, in which case all of the elastic ener-
gy is lost, and that tells us how much elastic energy was available. The other
way is to change the frame of reference so that the momentum of the racquet
is equal and opposite the momentum of the ball. Then the total momentum is
zero before, during and after the collision. During the collision, both the rac-
quet and the ball come to a complete stop, in which case all of the original
kinetic energy gets converted to elastic energy. We will try the e = 0 method.
 Collisions 131

Suppose that e = 1, then, from Equations 12.4 and 12.5

y= MV Mm) and V,-= (Mam.

/ OMe tim)

The total energy before the collision is iMV,’, and the total kinetic energy after
the collision is the sum of the racquet energy and ball energy:

Total energy after collision = 1Mv,” * :mv,”

Substituting v, and V, from above, we get

(M-m)V, 7? 2MV, 72
became ees Re
Nile
(M + m)
|
(M + m)
|

(M - m)’V,’ 4M’V
Nileboreal
(M + my? ean
‘ ergs

foe -myVv, 4mM’vV/

ele Re |
j (M + m)* (M + m)”

3 (M? - 2Mm + m2 + 4mM)

MV, | -————" |
(M + m)

, (M? + 2mM + m° )
mv,;[ 3 : |
(M° + 2mM + m°)

2 2 hes
1 mv, + i inves 5MV, (energy after collision equals energy before)

If e = 0, then from Equations 12.4 and 12.5 we get

VV COM th) and V, = MV, / (M + m)

132 Chapter 12

so the total energy after the collision is

MVy a> MV, 772

1 eres = = et [{_—~ ]
aa ae: Mle + = 2™ La + m)

(MV,)°
= +(M +m) |
; (M + m)"

(MV,)° <a
= = Total energy after collision
2(M + m)

The energy lost, E, is given by the energy before the collision minus the ener-
gy after the collision:

ep agMD
E = Energy lost =; MV," - ————
(M + m)

M
E =3 MV," [: = —
- (M + m)

: m
E =i mv,’ [——]
(M + m)

Since we have assumed that e = 0, all the elastic energy stored during the col-
lision is lost. The elastic energy stored during a collision is therefore equal to
E. In any other head-on collision between a moving mass M and a stationary
mass m, the amount of elastic energy stored during the collision will be equal
to E, but the amount of energy that is subsequently lost will be less than E if
e > 0. The fraction, f, of the initial energy that is stored as elastic energy is
therefore

E m
= 12.6
1MV,? = (M+m) ate

For example, if M = m then f = 1/2, meaning that half of the initial kinetic ener-
gy of the racquet gets stored as elastic energy in the ball and the strings and
the racquet frame during the collision. If all that energy is recovered, then e
will be 1.0, but if all that energy is lost then e will be zero. If M = 2m then f =
1/3, meaning that only one third of the initial kinetic energy of the racquet gets
stored as elastic energy.

The ball is required to lose about 45% of its stored elastic energy and the
strings lose about 5%. The string plane stiffness is typically about the same as
that of the ball, so the elastic energy stored in the strings is about the same as
that stored in the ball (see Chapters 13 and 32). About one quarter of the total
elastic energy is therefore lost during a racquet and ball collision. If M = m, the
 Collisions 133

total energy loss amounts to about 1/8 of the initial energy of the racquet. If M
= 2m, it amounts to about 1/12 of the initial energy of the racquet. Changing
the energy loss in the strings or the racquet frame by 2 or 3% therefore has
very little effect on the ball speed (Figure 12.3).

Energy Flow Between Racquet, Strings, and Ball

KE in racquet

KE in ball
2/3 | g racquet 0.774 | KE in ball
1c

PE in ball
1/6 (45% is lost)
ow 0.143 |KE in racquet
PE in strings
(5% is lost) pe |1/12 |Energy lost

Before collision During collision After collision

Figure 12.3

Energy fractions when M = 2m, f = 1/3, e = .866 and when the elastic potential ener-
gy (PE) is shared equally between the ball and the strings.

You might wonder why we are assuming here that M = m or 2m when the actu-
al mass of a racquet (about 340 gm) is typically about 6 times larger than the
mass of the ball (57 gm). That’s because the effective mass of a racquet at the
impact point is less than the actual mass, as we will shortly explain. It will take
a fair amount of explaining, but consider this. If you hold a racquet by hand
so that the strings are horizontal, and drop a ball near the tip, then the ball will
bounce very weakly or not all. That’s because the effective mass near the tip
of the racquet is about 57 gm, about the same as the ball. We saw above that
when two billiard balls collide head-on, one of them stops dead. The same
effect occurs when you drop a 57 gm ball on a 57 gm racquet.

For an impact in the middle of the strings, the effective racquet mass is typi-
cally about half the actual racquet mass. For a 340 gm racquet, this gives M =
3m. There is no energy loss due to vibrations of the racquet for an impact in
the middle of the strings, so we only need to consider the losses in the ball
and the strings. If M = 3m then the fraction of the initial kinetic energy stored
as elastic energy during the collision is f = 1/4. If the ball and the strings are
equally stiff, half of the elastic energy is stored in the ball and half in the
strings. Given that the ball loses about half of its stored energy and the strings
lose almost none, then 1/4 of the stored elastic energy is lost. The energy lost
is therefore 1/16 or 6.25% of the initial racquet energy. From Equations 12.4
and 12.5, it can be shown that fraction of the energy lost when e is finite is
134 Chapter 12

given by

f. = 2E/(MV,) = 1 - e*) m/ (M + m)

which indicates that f, = 0 when e = 1 and it agrees with Equation 12.6 when e
= 0. If M = 3m and e = 0, thenf, = 1/4. However, if M = 3m and if e is not zero,
then f, = 1/16 for realistic losses in the ball and the strings. Hence

1-e? = 4/16 = 0.25 so e = 0.87

which agrees closely with measurements of the COR, at least for low speed
impacts (see Note 1).

ROTATION
We assumed above and in Chapters 10 and 11 that racquets move in straight line
paths before and after a collision with a ball. In fact, the motion of a racquet
usually involves rotation about some axis as well as translation along a straight
line. A racquet that moves in a straight line can be described in terms of its mass,
its momentum and ‘its kinetic energy. A rotating racquet is described in terms of
its rotational inertia, its angular momentum and its kinetic energy. The first two
quantities are analogous to mass and momentum but there are differences. The
differences are due to the fact that when a racquet rotates about an axis, points
close to the axis rotate at a lower speed than points further from the axis.
Consequently we cannot describe a rotating racquet simply in terms of its speed
in m/s. Instead we describe a rotating racquet or a rotating ball in terms of its
angular speed, measured in revolutions per second or revolutions per minute or
in radians/sec.

If a racquet is rotating at one revolution per second, then it rotates at 60 revo-

lutions per minute and it rotates through 360 degrees per second. In physics we
use another measurement of angle called the radian, and we do so for a good
reason. One radian is the same as 57.3 degrees, and 2m radians is equal to 360
degrees. One revolution/sec is therefore the same as 2m = 6.28 radians/sec.

Most students of physics are uncomfortable with radians when they first hear
about them, but there is no logical reason to be afraid of radians. In fact, hav-
Figure 12.4
ing 360 degrees in a circle instead of say 100 or 1000 is even less logical. We
Relation between radius can blame the Babylonians who had troubles with fractions, and who found that
and arc length. 360 can be divided by lots of numbers without having anything left over.

The reason that we use radians is that it makes it easier to calculate

iy, BD
the length of an arc of a circle. The length of the circumference of a
circle of radius r is 2mr and it spans an angle of 360 degrees or 2m radi-
ans. The length of any section of this circle is found by simple pro-
portion. For example, the length of a quarter of a circle is mr/2 and it
spans an angle of 90 degrees or x/2 radians. This means also that the
length of an arc spanning one radian is r, and the length, s, of an arc
spanning 9 radians is given by s = r 8 (see Figure 12.4).

een An o.
: Relation Betw

If a racquet or any part of the racquet is rotating in a circle of radius r, and if it

rotates by an angle 0 in a time t, then it traces out a path of length s = r 0 in
 Collisions 135

time t. The speed is v = s/t = r0/t and the angular speed is defined as w = 0/t
(in radians/sec). Hence we have the useful results that

s=r6 and v=ro (12.7)

Consider a racquet of length 70 cm that is rotating

at 60 rad/sec about an axis 10 cm beyond the end
of the handle as shown in Figure 12.5. The butt end
of the handle is moving in an arc of radius 10 cm
at a speed v = rw = 0.1 x 60 = 6 m/s, and the tip is
traveling in an arc of radius 80 cm at speed v = ro
= 0.8 x 60 = 48 m/s. It doesn’t make much sense to
talk about the speed of the racquet since every
point is traveling at a different speed. But every
point is rotating at the same angular speed (60
rad/s) and every point rotates through the same
angle during the same interval of time (60 x 57.3 =
3438 degrees/sec or 3.438 degrees every 1 millisec-
ond).

Armed with information on the angular speed of a

racquet, its axis of rotation and values of the COR
(e) or the ACOR (e,) for that racquet, it is easy to
Figure 12.5
locate the maximum power point of the racquet.
Rotation of a
An example is shown in Figure 12.6 where we
tennis racquet.
have assumed that the racquet is swung as shown
in Figure 12.5 with @ = 60 rad/sec. Note that values of e, are shown in Figure
12.6 rather than values of e, since it easier to measure e , than it is to measure
e. The racquet speed is shown at three different points and is given by V = 60r
since we are assuming that w = 60 rad/sec.

As explained in Chapter 11, the serve

speed at any impact point is given by
v = (1 + e,)V where V is the speed of
the racquet at the impact point. In the V=ro v=(1+e,)V
example shown in Figure 12.6, the Pe ee a —mi— e, =0.1,|V= 42 mis, serve speed = 1.1 x 42 = 46.2 m/s
serve speed is about 8% larger when
serving from the middle of the strings w=60 r=0.6m —a—- @, =0.4,|V= 36 mss, serve speed = 1.4 x 36 = 50.4 m/s
than when serving from near the tip.
Top players tend to serve from near ean er —~—a— _e, =0.5,/V= 30 ms, serve speed = 1.5 x 30 = 45.0 m/s
the tip because the added _ height
advantage reduces the chance of serv-
ing a fault. However, they usually cus-
tomise their racquets by adding lead
tape at the tip, which acts to increase
e, in that region. An increase in e,
from 0.1 to 0.2 would increase the
serve speed near the tip to 1.2 x 42 =
@ Axis of rotation
50.4 m/s, in which case the serve
speed near the tip would then be the
same as the serve speed off the center Figure 12.6
of the strings. Further details on cus-
Serve speed at various impact points.
tomizing racquets in this manner are
described in Chapter 106.
136 Chapter 12

_ __ _ ® MomentofInertia
A rotating racquet is moving and therefore has kinetic energy. The kinetic
energy is 4 Mv~ where M is the mass of the racquet, but what is v when every
part of the racquet has a different speed? The answer is that v is an average
speed, but it is not a simple average since we need to find the average value
of v’ rather than the average value of v. If the racquet is rotating about a fixed
axis, then the average value of v is actually zero since the whole racquet is
going nowhere — it is just stuck on its axis and is spinning round and round
on the same spot.

Suppose we mentally divide the racquet into 100 small pieces all of mass m =
M/100. The kinetic energy of each piece is 5my’, where v = ro is the speed of
each piece, and r is the radius of each piece G.e., the distance to the axis) .
The kinetic energy of each piece is then 5mr’, so the total kinetic energy of
the whole racquet is

KE = im(ry + ty” + oe + Fy99) © HSI 0° (12.8)

where

Tem(h. et ence e ee (12.9)

Ican be calculated for simple objects of simple shape and is called the moment
of inertia of the object. For example, the moment of inertia of straight rod or
. ‘ iD} - : : .
beam of length L and mass M is I = ML*/12 when the axis of rotation is at right
angles to the rod and passes through the middle of the rod, as shown in Figure
12.7. The moment of inertia for several other objects is also shown in Figure
WAFe

A racquet is not a simple shape but its

moment of inertia is almost the same as
that of a straight rod of the same mass
Axis and length — at least when it is swung
as l=ML7/12 cb , about its usual axis. In that case we often
|= 2MR/3 for thin say that the moment of inertia is the
spherical shell swingweight of the racquet, despite the
fact that the moment of inertia is not a
weight and it is not even a mass. So, what
is this thing called “moment of inertia”?
cb There are two answers, and you can take
your pick. The first is that it is the thing
|= 2MR2/5 for defined by Equation 12.9. The second is
solid sphere that is a measure of the resistance or iner-
tia of an object to rotation. Consider two
tennis balls, both of the same mass and
diameter. One has all its mass near the
Uniform rods or racquets Uniform balls outside of the ball and is hollow inside,
just like a normal ball. The other is made
from a small, 40 gm ball of lead in the
Figure 12.7
middle and is surrounded by 17 gm of
Moment of inertia for common objects. rubber foam designed so that both balls
 Collisions 137

bounce the same. The second ball has most of its mass in the middle and has
a lower moment of inertia. The result will be that the second ball spins faster
than the first since it has a lower resistance to rotation. Maybe I should patent
that idea. If Imade a golf ball like that I could hit it a lot further (due to the
increased lift force).

Angular Momentum _ slid

Z

A racquet that is rotating about a fixed axis is moving and therefore has
momentum. If we divide a racquet of mass M into 100 small pieces each of
mass m = M/100 then the momentum of each piece is mv. However if we add
up the momentum of all 100 pieces the answer will be zero since for every
piece moving left to right there is usually another piece moving right to left.
Alternatively, if we look at the racquet in Figure 12.5 we see that the racquet
spends half its time moving left to right and the other half (when it is below
the axis) moving right to left. This is inconvenient since the momentum of any
object rotating about a fixed axis is either zero, or its time average value is zero.

Consequently we need to define the momentum of a rotating object in a slight-

ly different way to the momentum of an object moving in a straight line, and
we give it a different name. The angular momentum of an object that is rotat-
ing about a given axis at angular speed w is given by Im, where I is the moment
of inertia about that axis. This is analogous to ordinary momentum Mv in that
we replace the linear velocity v by the angular velocity wm and we replace iner-
tia or mass M by rotational inertia I. The driving force in translational motion
is an ordinary push or pull. The driving force in rotational motion is called a
torque or a “moment”, which acts to twist or rotate an object. When you use a
screwdriver or a spanner or a steering wheel you apply a torque to rotate it.
The torque or moment is given by the force exerted multiplied by the distance
to the axis of rotation. The torque that spins a tennis ball when you hit it is the
friction force between the ball and the strings (acting parallel to the string
plane) multiplied by the radius of the ball. Everything concerning rotational
motion is analogous to motion in a straight line except that we need to multi-
ply or divide everything by the radius or the radius squared.

The formula for the angular momentum of any object is worked out by sup-
posing that an object of mass M is made up of lots of little pieces each of much
smaller mass m. If one particular piece has a mass m, a speed v, and is at a
distance r from the axis of rotation, then its linear momentum is mv and we
define its angular momentum by mvr. In other words, we take the linear
momentum, multiply it by the distance to the axis and call it the angular
momentum. From Equation 12.7 we know that v = ra so the angular momen-
2
tum of each piece is mvr = mr°w. The angular momentum, AM, of the whole
object is then

AM = m(t,? + fy) + oe + Fyq9) © = To (12.10)

where we have assumed that the object is divided into 100 pieces each of mass
m = M/100. Dividing the object into an infinite number of pieces will give a
more accurate result. That is easy to do if you understand calculus, but we are
assuming that most readers are not familiar with calculus. If you are, then you
will probably know a lot about moments of inertia and angular momentum
anyway.
138 Chapter 12

Like linear momentum, angular momentum is conserved in a collision of two

objects, provided there is no other force or torque acting on the colliding
objects. If a diver does a somersault into a swimming pool, the diver will spin
faster if he or she tucks in the arms and legs to make sure that all the differ-
ent bits of the body are close to the axis of rotation. Adding up all the mr’ val-
ues of each bit results in a small moment of inertia. While the diver is in the
air there is no torque on the body (given that gravity acts through the center
of mass) so angular momentum is conserved. The diver spins faster by reduc-
ing I and then stretches his or her arms and legs out to increase I and to reduce
w just prior to entering the water. During this time Im remains constant so it is
“conserved.”

In Chapters 10 and 11 we confined our attention to racquets that impact the

ball at the center of mass of the racquet, somewhere near the throat of the rac-
quet. We did that for two reasons. One was to simplify the maths and the other
was because we knew that we would eventually get the correct result at any
other impact point if we replaced the actual mass of the racquet by its effec-
tive mass. It is now time to set the record straight and prove that we can indeed
get such a result. However, we will do so only for a case where there are no
energy losses, since the maths gets a bit too complicated if we allow for ener-
gy losses in the ball and the racquet. Energy losses are best treated in terms of
the measured values of the ACOR for any given racquet, but for now we just
want to look at the essential features of a racquet that rotates when it collides
with a ball. We will allow the ball to strike the racquet at any point along the
length of the racquet, but to simplify the calculations, we will assume that the
racquet is initially at rest (see Note 2).

The situation is shown in Figure 12.8. A ball of mass m is incident at speed vi

and strikes the racquet at a point at a distance b above the center of mass (CM)
of the racquet. The ball bounces off at speed v,, the CM of the racquet recoils
at speed V.., and the racquet rotates clockwise about its CM at angular speed
0),.

A i

O-- +t ~© |L-
V1 Vo ee

: b b

: 1 -+'1CM CM Vem

BEFORE -_ AFTER ae
Collision collision

At rest ip,
eles 6
Figure 12.8
A ball incident on a racquet at rest.
 Collisions 139

Applying the conservation laws of linear and angular momentum as well as

kinetic energy, we find for the situation in Figure 12.8 that

mv,=MV,,-mv, (12.11)

mv,b=Io,-mbv, (12.12)

and

pMv, = 5mv, + MV,

+ g1 aw,” (12.13)

The ball is not rotating about its own axis, but it has angular momentum mvb
about the CM of the racquet. In applying the conservation of angular momen-
tum law, we need to consider the total angular momentum of the ball and the
racquet about a common axis. In this case, the best axis to work with is the
axis through the CM of the racquet. Like linear momentum, angular momen-
tum is taken to be positive or negative depending on the direction of rotation.

From Equations 12.11 and 12.12 we find that

Vem = M(v, + Vv.) /M (12.14)

and

@, = mMb(v, + v,)/1 (12.15)

Substituting Equations 12.14 and 12.15 into 12.13 gives

2 2 2 2 mad 2
mv, =mv, + m(v,+ v,)/M+m Biv, + Vv)

SO

V,? =v," + m(1/M + b?/ I)(v, + v,)° (12.16)

The quantity 1/M + b’/I is not easily recognisable as anything of importance

but it is equal to 1/M, where M, is the effective mass of the racquet at the
impact point. Consequently we will define M, by the relation

ery (12.17)

in which case, Equation 12.16 can then be written as

ee rel? Z
Van 95, FIV atv) ML

Expanding this out gives

vl +m/ MM). 2 m.va/ Mv, = C1. - m/Mv," = 0)

140 Chapter 12

which is a quadratic equation with solution

- m/M
gop Mala (12.18)
2" (1+ m/M,)

There is another solution, v, = - Vj, but this is of no interest unless the ball
passes straight through the racquet and emerges unscathed on the other side.
Equation 12.18 can be compared with Equation 12.5 and it will be seen that
Equation 12.18 is the result that we would obtain if the ball collided head-on
with a mass M.. It is for this reason that we identify M, as the effective mass
of the racquet.

Substituting Equation 12.18 into 12.15 gives

®,=2mbv,/[1(1+m/M,)] (12.19)

An interesting result is obtained if we consider the impact point on the racquet.

This point recoils at a linear speed V = V.,, + b ,. From Equations 12.14, 12.18
and 12.19 we find that

V = 2m, 7 Gee My)

and hence the relative speed of the ball and the impact point after the colli-
sion is V + v, = v,, which is the same as the relative speed before the collision.
Consequently, the COR for this collision is 1.0, meaning that there is no loss of
energy during the collision, which is what we assumed at the start. More
importantly, it shows that in this situation the COR must be defined in terms of
the relative speed of the ball and the racquet at the impact point, and not in
terms of the speed of the CM of the racquet.

_Effective Racquet Mass

The formula for the effective racquet mass is given by Equation 12.17. This can
be calculated for any given racquet if we know or measure its moment of iner-
tia, I, but most racquets of mass M and length L have a value of I that is close
to that for a uniform rod of mass M, and length L, in which case I = ML?/12. If
we substitute this value of I into Equation 12.17 we find that

M, = M((1 + 12 b7/L?) (12.20)

In the middle of the racquet, b = 0 and M, = M. At the tip of the racquet, b =

L/2 and M, = M/4. Halfway between the middle and the tip, b = L/4 and M, =
M/1.75. The center of the strings is about 20 cm from the CM of a 70 cm rac-
quet where M, = M/1.98.

The effective mass is therefore smaller than the actual mass at all impact points
other than the balance point in the middle of the racquet. The reason is as fol-
lows. If a force F is applied at the middle of the racquet then the whole rac-
quet will accelerate in the direction of the force according to the relation F =
Ma, where M is the actual mass of the racquet. The acceleration a = F/M is
inversely proportional to the mass M. A light racquet will accelerate rapidly to
 Collisions 141

high speed and a heavy racquet will accelerate slowly at a lower speed. Now
suppose the force is applied at a point closer to the tip of the racquet. The
whole racquet will still accelerate in the direction of F with a = F/M, but in addi-
tion, the racquet will rotate about its center of mass. As a result, the point on
the racquet at which F is applied will move in the direction of F at a speed that
is faster than when F is applied at the center of mass. In other words, that point
behaves as if it had a mass that was less than the actual mass of the racquet
since its acceleration is greater than F/M.

We can quantify this effect as shown in the Appendix of this chapter.

Example

Suppose that a 60 gm tennis ball traveling at 18 m/s collides with a 300 gm ten-
nis racquet at rest, as shown in Figure 12.9. The racquet could be held in the
hand, or it could be balanced with its butt end on the floor, or it could be sus-
pended by a long length of string tied to the ceiling. If the ball hits the middle
of the racquet at its center of mass, and if there is no loss of energy during the
collision, the ball would bounce off the racquet at 12 m/s and the racquet
would recoil at 6 m/s. The momentum before the collision is 18 x 60 = 1080
and the momentum after the collision is 300 x 6 - 60 x 12 = 1800 - 720 = 1080.
The relative speed after the collision (12 + 6 = 18 m/s) is equal to the relative
speed before the collision.

BEFORE collision AFTER collision

300 gm racquet at rest

Figure 12.9

Now suppose the ball hits higher up, in the middle of the strings, as shown in
Figure 12.10. The racquet recoils as before, but it also rotates around an axis
near the end of the handle. As a result, the recoil speed of the racquet varies
from one end to the other. This represents the combined effect of a recoil at
5.14 m/s (the speed of the center of mass) plus rotation around the center of
mass.

These results were calculated assuming that linear momentum, angular momen-
tum, and kinetic energy are all conserved. The ball ends up with 18% of its
original energy and the racquet ends up with the rest. As far as the ball is con-
cerned, it collided not with a 300 gm mass, as in the first example above, but
with a 150 gm mass. The effective mass of the racquet is therefore 150 gm at
this point, but it is the full 300 gm at the middle of the racquet. The relative
142 Chapter 12

RUE Ie BEFORE collision AFTER collision

300 gm racquet at rest

speed at the impact point is 18 m/s before the collision and 10.3 + 7.7 = 18
m/s after the collision. The linear momentum of the racquet after the collision
is 150 x 10.3 = 1545 (which is also equal to 300 x 5.14). The total (linear)
momentum is 1545 - 60 x 7.7 = 1080 as before.

The effective mass near the middle of the strings is about half the actual rac-
quet mass for all racquets. About two inches from the tip, the effective mass is
about 1/3 of the actual mass.

Estimating the ACOR from the Effective Racquet Mass

We explained in Chapter 11 how the ball speed can be calculated by measur-

ing the COR (e) or by measuring the ACOR (e,), but we side-stepped the issue
of how the ACOR is related to the COR. This relation is easily derived. Consider
the serve situation described at the beginning of this Chapter where a racquet
of mass M and speed V, strikes a ball of mass m, and where the ball is initial-
ly at rest. In this case the ball is served at speed v, given by Eq. (12.4)

v, = (1 + ©) yam)

We have since learned that M is not the actual mass of the racquet. Rather, it
is the effective mass at the impact point, so M should be replaced by M,. We
also found, in Chapter 11, that v, can be calculated from a measurement of ee
and is given by the serve speed formula

V> = let eave

If we equate these two expressions for v, and replace M by M, we find that

e (eM, - m)
en

This relation shows how e, depends on the COR and the effective mass and it
agrees nicely with experimental data on different racquets. For an impact in the
center of the strings, e is typically about 0.88 and M, is about half the actual
 Collisions 143

racquet mass. If m = 57 gm, M = 300 gm and M, = 150 gm, then e, = 0.36. For
a heavier racquet with M = 400 gm and M, = 200 gm, e, = 0.46. Near the tip
of a racquet e drops to about 0.7 due to energy lost in frame vibrations and M,
is about 1/3 of the actual racquet mass. If e = 0.7 and hs 300/3 = 100 gm,
thence, = 0-08 Ie = 07" “and M_ = 400/53 = 133 em, then €, ="0.19.

Racq
As described in Chapter 6, the motion of a racquet can be regarded as con-
sisting of two parts, one being a translation of the whole racquet in a straight
line at the speed of its center of mass (CM), and the other being a rotation
about an axis through the CM at angular speed w. The combination of this
translation and rotation means that the actual axis of rotation is shifted away
from the CM. Just before a collision with a ball, the rotation axis is typically
about 10 to 20 cm beyond the butt end of the handle. The location of the actu-
al axis of rotation before the collision depends on the type of stroke and the
amount of wrist action. During a collision, the rotation axis shifts to a point
closer to the butt end or to some point under the hand, depending on the
impact point on the strings. There is nothing a player can do to change the
rotation axis during the collision, since the force of the ball on the strings is
much larger than the force of the player's hand on the handle.

There is a simple formula, which we will now derive, that can be used to cal-
culate the position of the axis of rotation during a collision. Suppose that, dur-
ing the collision, the actual rotation axis is somehere near the butt end of the
handle at a distance x from the center of mass. As shown in Figure 12.8, the
whole racquet translates forwards at speed V_,,, while any point in the handle
at a distance x from the center of mass rotates backwards at a linear speed xo
with respect to the center of mass. If that point is at the axis of rotation, the
net linear speed at that point is zero, and hence, xw = V_.,. From Equations
12.14 and 12.15, we find that

xmb(v, + v)/I = m(v, + v,)/M

Xe)
x = I/Mb

where I is the moment of inertia of the racquet about its CM. Note that if b =
0, then x = infinity, meaning that if the ball impacts the racquet at its CM, the
racquet does not rotate at all. For most racquets, I is about ML?/12 where L is
the length of the racquet, so x = L’/(12b). In that case, the axis of rotation will
be at the butt end of the handle, at x = L/2, if b = L/6. If L = 69 cm, then L/6
=11.5 cm, meaning that the center of percussion in this case is 11.5 cm from
the middle of the racquet or 23 cm from the tip.

The actual rotation axis passes through a point known as the conjugate point
with respect to the impact point, and the impact point is said to be the center
of percussion for that particular conjugate point. Every point in the handle has
a different center of percussion on the strings.

In deriving the formula for the effective mass of a racquet we have implicitly
assumed that the racquet rotates about an axis through the conjugate point. It
144 Chapter 12

does so during and immediately after the collision, but the axis of rotation
before the collision is usually somewhere else. As a consequence, the kinetic
eee 2
energy of a racquet before the collision is not necessarily equal to 5M.V and
its momentum is not necessarily equal to M.V, where V is the speed of the
impact point. These expressions are exact if the racquet rotates about the con-
jugate axis defined above, or if V = 0, but they are only approximately correct
if the racquet rotates about some other axis. Nevertheless, the collision
between a racquet and a ball is correctly described, and the speeds of the ball
and the impact point are correctly determined, if we assume that the impact
point on the racquet is an object of mass M., traveling at speed V, and having
kinetic energy : MV’, momentum M.V and angular momentum M,Vb. The
actual kinetic energy of the whole racquet before the collision is typically about
10 or 20% greater than iM.V for a serve or a groundstroke. The correct for-
mula for the kinetic energy of a racquet before the collision is given in Note 3
of Chapter 32.

Notes

1. The COR is normally measured by firing a ball at a racquet whose head is

firmly clamped. In that case, all the initial ball energy is converted to elastic
energy in the strings and the ball. If 1/4 of that energy is lost, the ball will
bounce with We i.We = 0.75 and hence COR = e = 0.87. The COR varies slight-
ly with string tension and it also varies with ball speed. For example, at ball
speeds around 10 m/s, the COR is typically about 0.9 on a racquet strung at
low tension and about 0.85 on a racquet strung at high tension. Measurements
at ball speeds around 40 m/s show that the COR decreases to about 0.80 on
racquets strung at either low or high tension. The decrease in the COR is due
mainly to the fact that the ball loses a greater fraction of its energy in a high
speed collision.

2. A similar calculation can be done for a serve where the racquet is swung
towards a ball at rest. In that case the resulting ball speed is the same as that
for a head-on collision between the ball and a mass M, incident at speed V,
where M, is given by Eq. (12.17) and V is the speed of the racquet at the
impact point.

The effective mass M., of the racquet at an impact point lying along the longi-
tudinal axis can be determined from the following relations (see Figure 12.11):

dv
Bee Ma Mee and F = M —
cm

dt dt

dv.
V=V._ + bo sO Se ee
 Collisions 145

F = force of ball on racquet Figure 12.11

V = speed of impact point

CM = center of mass
cm |Le Vom Vom = speed of CM
= M = actual racquet mass
w = angular speed of racquet
I = moment of inertia about CM

w
Fb =I— (where Fb is the torque about the center of mass.)
dt

F F Eb* i b?
Hence, Se SO, =a be
M MI M. M I

Or

+ M
2 (12.22)
© 14+ Mb7/I

The effective racquet mass, M., of the racquet at any impact point can be deter-
mined from the following relations (see Figure 12.12):

Figure 12.12

P = impact point

y = long axis

x = axis through CM

R = distance from P to y axis

Xx b = distance from P to x axis

M = actual racquet mass

I = moment of inertia about x axis

ly = moment of inertia about y axis

146 Chapter 12

dv cm.
Fs MAN ea PSM
dt dt

dv dV.cm + b dw +R dw y
V=V__ + ba + Rw sO ar i — —
ss i dt dt dt dt

dw
Fb = 1— (where Fb is the torque about the center of mass.)
dt

dw
2 aes 6 — (where FR is the torque about the center of mass.)

eae (12.23)

When R = 0, this formula is the same as Equation 12.22. If the impact point is
closer to the edge of the frame, then R is not zero, and hence, R increases, sO
1/M, increases, so M, decreases. As a result, the ball will not bounce as well,
and the speed of the ball will be reduced, regardless of whether the shot is a
serve or a groundstroke.
 Racquets,
Strings, and
Balls as Springs

Arantxa Sanchez-Vicario, Spain

By Rod Cross

ennis balls and tennis strings can be regarded as springs, for the simple reason that they ARE springs.
They are not the usual type of metal spring wound into a coil, but they function in essentially the same
™ way. In fact, all objects behave as springs, but not all objects are designed to compress or stretch as their
main function. Strings are designed to stretch, balls are designed to compress and racquet frames are designed
to be stiff so that they will not bend too far. In that respect, racquet frames are designed as stiff springs

PA SCUTTLE SRLS
Springs come in all sorts of shapes and sizes, but the most common variety is a metal coil. If the turns are wound
tightly together, then they are called extension or tension springs, since they cannot be compressed further. If
the coils are spaced so that there is a gap between each turn, then they are called compression springs. A com-
pression spring is designed to be compressed. A compression spring can also be stretched in the same way as
a tension spring, but it can’t be stretched as far since the turns are already separated a certain distance, and since
a spring will not recover properly if it is stretched too far.

Suppose that a spring of length L stretches by an amount x when subject to a pulling force F, as shown in Figure
13.1. The total length of the spring is then L + x. We define the stiffness (also called the spring constant) of the
spring by the ratio:
148 Chapter 13

Figure 13.1
k = Fix (13.1)

If x is large, then k is small, and we

say that the spring is soft. If x is
small, then k is large, and we say that
Normal
the spring is stiff. For example, sup-
length
pose that a force of 10 Ib causes the
spring to stretch by 2 inches. Then k
= 10/2 = 5 lb/in. If the spring stretch-
es by only 0.2 inches, then k = 10/0.2
= 50 lb/in. The same formula applies
if we compress a spring, but x then
Compressed represents the decrease in length of
F the spring.

Coiled metal springs usually have the

property that their stiffness is the
x same regardless of how far they
Es
stretch or compress. A typical force
| vs compression curve for a_ coil
spring is shown in Figure 13.2. This
E FP spring compresses by 0.1 inch when
SS ———— — te SSS
the force is 10 Ib, so k = 100 lb/in. It
compresses by 0.2 inches when the
force is 20 lb, so k = 20/0.2 = 100
lb/in as before. The graph of F vs x
is a straight line passing through the origin, and k has the same value
at any point on this line. This is typical of the behaviour of a metal
spring, but tennis balls and tennis strings are not as obliging. Balls get
Figure 13.2
stiffer as the force increases, provided the force increases slowly with
time. If the force on a ball increases rapidly (as it does when the ball
Force vs Compression
bounces), then the ball starts out being very stiff and becomes softer
F (Ib) when the wall starts to bend.

Tennis strings also get stiffer as the force increases. There is no

20 increase in stiffness for a rapid stretch, but there is decrease in stiffness
for a slow stretch. A nylon tennis string will stretch by about one inch
when a pulling force of 60 Ib is first applied, but it keeps stretching
further over time, even when the force is held constant. A string that
X (inch) stretches by say 20 mm initially will stretch about 2 mm further over
the next few minutes, and then more slowly after that. As a result, the
stiffness decreases by about 10% in the first few minutes and by anoth-
er 10% over the next year or so.

If two identical springs are joined end to end (in series), then the stiffness of
each spring is unaffected, but the combination is only half as stiff as each sep-
arate spring. The reason is that a pulling force F at each end will stretch each
spring by an amount x, so the combination stretches by 2x. For example, if a
12 inch long tennis string stretches by one inch, then a 24 inch length of the
same string will stretch by 2 inches when subjected to the same force, so it will
be half as stiff as the 12 inch length of string. Even though it is the same string,
stiffness depends on the length of the string.
 Tennis Springs 149

The stiffness of a length of string also depends on its diameter. A thick piece
of string will not stretch as far as a thin piece of string made from the same
material. Suppose that you twist two identical lengths of the same string to
make a single fat string. The strings are then in parallel. If a force F stretches
each individual string by a distance x, then you will still need a force F on each
individual string to stretch it by a distance x when the two strings are twisted
together. However, the total force at each end of the twisted pair is then 2F, so
the combination is twice as stiff as each individual string.

A complication with tennis strings is that the stiffness of a string increases as it

stretches further. If a pulling force of 30 Ib stretches a string by 1/2 inch, then
a force of 60 lb will stretch it not by 1 inch, but by about 3/4 inch. When ten-
sioned to 60 lb, a thin string will always stretch further than a fat string made
from the same material. But the extra stretch when hitting a ball is about the
same for both fat and thin strings since thin strings start to get stiffer at a small-
er tension than fat strings. Thin and thick strings made from the same materi-
al start to get stiffer at the same stress = tension/area value. Chapter 35, String
Gauges, explains this in more detail.

An important property of a spring is that it can store energy. Energy stored in

this manner is not kinetic energy due to motion but is classified as potential
energy. A good example of this is the potential energy stored in the string and
the arm of a bow. When an archer shoots an arrow, the string and arrow are
drawn back by pulling on them. The work done by the archer is then stored
as elastic potential energy in the bow. When the string is released, all (or most)
of that energy is used to shoot the arrow. The stored elastic potential energy is
converted to kinetic energy of the arrow.

The archer could simply throw the arrow by hand, but the arrow goes much
further when it is released from the bow since more energy can be stored in
the bow than can be given to an arrow just by throwing it. That’s because the
force applied by a person when throwing a light arrow is small compared to
the force needed to stretch the string in a bow. As a result, more work is done
by the archer in stretching the bow and more energy is thereby stored in the
bow. For a similar reason, one can serve a tennis ball with a racquet faster than
one can throw the ball, although in that case the serve speed is only about
twice the thrown speed. The racquet is swung at the ball using an action sim-
ilar to throwing the racquet, so the situation is not the same as the bow and
arrow case. However, the force applied when swinging a racquet is larger than
the force applied when throwing a light ball (since the racquet is heavier) with
the result that more energy is given to the racquet.

Stored ina Mass |

Potential energy is stored in any object when it is lifted vertically since work is
required to lift it. When you let go, the potential energy is converted to kinet-
ic energy as the object falls to the ground. The work W done by a force F in
moving an object through a distance d (in the direction of F) is given by

Work = force x distance or W = Fd. (13.2)

150 Chapter 13

When F is measured in Ib and d in ft, then the units of work or energy are foot-
pounds. When F is measured in Newton and d in metre, then the units of work
or energy are Joules. One foot-pound = 1.356 Joule.

The minimum force needed to lift a mass m is mg, so if it is lifted off the
ground through a height h, then the work done is mgh. If it is lifted by a force
mg, it does not accelerate and there is no increase in kinetic energy. The work
done is equal to the increase in potential energy. Hence, the increase in poten-
tial energy is mgh. A heavy object stores more potential energy than a light
object, assuming both are lifted to the same height, since more work is
required to lift the heavy object.

If the object is then dropped and falls to the ground, it gains an amount of
kinetic energy mwv’/2 equal to mgh. The object therefore hits the ground at a
speed given by v’ = 2gh. This speed does not depend on the mass of the
object. In theory, heavy and light objects fall to the ground at exactly the same
speed. In practice, very light objects (e.g., a feather or a bit of paper) fall slow-
er since air resistance has a relatively large effect on light objects. The force on
an object due to air resistance depends on the area and the speed of the object,
but not its mass. The force of gravity on an object depends on the mass of the
object, but not its area or its speed (see Note 1).

An object thrown or hit vertically upwards at speed v has kinetic energy mv’/2.
It will rise to a height h before stopping, at which point all the kinetic energy
is converted to potential energy. Since mv’/2 = mgh, the object rises to a height
h= v’/2g. The height does not depend on the mass of the object. An object
thrown twice as fast does not travel twice as high. It travels four times higher.

The work required to stretch or compress a spring is given by (force) times

(distance), but in this case the force required to stretch or compress the spring
keeps increasing the more it is stretched or compressed. For a spring, the force
F is given by F = kx. Since the distance compressed or stretched is x, one might
expect that the energy stored in a spring is Fx = kx’, but this is the wrong
answer. The problem is that in order to compress a spring through a distance
x, the force needs to start at zero when x = 0 and then it needs to keep increas-
ing until F = kx when the compression is equal to x. The average force need-
ed to compress a spring by an amount x is kx/2 and the correct answer for the
energy stored in a spring is

Energy stored = Average force x distance = kx2/2. (13.3)

This is the correct answer when the stiffness remains constant. If k varies as
the spring is stretched or compressed, then the answer is roughly correct if k
is taken as the average stiffness (see Note 2).

The Tennis Ball

As aSpr
An approved tennis ball must squash by about 1/4 of an inch (6.5 mm) when
subject to a load of 18 lb (8.2 kg or 80.4 N). It therefore has a stiffness k = F/x
= 18/0.25 = 72 lb/in. or k = 80.4/0.0065 = 12,400 N/m = 12.4 kN/m. At least,
 Tennis Springs Von

this is the stiffness for a slow compresion. As described in Chapter 37, k is actu-
ally larger than this if the ball is compressed rapidly.

The stiffness of a ball affects the way it interacts with the court and with the
strings of a racquet. Consider what happens when a 57 gm tennis ball is
dropped vertically from a height of 100 inches (2.54 m) onto a hard surface. It
hits the surface at speed v = V(2gh) = 7.06 m/s and with kinetic energy 1.42
Joules. When the ball squashes, we can assume that all the kinetic energy is
converted to potential energy so the ball squashes by an amount x given by

kx7/2=1.422 so x=15mm and = F=kx=188N

assuming that k = 12,400 N/m. The force of gravity on the ball is mg = 0.506 N,
which is tiny compared with the force between the ball and the surface com-
pressing the ball. The average force on the ball during its compression is 94 N,
and the average acceleration upwards (or deceleration downwards) is given by
a = F/m = 94/0.057 = 1650 m/s’. The ball comes to a halt very rapidly. Since
v =u - at, where v = final speed = 0 and u = initial speed = 7.06 m/s, the time
taken to come to a stop is t = u/a = 7.06/1650 = 0.004 seconds. The stored
potential energy is then converted back to kinetic energy as the ball expands,
and it takes another 0.004 seconds to accelerate back to 7.06 m/s if none of
the stored elastic energy is lost. This calculation indicates that the ball should
spend about 0.008 seconds in contact with the surface. In fact, the time is typ-
ically about 0.005 seconds since (a) the ball does lose some of its stored elas-
tic energy, and (b) the ball is stiffer during a rapid compression than during a
slow compression. The point of the above calculation was not to get an accu-
rate value of the collision time (see Note 3), but to explain how and why it is
determined by the mass and stiffness of the ball.

A surprising feature of the ball bounce is the fact that the force on the ball is
huge when it hits the ground. One could ask, “Where did that force come
from?” It is the same sort of force that arises when you hit a hammer with a
nail. It is very difficult, if not impossible, to push a nail into a block of wood
with a hammer, but it is easy to hit the nail into the wood since the force on
the nail is magnified when you hit it with a hammer. The force you exert on
the handle to swing it is typically 500 times smaller, so where does this extra
force come from? The simple answer is that it comes from the stiffness of the
ground or the stiffness of the nail. When you swing a hammer or drop a ball,
it acquires a certain amount of kinetic energy given by the force on the object
times the distance moved. The hammer or the ball is subject to a relatively
small force acting over a large distance before the impact. During the impact,
the hammer or the ball is brought to rest by a large force acting over a small
distance. The work required to stop the hammer or the ball is the same as the
work needed to accelerate it from rest, so force times distance before the
impact is the same as force x distance during the impact. The nail or the
ground cannot compress very far since both are very stiff, but the hammer or
the ball has enough energy to compress them a small amount which acts to
generate very large forces during the impact.
y

Strings

When a string is installed in a racquet, it stretches a certain distance and is then

tied to the frame so that it remains at that stretched length for the rest of its
life. The force used to stretch the string is typically about 60 lb (27.2 kg or 267
152 Chapter 13

N). Force and tension mean essentially the same thing here. We say that the
string tension is 60 lb when the force applied at each end is 60 lb. The tension
stays in the string when it is tied to the frame and after the stringing machine
is disconnected and when you take the racquet onto the court. The string
behaves as a permanently stretched spring. If you cut or break a string in your
racquet, the string rapidly snaps back to its original length (almost).

While the string remains stretched, it stores a certain amount of potential ener-
gy. However, none of that energy is used by a player when hitting a ball. The
work required to stretch all the strings when installing them in a racquet is
about 100 Joules. This is the potential energy stored in the strings. The kinetic
energy of a ball hit at 30 m/s is 25 Joules. In theory, there is enough energy in
the strings to launch four balls each at 30 m/s. If the original energy stored in
the strings could somehow be released, then the tension in the strings would
drop every time a ball was hit, and the tension would drop to zero after about
four hits. However, it is only when a string breaks that the tension drops to
zero and the stored potential energy is then suddenly released.

The strings are tensioned in order to give them a certain amount of stiffness in
a direction perpendicular to the string plane. A ball impacting on the strings
pushes the strings and stretches them a bit further than when they were first
installed. This adds to the stored elastic energy, and this extra energy is then
immediately given back to the ball as it bounces off the strings. This can be
repeated thousands of times, at least until the strings break. Players sometimes
complain that strings lose their resiliance when they get old and they can’t get
any power out of the strings. This is not consistent with actual measurements.
The string tension drops over time, but there is no loss in ball speed. If any-
thing, the ball speed should increase very slightly when the string tension
drops. What players may notice is that the ball stays on the strings a bit longer
at low tension, which might give the impression that the ball comes off the
strings more slowly.

The stiffness of the whole string plane, in a direction perpendicular to the

string plane, is about the same as the stiffness of a tennis ball. As a result, a
certain amount of elastic energy is stored in the ball when it squashes against
the strings, and about the same amount of energy is stored in the strings. The
strings always return about 95% of their stored energy, but the ball returns only
about 55% of its stored energy.

Is Shared
The fraction of the total elastic energy stored in
Ball —& = Strings the ball and the strings during a collision
depends on their relative stiffness. The ball and
the strings each behave like two springs pushing
against each other, as shown in Figure 13.3.
Despite what your intuition might tell you, the
Kp Ks force of the ball on the strings is exactly equal to
the force of the strings on the ball, regardless of
how far the ball squashes and regardless of how
far the strings stretch (see Note 4).

F on ball F on strings

Figure 13.3
 Tennis Springs 153

Suppose that the ball stiffness is k,, the string plane stiffness is k,, the ball com-
presses by an amount x, and the string plane is deflected by a distance x,.
Then the force F on the ball is F = k,x, and the force on the strings is F = k.x,.
Since the forces are equal, k,x,, = k.x.. This tells us that if the ball is stiffer than
the strings, then it will compress by a relatively small amount and the strings
will deflect by a larger amount. For example, if the ball is twice as stiff as the
strings and if the string plane deflects by 20 mm, then the ball will compress
by 10 mm. The energy stored in the ball is kee, and the energy stored in
: 2 ae ; ‘ 2 2
the strings is k x.“/2. Since the strings stretch twice as far, 2a 4 2 but k=
k,/2, so the strings will store twice as much energy as the ball. This seems like
it might make a big difference to the speed of the ball off the strings, but it
doesn’t. The reason is given below.

Suppose that a 57 gm tennis ball is incident vertically downwards on a hard,

heavy surface (such as the court) at a speed of 10 m/s. The kinetic energy of
the ball is then mv*/2 = 2.85 Joules. It loses about 45% of its energy during the
bounce (0.45 x 2.85 = 1.28 J) and the kinetic energy after it bounces is there-
fore 2.85 - 1.28 = 1.57 J. The bounce speed is given by mv-/2 = 157 JaSONe =
7.42 m/s. The coefficient of restitution in this case (COR = bounce speed/inci-
dent speed) is 7.42/10 = 0.742, as required by the rules of tennis. If the ball
bounced slower (and hence lower) or faster (and hence higher) than this, play-
ers would complain that there was something wrong with the ball.

Now suppose the ball is incident vertically downwards at 10 m/s on the strings
of a racquet and that the head of the racquet is clamped so that it can’t move.
If the strings and the ball are equally stiff, then the kinetic energy of the inci-
dent ball (2.85 J) is shared equally between the strings and the ball, so they
each store 1.425J at the instant the ball comes to rest. At that instant, all of the
initial kinetic energy of the ball is converted to potential energy in the ball and
the strings. The strings lose 5% of their stored energy (0.05 x 1.425 = 0.071 J),
the ball loses 45% of its stored energy (0.45 x 1.425 = 0.041 J) so the total ener-
gy loss is 0.071 + 0.641 = 0.712J. The ball will bounce with kinetic energy (2.85
- 0.712) = 2.138 J and at speed 8.66 m/s. The coefficient of restitution in this
case is 8.66/10 = 0.866. The ball bounces 16.7% faster and 30% higher off the
strings than off a hard surface (Table 13.1).
Table 13.1

Velocity Energy E to E to E Loss E Loss E Loss Vv

Conditions In(m/s) In(J) Strings (J) Ball (J) — Strings (J) Ball (J) Total (J) Out (m/s) COR
Concrete 10 2.85 0) 2.85 0 1.28 1.28 7.42 0.742
Ke=ik,. 10 2.85 1.425 1.425 0.0715 0.641 0.712 8.66 0.866
2k, =k, 10 2.85 1.9 0.95 0.095 0.428 0.523 9.04 0.904

If the strings are twice as soft as the ball, then the strings will store twice as
much energy as the ball. The ball stores 2.85/3 = 0.95 J and the strings store 2
x 0.95 =1.90J. The total energy stored is then 0.95 + 1.90 = 2.85 J. The strings
lose 5% of 1.90 J = 0.095 J, the ball loses 45% of 0.95 J = 0.428 J and the total
154 Chapter 13

energy loss is 0.523 J. The ball will bounce with kinetic energy 2.328J and at
speed 9.04 m/s. In this case, the coefficient of restitution is 9.04/10 = 0.904.
This is slightly higher than the previous case, but it is not a lot higher since the
ball already bounces quite well off the strings, even when they are as stiff as
the ball. This is an important point and it is often overlooked.

The point is that there is not much one can do to increase the bounce speed
off the strings, just by decreasing the string tension or using a different type of
string. To emphasise the point, suppose that the string plane is one million
times softer than concrete. Decreasing the softness of a concrete slab by a fac-
tor of one million would increase the ball speed by only 16.7%, from 7.42 to
8.66 m/s. If the softness was decreased by an additional factor of one million
so that there was no energy loss in the ball at all, then the ball speed would
increase by another 12.6%, from 8.66 to 9.75 m/s (some energy being lost in
the strings).

A factor of two decrease in string plane stiffness results in only a 4.4% increase
in the ball speed off the strings (from 8.66 m/s to 9.04 m/s), and this is when
the racquet head is clamped. The string plane stiffness is roughly proportional
to string tension, but a factor of two change in string tension is unrealistic.
Players might be prepared to change the string tension from 60 Ib to 50 lb, but
no one strings a tennis racquet at 30 Ib. If the string tension is decreased from
60 to 50 lb, the increase in ball speed would be only about 1.5%. Furthermore,
if the racquet head is not clamped, then the increase in ball speed off the
strings is even smaller, as explained in detail in Chapter 12 on collisions. The
basic reasons are that (a) only some of the energy of the incident ball (or the
incident racquet) goes into stretching the strings or compressing the ball while
the rest is used in changing the kinetic energy of the racquet, and (b) the speed
of the ball off the strings depends primarily on the speed of the racquet rather
than on the amount of energy lost in the ball and the strings.

The stiffness of a string, in a direction perpendicular to the string, is usually a

lot smaller than the stiffness for a lengthwise stretch. The transverse stiffness
can be calculated with the aid of Figure 13.4. Suppose that the string has a
length L, and is at tension T,. A transverse force F applied in the middle of the
string pushes the string sideways by a distance y. As a result, the string stretch-
es slightly, by an amount that can be calculated from the geometry.

If the string increases in length from L, to L3--then the

Figure 13.4 right-angle triangle on each side has sides L,/2, y and
hypotenuse L,/2. Pythagoras’ theorem tells us that

(Ly 2) = ye +a
If we multiply both sides by 4 then

L, = 1, + 4y =1 G + 4y7/ 19)
Now take the square root of both sides. Provided that y
is much smaller than L,, the answer is given to a good
 Tennis Springs 155

approximation by

Dnt 2
Tees Cb 2" 2, Qe oh 2ye7 Le

where we have used the result that the square root of (1 + N) is approximate-
ly ( + 0.5N) when N is much smaller than 1. For example, the square root of
1.04 is actually 1.0198 which is very close to 1.02.

The increase in string length is x = L, - L, = 2y*/ L, and the increase in string

tension is kx where k is the stiffness for a longitudinal stretch.

The tension increases from T, to T, where T=, Akxs sO

Tole + ky

The force F acting to push the string sideways is opposed by the effect of the
string tension pulling back in the opposite direction. The tension does not act
in the same direction as F, but it has a force component Fain the y direction
given by

smn) =i T= y7/ (5/2) so Fy = 2yT,/ Ly.

Each half of the string exerts a force F, opposing F, so the string moves side-
ways until F = 2F,,, and then

Beye (le k,, y

where k, = 4T,/L, is the perpendicular stiffness of the string. Given that L, is

essentially the same as L, when y is small, ist is given to a good approxima-
tion by

k, = G/L), + 2ky*/L,) = 47, / 1,

This equation shows that the string has a perpendicular stiffness even if y = 0,
given by k 7 4T,/L,. When y is very small, the stiffness depends only on the
string tension T, and the length L, of the string. The transverse stiffness of a
steel string would then be the same as that of a nylon string if both strings have
the same length and are at the same tension. However, as y increases, Kk
increases since the string tension increases. If y is larger than a few mm, the
transverse stiffness of a steel string would become larger than that of a nylon
string.

Consider a nylon string of length 30 cm = 0.3 m tensioned to 28 kg (274 N).

Such a string is typical of a main string in a tennis racquet. A 30 cm nylon string
has a longitudinal stiffness k of about 40,000 N/m. When y = 0, k, = 4x 274/0.3
= 3,653 N/m. When y = 1 mm, the tension increases by only 0.27N so Kk, is
essentially the same. If y = 10 mm then the tension increases by 27 N and ie
increases slightly, to 4,013 N/m. If y = 30 mm then the tension increases by 243
N (24.8 kg) and is increases to 6,893 N/m, about double the y = 0 value.
156 Chapter 13

Now consider a 30 cm kevlar string with k = 120,000 N/m, also tensioned to

28 kg. For small values of y, k,, is the same as that for a nylon string. If y = 10
mm, then the tension increases by 81 N and k,, = 4733 N/m. If y = 30 mm,
then the tension increases by 729 N and k, increases to 13,373 N/m. A high
speed ball impacting on a nylon string in a fully strung racquet will push the
strings sideways by about 30 mm. However, it will not push a kevlar string that
far since kevlar is a stiffer string. Just what happens can be calculated as
described in the following section.

Impact of a Ball on a Stretched String

OO TN EFT~ EF“

Suppose that a ball strikes a single string at right angles and in the middle of
the string, as shown in Figure 13.5. If there is no energy loss in the ball or in
the string, then the ball will bounce off the.string at the same speed that it hit
the string. This is not a good assumption for a tennis ball, but it is a good
approximation for a metal ball. Impacting with a metal ball is a good way to
test the properties of the string, since the change in ball speed off the strings
provides a measure of the energy loss in the string. Experiments show that a
metal ball or hammer bounces off a string at about 95% of the incident speed,
regardless of the type of string and regardless of the string tension and regard-
less of whether the string is new or old. The only exception is the very first
impact on a brand new string, in which case the ball bounces off the string at
about 92% of the incident speed. In that case, a few weak bonds in the string
get broken due to the large increase in tension during the impact. Once bro-
ken, they remain broken, and from then on the energy loss in the string
remains small.

Figure 13.5

String [OL al String

String

One might argue that the impact of a ball on a single string is not the same as
impacting a ball on the whole string plane in a racquet. However, the effect is
exactly the same. A 0.8 kg (1.8 Ib) steel boule (as used in the French or Italian
version of bowls) dropped from a height of 100 inches onto the strings of a
head-clamped racquet bounces to a height of 90 inches and therefore rebounds
off the strings at 95% of its incident speed. The bounce height is the same
regardless of whether the strings are old or new and regardless of whether the
strings are at high or low tension. Friction between the strings results in no sig-
nificant or additional energy loss.
 Tennis Springs 157

An interesting question is whether string tension or the type of string makes a

significant difference to the behaviour of a string when a metal ball bounces
off the string. At high tension, the string will be relatively stiff and the ball
should bounce off the string sooner (rather than faster, as in higher speed). At
low tension the string will stretch further. If it stretches further, the increase in
tension during the collision will be larger. The question of interest is whether
the large increase in tension in a string at low tension might be equivalent or
almost equivalent in effect to an impact on a string that is initially at high ten-
sion.

A ball impacting in the middle of a string experiences a backwards force

f= Ay as described above. The equation of motion is therefore

m d’y / dt? = -(4y /L,)(T, + 2ky?/L,) (13.4)

where m is the mass of the ball. This equation can be solved analytically only
when y is very small. This is not a very interesting case since all strings behave
the same when y is very small. The more general case, when y is large, can be
solved numerically. Typical results are described in Chapter 30.

Another interesting question is whether an impact towards one end of the

string differs from an impact in the middle of the string. Right at the very end
of a string, where it joins the frame, the string is infinitely stiff since the string
will not deflect at all if the frame is rigidly supported. Consequently, the stiff-
ness varies from a minimum at the middle of the string to a maximum at the
end of the string.

There is another interesting effect that can occur when a ball impacts away
from the center of the string. The tension in the string on each side of the ball
exerts a net force on the ball towards the middle of the string. As a result, the
ball will bounce back towards the center of the string, as shown in Figure 13.6.
It is difficult to do this experiment with only a single string, but it is easy to see
the effect by dropping a heavy steel ball on the strings of a head-clamped rac-
quet, or by dropping a steel ball bearing on the membrane of a drum. If one
drops a small steel ball or a golf ball on the strings of a racquet, the ball
bounces at strange angles since the strings are lumpy and uneven. The effects
of the lumps are evened out on a larger ball and when the strings deflect a
large distance. This effect means that when a player mis-hits a ball near the
edge of the frame, the ball will still bounce off the strings towards the intend-
ed direction.

Figure 13.6

Motion of ball
Force on ball

String
tng String
158 Chapter 13

Notes

Note 1. Since the force of gravity on an object increases with its mass, many
people assume that a heavy object will fall faster since the force on the object
is bigger. But force = mass x acceleration, so a heavy object needs a bigger
force on it to accelerate at the same rate as a light object. The acceleration due
to gravity, g, is the same for all objects, regardless of its mass.

Note 2. The amount of energy stored in the spring can be determined exactly
from the area under the F vs x curve since the work is given by the integral of
Fidx.

Note 3. When a ball bounces and compresses by an amount x, the force on

the ball is F = kx = -ma where a is the acceleration of its center of mass. If we
assume that a = d°x/dt’, then the ball undergoes simple harmonic motion and
is in contact with the ground for one half cycle. An accurate calculation of the
impact time is not possible since (a) the center of mass moves a smaller dis-
tance than x by a fraction that changes with time, and (b) k does not remain
constant during the bounce.

Note 4. For every action there is an equal and opposite reaction, but this idea
is sometimes hard to accept. For example, if you are standing at the net and
your partner serves a ball into the back of your head, the ball pushes on your
head with the same force as your head pushes on the ball. Your head didn’t
deliberately push on the ball, but it is obvious from the fact that the ball
bounces backwards that something pushed the ball backwards. That something
was your head. Similarly, if you drop a ball, it falls to the ground because the
earth exerts a force downwards on the ball. The ball exerts an equal and oppo-
site force on the earth, pulling it upwards, but the earth is a lot heavier than
the ball and its motion is undetectable.
 The Factors
%
Influencing Power

.““
7

Mark Philippoussis, Australia

By Rod Cross

acquet power is a quantity that is hard to measure since it varies with the impact point on the strings
and with the type of shot. Heavy racquets are more powerful than light racquets if they are swung at
~™® the same speed. If light racquets are swung faster than heavier racquets, then the power is about the
same. Most of the power in a shot comes from the player's arm, with the result that all racquets have about the
same maximum power, regardless of racquet mass, string tension, or string type. Racquet stiffness has no effect
on power when the ball is hit in the middle of the strings, but stiff racquets are more powerful when the ball
is hit nearer the tip of the racquet.

Racquet and string manufacturers like to describe their products using terms such as power, control, feel, and
playability. These quantities are commonly rated on a scale from 0 to 10 and are usually described as being
good, excellent, exceptional, or superb. Subjectivity and marketing aside, there is an assumption by the manu-
facturers and by many players that power, control, feel, and playability are quantities that vary quite significantly
from one racquet, or one string, to another, and that they are quantities that can be measured. However, all these
160 Chapter 14

terms and adjectives are meaningless from a purely scientific point of view
since none of them have ever been measured. Racquet power could, in prin-
ciple, be measured, but power is a quantity that is hard to pin down in terms
of a clear and simple definition. In this respect, power is a bit like happiness.
Everyone knows what it is, but no one knows how to measure it.

Over the last few years, I have measured the properties of about 300 different
models of strings. Every string differs from the next, sometimes by a large fac-
tor in terms of the amount they stretch, but there is one factor that is essen-
tially the same for every string. Every string stores elastic energy very efficiently
when it is stretched rapidly, and it returns between 95% and 97% of that ener-
gy when it returns to its original length. As a result, the difference in power
between any two strings is almost undetectable. Any difference in power
between different strings must therefore be related to the effect of the strings
on energy loss in the ball. At low string tension the impact is softer and the
ball is less deformed, so there will be a smaller energy loss in the ball. Different
strings stretch different amounts on impact with a ball, with the result that the
increase in tension during the impact varies from one string to the next. For
example, the increase in tension is small with a natural gut string, larger with
a nylon string, and largest with a kevlar string. This affects the softness of the
impact and it affects the energy loss in the ball (as well as energy lost in frame
vibrations), but the energy loss in the string itself is negligible, regardless of the
type of string.

As to racquets, research is equally clear that, (a) most of the power comes from
the player's arm rather than the racquet, and (b) the difference in power
between any two racquets is likely to be less than 5%. My own tests, and tests
by others (see Tennis Magazine, August 1997), have shown that players can
serve a ball almost as fast with an old wood racquet as with a modern graphite
racquet. The advantage of modern racquets is not that they are a lot more pow-
erful, but they allow players to use or control the power more effectively.
When combined with the increased strength and fitness of modern players, the
result is that the modern game of tennis is played at a much faster pace than
in the old wood racquet era.

1. Racquet power refers to the speed of the ball off the strings. But the speed
of the ball depends on many factors, not just the power of the racquet. For
example, the speed of the ball depends on how hard a player chooses to hit
the ball and on how hard the player is able to hit the ball. A sensible defini-
tion of racquet power therefore requires some stardard way of hitting the ball,
keeping either the racquet speed or the racquet energy or some other param-
eter constant. Just what this parameter should be has not yet been determined.

2. If one were to measure racquet power by swinging all racquets at the same
speed, then heavy racquets would all show up as being more powerful than
light racquets. And the same could be said if balls were fired at a stationary
racquet. In each of these specific cases, more mass means more power, but
there are other circumstances (such as an unwieldy 5 Ib racquet) where more
mass could mean less power.

3. For a serve, the speed of the ball is directly proportional to the speed of
the racquet. Power therefore depends on both racquet mass and racquet speed.
 Racquet Power 161

For a groundstroke, the speed of the ball off the strings also depends on the
speed of the incoming ball. This is another parameter that would have to be
kept constant in order to compare the power of different racquets.

4. Under normal playing conditions, the speed of a swung racquet depends pri-
marily on how fast the player chooses to swing it. If the player chooses to
swing it as fast as possible (at the risk of miss-hitting the ball) then light rac-
quets can be swung faster than heavy racquets. Just how much faster is some-
thing that has never been measured precisely, and is something that will
depend on the type of stroke and the individual player. For example, players
who use a long backswing on groundstrokes can accelerate the racquet to a
higher speed than players who use a short backswing. Players with a long
backswing are likely to get more power with a heavy racquet and players with
a short backswing are likely to get more power with a light racquet. However,
this is also something that has never been measured. Some players seem to be
able to generate very high racquet speeds using a short backswing and a rapid
punching action when they strike the ball.

5. Despite the fact that light racquets can be swung faster than heavy racquets,
players don't necessarily do so. Swinging a racquet as fast as possible is the
best way to miss-time the shot or miss-hit the ball. A player swinging a racquet
at a controlled speed will get more power from a heavy racquet if the racquet
speed is kept constant. In general, players tend to swing a racquet at a con-
trolled speed for groundstrokes, volleys, and second serves, but they swing
faster for a first serve or an overhead smash.

6. If a ball is dropped on a stationary, hand-held racquet, the ball bounces best

at the throat and worst at the tip. If one swings a racquet, the racquet travels
fastest at the tip and slowest at the throat. These two factors (racquet speed
and bounce efficiency) must be factored together to determine the ball speed
off the strings, and the answer will vary depending on the relative speed of the
ball and the racquet at the impact point. For example, the point of maximum
power is near the throat of a racquet when the ball is traveling a lot faster than
the racquet prior to the impact, since that is where the ball bounces best. At
the opposite extreme where the racquet is travelling a lot faster than the ball,
the maximum power point is closer to the tip of the racquet, since that is where
the racquet is traveling fastest.

7. The point of maximum power depends on how the racquet head speed
varies from the tip to the throat, and that depends on the type of shot and the
amount of wrist action. In a wristy shot, the tip might travel twice as fast as the
throat. In a volley, the tip and throat might travel at the same speed.

Racquet mass, speed, and bounce are the key elements in power. Despite the
complications outlined above, some definite statements can be made about
racquet power by combining existing experimental data with the laws of
physics. The essential physics is that, (a) when a player swings a racquet, a cer-
tain amount of energy is given to the racquet, and (b) the amount of energy
transferred from the racquet to the ball depends mainly on the energy loss in
the ball and the energy lost to vibrations of the racquet frame. A surprising
result is that there is very little one can do to change a racquet so that it works
more efficiently in transferring energy out of the arm and into the ball. If the
ball is hit in the middle of the strings, then there is no energy loss due to vibra-
tions. One could get more power by eliminating most of the energy loss in the
ball, but that is not allowed by the rules of tennis.
162 Chapter 14

OT
NT CF sCTFN

The essential physics of racquet power is described in Chapters 10 to 13. We

can summarise the main results as follows:

Suppose that a ball of mass m is served from a point on the strings where the
effective mass of the racquet is M,. The effective mass is about half the actual
racquet mass in the middle of the strings, and it drops to about 1/3 of the actu-
al racquet mass at a point about 5 cm from the tip. The ball can't be served
any closer to the tip since it would impact on the frame of the racquet. If the
impact point is headed towards the ball at speed V and if the ball is served
without any significant spin, then the ball will be served at a speed v given by

v =(1+e,)V (14.1)

where e, is a number less than 1.0 called the apparent coefficient of restitution
(ACOR). It is a measure of how well the ball bounces off the strings, and it
depends on the effective mass of the racquet at the point where it bounces.
The bigger the effective mass, the better it bounces. The formula for serve
speed shows that serve speed is equal to the racquet speed V plus a small
amount that depends on how well the ball bounces off the strings.

The ACOR decreases from the throat area to to the tip area and is typically
about 0.5 in the throat, 0.4 in the middle of the strings and about 0.1 near the
tip. The ACOR is measured by firing a ball at right angles to the strings of a
stationary racquet, and the values obtained using this method agree closely
with the expected result that

(e- m/M,)
pS ee (14.2)
(1 + m/M,)

where e is the coefficient of restitution (COR). The COR is typically about 0.88
in the middle of the strings, but it decreases to about 0.70 near the tip and
throat of the racquet due to the energy loss caused by vibrations of the frame.

When a racquet is swung to serve a ball, the speed at the tip is greater than
the speed at the middle of the strings. The decrease in e, near the tip is large-
ly compensated by the increase in V, with the result that the serve speed does
not depend strongly on the impact point. Nevertheless. the serve speed does
vary with the impact point. The maximum ball speed results for an impact
somewhere between the middle of the strings and the tip, depending on
whether the racquet is head heavy or head light and whether any lead tape has
been added near the tip of the frame.

Suppose that the actual racquet mass M = 0.32 kg, the effective mass Me =0016
kg, and e = 0.88 for an impact in the middle of the strings. Then C= 00 FOr
such a racquet, the serve speed is given by v = (1 + e,)V = 1.39V. In order to
serve a ball at 200 kph, the racquet speed must be V = 200/1.39 = 144 kph. If
you use a heavier racquet with, say, M = 0.36 kg and Ma = 0.18 ke, then ie e=
0.43 and v = 1.43V. In this case the racquet speed needs to be 140 kph to serve
a ball at 200 kph. It is clear from this example that the serve speed depends
mainly on the racquet speed and it depends only slightly on the mass of the
racquet.
 Racquet Power 163

Similarly, the ball speed does not depend very much on the strings or the string
tension. Changing the string tension has only a small effect on e or e, , anda
small change in e, has an even smaller change on the ball speed, as described
in Chapters 13 and 32.

In theory, maximum possible racquet power would be obtained by eliminating

all energy losses in the ball and the racquet. In that case, e = 1 and then

(1-m/M,)
——————— 14.3
°a (1 +m/M,) ey

Given that the serve speed is v = (1 + e,)V, one would like to make e, as large
as possible. The largest possible value of e, is 1.0, which means that M, must
be a lot bigger than m. For example, if m/M, = 0.1 then e, = C1 - 0.1/1 + 0.1)
= 0,82, and if m/M, = 0.01, then e, = C1 - 0.01)/ + 0.01) = 0.98. In the limit
where e, = 1.0, the serve speed would be 2V. However, the racquet would be
too heavy to swing very fast. If we consider a lighter racquet with M = 0.32 kg,
M, = 0.16 kg, and if e = 1.0 for an impact in the middle of the strings, then e,
= 0.475. For such a racquet, the serve speed is given by v = (1 + e,)V = 1.475V.
This is the maximum possible serve speed for such a racquet, using an illegal
ball with no energy losses. The serve speed is then 6% faster than that for a
legal ball where v = 1.39V. The serve would be almost impossible to return
since the ball would bounce off the court about twice as high as a legal ball.

At the other extreme, one could reduce the serve speed by using a ball with e
= 0. Such a ball would act like plasticene and would not bounce at all off the
court. Nevertheless, it would bounce off the strings since the strings will stretch
and eject the ball. If the ball and the strings have about the same stiffness, then
e will be about 0.5 for an impact on the strings, in which case e, = (0.5 -
m/M,)/( + m/M.,). For example, if m = 57 gm and M, = 160 gm, then e, =
0.106 so v = 1.106V. The serve speed would then be 20% slower than for a
legal ball where v = 1.39V, but the ball would be impossible to return since it
would roll along the court instead of bouncing.

_ Estimating Racqu
We have shown how racquet power depends on racquet mass and the coeffi-
cient of restitution, but it also depends on racquet speed. It should be possi-
ble to measure the ball speed for any given racquet speed and impact point,
but no one knows how fast one should swing a racquet or what string to use
and at what tension in order to measure power. Power depends on who swings
the racquet. If the same racquet is given to 10 people and they are asked to
swing the racquet as fast as possible, then 10 different answers will be
obtained. Similarly, person A might swing a lighter racquet 10% faster, but per-
son B might swing the lighter racquet 5% faster.

The only racquet speed that could be used to test racquet power without ambi-
guity is zero speed. One can fix the speed of the incoming ball and measure
the speed of the outgoing ball. But the result will still not be a measure of rac-
164 Chapter 14

quet power since power depends on the speed at which any given racquet can
be swung.

One can estimate the effect of racquet speed on racquet power by making
some assumptions about racquet speed vs racquet mass. For example, one
could assume that a racquet that is 10% lighter can be swung either 5% or 10%
faster and then calculate the increase in ball speed in both cases, for both a
serve and a groundstroke. This is really only guessing what might happen, but
we can improve the guess using some simple physics. For example, if the ini-
tial energy of the racquet is the same for all racquets, then a 10% decrease in
M will require a 5% increase in racquet speed given that the racquet energy is
proportional to M x speed squared. However, this suggests that if M is reduced
to zero, then the racquet speed will increase to infinity. Reducing the racquet
mass to zero will not increase the racquet speed enormously since the mass of
the arm places a limit on how fast one can swing a racquet. Consequently, if
the racquet mass is decreased by 10%, then the increase in racquet speed will
be less than 5% and probably around 2% or 3%. An extreme example would
be to decrease the racquet mass from 20 gm to 10 gm. In this case, the mass
is decreased by a factor of 2, but there would be almost no change in the rac-
quet speed.

When a player serves a ball, the player swings the racquet in a complicated,
loopy path over the shoulder, behind the back and then upwards and forwards
back over the shoulder. The initial action is to position the racquet for a sud-
den explosive acceleration just before impact. The essential features of the
acceleration phase can be described by assuming that the racquet is swung
through a distance d by applying a force F. The work done is Fd, and hence,
the kinetic energy of the racquet just before impact is Irv? = Fd. The force is
therefore given by F = MV’/(2d). If the player serves with a heavier racquet
then the force that is applied to the racquet will not necessarily be the same.
Try pushing a tennis ball through the air and then push a brick wall. It is obvi-
ous that the force you can exert on the wall is much larger than the force you
can exert on a tennis ball. If the force applied to swing a heavy racquet was
the same as the force on a light racquet then MV’ would remain the same and
hence the kinetic energy of the racquet would be the same. In that case, a play-
er could swing a racquet of zero mass at infinite speed. A more likely result is
that the total kinetic energy of the arm plus the racquet will remain roughly
constant when the racquet mass is varied.

Suppose that when M is increased, F increases and V decreases in such a way

that FV stays constant. In physics, force times velocity is equal to power and is
measured in watts. The power supplied to a moving racquet (or any other
object) is defined to be the rate at which its kinetic energy increases. In other
words, power = force x distance/time = force x velocity. If FV remains constant
when the racquet mass is increased, then Mv?/ (2d) = constant, so M x V cubed
will remain constant. In that case, a 10% decrease in M would result in a 3.3%
increase in V.

As defined in physics, “power” is not the same as the “power” of a racquet

since racquets don't deliver any power on their own. We can talk about the
power of an engine or the power of a player since engines and players act to
do useful work. A racquet on its own can't jump up and serve a ball. It is just
an instrument that helps the player transfer energy out of the arm and into the
ball. Consequently, when we talk about the power of a racquet we are really
 Racquet Power 165

talking about the relative efficiency by which the racquet

acts to transfer energy from the player's arm into the The Moving Racquet versus
ball. In that respect we can say that a racquet is more the Free Racquet at Rest
efficient than say a pencil or a table tennis bat, but com- By Howard Brody
mon useage is to say that it is more powerful.
Abstract
The exact relation between racquet speed vs racquet
mass is not known, and probably varies from one play- |s there anything that can be learned about the inter-
er to another anyway. However, we can model the effect action of a ball and a tennis racquet that can be deter-
mined using a moving racquet (mechanically swung)
on serve speed by assuming that the racquet speed is
that cannot be determined from firing balls at a free rac-
given by V = k/M” where M is the actual racquet mass, quet at rest? It has recently been proposed that a
k is a constant and n is some number between 0 and machine that swings racquets be constructed as a
about 0.5 that describes how V varies with M. For exam- method for measuring the “power” of a tennis racquet.
ple, if n = 0.5 then the racquet energy will remain con- This sidebar will claim that tests involving a moving rac-
stant regardless of the racquet mass. If n = 0 then the quet can provide NO information that tests using a free
racquet speed will remain constant regardless of M. If n racquet at rest cannot provide.
= 0.33 then the power input from the arm to the racquet
Introduction
will remain constant. Calculations for various values of n
are shown in Figure 14.1. It was assumed arbitrarily that
Tennis racquets are tested by measuring the speed of
V = 80 mph when M = 300 gm, regardless of the value rebounding tennis balls that have interacted with the
of n, simply so we could see how different values of n strings. This can be done in a number of ways — with a
affect the serve speed. It was also assumed that M, = racquet at rest being hit by a ball of known speed, with
M/2 and e = 0.85 in order to calculate some typical val- a racquet of known speed hitting a ball at rest, or having
ues Of e,. a moving ball strike a moving racquet. In each case the
velocity of the rebounding ball and the location of the
impact on the strings are recorded. If all the desired
information can be obtained using a system with the rac-
quet at rest, it is clearly the easiest method for testing
the racquet, and it is the method that has been used for
Constant years by racquet manufacturers and sports scientists.
racquet energy Swinging the racquet gives no new information and it
adds to the experimental difficulties as well as adding
possible measurement errors.

Frames of Reference Argument

When a tennis ball rebounds off of a racquet at rest, the

ratio of the rebound ball speed to the incident ball speed
Serve
(mph)
speed
Constant is called the apparent coefficient of restitution (ACOR).
racquet speed The word “apparent” is used to distinguish this quantity
from the actual coefficient of restitution (COR), which
200 250 300 350 400 450 takes into account the recoil speed of the racquet. It has
Racquet mass (gm) been shown’ that the value of ACOR obtained with a
freely suspended, static racquet (the racquet is at rest in
the laboratory frame of reference) can be used to find
Figure 14.1
the ball rebound speed for a moving racquet hitting a
moving ball (both in the tennis court frame of reference).
This can be done using Newtonian transformations from
The biggest variation in serve speed results if we assume one frame of reference to another and the result is the
that the racquet speed remains constant when the rac- ball speed formula:
quet mass is increased. In that case, increasing the rac-
quet mass from 250 gm to 400 gm results in a 15% Vion =O,’ Vint (1 +e,)-V,
increase in serve speed (from 100 mph to 115 mph). If
we assume that the racquet energy stays constant, then where V,,,, is the ball speed coming off of the racquet,
V,, is the ball speed approaching the racquet, and V, is
there is an 11% decrease in serve speed. However, it is
the racquet head speed at the impact point, all in the
more likely that n will be about 0.3, in which case there
Continued on page 166
is a 1.9% drop in serve speed when the racquet mass is
166 Chapter 14

increased from 250 gm to 400 gm. If n = 0.2 there is a

tennis court frame of reference. This is an exact formula 2.3% increase in serve speed.
for the normal component of ball and racquet head
velocity, and no approximations have been made in
We conclude from the above analysis that there is
deriving it.
unlikely to be any significant variation in serve speed
The value of ACOR used in the ball speed formula with racquet mass. This conclusion is based on the fol-
depends upon the impact location and the relative ball- lowing observations:
racquet velocity, so data must be taken in the lab over a
range of incident ball velocities and impact locations. e It is reasonable to assume that racquet speed will
The value of ACOR will depend, among other parame- decrease as racquet mass increases, hence n > 0.
ters, on the racquet weight and weight distribution,
which will also affect the racquet head speed, V., when e It is unlikely that the racquet energy will remain con-
swung by a player. stant, since this assumes that the arm mass is negligible,
Racquet Head Speed hence nis 0:5.

How the racquet head speed varies with racquet weight e Players who like light racquets are unlikely to switch
and racquet weight distribution is at present poorly to heavy racquets (or vice versa) just to get a few per-
known for individual players or for players in general. It cent increase in serve speed, so a change in racquet
is felt that players swing heavier racquets at a slower mass from 250 gm to 400 gm is not a realistic option for
speed, but there is no data to confirm this. Research of - most players.
this type is presently under investigation § at
Loughborough University, and a report which measured
If a player wants to serve x % faster, the best solution
a few players serving is available’. If a machine were
will be to swing the racquet x % faster, and that might
constructed to swing a racquet, it would probably be
designed to input constant energy, constant impulse,
mean more gym work or a better diet rather than a heav-
constant force, constant torque, or some function of one ier racquet. The same conclusion does NOT apply to
of these physical characteristic that tests reveal for typi- groundstrokes. In that case, players are more likely to
cal players. In addition, the effective pivot point of the swing at a relatively controlled or constant speed, in
swing would have to be determined to determine how V, which case heavier racquets will be more powerful, less
varied along the face of the racquet. If such information maneuverable and more stable in terms of the amount
were available, then it would be possible to predict the of rotation during impact with the ball.
racquet head velocity from the known racquet’s physical
parameters and the laws of motion without resorting to
the machine. Therefore, the racquet swinging machine
would not be needed since a simple calculation using
the parameters used in designing the machine would Equations 14.1 and 14.2 remain valid for an impact any-
give the value of V.. where on the strings, not just at the center. The only
things that need to be changed are the values of e and
The results of this calculation for V, and the measured M.. For an impact near the tip or throat of a racquet, e
values of ACOR could be substituted into the ball speed
drops to about 0.6 for most graphite racquets due to the
formula, and the value of V,,,, would be obtained for any
energy loss caused by frame vibrations. Very stiff frames
value of V,,, the incident ball velocity. A similar argument
will hold for any type of bat, racquet, or paddle used in
don't vibrate as much, in which case e is about 0.7 at the
a sporting event. tip and throat, but it is still about 0.88 in the middle of
the strings. For an impact near the tip, M, is about 0.3M
and e, is close to zero. For an impact near the throat, M.
1.) Brody, H., The Physics of Tennis Ill, American
is about 0.7M and e, is about 0.5. In each of these cases,
Journal of Physics, 65, 981-987, Oct 1997.
Brody, H, How Would a Physicist Design a Tennis V is the speed of the racquet at the impact point, not just
Racquet? Physics Today, 48, 26-31, Mar 1995. the speed at the middle of the strings. The best proce-
dure is to actually measure e, at various points on the
2.) Mitchell, S.R., et al, Head Speed versus Racquet
strings rather than calculate it, and then calculate v from
Inertia in the Tennis Serve, Journal of Sports
Equation 14.1. Typical results are described in Chapter
Engineering, 3, 99-110, 2000.
10.

Stiff racquets are generally more powerful than flexible

racquets since less energy is wasted in frame vibrations.
 Racquet Power 167

The exception to this general rule is that frame stiffness makes no difference
to power for an impact at the vibration node near the middle of the strings
since frame vibrations are absent for such an impact. For example, if the frame
stiffness is increased so that the frame vibrates at 200 Hz instead of 125 Hz,
then e, near the tip will increase from about 0.10 to about 0.18 for a racquet
of mass M = 340 gm. This translates to an increase in serve speed of 7% since
(1 + e,) is increased by 7% (Cross, Sports Engineering, 2000). The serve speed
from the tip is slightly less than the serve speed from the middle of the strings,
but elite players tend to serve from a point near the tip due to the added height
advantage.

References

R. Cross, Flexible beam analysis of the effects of string tension and frame stiff-
ness on racquet performance, Sports Engineering 3, 111-122 (2000).

A. Nathan, Characterising the performance of baseball bats, American Journal

of Physics, in press (2003)
 Do-It-Yourself
Measurements
of Ball Speed
Anastasia Myskina, Russia

By Howard Brody

_Measuring Serve Speeds

adar Gun. As soon as you hear the words “How fast is the ball moving?” the first thing you think of is
the radar gun and the serve speeds you see displayed in lights at big tournaments. Radar guns are tra-
® ™ ditionally used for measuring serve speeds, but you never seem to see them measuring the speed of a
groundstroke. If they are calibrated correctly and aimed properly, radar guns give you the speed of the ball as
it leaves the racquet. Clearly, if the gun is not pointing at the ball impact location, it cannot give you the cor-
rect speed off the racquet. A tennis ball loses speed due to air resistance, and if the gun does not see or pick
up the ball’s trajectory until it crosses the net, for example, the answer will be 10 to 20 miles per hour slower
than the ball’s speed off the racquet. There is a second effect due to the ball’s direction of motion. If the ball is
hit at an angle relative to a line between the radar gun and the racquet (for example, the gun is aiming down
the center and the shot is hit cross court), the gun reading will be lower than the true ball speed, and this cor-
rection is given in Table 15.1.
170 Chapter 15

Table 15.1
ball
Change in radar gun reading due to pointing of gun at an angle to the
flight direction.

Reduction in
Reading
Angle Degrees Cosine of Angle of 100 mph serve

0 1.000 0
5 .996 -0.4
10 985 -1.5
15 .966 -3.4
20 .940 -6.0
25 .906 -9.4
30 .866 -13.4

As tennis coaches, many of you would like to be able to tell your students
exactly how many miles per hour they are hitting their shots. But you don’t
own a radar gun, and you have no intention of buying one, considering its lim-
ited use and high cost.

VCR/Camcorder. On the other hand, many of you already own a camcorder

with a high-speed shutter and a VCR capable of single frame or single field
advance. You can use this TV equipment to determine the speed of a serve or
a groundstroke with acceptable accuracy.

The major disadvantages of this proposal are that you do not get an instant
answer as you do with a radar gun (you have to play the tape back in the VCR)
and you have to do some work, where the radar gun just has to be pointed
and the speed is displayed.

There are two ways to measure ball speed with the tape you obtain from your
camcorder. The direct method of measuring exactly how far the ball moves for
each frame or field advance of the tape is difficult. It requires special camera
placement in addition to having the area around the hitter calibrated as to size
versus picture image. It is not a recommended method unless you are knowl-
edgeable about optics and physics.

Stop Watch. There’s a second method which will determine the ball speed off
the racquet. It’s done by measuring the average ball speed and applying a sim-
ple formula. For many years, I have been using a computer program that deter-
mines a tennis ball’s trajectory and I have noticed an interesting result.

The ball speed at the receiver's baseline is approximately half the speed of the
ball coming off the server’s racquet. (There are slight modifications depending
on court speed and type of shot, but to the accuracy that we need, these can
be neglected.)

The distance from baseline to baseline is 78 feet. So if you could measure the
time it took the ball to travel that distance, you would have measured the aver-
age ball speed. My computer output has shown that the ball speed off the rac-
quet head is 4/3 times this average ball speed.
 Ball Speed 171

If the time it took for the ball to go from baseline to baseline is 0.85 seconds,
the average speed would be 78 feet/0.85 seconds = 91.8 feet/second. Using the
4/3 factor predicted by the computer, this would correspond to a speed of 122
feet/second (4/3 x 91.8) off the racquet head.

To convert feet/second to miles/hour, you divide by 1.467, and this would give
you 83.4 miles/hour. Doing all the above arithmetic and combining it into one
formula (V = (4 x 78)/@ x 1.467 x t) = 70.9/T), the ball speed off the racquet
head Gin mph) would be 70.9 divided by the time (in seconds) for the ball to
go from baseline to baseline. In this case, the formula would be: ball speed =
70.9/time.

Now if you can handle a stopwatch sufficiently well enough to be able to time
the ball between when it’s hit and when it reaches the baseline, then you can
get the ball speed off the racquet head immediately. Just divide 70.9 by the
stopwatch time in seconds.

The problem is that most of us cannot do that with the needed accuracy, so
you should use the camcorder and VCR to determine the time. During single
advance of the tape during playback, you just count the number of frames or
fields between when the ball is hit and when it reaches the baseline. Each
frame is equivalent to 1/30 of a second and each field is equivalent to 1/60 of
a second (0.01666 sec).

If you advanced your VCR 51 times in tracking the ball from baseline to base-
line, and it was a single field advance, then 51 x 1/60 = 0.85 seconds. If you
have a single frame rather than single field advance, 25 advances would cor-
respond to 0.833 seconds and 26 advances would correspond to 0.867 seconds.
Just plug the time you measured into the (70.9/time) formula and you have the
ball speed as it leaves the racquet.

Corrections. There are some small numerical corrections that you may want
to make since not all balls travel exactly 78 feet each time they are hit, as some
go crosscourt, some are hit from behind the baseline, etc.

The numerical constant used (70.9) must be modified for those cases. If the ball
were hit from three feet behind the baseline, the distance would increase from
78 to 81 feet, and the numerical constant (to be divided by the time) would be
70.9 x 81/78 = 73.6.

As an example, a wide serve hit from the T, but returned from the baseline at
the doubles line, would have traveled exactly 80 feet, not 78 feet. For this case,
the ball speed (in mph) off the racquet head would be 70.9 x 80/78 divided by
the time in seconds. If the time was again 0.85 seconds, the initial ball speed
leaving the strings would have been 85.6, not 83.4 mph.

If you compare the results you get from this method to the results obtained
with a radar gun, you may discover it may give somewhat higher serving
speeds. This could be caused by a number of things, among them: 1) a radar
gun out of calibration, 2) a radar gun being used at an angle, or 3) the fact that
the radar gun does not pick up the ball until it has moved a certain distance
off the racquet. (A typical serve will lose 0.6 mph for each foot it moves
through the air.)
172 Chapter 15

If you think about it, this time interval you have measured is actually more
important to you as a tennis player than the ball speed. A discussion of the
importance of this time interval will be provided in Chapter 24 on Strokes and
Time.

; Measu ring Groundstroke S|

You can also use the VCR/Camcorder method to time the speed of a ground-
stroke, but as with the serve, it is a bit complicated. However, there is also a
very simple method for measuring the speed of the ball as it leaves the racquet
during a groundstroke rally that only requires a stopwatch. It also has the
advantage of giving an immediate answer (you don’t have to get out the VCR
and count frames).

If two players are having a groundstroke rally (in a practice session or in a

match), and you time how long it takes, on average,-for the ball to go from
racquet to racquet, Figure 15.1 tells you how fast the ball was moving when it
left the racquet. That’s all there is to it — except for a few details, which follow.

1) It is very difficult to time a single ball pas-

Figure 15.1 sage with enough accuracy using a stopwatch
to make this method practical. To get around
this problem, you should measure the time it
Time Between Hits vs. Ball Speed takes for a 3-, 4- or 5-shot (or even more) rally,
and then divide by 3, 4, or 5, as the case may
be.

This will reduce the timing error by a factor of

3, 4, or 5. If a 5-shot rally (ball crosses the net
5 feet behind five times) takes 6.9 seconds with a timing
baseline
uncertainty of +/-0.2 seconds in measurement
error, a single shot takes 6.9/5 = 1.38 seconds
)Time
(seconds pacsine with an uncertainty of only 0.2/5 = 0.04 sec-
onds.

You have the same total uncertainty or error in

timing whether you time one shot or 10 shots.
Ball Speed ( miles
/hour ) However, when you measure a multiple shot
rally, the uncertainty for a single shot is divid-
ed by the number of shots hit, greatly reducing
the error.

2) The time of travel depends on where the players are standing when they hit
the ball (on the baseline, behind the baseline, etc.). So the calculation has been
done (and the chart shows the results) for both a baseline to baseline rally and
for a rally with both players five feet behind the baseline. If they are at some
other position, such as 2-1/2 feet behind the baseline, then the curve will be
halfway between the two curves shown.

As an example, if you had timed a four-hit rally (four complete ball transver-
sals) and it took 5.2 seconds, the time between hits would be 1.3 seconds
(5.2/4 = 1.3). Looking up 1.3 seconds on the chart, you would see that the orig-
inal ball speed off the racquet would have been 55 miles per hour if the play-
ers were both on the baseline, 65 miles per hour if both players were five feet
 Ball Speed 173

behind the baseline, and 60 miles per hour if they were both 2-1/2 feet behind
the baseline (or one was on the line and the other five feet behind it).

Note that you cannot take the 78 feet (baseline to baseline), 88 feet, or any
other number of feet, and simply divide it by the time to get the ball speed off
the racquet. That calculation will give an average speed of the ball over its
entire flight and not the speed as it leaves the strings. This is because the ball
slows up considerably due to air resistance and its bounce. These factors,
which produce Figure 15.1, have been included in the computer program.

3) If instead of hitting shots parallel to the sidelines, the rally is going corner
to corner. That distance is 82.5 feet, which is halfway between the two curves.
(A cross-court rally from 2-1/1 feet behind the baseline would be similar to the
five feet behind the baseline data.)

Court surface and spin will make only a slight difference. All calculations were
done assuming the ball bounces about six to eight feet inside the baseline
(which is typical for most groundstroke rallies). It was also assumed that both
players are hitting the ball equally hard (which is the usual case for a rally) and
they are both hitting with the same pace from shot to shot. If these last two
conditions are not met, the result you get will be approximately the average of
the ball speed off the racquet.

This measurement of groundstroke ball speed does not have to be done “live,”
as it will give the correct results when you use a video of a rally. (You could
also then measure the speed of a single shot, as opposed to the average speed,
by the single frame/field advance method previously described.) From the
video replay, you can also determine the average groundstroke speed for some
of the old greats, such as Rod Laver. If you could trust the speed of your movie
projector (your replay was at the same number of frames per second as it was
when taken), you could also determine groundstroke speed for Bill Tilden and
Don Budge from old movie clips.

As a test of all this, I timed two touring pros as they were warming up. When
they started, the time for five shots ran from about 7.6 to 8.3 seconds. As they
warmed up, the five-shot time was reduced to about 7.2 seconds. After a few
more minutes, they reached times like 6.95 seconds for the five hits from about
three feet behind the baseline.

Michael and Carl Chang were hitting from about four feet behind the baseline
and the times were about 6.9 seconds for the five shots. Then MaliVai
Washington replaced Carl and the time for five shots dropped to about 6.5 sec-
onds, but they were a bit closer to the baseline.
 acquets
Adding Lead Weight
to Affect Performance

Bue

Jimmy Conner s, United States

By Rod Cross

ustomizing a racquet means modifying it in some way to suit the user. The most common way is to add
a string dampener. This deadens the sound of the strings, but that is all it does. It does not have enough
mass and it is not located in the right spot to reduce vibrations of the racquet frame. The human hand
is a much better vibration dampener. The next most common way to modify a racquet is to change the grip.
Players can replace a grip with a better one at low cost, or they can add tape or plastic heat-shrink under the
grip, or add an overgrip, to increase the grip size. Professionals may even have their grip custom made to suit
their exact hand shape and to suit the way they grip the handle.

The term customizing most commonly refers to the fact that self-adhesive lead tape can be added at one or more
spots on the frame to change the way the racquet feels and performs. The average player rarely bothers to add
lead tape, but the professionals insist on it, partly to ensure that every one of their racquets feels exactly the
same. Professionals are provided with racquets for free, but each racquet is usually a bit different from the next.
For example, it is usual for racquets to differ in weight by up to 5 gm, even if they are identical models. Rather
than shaving 5 gm off the heaviest one, it is simpler and better to add 5 gm to the lightest one. Alternatively, a
player might like to add 10 gm to the heaviest racquet and 15 gm to the lightest one so that all the racquets are
a bit heavier than the racquets provided. Extra weight should be added in the right spot, not just anywhere in
176 Chapter 16

the handle or around the tip. Adding weight changes the balance of the rac-
quet, which changes the way it feels when you swing it. It changes a whole
bunch of other things as well, and we will examine each of these things in
turn.

The balance point of a racquet refers to the location of its center of mass (or
its center of gravity to use an older term). It is easy to locate the balance point
by balancing a racquet on one finger, or on the edge of a ruler, or some other
narrow edge. It is usually close to the middle of the racquet, about 27/2 = 13.5
inches (343 mm) from the butt end of the handle. If it less than 13.5 in. from
the end of the handle then the racquet is said to be head-light. More than 13.5
in, then the racquet is head-heavy. The balance point will shift towards the tip
of the racquet if extra weight is added at the tip, by an amount that is easy to
calculate. Suppose the mass of the racquet plus the strings is M, its length is L,
and its balance point is located at a distance x from the butt end of the han-
dle. Now, suppose we add a small mass m at the tip. The balance point will
then be at a distance B from the butt end, where B is given by
(mL + Mx)
Be (16.1)
(M + m)

The balance point shifts a distance y = B - x = m(L - x)/((M + m).

For example, suppose that M = 320 gm, L = 690 mm, x = 340 mm and m = 10
gm. Then B = (10 x 690 + 320 x 340)/330 = 350.6 mm and y = 10.6 mm. The
balance point, therefore, shifts towards the tip by 10.6 mm, changing the rac-
quet from head-light to head-heavy in this case, the point halfway along the
racquet being 345 mm from the butt end. Since these values are typical of most
racquets, it is a general rule that an extra 10 gm at the tip will shift the balance
point by about 10 mm. If the racquet is heavier than 320 gm then 10 gm at the
tip will shift the balance point by a slightly smaller amount.

What if we add a mass m at the butt end instead of the tip? Then the balance
point will be at a distance
M
B= (16.2)
(M + m)

Using the above numbers, B = 320 x 340/330 = 329.7 mm so the balance point
is shifted towards the butt end by 10.3 mm. It doesn't matter whether m is
added at the tip or the butt, the shift caused by adding m is about the same,
except that the shifts are in opposite directions. Adding only 5 gm, the shifts
will be half as big.

Suppose you want to match two racquets so they have exactly the same mass
and the same balance point. Suppose that one racquet is heavier than the other
by G gms, and the balance points are also different. There are many ways to
match the two racquets, but the simplest way is to add G gms to the lightest
racquet. If you add it either at the tip or the butt end, then you might shift the
balance point too far, or not enough. If the shift is too big, then you will need
to add the extra mass at a point closer to the middle of the racquet. A good
result is obtained by adding equal weights at the 3 and 9 o'clock positions since
that makes the racquet more stable in terms of rotation around the long axis.
But what if the shift is not enough? Then you will need to add more than G at
 Customizing Racquets 177

one end of the lightest racquet and then add a bit extra to the heavier racquet.
If you add mass at the balance point of the heavier racquet, then the balance
point of the heavier racquet is unaltered. However, the aim would probably be
to shift both balance points closer together, since that way you would need to
add only a relatively small mass to the heavier racquet.

The swingweight of a racquet refers to its resistance to rotation when you

swing it. The actual weight itself represents the resistance to motion in a
straight line. If you didn't support the racquet, it would fall in a straight line to
the ground. To stop it falling you need to lift or push upwards with a force
equal to the weight of the racquet. Swinging a racquet is a different action. If
you swing a racquet with a certain force, then a racquet with a small swing-
weight will rotate faster than one with a large swingweight. The best way to
appreciate the difference between swingweight and actual weight is to tape a
small metal object, like a bolt, to the racquet, then swing it. It doesn't matter
where you tape the object, the total weight of the racquet plus the object will
be the same. But if you tape the object to the tip of the racquet, then the rac-
quet will feel a lot heavier when you swing it than if you tape the object onto
the handle. The reason is that an object at the tip travels faster than an object
on the handle and a bigger force is needed to accelerate it to a higher speed.

The swingweight of a racquet is known technically as the moment of inertia or

the rotational inertia. It is approximately equal to ML’/3, but it will be a frac-
tion higher if the racquet is head-heavy, and it will be a fraction lower if the
racquet is head-light. For example, if M = 0.3 kg and L = 69 cm, then the swing-
weight is approximately 0.3 x 69 x 69/3 = 476 kg cm’. If you add a mass m at
the tip of the racquet, then the swingweight will increase by an amount mL’.
Consequently, the swingweight will increase by a fraction 3m/M or by
300m/M%. For example, if M = 300 gm and m = 10 gm, then the swingweight
will increase by 300 x 10/300 = 10%. Most people would notice the difference
easily. However, if extra mass is added in the handle, the increase in the swing-
weight is much smaller and would not be noticeable. If you add a mass larger
than 10 gm in the handle, then the racquet will definitely feel heavier since the
actual weight of the racquet is increased even if the swingweight is not affect-
ed.

The above formula for swingweight refers to rotation of the racquet about an
axis through the butt end, and it gives an accurate measure of the effect of
swinging a racquet about an axis through the wrist. However, swingweight is
normally measured or specified for a rotation axis 10 cm from the butt end of
the handle, at a point near the first finger. The reason is not because players
swing the racquet about that axis, but because the racquet needs to be clamped
by the handle in a machine to swing it when measuring the swingweight. The
swingweight measured this way is less than the value ML’/3, by an amount
equal to 100M. For example, if ML?/3 = 476, and M = 0.3 kg, then the meas-
ured swingweight will be 476 - 30 = 446 kg-cm*’.

When a racquet is swung by a player (instead of by a machine), the racquet

rotates about several different axes all at once. There is rotation about axes
through the wrist, the elbow, and the shoulder, as well as rotation about the
hips. For each of these axes, the racquet has a different swingweight. Adding
178 Chapter 16

a weight at the butt end of the handle has no effect on the swingweight about
an axis through the wrist, but it increases the swingweight about an axis
through the elbow or shoulder. The effect is fairly small since the total swing-
weight about the elbow includes the swingweight of the hand and forearm as
well as the racquet.

Racquet Power _
Adding a small weight at the tip of a racquet increases the power of a racquet
by a small amount. Adding weight anywhere in the head will increase racquet
power, but adding weight at the tip is the most effective. The trade-off is that
the racquet is harder to swing and it is less maneuverable. But if the racquet is
very light to start with, or if you have strength to spare, then the extra power
could be useful. The extra power is relatively small since you won't be able to
swing the racquet as fast if you increase its swingweight. For a controlled shot
where racquet head speed is kept the same, the extra weight will have a more
noticeable effect on racquet power.

A weight added at the tip of a racquet shifts the point of maximum power clos-
er to the tip. For a serve, this gives the player an added height advantage. For
a groundstroke, the point of maximum power is closer to the center of the
strings. The point of maximum power depends on the speed of the incoming
ball as well as the speed of the racquet. If the racquet is initially at rest or mov-
ing slowly compared with the speed of the ball, then the maximum power
point is near the throat of the racquet, since that is where the ball bounces best.
For a serve, the maximum power point is near
the tip, since that is where the racquet is mov-
ing the fastest. Calculations of the ball speed
210 with 30 gm added at various spots are given in
Figures 16.1 and 16.2. These calculations are
200 based on measurements of the ball speed off an
7. S No added initially stationary racquet of mass 332 gm
190 ie a Jag mass (Cross, 2001).

180
(km/hr)
v 170 30g atx=16cm
If a player hits a groundstroke near the top or
bottom edge of the racquet, then the racquet
160 30 g at tip
will tend to rotate about its long axis. The hand
will stop it spinning around this axis, but the
racquet will rotate at least a few degrees,
enough to alter the intended trajectory of the
ball. The best solution is to hit the ball in the
middle of the strings, but no one is perfect. The
next best solution is to use a racquet that does-
Figure 16.1 n't rotate easily around the long axis. The way
to do that is to make sure that the material in
Serve speed with 30g added at the tip or 16 cm
from the tip. The serve speed is shown vs the
the frame is as far from the long axis as possi-
impact distance, x, from the tip. ble and it is as heavy as possible. The extreme
case would be a very heavy racquet with a
The racquet speed was decreased slightly as the huge head. However, it is not necessary to go
swingweight increased, keeping the ball speed to that extreme. A more practical solution is to
fixed (200 km/hr) at x = 15 cm. use a moderately large head, or to add small
weights at the 3 and 9 o'clock positions. Adding
 Customizing Racquets 179

a weight at the tip will have no effect at all in

this respect, since the axis passes straight
160
through the 12 o'clock position. For the same
reason, additional mass in the handle will have
no effect. ee
mS ; aes a= 140
An ice skater can spin around rapidly on one =
spot by keeping the arms close to the body. In £ 130
order to slow down, the skater pushes the = No added
arms out, away from the body. Similarly, a 420 mass
diver doing a triple somersault tucks the arms
and legs close into the body in order to spin 410
as fast as possible, and then stretches out to
slow the rotation speed. It's the same with a
racquet. Mass close to the spin axis has very 0 5 10 15 20 25 30 35
little effect on reducing racquet rotation, but
mass well away from the axis reduces the rota- X (cm)
tion speed. The moment of inertia about the
long axis is the important factor. Old wood
racquets with small heads had a small moment Figure 16.2
of inertia, which probably explains why play-
ers using wood racquets did not hit the ball as Ball speed vs impact distance from the tip, x, for
hard as the modern players. The larger head in a groundstroke when the racquet speed is kept
constant. The ball is incident at 20 m/s (72
modern racquets gives much better rotational
km/hr) and the racquet is swung so that the tip is
stability and allows the player to hit the ball
rotating at 27 m/s (97 km/hr) and the throat (at x
hard with reduced rotation about the long = 30 cm) is rotating at 18 m/s (65 km/hr).
axis.

Another method of reducing the rotation angle

slightly is to increase the string tension. The
rotation angle increases with the rotation speed and it increases with time. At
high tension the ball spends a shorter time on the strings and hence the rota-
tion angle is reduced. This is possibly the main reason that players notice an
increase in ball control at high string tension.

The sweetspots of a tennis racquet are near the middle of the strings. Adding
a weight at the tip of the racquet shifts the sweetspots closer to the tip. One of
the spots is technically known as the fundamental vibration node and the other
is called the center of percussion. If a racquet hits a ball at the vibration node,
then the racquet will not vibrate, regardless of how stiff or flexible the racquet
is. That point will shift by about 16 mm closer to the tip if an extra 10 gm is
added to the tip, or 32 mm if 20 gm is added to the tip. If weights are added
to the 2 and 10 o'clock positions on the frame, then the vibration node is not
shifted at all since those points on the frame never vibrate, regardless of where
the ball strikes the strings. Adding weights at 3 and 9 o'clock has only a tiny
effect on shifting the vibration node in the middle of the strings. Adding weight
in the handle has essentially no effect on shifting the vibration node in the
middle of the strings.

An impact on the strings at a point on the long axis causes the racquet to rotate
about an axis located at a point in the handle. That point is called the conju-
gate point. Every point in the handle has a different COP in the strings. For
180 Chapter 16

example, if the conjugate point is right at the butt end of the handle, and if the
racquet has length L and has a uniform mass distribution, then the COP is locat-
ed at a distance L/3 from the tip. The center of percussion (COP) is located
about 40 mm closer to the throat than the vibration node. If a racquet strikes
a ball at the COP, then the racquet will vibrate a small amount, but there is no
sudden movement of the butt end of the handle one way or the other (apart
from the back and forth vibrational motion). As a result, the shot feels nice.
The shot feels nice anywhere around the center of the strings, not just at one
or two tiny spots, since the level of shock and vibration drops significantly in
this region.

Adding 10 gm at the tip will shift the COP about 5 mm closer to the tip. Adding
weights in line with the COP (near the 3 and 9 o'clock positions on the frame)
does not shift the COP at all. Adding weight in the throat will shift the COP
slightly closer to the throat. Removing weight from the throat (when the rac-
quet is manufactured) shifts the COP slightly towards the tip. Adding weight at
the conjugate point in the handle does not shift the COP at all.

The addition of weights to a racquet changes almost every physical parameter

of the racquet. However, there are particular locations where an added weight
may have no effect at all on some of these parameters. For example, (a) a
weight added at the balance point does not shift the balance point, (b) a
weight added in the handle does not change the swingweight of the racquet
about an axis through the handle, (c) a weight added at the center of percus-
sion or the corresponding conjugate point does not shift the center of percus-
sion, and (d) a weight added at a vibration node on the frame does not shift
the location of the vibration node and it does not change the vibration fre-
quency of the racquet.

Reference

R. Cross, Customizing a tennis racket by adding weights, Sports Engineering 4,

1-14 (2001).
 How Equipment Has
Changed the Game

Daja Bedanova, Czechoslovakia

By Howard Brody

hen you watch tennis being played today, if you are as old as the authors of this book, you imme-
diately notice that the tennis racquets are quite different from the wooden racquets that were in use
30 or more years ago. The modern racquet is larger in both head length and head width, consider-
ably lighter, less flexible, made of a reinforced plastic material, and it is possibly an inch or so longer. When you
watch the more proficient young players hitting, you also see that the style of tennis strokes has also changed
over that period of time. The classic, smooth, flowing groundstrokes that were displayed by the tennis champi-
ons of the first 2/3 of the 20th century have been replaced by a game that emphasizes power from the baseline
and the ability to end the point with a single swing anytime the opponent is slightly out of position, or hits a
shot that bounces short. Is it possible that the changes in stroke mechanics are a direct result of the changes in
racquet technology, or did the strokes evolve independently of the racquet changes? What this chapter will
explore, is how the new technology in tennis racquets has allowed players to modify the way they hit the ball,
and get away with it.

nal Racquets

The old standard wood racquets that were used by most players up until the 1970 era, weighed at least 14 to
15 ounces, had a neutral balance, and had a small head in both length and width. Because of the structural lim-
itations of wood, in order to fabricate a racquet light enough for a player to comfortably swing, the frame could
184 Chapter 17

not be made very thick, which resulted in some degree of flexibility, particu-
larly near the tip. In addition, the strength-to-weight ratio of wood dictated that
the head could not be too large if you wanted to string it at a reasonable ten-
sion.

The modern racquets are molded from a graphite-reinforced material (plus, on

occasion, other exotic reinforcing materials), weigh 9 to 11 ounces, and are
often handle light. It is even possible to construct racquets with a weight as
low as 7 ounces, yet have a head that is quite large and produce a frame that
is exceedingly stiff and durable.

The classic groundstroke used by most players up until the middle of this cen-
tury was a long flowing swing with good, early preparation, and a long follow
through. The players stood sideways and smoothly accelerated the racquet
through to the hitting region as the body weight was transferred forward (you
stepped into the ball). The shot was often hit flat, or with some topspin, but
occasionally players preferred to use a small amount of slice (underspin) on
the backhand.
ew Stro } ‘

The modern forehand groundstroke that many players have adopted has an
open stance with a semi-western or western grip. The players coil the body,
and then uncoil it as the shot is hit, bringing the big trunk muscles into action.
This uncoiling, rotational motion often causes the player to leave the ground
during, or at the end of, the shot. The result is that the ball is hit with exces-
sive topspin and usually very hard.

In today’s tennis, even the youngest juniors seem to be whipping their racquets
around with abandon. Years ago, probably only the exceptional player would
have had the physical ability to have done that with the old, heavy wood
frames. The average player and the recreational player needed the long, flow-
ing swing to accelerate the old, heavy racquet up to hitting speed. In addition,
this gradual acceleration gave the player much more control of the racquet
head and allowed the player to hit the ball at approximately the same location
on the racquet face each time it was swung. This was essential, since various
impact locations on the old racquets often responded quite differently (the
“Sweetspot” was small). A ball hitting an inch or so away from the desired loca-
tion on the head might easily end up in the net or go long. With these old rac-
quets, to get the ball to land consistently in the court required a consistent
impact location on the strings, which could only be accomplished with a con-
trolled swing.

With the quick, whipping swing that many players use today, it is more diffi-
cult to hit the ball at exactly the same location on the head each time the rac-
quet is swung. However, due to the characteristics of the modern racquet and
the heavy topspin strokes used, the resulting ball trajectory is much less sensi-
tive to the exact location of the ball impact on the strings. (The racquet is more
 Racquet Technology and Strokes 185

“forgiving” or it has a large “Sweetspot”.) If there is a preferred location to hit

the ball (such as the center of the strung area), the new racquets give you more
latitude for error of impact location, both in a direction across the racquet face,
and along the main axis of the frame.

Because the new frames are much wider than the old wood racquets, they are
much more stable against twisting when the ball impact point is not along the
principal axis. The physical property of the racquet that produces this stability
is called the polar or roll moment of inertia. The larger this moment, the less
the racquet will twist on off-center hits and the less the power degrades as the
ball impact point moves off the axis. This moment of inertia is proportional to
the weight of the racquet and to the square of the width of the racquet head.
A 10-inch wide head (the size of a typical oversize racquet) is 25% wider than
the old 8-inch wide wooden frame, so it has a moment of inertia that is over
50% greater. This more than makes up for the 25% reduction in weight that
comes with the newer racquets. This increase in polar moment reduces the rac-
quet twist on off-axis impacts (hence reduces the ball's errant angle due to the
twist), and keeps the rebound ball speed from changing too much on such off-
center hits. Both of these effects give the player a larger margin for impact loca-
tion error in the striking of the ball. In addition, the use of topspin gives the
player a much larger “window” of acceptable angles to hit into, if the shot is
to land in the court.

When the power of a racquet is measured in the laboratory, balls are fired at
the frame and the ball rebound speed is measured for various impact locations.
For a typical racquet, the ball rebound speed is a maximum for impacts near
the throat and the ball rebound speed falls off as the impact location moves
toward the tip. As a general rule, the further the impact location is from the
balance point, the lower is the ball rebound speed in the lab. In addition, stiffer
racquets tend to have more power than flexible racquets, particularly closer to
the tip. Many of the new racquets are head heavy, which means their balance
point is further up in the head. Because of this, the maximum power point
moves up, away from the throat. As the racquet head is made stiffer, the power
degradation near the tip is reduced. These results are for a situation where the
racquet is not being swung, but is at rest in the laboratory.

When the racquet is swung, the tip is moving somewhat faster than the throat
and the ratio of these two speeds depends upon the exact nature of the swing.
A very wristy, whipping motion will have a greater tip to throat speed ratio
than the old classic swing. Because the tip is moving faster than the throat, the
maximum power point moves up higher in the head. When the physics of all
of these factors (the racquet response and player swing) is combined to pre-
dict the actual playing characteristics of the racquet, the new frames with the
modern style of swing show a uniformity of power response over a large area
of the head. This is contrasted by the old frames, coupled with the classic
swing, where the ball had to be hit at exactly the same place every time to get
a uniformity of response.

For the classic racquet and classic swing, if the ball impacted beyond the cen-
ter of the head, there was a loss of power and the ball might not clear the net.
186 Chapter 17

For those old wooden racquets, if the ball impacted closer to the throat than
the center of the head, the result was more power, and it was likely that the
shot would go over the baseline. With the modern racquet and modern swing,
these sorts of variations in ball impact location will result in very little variation
in resultant ball speed, compared to the ball speed when the impact is at the
center of the head. This means that if the ball is struck an inch or two from the
center of the head, the result will still be a good shot.

If a player had the physical prowess to swing a heavy, classic racquet in the
modern, wristy manner, any small resulting miss-hits might end up spraying the
ball all over the court. The modern racquets, due to their forgiving nature and
their lighter weight, have allowed players to adopt a new style of swing and
still get the ball to land in the court where they want it.

_Court Surfaces
and Strokes _
There is a second reason why strokes using the western and semi-western
grips are seen more frequently these days than 30 years ago. From the begin-
ning of tennis, the game was played on grass, which is a low bouncing, fast
surface. Three of the four “Grand Slam” tournaments were on grass, as were
many of the other tournaments. (For example, the USTA was known as the
USLTA, where the L stood for lawn.) The western and semi-western grips are
not optimized for fast, low bounces, but are ideal for medium or slow paced,
high bouncing surfaces. Today the professional grass court circuit lasts at most
a month, so most players learn, practice, and play on either hard courts or clay
courts. Consequently, the new grips and the new types of strokes have evolved
along with the new racquet technology to match the predominant court sur-
faces.

_Bring Back the Wood Racque'

Over the years there has been a cry (particularly by a certain TV commentator)
to bring back the old wood racquets and return tennis to the sport it was 30
or 50 years ago. There are complaints about the speed of the game, the short-
ness and rareness of a rally, and the new style of stroking that has developed.
This may sound fine on television, but when it comes to drafting the rules of
tennis, there are some problems with this suggestion. Exactly what is a wood
racquet? The so-called wood racquet in use 40 years ago was actually a frame
made of wood, glues, fibers, and other materials. In other words, it was a com-
posite, just as today’s racquets are composites. Would the International Tennis
Federation (the group responsible for preserving the basic character of the
game of tennis and enforcing the rules of tennis) have a rule saying the frame
must be 75% or 90% wood? Would graphite fibers and epoxy be outlawed,
even if the remaining 90% of the frame were wood? How would the ITF
enforce any such rule? Would they have to take apart each and every racquet
model from each and every racquet manufacturer and do an analysis of the
composition? This does not look like a promising solution. That is why the ITF
has opted for a different solution to the perceived problem.

There is no doubt that the new, graphite composite racquets, having only
between half and two-thirds the mass of the old wood racquets, allow you to
generate more power (rebounding ball speed off the head). It is also clear that
the new lightweight, oversize, extremely stiff frames have a bigger “sweetspot”
in that they allow you to miss-hit the ball a bit more without severe penalty (in
other words, they are more forgiving of player error).
 Racquet Technology and Strokes 187

There is general agreement that the new racquet technology has allowed the
game to change. Beginners find it easier to learn to play and they reach a level
of competence where they enjoy the game sooner — at least before they give
up in disgust because they perceive that tennis is too hard to learn. The aver-
age recreational player seems to get more enjoyment when playing with the
new equipment that is available. The women’s professional game has gone
from a slow, long rally contest to an exciting game. The women (as repre-
sented by the WTA) absolutely do not want the game to change back to the
way it was 50 years ago.

Then what is the problem that is bothering so many? It seems to be centered

around the men’s professional circuit, particularly on fast courts, such as the
grass at Wimbledon, some indoor carpets, and certain acrylic surfaces, such as
used at the US Open. People claim the game is getting too fast, and this often
means that it is becoming serve dominated and rallies are rare. Is this true? Has
the game become faster or more serve dominated over the years? This was a
question the ITF asked, and the ITF Technical Commission attempted to
answer, with a research project reviewing decades of Grand Slam data.

It was hypothesized that if the serve were becoming more dominant in the
men’s professional game, then the rate of tie-breaks would be increasing and
this rate should depend on the surface being played on. It was also hypothe-
sized that none of these tie-break correlations would appear in the women’s
game, where the serve is not as important. The four Grand Slam tournaments
were chosen for analysis since they have four different court surfaces, have
both men and women competing, and probably have the least turnover of
players from tournament to tournament during a given year.

The results of the analysis of the past 20 years of Grand Slam events (over
90,000 sets) showed a strong correlation of rate of tie-breaks to perceived court
surface speed for the men and no correlation for the women. Among individ-
ual men playing on the grass at Wimbledon (a very fast court surface), there
was a strong correlation between tie-breaks played and both average first serve
speed and maximum first serve speed. There was also a slow but steady
increase in the rate of tie-breaks over time at all four Grand Slam tournaments
for the men, but no such trend for the women.

This clearly shows that the men’s game is getting faster and becoming more
serve dominated. Is this a problem? That is a matter of opinion. If this trend
continues for 10 or 20 more years, will it harm the game of tennis and will
spectators begin to lose interest? There is some chance of this happening. Will
people want to see the serve and volley game at Wimbledon become a serve
ONLY game? Tie-breaks are exciting, but if more than half of all the sets played
end up in a tie-break because no one ever has their serve broken, is that good,
exciting tennis? Having done the research, the ITF Technical Commission had
to see what it could do to control the speed of the game, if it became neces-
sary in the future.

A number of possible means to slow the game down were carefully analyzed.
Many of these ideas, when subjected to scientific scrutiny, were discovered to
not be effective. Other means (such as changing the dimensions of the tennis
court) were considered too expensive, since there are over 100,000 tennis
courts in the world. Many possible solutions to the perceived problem would
have adversely impacted the recreational game of tennis (the ITF wants one set
of rules for the game — both professional and recreational). The solution adopt-
188 Chapter 17

ed after several years of testing and analysis was to allow (not mandate) the
use of a larger (6% in diameter) tennis ball. The use of this larger ball will slow
down the game and reduce the serve dominance, and at the same time prob-
ably improve the quality of the tennis played by the recreational player. In
addition, due to its increased cross-section (12%) the larger ball should play
better at higher altitude than a standard size ball.

The ITF at its 2001 annual general meeting, approved the use of the larger ball
(called a Type 3 ball) for use on any surface, in any tournament or match,
recreational or professional. The “old” standard ball (called a Type 2 ball) is
still allowed to be used in any tennis match as well.
 Where To Hit
on the Racquet

Hernan Gumy, Argentina

By Howard Brody

uet Head Should You Hitthe Ball?

ost players have no idea as to exactly where on their racquet head they hit the ball. And, if they were
interested, they would probably have no way of determining the ball impact location, let alone mod-
“ ’ ifying it. They might have an idea of where they hit the ball ON AVERAGE from the wear on their
strings or where the yellow ball fuzz is caught on the strings, but not the location of a specific single shot. Is
the question of where the ball strikes the strings only of academic interest, or can it be of some practical use to
a tennis player?

Ask a player where they aim to strike the ball, and they will often say “the sweetspot — wherever that is.” Some
players aim to hit the ball where it feels good, while other players aim for the center of the head because that
gives them the largest margin of error for hitting a clean shot. Wouldn’t it make sense for racquet designers to
have the sweetspot of their racquet in the center of the head? Then many players would be satisfied and part
of the problem would be solved. But not really!

There are three distinct sweetspots on a racquet, where it “feels good” when you hit the ball there. They are the
ball impact location that produces the least shock or jar to the hand (called the center of percussion or COP),
190 Chapter 18

the ball impact location which results in the least racquet frame vibration
(called the node), and the ball impact location where the player gets the max-
imum “power” on the shot. The location of minimum shock, or jar, and the
location that minimizes vibrations are two obvious places where it feels good.
The maximum power location is the place where the ball zings off the racquet
with the least effort on the player’s part, so it also feels good.

This maximum power point may also tend to produce shots that also maximize
control and minimize errors, so it may be the sweetest of the sweetspots. This
“power” spot is the location that a player should aim for. Unfortunately, you
can’t paint a spot on the strings at that location because it is not a fixed point
since it depends on the racquet swing. On the other hand, the location of the
node and COP are fixed by the racquet parameters (length, weight distribution,
etc.), and with modern, oversize racquets, they are usually near the center of
the head. With the old, small headed wood racquets, these two sweetspots
were near the racquet throat. When Howard Head (who invented the Prince
racquet) enlarged the size of the racquet head, the center of the head came out
close to the location of these two sweetspots, not by moving their position, but
by moving the location of the center of the strung area. This was one of the
reasons for the immediate popularity of the oversize racquet. The maximum
power point is determined by those same racquet parameters Cength, weight
distribution, etc.), plus the type of shot that is being hit, the incoming ball
speed, and the peculiarities of an individual’s stroke. It can be anywhere from
the throat area to near the tip, depending on the details of the stroke. That’s
why you cannot nail down its location and paint a spot on the strings there.

If you add weight to your racquet by placing lead tape near the tip, then you
will move all three sweetspots up, away from the throat, and closer to the tip.
Placing extra weight at the 3 and 9 o’clock locations of the head will probably
move the power point up a bit, but will not markedly change the location of
the node or COP. Ct will give better stability against twisting for off-axis
impacts. )

In one well known study’, where the ball impact location was measured for
the groundstroke of a number of good players, it was reported that the node
location was struck rather than the COP location by most of the players tested.
However, the study was faulty in that it did not determine the location of the
power point, so it might have been a preference for that point over the COP
point that was being observed.

As a general rule, the maximum power point moves away from the hand as
the incoming ball speed decreases relative to the racquet head speed or as the
stroke becomes more “wristy” and the head is moving much faster than the
throat. If your opponent hits a hard shot, you want to hit the ball on your rac-
quet a bit closer to your hand (see Note 1). The same holds true with respect
to court speed. You will probably do better to hit the ball closer to your hand
on grass courts and further away on clay courts. The limiting case is the serve,
where the ball speed is zero and the racquet snap gives you effectively a very
wristy stroke. For a flat serve, you want to strike the ball well above the cen-
ter of the head of the racquet to get the maximum ball speed. As a bonus, this
extra height at which the ball is struck allows you to get in more serves.

With a modern racquet, groundstrokes should be hit not too far from the cen-
ter of the head. It is even possible to have a slightly different optimum ball
 Impact Location 191

impact location on the strings for forehand and backhand, depending on the
mechanics of your stroke. With the classic groundstroke that was seen 40 or so
years ago, when used with the wood racquets that were popular then, the ideal
impact location for maximum power was below the center of the (small) head,
near the throat. With today’s modern lightweight racquets and strokes using a
more flexible wrist, the power point has moved up somewhat and its location
is now nearer the location of the other sweetspots.

That still leaves the problem of knowing exactly where on the strings the play-
er actually hits the ball. For the serve, that is an easy experiment to do. Take
the ball, chalk it Qwith school blackboard chalk), and then serve that ball. There
will be a clear pattern of chalk residue left on the strings showing exactly
where the ball was struck. On a flat serve, if that mark is not between the cen-
ter of the head and the tip, you are probably not serving to your full potential.
Groundstrokes are more difficult to assess, since chalking the ball may not
work too well. You might inspect your strings for wear, but this will only pro-
duce meaningful information if you are very consistent in your stroking and
you get a lot of hits out of a set of strings. Another method is to coat or paint
the strings with something and see where the wear in the coating occurs. A
final method is to use a video camera (with a fast shutter) and watch a num-
ber of shots taken from behind or in front of the player.

ransverse Direction
Up to now in this chapter, only locations along the long axis have been dis-
cussed. Is there a problem in the transverse direction — away from the long
axis? For a flat shot, the closer the ball impacts to the axis, the better the shot.
Then there will be no twisting or turning of the racquet leading to a ball exit-
ing at an unwelcome angle. There will be no loss of power, since off axis
impacts degrade the rebounding ball speed. But this is all for flat shots.

What about spin (topspin, backspin, and slice) shots? When you hit a shot with
spin on it, the ball does not enter and leave in a direction normal to the string
plane. The ball impacts at a particular point, then slides or rolls a bit, and
comes off of the strings at a different location. If the ball initially impacts close
to the racquet axis, then rolls or slides for a distance, the ball may strike or
come close to striking the frame before fully exiting from the strings. This
would greatly reduce your margin of error for miss-hitting a spin shot.

To try to increase the margin of error, the ball should initially impact off-cen-
ter, roll or slide across the racquet axis, then exit off-center the other way. This
will allow you to hit harder shots with more spin while reducing the chances
of the ball hitting the frame.

How can all of this help you? If you are like most players, you just hit the ball,
and over the years you have developed a hand-eye pattern that probably has
the ball impacting close to the location that this chapter is recommending. But
you can do somewhat better. The next time someone hits a particularly hard
shot at you, make a conscious effort to strike your return just a little closer to
your hand, toward the throat of your racquet. If the shot coming toward you
is somewhat slower than average and you want to hit your return hard, then
try to meet the ball a little bit further out on the strings, in the direction of the
192 Chapter 18

tip. Both of these suggestions will tend to optimize your racquet’s performance
over the performance you will get if you are hitting the ball at the same loca-
tion on the strings every time you swing. On your serve, make sure that you
are hitting the ball above the center of the head.

Notes

1. It is well known in baseball, that when the pitch is a fastball across the out-
side of the plate, you should “go with the pitch” and hit to the opposite field
rather than try to pull the ball down the line. This is because batters do not
have the luxury (or time) to reposition themselves and turn an outside pitch
into one closer to their body. Tennis players can move and have the option of
hitting the ball on the racquet closer or farther away from the throat.

References

1. Impact probability distribution, sweet spot, and the concept of an effective

power region in tennis racquets. Herbert Hatze Journal of Applied
Biomechanics, 1994, 10, pgs. 43-50
 Physics Principles
for Getting the Ball In

By Howard Brody

Introdu

E
sing a knowledge of the laws of physics, a player can improve the serve percentage (get a larger frac-
tion of the serves to go in) by changing parameters so as to increase the vertical angular acceptance of
the shot. One obvious way to do this is to reduce the speed of the serve. However, a slower serve may
be counter-productive, in that it could lead to a net loss of more points than it gains, depending on the oppo-
nent. If it is desired to maintain a high speed serve, detailed examination of ball trajectories and an analysis of
racquet dynamics show how to get more serves to go in. There will be an improvement in serve percentage if
the ball is hit when it is higher, which is why taller players, being able to reach higher, have a distinct advan-
tage when serving. There also will be an improvement in the vertical angular acceptance (or "window") when
the ball is hit with the racquet moving upward in order to produce topspin. A third method of increasing the
serve percentage is to toss the ball up higher and hit it as it is falling downward. It is shown how each of these
suggestions can increase the vertical angular acceptance for the serve being good, which, in turn, should lead
to an increase in serving percentage.
194 Chapter 19

Many players love to hit their first serve as hard as they can, but unfortunate-
ly, it doesn’t go in very often when they really smash the ball. The reason this
happens may be caused by several separate effects; the biomechanics of their
swing, the physics of the ball trajectory, and/or the racquet they may be using.

When players try to hit the ball harder, they tend to lose some control of their
racquet and end up spraying their shots around. The result is that the ball hits
the net or goes long more often than it goes in. To remedy this problem
requires athletic skill, correct stroke mechanics, and practice. Lots of practice!
The other causes of a poor serving percentage, the physics of the ball trajec-
tory and not having optimum equipment, are the subject of this chapter. Using
the information developed here, it might be possible to have a better serve
without additional hours of practice.

On the serve, once the ball impact height and location, as well as the ball
speed and spin have been fixed, there is a minimum angle the ball must have
as it leaves the racquet just to clear the net. Keeping all the conditions fixed,
there is a maximum angle that the ball can have and not bounce beyond the
service line (Figure 19.1). The difference between the minimum and maximum
angle is the vertical angle of acceptance for that serve to be good. At the max-
imum angle (corresponding to bouncing on the service line), the ball clears the
net by a certain margin. This margin of clearance over the net can be consid-
ered a window of acceptance that the ball must be hit through if the serve is
to be good. If a ball reaches the net below the window, the net is hit. If the
ball crosses the net above the window, it will bounce beyond the service line,
and be a fault.

If this window (or clearance over the net) is large, the serve is more likely to
go in, compared to a serve that must pass through a small window. This is par-
ticularly true when the variation in the initial angle from one serve to another
is large, due to biomechanics of the stroke. The variation in vertical angle of
the ball off of the racquet on the serve can be reduced by improving the stroke
mechanics, taking lessons, practicing, and not being too ambitious. However,
there is a limit for each player beyond which they cannot go, regardless of how
much they practice. At this point, the only way to increase the serve percent-
age is to increase the window.
Figure 19.1

Trajectories for a ball that just

clears the net and a ball that
lands on the service line.
 Improving Your Serve 195

The size of the window, or clearance at the

net through which the ball must pass, Acceptance vs Serve Speed
depends on the ball speed off of the rac- 2
quet, the height at which the ball was hit
(impact height), the spin of the ball, and the
server's location on the court. It is clearly
advantageous to arrange these parameters
so as to maximize the window which will
allow more serves to go in, if at the same
time, a serve can be hit that will win points (degrees)
(or prevent your opponent from winning
Window
Acceptance
points). We will examine each of these | 7.5 feet |
Sa
parameters and see how to modify the serve
mechanics, equipment, or strategy in order
to improve effectiveness.
80 90 100 110 120 130 140
Serve Speed (mph)

A computer program has been written that :

calculates the vertical clearance window of Ales dice
a serve as a function of each parameter. Angular acceptance versus ball speed for serves hit at
First, we will examine what happens as the 7.5 and 9 feet above the ground (impact height).
ball speed is varied. The results of this cal-
culation are shown in Figure 19.2. From
these data, it is apparent that the harder the
ball is hit Gncreased ball speed off of the racquet), the smaller the window
through which the ball must pass in order to be a good serve. The physical
reason for this result is that the force of gravity is pulling the ball down into
the service box, and on slower shots, gravity can more readily change the ball’s
direction and get it to bounce in the box.

Many players, finding that their hard serve is not going in a reasonable fraction
of the time, switch to a slower serve in hope of raising their serving percent-
age. Just reducing the ball speed from 100 mph (with a clearance over the net
of 7.5 inches) to 90 mph (with a window of 9.8 inches) will increase the
chances of the serve going in by 30%. It is obvious that more serves will go in
through a 9.8 inch window than a 7.5 inch window. If only 50% of the first
serves go in at 100 mph, 65% of the first serves will go in at 90 mph. This cor-
responds to about 5 EXTRA first serves going in for each set of play.

However, there is a penalty for this action. All the first serves are now at 90
mph, and the result might be a loss of more points per set than when the 100
mph serve is used. Even if the reduction in ball speed leads to a net gain in
points won, many players would be opposed to the idea, since they have pride
in how hard they can hit that first serve, even if it rarely goes in. Therefore, we
will examine another parameter.

The ball’s height, when struck by the racquet (impact height) also influences
the size of the clearance over the net. Figure 19.3 shows the result of a com-
puter calculation of the window size versus the impact height. (Note, this is the
ball’s height when struck, not how high you throw the ball on the service toss.)
The graph clearly shows that the window opens up as the impact height
196 Chapter 19

increases. All players know this fact, and

Accetance vs Serve Impact Height make use of it when they serve overhand
rather than underhand. These data also
show the severe disadvantage a short
player has when attempting to get hard
———— = serves to land in the court. If we look at
100 mph the graph for the 100 mph serve, origi-
nally hit from a height of 8 feet, and if it
is desired to continue serving at 100 mph,
to get the 30% extra serves going in will
(degrees) require that the ball impact point be 6
inches higher. Is that possible?
Window
Acceptance

A measurement of serve impact heights

was made using members of the
7 75 8 8.5 9 9.5 10 University of Pennsylvania tennis teams.
Impact Height (ft) Each player was given a racquet with a
piece of chalk taped to the head. They
were asked to stand in front of a wall and
Figure 19.3 then to mark the wall with their arm
Angular window versus impact height for 100 and 120 mph extended upward as if they were serving.
serves.
Several players were then asked to repeat
the measurement, but reaching as high as
they could (on their toes, fully extended,
etc.). The new marks on the wall were 5
to 6 inches above the older, flat-footed marks. Since many players actually
leave the ground when they serve, this 5 to 6 inch increase in impact height is
the minimum that one might expect from a good server over a flat-footed serv-
er. As a general rule, a good server impacts the ball at a height above the
ground of approximately 1.5 times that players height.

If a player is hitting the ball flat-footed and also not stretching and extending
fully upward, they may be taking as much as a 30% penalty in first service per-
centage. The first serve percentage can also be increased by hitting the ball
slightly higher on the racquet head, holding the racquet a bit lower down on
the grip, or using a longer racquet (if at the same time the impact height is
increased by that amount). Every extra inch of impact height corresponds to
about a 5% increase in the first serve percentage for the average player hitting
a 100 mph serve, so even an increase of a single inch will be significant in the
course of an entire match.

Some players may already be hitting the ball at as high an impact location as
is physically possible. What can they do now to get more serves to go in? Hit
with topspin. Figure 19.4 shows how topspin opens the acceptance window
and allows more serves in play. That is great advice for some, but for many
players, trying to consistently hit serves with topspin is more difficult than get-
ting that hard first serve to go in. In addition, because the biomechanics of the
topspin serve usually has the racquet moving upward (as well as forward) at
impact, the ball must be contacted well below the peak of the swing. Unless a
fair amount of topspin is put on the shot, some of the extra window opening
added by topspin may be lost to the lowering of impact point, and in addition,
for some players, the loss of control may negate any other advantages.
 Improving Your Serve 197

Higher Ball Toss

But there is a way of getting some topspin
Acceptance vs Topspin for 100 mph Serve
on serves and opening up the acceptance
window — almost for free. Throw the ball
up higher on the toss! If the serve is hit
perfectly flat, but the ball is falling when
struck, the resultant shot will acquire some
topspin. The amount of topspin is not
large when compared to the big kick the
top pros give to their serve, but every bit
helps. (And the more marginal the serve is (degrees)
now, the more it will help.) The faster the
ball is falling when struck, the more spin it Window
Acceptance
gets. Therefore, the higher the toss, the
more spin and the bigger the window will
be.
0 20 40 60 80 100 120 140

Figure 19.5 shows how the acceptance Topspin (rev/sec)

window is increased as topspin is added to
the serve because the toss is higher, but Figure 19.4
keeping everything else exactly the same.
Angular window of acceptance for a serve versus topspin on
(The timing probably will have to change ball for a 100 mph serve hit at 8.5 feet above the ground.
a bit to compensate for the higher toss.)
Note that a simple toss of only 6 inches
above the impact point will increase the
number of serves going in by 12% over a
toss to exactly the impact point. The next
time you see videos of Steffi Graf and her
high toss serve, think about this.
Percent of Extra Serves In
vs Ball Overthrow
For years the advice has been, toss the ball
to exactly the correct hitting height, then
crush it while it was essentially at rest.
Now the advice is to throw the ball above
the eventual contact point and hit it on its
way down. Which is better? What are the
disadvantages and advantages of each rec-
ommendation?

It is clear that it is easier to hit a stationary Extra

of
Percent
Going
Serves
In
ball than a moving ball, but on a higher
ball toss, is the ball moving sufficiently fast
so that this is a problem? Probably not,
particularly if the toss is only 6 inches to a
foot or so above the eventual impact
point. It all depends on how good is a Ball Overthrow (inches)
player’s timing. You can attempt the fol-
lowing experiment and find out.
Figure 19.5
Try hitting a forehand with the frame of
the racquet, not the strings, by turning the Increase in the acceptance window versus ball toss above
racquet 90 degrees (so the string plane is impact point for 90 mph serves hit 8 feet above ground (5’
4" server).
parallel to the ground). Throw the ball up
198 Chapter 19

a foot or so over the anticipated hitting point and see if you can hit the ball on
the way down with the edge of the frame. If you can, then your hitting accu-
racy on a falling ball is as good as the racquet is thick (+/- 0.5 inch), which
should be sufficient for most servers. It is interesting to note that this is exact-
ly how baseball players hit "fungos," and an experienced batter can hit fungo
line drives all day.

Are there any obvious disadvantages to a high ball toss on the serve? On a
windy day, the compact toss is a great benefit, particularly if the wind is gusty
as opposed to being steady. A brisk breeze can easily push a high toss by a
foot or more from its desired hitting point. It is possible to compensate for this
on the toss if the wind is steady, but swirls cannot be anticipated.

Are there any obvious disadvantageous to a toss that reaches exactly the
desired hitting height, but no higher? If the. toss is perfectly grooved and it
always goes to exactly the correct height, it is a fine technique if the addition-
al topspin is not desired. If, however, there is some variation in the height of
the service toss, trying to hit at the peak of the toss can lead to problems. If
the ball is thrown a bit too high, it can be compensated for with timing (wait
for it to come down) or can be hit when it is higher. If the ball is not thrown
up quite high enough, the impact point will be too low, which will close down
the window, reducing the chances of the serve going in, and possibly reduc-
ing the ball speed.

The next parameter to examine is the location of the server on the court. The
server can stand as far behind the baseline as desired and can move away from
the center line to as far as the sideline. The displacement toward the sideline
opens the window slightly on wide serves, and in this respect it is advanta-
geous. However, it does give the ball a longer flight path which allows the
receiver more time to react. The result is probably no net benefit unless it is
desired to pull the receiver wide, or in doubles where the server may want to
begin wide for better court coverage.

In the other direction, the answer is clear. The closer the server is to the net,
the larger the acceptance window, but it is not very essential to be within a
fraction of an inch of the baseline and run the risk of a foot fault. An inch in
this direction makes very little difference in the probability of a serve going in.
What the server can (and should) do is to toss the ball well into the court so
that contact is made when the ball is a foot or more inside the baseline. This
has many direct advantages. It will increase the number of serves that go in by
increasing the window opening. It will allow the player to get to net quicker
for the serve and volley. It will allow you to hit a serve slightly wider, pulling
your opponent off the court. It will give the opponent less time to react to the
ball.

A recent article in Tennis Magazine’ noted that Pete Sampras strikes the ball
when it is two feet inside the baseline. Figure 19.6 shows the increase in win-
dow at the net as the striking point moves inside the baseline for both a
Sampras-like serve and the serve of someone such as 5 foot 2 inch tall Amanda
Coetzer, who does not hit the ball quite as hard. Note that it is advantageous
to both of them to lean into the court when serving. In addition, the reduction
in flight path of 2 feet gives the receiver less time to react or get to the ball. A
 Improving Your Serve 199

110 mph serve, hit from 2 feet inside the base-

line has the same flight time to the opposite Vertical Window of Acceptance
baseline as a 113 mph serve hit from directly vs Distance Inside Baseline
above the server's baseline. This is almost a 3%
increase in effective serve speed, and a larger
percentage of the serves go in at the same
time. It is a win-win situation.

It is also of interest to note that the size of the

acceptance window does not change very
much between serves hit down the middle
(degrees)
(aimed at the TEE) and serves that bounce
near the corner of the service box. Acceptance
Vertical

For the player who is presently reaching up as 0 1 2 3 4 5 6

high as possible and as far into the court as Inside Baselline (ft)
possible, it is clear that a longer racquet or a
higher impact point on the racquet would
improve the serving percentage. This will only Figure 19.6
be advantageous if, at the same time, the ball Angular window of acceptance for a serve versus impact
speed of the serve is not decreased by these location distance inside the baseline that ball is hit.
changes. This may not have been the case Upper curve is at 90 miles/hour and 8 foot high impact
when the classic wood racquets of 30 years (Amanda Coetzer) and lower curve is 120 miles/hour
ago were used. Racquet technology can now and 9 foot high impact (Pete Sampras).
produce a frame that will give maximum
power (ball speed) on the serve for impacts
close to the tip and allow a longer racquet to
be used with none of the disadvantages that
an extended old wood racquet would have
had.

To move the maximum rebound ball speed

impact location higher in the head of the
newer frames, manufacturers have done a
number of things. They have made the frame
much stiffer (with a thicker cross-section of
graphite reinforced composite material), and
they have moved the center of mass up high-
er on the head (by reducing the weight in the
handle). In addition, the head shape has been
modified to make the racquet wider near the
tip and a fan-like string pattern is being used.
It is now possible to get maximum ball speed
on a serve when the impact point is very close
to the top of the head. Figure 19.7 illustrates
this. This advance in racquet technology can
provide both maximum ball speed and the
maximum window of acceptance at the same
impact location, high on the head. Figure 19.7
Computer generated power contours of the original
Making the frame longer, which allows the Prince racquet and a modern, very stiff racquet for
player to hit the ball when it is higher, will the serve.
200 Chapter 19

clearly produce a racquet that will allow more serves to go in. This will only
be true if the player takes advantage of the extra length, by hitting the ball
when it is higher. It is possible to determine exactly where on the head of the
racquet the ball is impacting the strings, by chalking the ball before the toss
and observing the chalk mark on the strings after impact.

However, maintaining (or bettering) the playing characteristics of the standard

27 inch frame with a longer racquet was a challenge for the racquet manufac-
turers. If they were to just scale up the standard, neutral balance (center of
mass at geometrical center) racquets by another few inches, or add two inch-
es to the butt of an existing racquet, the result would be a poor serving frame.
The impact point near the tip would be farther from the balance point and the
speed at which the racquet could be swung might be lower because of the
increase in swing weight (moment of inertia about the butt end). Both of these
effects would contribute to slower serves. Many of the new, longer racquets
have been made much lighter, their balance point moved toward the tip, and
their swing weights have been kept similar to the older 27 inch frames. This
allows the new frames to be swung with the same angular velocity as the old
frames, but the longer lever arm can result in higher ball speeds.

Table 19.1 gives an estimate of how much the vertical acceptance angle (win-
dow) has been increased by each of the suggestions made in this article. These
results are based on a 100 mph serve hit initially at a height of 8 feet above the
ground by a six-foot tall server, hitting flat-footed, and using a standard (27
inch) frame with an impact point at the center of the racquet head.

Table 19.1

Increase in serve window for various changes

2 inch longer racquet 10%

Hitting 2 inches higher on head 10%

Stretching up 6 inches 30%

Spin on ball due to higher toss 20%

Reaching into court 10%

Jumping up 4 inches 20%

References
1.) The Sampras Serve, Tennis, pg 45 February 1994
2.) How Would a Physicist Design a Tennis Racquet, H. Brody
Physics Today pg 26-31 March 1995
 Theory, Motion,
and Strategy

Gred Rusedski, Great Britain

By Rod Cross and Howard Brody

Serve Speed vs Racquet Speed __ Bs:

n order to serve a ball at 100 mph, a player must swing the racquet at relatively high speed. Most people
would probably guess that the racquet needs to be swung at about 60 mph or so depending on the power
of the racquet. In fact, the tip of the racquet needs to be swung at about 100 mph, give or take a few mph
depending on the weight and the weight distribution of the racquet. There is an important and surprising les-
son in this result. That is, the power of the racquet and the power of the strings has only a small effect on serve
speed. The serve speed is always about the same as the tip speed regardless of the racquet or the strings. A dif-
ferent racquet or a different string will change the serve speed by a few mph but almost all the power comes
directly from the player's arm.

Figure 20.1 shows a satellite tournament player serving at 162 km/hr (100.7 mph). The action was captured on
a digital video camera at 100 frames/sec or one frame every 10 ms (0.01 s). Digital video camera images can be
transferred in real time to a computer for analysis, usually at 25 or 30 frames/sec. This is not fast enough to
record a high speed serve and to measure everything of interest. The camera used to capture the images in
Figure 20.1 was a JVC 9600 which records at 100 frames/sec for smooth slow motion playback. Not many video
cameras have this feature. A shutter speed of 1/500 sec or smaller is necessary to avoid blurry images.
202 Chapter 20

-380ms_ (Peakof ball toss) ©

162 km/hr
ea

tO"
Oe ae
70 60 ms

(m)
Height

-100 ms

220 ms

Distance from baseline (m)

Ball speed= 45 m/s = 162 km/hr= 100 mph

Angular velocity of racquet at impact= 80 rad/s
Speed of tip atimpact = 43 m/s
Speed of impact point (10 cm from tip) = 35 m/s
Hence ACOR = 0.29 at impact point

Figure 20.2

The picture quality in Figure 20.1 is not brilliant so a white

dot has been added to show the tip of the racquet. The
positions of the ball, the racquet tip and the player's hand
are shown in Figure 20.2 at various times before and after
impact. The racquet impacts the ball at time zero. The ver-
tical and horizontal distances are measured from the center
mark on the baseline. The distance scales on the video
image were calibrated from the height of the player (6' 2").
This was checked against the drop height and speed of the
Figure 20.1 ball during the ball toss, and both methods gave the same
calibration. The impact location and time was determined
accurately from the intersection of the ball trajectories
before and after impact.

The racquet was almost vertical at impact, almost horizontal 20 ms before impact
and was at about 45 degrees to the horizontal 10 ms before impact. The racquet
therefore rotated by about 45 degrees each 10 ms prior to impact, at an average
rotation speed of 4,500 degrees/sec = 80 radians/sec (1 radian = 57 degrees). The
tip and the hand were both moving horizontally at impact, at speeds of 43 m/s
and 12 m/s respectively, and the ball came off the strings at 45 m/s. The ball did
not impact at the tip of the racquet. It impacted about 10 cm from the tip. The rac-
quet speed at the impact point was 35 m/s (78 mph).

Formu
Serve Speed
ineispeed Vp at any point on the racquet can be calculated if one knows the rota-
tion speed (in radians/sec) and the speed at the tip (V,). The formula is Vp =
V, - Dw where D is the distance from the tip. In this case, Vp = 43 - 0.1 x 80 = 35
m/s. The racquet slowed down immediately after impact. In Figure 20.1, the rac-
 Serve Speed Strategy 203

quet rotated by about 45 degrees each 10 ms before impact but it rotated only
about 10 degrees each 10 ms immediately after impact. This means that energy
was taken out of the racquet and given to the ball.

The serve speed V is given by V = (1 + e,)V, where e, is the apparent coeffi-

cient of restitution. e, is typically about 0.4 in the middle of the strings and about
0.1 near the tip. In this case, we find that 1 + e, = V/V, = 45/35 = 1.29 so e, =
0.29. The best position to serve the ball is about 10 cm from the tip since the rac-
quet travels faster there than at the center of the strings and since the server gains
a height advantage by serving closer to the tip. eg is larger at the center of the
strings but V,, is lower, with the result that the serve speed varies only slightly
(about 5%) with impact point. The impact point at which the serve speed is a max-
imum depends on the distribution of racquet weight but is typically about 10 cm
from the tip.

The time spent by the ball on the strings is only about 5 ms. The impact duration
can be shortened from say 5 ms to 4.5 ms by increasing the string tension or by
an increase in ball stiffness. The impact time is so short that one might think that
a difference of 0.5 ms is not very significant. However, Figure 20.1 shows that the
racquet travels a relatively long distance and rotates through a large angle during
the last 20 ms. The racquet slows down during the impact but it will still travel a
respectable distance during the impact. Changing the impact duration by 0.5 ms
can therefore make a significant difference to the angle of the ball off the strings.
To serve accurately at a different string tension, a player would need to need to
change the actual impact time or change the ball toss slightly.

What if the player hits the ball 0.5 ms too early or 0.5 ms too late? Will the ball
come off the strings at the wrong angle? If the racquet rotates at 80 rad/s then in
0.5 ms it will rotate through an angle of 80 x 0.0005 x 57 = 2.3 degrees. This is
more than enough to change a good serve into a fault. In fact, the whole range
of good serve angles is only about 2 or 3 degrees at 160 km/hr. If a ball served 5
degrees down from the horizontal lands on the service line then at 7 degrees the
ball will only just clear the net. This seems to indicate that a player must hit the
ball at exactly the right time, with less than 0.5 ms room for error. The problem
is, no one can judge time that accurately. For example, if you try to hit an object
at exactly the same time with your left hand and your right hand, then measure-
ments show that you would be lucky to hit within 5 ms.

Fortunately, the serve problem is not one of timing but one of positioning. If you
toss the ball in the right spot, then the racquet will hit the ball at that spot almost
regardless of the exact timing. The ball in Figure 20.2 dropped from a maximum
height of 3.6 m and it was dropping at 3.7 m/s when it was hit. If it was hit say
1 ms later, then it would have dropped an extra 3.7 mm. The angle at which the
racquet strikes the ball depends mainly on the ball toss position rather than the
exact timing. If the ball is hit 1 ms later than intended, the ball will be hit a few
mm lower on the strings, but the angle will be the same if the ball toss and arm
swing are the same.

During the last 20 ms prior to impact, the hand and the racquet both rotate at 80
rad/s. It might appear that this is all due to deliberate wrist rotation on the part of
the player. However, every joint is involved, including the elbow, shoulder, hips,
204 Chapter 20

knees and ankles. Each joint adds rotation speed to the previous one. For
example, if the only body segment that rotated was the hand (about an axis
through the wrist) then the rotation speed of the racquet would be only about
10 radians/sec. If the forearm is also rotating at say 40 radians/sec about an
axis running along the upper arm then the combined rotation speed would be
40 + 10 = 50 radians/sec. The upper arm rotates about an axis through the
shoulder, adding further rotation speed to the racquet.

Detailed studies of body segment rotation in tennis are described by Elliot et

al. (1995) and by Bahamonde (2000). An interesting feature is that the
sequence of rotations starts from the bottom and works its way to the top. For
example, most of the hip and shoulder rotation occurs from -380 ms to -60 ms,
most of the elbow rotation occurs from -60 ms to - 20 ms and most of the wrist
rotation occurs from -20 ms to impact time. These features are all clearly evi-
dent in Figure 20.1. Similar results are found in other sports. For example in
the javelin throw, the hip, shoulder, elbow and javelin reach their peak veloc-
ities in that order and all within about 100 ms of each other. This is described
in biomechanics books as the kinetic link principle. A similar sequence of
events is observed if one swings two or more rods joined end to end by pivot
joints. If the rods are swung correctly, the last rod in the chain will rotate last
and it will rotate the fastest.

ve Speed Strategy _
Many tennis players have the ability to hit their first serve hard, very hard, or
really smash the ball — hitting as hard as they can. There is a small problem
with hitting as hard as you can — the ball usually does not go in as often as
you would like. This poor first serve percentage is due to two separate effects.
As you swing harder you may lose some control over your racquet, and the
higher the ball speed, the smaller is the "acceptance window" for a ball going
in. However, when your really hard serve does goes in, you are more likely to
win the point (Table 20.1).

What strategy should you adopt? Smash the ball as hard as you can? Play it safe
and get your serve in as often as is possible? Hit the ball hard, but not as hard
as you can? The answer is not obvious, and depends on a number of factors.

Let us assume that the winning percentage on your

first serve depends on how hard you serve and that
Table 20.1
the percent of serves going in also depends on how
hard you serve. Let us take some typical examples.
Percent first Percent win Total win percentage
serves in if goes in on first serves ; ;
On this basis, you might advocate the softer serve
50 90 45.00 (which goes in more often) since it would have the
55 85 46.75 highest winning percentage of points on the first
60 80 48.00 serve. This would be absolutely the wrong strategy.
65 US 48.75 Under these assumptions, you should hit your first
serve as hard as you can and you will win more total
points, because you also get a second serve. If you get
in 95% of your second serves and you win 55% of the points when that sec-
ond serve goes in, then (100 - first serve in) x .95 x 0.55 must be added to the
first serve numbers. The result of this is shown in Table 20.2.

In this case, it pays to hit the first serve hard, as you will win a larger per-
centage of the points. These numbers may not hold for every player (if you
 Serve Speed Strategy 205

never get your first serve in Table 20.2

when you really smash it, or if
your second serve is so weak Percent first Percent win Win percentage on
that you never win a point on _—-serves in if goes in First serve Second serve Total
it, then the other answer is 2 aC eG 26.1 711
obvious), but d th ; 3 :
assum oeae Mipe
an te Aa ae oe 60
ae 80
= aos
48.00 oe
20.9 ee
68.9
20.1 and 20.2, hitting the first 65 75 48.75 18.3 67.1
serve really hard is the best
strategy. Table 20.3

This range of winning percent- Percent of first Winning percent Percent of Chances of
ages seems very narrow (67.1 serves in when serving games won winning 6 games
to 71.1), but it has some effect 50 714 915 58.9
on the Soa e ae a 55 70.3 90.4 547
service game and a ae Dig 60 68.9 88.7 48.7
effect on the chances of not 65 67.1 86.2 41.1
being broken during an entire
set, as is shown in Table 20.3.

Case Study #1: Second Serve Speed By Howard Brody

he final point of the Wimbledon Championship Match between Sampras and Agassi ended on a “second” serve ace. You
| are seeing more and more second serve aces, as well as serve speeds in the 110 to 120 mph range for the second deliv-
ery. Is this foolishness and bravado by the players or does it make real sense and win more points?

Let us examine Pete’s serving statistics in the Championship Match. Sampras got in 60% of his first serves and he won 87.5%
of the points when his first serve went in. Multiplying these two numbers together (60% x 87.5% = 52.5%) gives you the prob-
ability of Sampras winning a point if the rules of tennis allowed only one serve and Pete hit his "first" serve every time.

On his second serve, Sampras won 18 of 37 points for a 48.6% winning rate. Had Pete used his "first" serve instead of his "sec-
ond" serve after a single fault, he would have won MORE points (52.5% versus 48.6%). That’s why Sampras will occasionally
go for the ace or the big one after missing the first serve, rather than use his conservative serve, particularly against an excel-
lent returner like Agassi. The odds are actually in his favor to play aggressively rather than conservatively.

There are people who advocate changing the rules of tennis and allowing only one serve per point as a way of slowing down
the game and reducing the dominance of the serve. If that rule was in effect at the Wimbledon Championship Match, Sampras
might have served his big one every time rather than using his second serve. He would have ended up winning more points,
which would increase rather than decrease the importance of the serve. In addition, he would miss his first serve (equivalent
to a double fault) about 50 times in order to win the 67 service game points that he needs to win the match. If 50 of the 127
points served by Sampras ended in a fault, interest in tennis might deteriorate quickly.

There are other matches that might be examined to see how the strategy of using only first serves would work.
If Rafter had used his "big" serve for his second serve as well as his first serve, he would have won the point on 29 rather than
22 second serves. This is an additional 7 points.

Rafter - Agassi Australian 2001, semifinal. Sampras - Martin Australian 2001, round of 16

Rafter, first serve average speed 161 km/hour Sampras, first serve average speed 170 km/hour
96 in/153 points = 63% 76 in/123 points = 62%
79 won/96 in = 82% 63 won/76 in = 83%
points won 79/153 = 51.6% points won 63/123 = 51.2%

Rafter second serve average speed 147 km/hour Sampras second serve average speed 158 km/hour
points won 22/57 = 38.6% points won 18/47 = 38.3%

If Sampras had used his "big" serve for his second serve as well as his first serve, he would have won the point on 24 rather
than 18 second serves. This is an additional 6 points.
206 Chapter 20

Case Study 2: Serve Speed By Rod Cross

\ ll tennis players serve their first serve faster than their second serve. A fast serve is less likely to go in, but if it does go
in it is more difficult to return. An interesting and surprising consequence is that most players win a greater fraction of
their second serves than their first serves, because a larger fraction of first serves are faults. At least, this is the case
at all four Grand Slam events where statistics are readily available on the web. In that case, would it not be better to serve a
slow first serve and a fast second serve? And would it be even better to serve both serves slowly?

We can answer these questions by looking at a few examples. Suppose that a player serves 100 first serves during a match,
and suppose 50 are faults. The player therefore needs 50 second serves and may serve 5 double faults. The first serve % is
50% (half the first serves went in) and the second serve percent is 90%. Now suppose that the player wins 35 of the 50 good
first serves (winning percentage on first serve = 100 x 35/50 = 70%) and 25 of the 45 good second serves (winning percent-
age on second serve = 100 x 25/45 = 55.5%). These figures are typical of any match and are not unusual. Out of the100 points
played, the server won 35 points on the first serve and 25 points on the second serve. The chance of winning a point on serve
is therefore 35 + 25 = 60%, meaning that the server is more likely to win the game than the receiver.

FAST FIRST SERVE SLOW SECOND SERVE TOTAL

Total 1st Total 2nd Total Total

Serves In % In Win % Won Serves In % In Win % Won Won %

100 50 50 70 35 50 90 45 55:5 25 60 60%

In the above example, the server won 35 first serves out of 100, and 25 second serves out of 50, so the chance of winning a
second serve is actually greater than the chance of winning a first serve. So, lets suppose the player reverses the serve order
and serves a slow first serve first and a fast second serve. This time, the player will serve 90 good first serves and 10 first serve
faults and the first serve percentage increases to 90%. Using the same averages as the first example, the player will win 50
points on the first serve. Out of the 10 second serves, the player will serve 5 faults and will win 3 or 4 good second serves (since
70% of 5 is 3.5). Out of the 100 points played, the server wins 50 points on the first serve and, at most, 4 points on the second
serve. The chance of winning a point on serve will therefore drop to 50 + 4 = 54%.

SLOW FIRST SERVE FAST SECOND SERVE TOTAL

Total 1st Total 2nd Total Total

Serves In % In Win % Won Serves In % In Win % Won Won %

100 90 90 55.5 50 10 50 5 70 4 54 54%

Serving a slow first serve doesn't help since the chance of winning a good first serve is too low, especially when most of the
first serves are going in. It would actually be better to serve a slow second serve if the first serve is slow. Out of the 10 second
serves, the player would then win 5 instead of 3 or 4. The chance of winning a point on serve, using two slow serves, is then
50 + 5 = 55%.

SLOW FIRST SERVE SLOW SECOND SERVE TOTAL

Totalist Total 2nd Total Total

Serves In % In Win % Won Serves In % In Win % Won Won %

100 90 90 55.5 50 10 90 9 55.5 iS) 55 55%

 Serve Speed Strategy 207

How about serving two fast serves instead of a fast and a slow serve? In that case, the player would win 35 points out of 100
on the first serve, and would need 50 second serves as before. Applying the same ratio, the player would then win 17 or 18 of
the 50 second serves. The chance of winning a point on serve would drop to 35 + 18 = 53%, which is even worse than serving
a slow first serve.

FAST FIRST SERVE FAST SECOND SERVE TOTAL

Total 1st Total 2nd Total Total

Serves In % In Win % Won Serves In % In Win % Won Won %

100 50 50 70 35 50 50 25 70 18 50 53%

Can one increase the chance of winning a first serve by serving a bit slower or a bit faster? The problem here is that if you
serve slower you will get more serves in, but your opponent is more likely to win the point. If you serve faster, you are more
likely to serve a fault. It seems that there is nothing you can do to improve your chances apart from practicing more to increase
your first serve percentage. This is in fact borne out by Grand Slam statistics. Players with a high average first serve speed gen-
erally have a lower first serve percentage than players with a low first serve speed. As a result, slow servers and fast servers
win essentially the same fraction of points on their first serve when averaged over all players. Put another way, the players who
win the most matches are not necessarily the fastest servers. The players who do win the most matches not only serve rela-
tively fast, but also have a high first serve percentage and an ability to hit a good return when a fast ball is served by their oppo-
nent.

Case studies 1 and 2 seem to contradict each other. In case study 1, the play-
ers won a bigger fraction of their first serves (51%) than their second serves
(38%) and would therefore have achieved a better result by serving two: fast
serves. In case study 2, the player won a bigger fraction of his second serves
(50% counting all second serves) than his first serves (35%), in which case the
best result is achieved with a fast first serve and a slow second serve. In case
2, serving two fast serves gives the worst possible result.

References

B.C. Elliott, R.N. Marshall and G.J. Noffal (14995), Contributions of upper limb
segment rotations during the power serve in tennis, Journal of Applied
Biomechanics, 11, 433-442

R.E. Bahamonde (2000), Changes in angular momentum during the tennis

serve, Journal of Sports Sciences 18, 579-592
 Where Should
You Stand?

EES Mili

By Howard Brody

here should you stand when you receive serve? On the baseline? Inside the baseline? Well behind
the baseline? It depends on the type of serve you are returning, whether you are playing singles or
» ™ doubles, whether the server can swing the serve wide, and whether the server is coming to the net
behind the shot.

I have calculated the time you have to respond to a flat, down-the-middle serve as a function of where you posi-
tion yourself on the court (Table 21.1).

Assume your opponent hits a serve down the middle that leaves the racquet at 110 mph. If you stand five feet
behind the baseline to receive that serve, you will have as much time as if you were on the baseline and your
opponent is only hitting 102-mph serves. Moving five feet behind the baseline gives you 10% more time to react
and get to the ball.

If you move in so that you are five feet inside the baseline, you have effectively transformed your opponent's
110-mph serve into a 120-mph blast, when you consider the time that you have to respond.

Since it is 78 feet from baseline to baseline, why does five feet make such a large difference (10%) in the time
you have to return the ball?
210 Chapter 21

Table 21.1

Serve Position Relative Time

Speed to Baseline (sec)

120 mph 0 0.59

110 mph -5 ft (inside) 0.59
110 mph 0 0.65
110 mph +5 ft (behind) 0.71
102 mph 0 0.71

The reason is that the ball slows down quite a bit due to air resistance and the
bounce. By the time a 110-mph serve gets.to you at the other baseline, it has
slowed down to about 55 mph; therefore, each foot added or subtracted at the
slow end of the trajectory gives you proportionally more (or less) time than the
actual distance itself might lead you to anticipate.

Based on this information, should you then routinely stand five feet behind the
baseline when returning serve? If your opponent is booming in big, flat serves
down the middle, and is not coming to the net behind them, it seems like a
good idea. If, on the other hand, your opponent often slices one wide or kicks
up high, then you may not be happy with your deep position.

To return the top spin serve from very deep, you may almost have to hit an
overhead, and a wide slice may pull you into the next court. If your opponent
is charging the net, your return from well behind the baseline gives the server
more time to get to your return and hit a very good volley.

At the other extreme, if your opponent is plopping in serves at 70 mph, then

where you choose to stand depends on what you are trying to do. If your prin-
cipal concern is to get the serve back, then you clearly want to stay at or
behind the baseline. If, on the other hand, you want to be aggressive and
attack the serve, you should move in and be well inside of the baseline.

In doubles, there always is a player facing you at the net when you return
serve, and often the server is moving up as well. How does that change your
choice of position on the return?

Again, assuming it's a 110-mph serve, the speed at which your return leaves
your racquet depends on the incoming ball speed (55 mph by the time it gets
to you) and your racquet head speed. If, instead of taking a full swing, you just
punch or chip at the ball with a racquet head speed of about 30 to 55 mph,
the return ball speed off of your racquet will be about 70 mph.

If the opponent is crowding the net (less than six feet behind it), then you have
an interesting option. When you move in and hit your return from five feet
inside the baseline right at that opponent, you might handcuff the player at the
net. The time that your opponent at the net has to return your shot depends
on how hard you hit the ball and where you are standing when you hit it.

Assuming your return leaves your racquet at 70 mph, the result will be the fol-
lowing times for the player at the net (Table 21.2).
 Receiving Serve 211

Table 21.2

5 ft inside On the 5 ft behind

the baseline baseline the baseline

0.42 sec. 0.49 sec. 0.56 sec.

Some of the best professional athletes can just barely react (catch a ball, block
a puck, etc.) in about 0.35 seconds, so if you step inside the baseline and
punch your return at that opponent who is crowding the net, you will proba-
bly win the point (and force that player to retreat on future shots).

Another way of looking at these numbers is to realize that the time you have
to react to a 150-mph serve (baseline to baseline) is 0.49 seconds. If the per-
son at the net tries to return a shot, but has only 0.42 seconds of time, the odds
are that you will win the point. Just think about how you would react to a 150-
mph serve!
 Factors in Hitting
and Receiving
Spin Shots

Martina Hingis, Switzerland

By Howard Brody

Topspin vs Backspin

ver the years, we have been told by many tennis experts that topspin provides a much better shot than
|a flat shot or one with backspin on it. Then why is it many players still like to hit shots with backspin,
— =suusing a chopping motion for the stroke, or hit flat shots with very little spin on the ball? Why is it that
most tennis coaches teach the flat stoke to the beginner rather than start them with topspin?

When a computer is used to calculate ball trajectories, it is very clear that topspin shots have a much higher
probability of clearing the net and landing in the court than do flat shots or shots with underspin. In fact, it has
been shown by this author (in the book Tennis Science for Tennis Players), that for a ball leaving your racquet
at 60 miles per hour, the window or vertical acceptance angle (the variation in exactly what angle the ball can
leave your racquet and still be a good shot) was two and a half times larger for a topspin shot compared to a
ball hit with backspin. This means that if all else is the same, you are two and a half times more likely to make
an error with the ball going into the net, or long, if you chop at the ball rather than hit it with topspin. Then
why do players still slice at the ball instead of hitting with topspin?

The answer is that everything else is not the same. The variation or error in the vertical angle of the racquet
(hence the ball leaving your racquet) is different when you attempt to hit a flat shot or one with backspin on it
214 Chapter 22

Figure 22.1 as compared to attempting to hit with topspin. The

Friction Force on a Bouncing reasons for this are due to physics and the mechan-
Ball. The friction force to the left ics of stroke production.
will cause a torque on the ball
giving it a clockwise spin, which Since ground strokes (by definition) take place
is topspin for a ball moving to after the ball has bounced, the spin the ball
the right. acquires due to the bounce is important. After the
bounce, the ball is usually coming at you with con-
siderable topspin even, if the shot was hit as a
Friction Force On Bouncing Ball normal, flat drive with no initial spin. This is
because of the friction force on the ball as it
bounces. This friction force produces a torque on
the ball that causes it to hav topspin (Figure
22h),

7™ Direction of
Ball Motion
If you want to return that shot with considerable
topspin of your own, not only must you turn the
Friction
direction of the ball’s velocity around Cit is coming
Force toward you and after you hit it, it is moving away
from you), but you also must reverse its spin direc-
tion by applying a large angular impulse. (A top-
spin ball coming toward you has its spin in the
Surface
opposite direction than a topspin ball moving away
from you.)

If you desire to return the shot with underspin, you

only have to reverse the ball’s direction because
the ball is already spinning in the way that you
Ball Direction and Spin
Figure 22.2 want it to. (The topspin on a ball coming toward
you has the same absolute spin direction as the
Ball Direction and Spin
Spin. (a) and (c)
Diréction backspin on a ball moving away from you.) Figure
are balls moving to
Topspin 22.2 shows the direction of the spin for balls mov-
the right with top- ing to the right and to the left. If you assume the
spin. (b) and (d) are ball moving to the right is coming toward you after
balls moving to the the bounce, then it is clear that the ball’s actual
left with backspin spin direction must be reversed in order for it to
and topspin respec- Spin
Direction have topspin after you hit it.
tively. Note that the ~ ™
actual spin of the
This has been run through a computer program
ball must be
reversed to get top-
that was written to determine ball velocities and
spin on a ball mov- spins for various hitting speeds and racquet angles.
ing to the left. Under these conditions (a groundstroke return), to
Spin
Direction get comparable ball speeds on your return, you
x ™ Topspin
must swing the racquet much harder to produce a
topspin shot than to get underspin or hit the ball
flat. At first glance, you would expect the racquet
speed to be the same, but the fact that the ball
coming toward you has topspin (due to its bounce)
Spin
Direction breaks the symmetry and means you must swing
considerably harder to get your return to have top-
spin.

Assume the ball is coming toward you at 60

feet/sec (41 mph), and you attempt to return it
 Strokes and Spin 215

with underspin and a racquet head speed of 45 feet/sec (31 mph). With that
incoming ball speed, the ball will leave your racquet with a speed of about 90
feet/sec (60 miles/hour). If you want your shot to have topspin, to get the same
outgoing ball speed, it will require a racquet head speed of 70 feet/sec (48
mph). You really have to swing much harder to get the same speed on your
topspin shot as on your backspin shot, and this probably will cause you to
spray the ball around quite a bit. It is not easy to swing a racquet at 70 feet/sec
and maintain the degree of control that you would get when you only have to
move the racquet head at 45 feet/sec. A top player can do this, but a beginner
or intermediate player cannot. To hit the ball harder requires greater racquet
head speed, hence better preparation. To slice at the ball requires little racquet
head speed, therefore little racquet preparation. It is a great stroke for an emer-
gency. To a beginner, every shot is an emergency.

Topspin shots, because of their trajectory in the air, give you a larger margin
of error in your stroke, allowing you to miss-hit the ball Gn angle) a bit.
However, the argument presented in this note shows that the inherent errors
associated with production of topspin shots may be considerably larger than
those associated with backspin or flat shots because you must swing the rac-
quet so much harder. If your timing is superb and your strokes really well
grooved, topspin may be fine. For the elite, high performance player, topspin
is the stroke of choice. For the beginner and the occasional player, a flat shot
or underspin probably will lead to better depth control, less errors, and fewer
points lost.

Most players adjust their strokes for the speed of the incoming ball. If their
opponent hits a really hard shot, they know that they can return a ball with
good pace without swinging their racquet very hard. When their opponent is
hitting softly, most players realize that they must swing out if they want their
shots to have good pace. Not only is it difficult to swing out when your oppo-
nent is pounding the ball, but if you successfully connect, the ball may come
off your racquet like a rocket. In such a case, it is unlikely to stay in the court.
An experienced player usually makes these adjustments without thinking about
iC.

What many players do not do, is adjust their swing and strokes to compensate
for the spin their opponent puts on the ball. Just as you can often anticipate
the speed of your opponent’s shot from the racquet speed you observe, you
should watch your opponent’s racquet trajectory Cow to high, high to low, or
level) and adjust your swing and court position to compensate for it.

When your opponent hits a flat shot (not much spin on the ball as it leaves the
racquet), your eyes track the ball’s initial trajectory and your experience at ten-
nis allows you to anticipate the bounce location. If you have been playing ten-
nis for years, you know just where to move so that the ball will be at a com-
fortable hitting position for you Gn baseball talk, in your “strike zone”). When
you see your opponent hit with a very pronounced low to high motion of the
racquet, you can anticipate a topspin shot that will dip in the air, bounce short-
er, rebound higher, and may not slow down as much when it bounces. Your
tennis experience with flat shots is such, that when a topspin shot does not
slow down very much on the bounce, you may (incorrectly) perceive that it
has gained speed. When you see your opponent slice down at the ball (or
216 Chapter 22

come under it), you know that the ball will have a flatter trajectory, bounce a
little deeper in the court and then, if it was hit with good pace, it may skid or
slide, staying low. If it was hit without great pace, the sliced shot may float a
bit and then "sit-up" when it bounces.

However, the problem most players are not aware of concerns the spin of the
ball AFTER the bounce. For a typical groundstroke that leaves the racquet at
60 miles an hour, the spin after bouncing depends on the initial spin given to
the ball by the opponent’s racquet. A ball with a backspin of 30 revolutions
per second will skid off of the ground after the bounce and end up with essen-
tially no spin. A shot hit flat (no initial spin) will have a forward or topspin of
40 revolutions per second after bouncing. A 60 mph shot hit with a topspin of
30 revolutions per second will end up with a forward or topspin of 70 revolu-
tions per second after the bounce. This is all on a typical mid-pace tennis court.
On a slow clay or fast grass, the spins after. the bounce may be different.

Topspin (after the bounce) will interact with your racquet strings and tend to
make the ball come off of the racquet at an increased angle. As an example, if
you swing your racquet perfectly level and have the face perpendicular to the
ground (neither open or closed), when you hit a spinless ball that is moving
horizontally, the ball will rebound perfectly horizontally also. If you try to
return your opponent’s flat shot (having 40 revolutions per second topspin
after bouncing) with the same flat swing, the ball will leave your racquet at
about 7 degrees above horizontal. This is because the spin on the ball will bite
into your strings and give the ball some upward velocity. It turns out that the
7 degree "lift" the ball acquires from this interaction with the strings is just the
correct amount for the ball to clear the net and land in the court.

If you attempt to return your opponent's slice (very little spin after the bounce)
with exactly the same stroke and racquet orientation as you did for your oppo-
nent’s flat shot, the ball will leave your racquet moving horizontally and it will
end up in the net. For your shot to clear the net, you must give the ball extra
lift and have it leave your racquet with some upward angle. You don’t want to
slice down at the ball unless you also give the ball a great deal of lift. Be very
careful if you want to try to return a slice with a slice, or the ball will end up
in the net.

If you take the same flat swing at a shot that your opponent has put topspin
on, the 70 revolution per second topspin the ball has after the bounce will dig
into your strings and make your return fly up into the air. Here is a good
Opportunity to slice down at the ball.

REMEMBER. When your opponent hits with spin, not only does it change the
ball trajectory while in the air (which you will observe) and the ball speed and
angle as it bounces (which you can usually anticipate), but it also will change
the way the ball interacts with your racquet. So watch your opponent’s swing
pattern carefully and adjust your strokes and racquet head angle to counteract
the after-the-bounce spins.

A shot hit at 60 mph with top, back, and no spin is returned perfectly flat. The
vertical angle of the return is given in Table 22.1.

If all three types of shots are moving near 30 mph as they cross the baseline
and all three types of shots are returned exactly the same way, only the return
of the flat shot will land in the court.
 Strokes and Spin 217

Initial spin Spin after Ball angle (degrees) Table 22.1

(rps) bounce (rps) for flat hit
A shot hit at 60 mph with top,
back, and no spin is returned
top 30 top 70 14 perfectly flat. The vertical angle
of the return is given in the last
none top 40 8 column.

back 30 none 0

If you decide to chop or slice back your return (instead of hitting it flat), you
will get the results in Table 22.2.

Initial spin Spin after Ball angle (degrees) Table 22.2

(rps) bounce (rps) due to your chop
Results if you decide to chop or
slice back your return (instead
top 30 top 70 11.0 of hitting it flat).

none top 40 ZO)

back 30 none -6.3

If you return your opponent’s topspin or backspin shot with EXACTLY the
same swing as you would return a flat shot, the ball will end in up the net or
go long. It is clear that you must adjust your swing or racquet head angle to
compensate for the spin your opponent puts on the ball.

Over the years, you may have heard that it is difficult to return a sliced ground
stroke with a slice. That it is much wiser to hit flat or even use some topspin
when your opponent has sliced (put backspin on) a shot to you. Is this true —
and if it is, what is the reason?

When a flat or topspin shot bounces, it acquires topspin due to the court fric-
tion acting on the ball. A moderate drive of 60 mph, with little or no spin, will
gain about 40 revolutions/sec of topspin after the ball bounces. If a ground
stroke is hit at 60 mph with a moderate topspin of 30 revolutions/sec, after
bouncing it will have acquired 65 revolutions/sec of topspin. These shots will
be relatively easy to return with backspin (sliced), since they are spinning in
the correct way already (see diagram). All your racquet has to do is turn the
ball's direction around — the spin will take care of itself.

When a 60 mph drive with moderate backspin bounces, it ends up with essen-
tially NO spin and stays low after the bounce. If you try to slice back such a
shot, not only must you turn the ball's direction around, but you must also sup-
ply the energy needed to get the ball to spin. If you want to slice back a slice,
218 Chapter 22

you must hit the ball harder (swing your racquet with greater speed). In addition,
since sliced shots bounce lower, you will have to lift your return higher to clear the
net. To make matters more difficult, the spin on flat and topspin shots after the
bounce tend to bite into your strings and lift the ball to a larger angle. Since sliced
shots have little or no spin after the bounce, to get your return to go over the net,
you must lift it even more.

If you want to return your opponent's sliced backhand with a slice of your own,
you must hit the ball harder and lift it higher over the net than you would if you
were returning a flat shot by your opponent.

The books, the articles, and the TV announcers all tell you to hit your approach
shot deep when you want to come to the net. ‘Some add the comment that you
should also keep the ball low. They go so far as to recommend a slice approach
shot rather than topspin to keep the ball as low as possible. Is this advice correct?

Keeping the ball low has two immediate and obvious advantages. A low shot, such
as a slice, because of its trajectory, gives your opponent slightly less time to get to
the ball compared to a higher topspin shot that bounces in the same location. The
more you can hurry your opponent, the less likely that you will be passed or hand-
cuffed by a strong return. This timing effect is actually of less importance than the
second reason for keeping the approach shot low.

The higher the ball is when your opponent hits it, the better are the chances that
the return will go in. Computer analysis shows that your opponent can hit a high
bouncing ball much harder than a low skimming ball, and have the same chance
of the shot going in. When a ball is hit while it is quite low, there is much less mar-
gin for error in the shot, particularly if the ball is being hit hard. You have more
than twice the chance of getting a 70 mph groundstroke to go in if you hit it when
it is four feet off the ground as opposed to 1 foot off the ground.

The chances of you being passed or handcuffed when you are at net depend on
how hard your opponent can hit the ball and how fast the ball gets to you. A high-
bouncing approach will allow your opponent to tee off on a shot and still have a
good chance of the ball going in. You will have trouble returning this type of shot.
If your opponent has to hit the ball when it is quite low, a hard hit will have a
small probability of both clearing the net and bouncing inside the baseline. This
means your opponent will either have to ease up on the shot or attempt a risky,
poor percentage shot. Therefore, keeping your approach shot low increases your
chances of winning the point.

But what about depth? This is a matter of ball speed and reflexes. The baseline is
39 feet from the net, and probably between 45-50 feet from you when you are at
the net. This means that there is approximately half a second between the time a
hard shot is hit and when it gets to you at net. Is this enough time to react to a
ball, determine where it is going, move the racquet, and volley the ball into the
court where you want it to go?

Yes, if you have reasonable reflexes. Plagenhoff has studied high-speed films of
tennis players, hockey goalies, etc., and he has discovered that if there is not at
least 0.37 seconds between the hit and the block, the ball (or puck) will not be
blocked unless you are lucky or guess right (it's called anticipation). This means
that if you allow your opponent to hit the ball hard from within the baseline, it will
 Strokes and Spin 219

be marginal as to whether you will be able to react and get to it, even if it is
within your reach.

By forcing your opponent back behind the baseline, you have bought yourself
time to react to the return and pounce on the ball. Therefore, it will take a
much better shot to pass you or handcuff you if you force your opponent to
hit the ball from behind the baseline.

This leads to an interesting question. You are generally advised to close in tight
to the net. However, if you find yourself at net and have hit a short volley,
should you retreat a little to buy yourself some time for your return? The analy-
sis above says “Yes.” If your opponent can get to it with some ease. This is
illustrated by Table 22.3 of time versus positions on the court. This table shows
the amount of time you have for well-hit (70 mph) shots hit from 10 feet
behind the baseline, on the baseline, and 10 feet inside the baseline for vari-
Ous positions you can have at net. Since very few of us have the 0.37 second
reflexes of a hockey goalie, a retreat is clearly recommended.

Your Position Hit From 10' Hit From On Hit From 10' Table 22.3
Behind Net Behind Baseline Baseline Inside Baseline
Time you have to return a
(feet) (seconds) (seconds) (seconds)
shot when you are at net
(for a ball speed leaving
2 56 44 32
opponent's racquet at 70
7 .63 50 .38
mph).
12 .70 56 44
le BG 562 .50
 =
=
- o
OC

Hitting and
Returning
Drop Shots
Richard Krajicek,

By Howard Brody

_Hitti Drop Shot us

his short chapter is an attempt to clear up one common misconception about how to hit and how to
think about the drop shot and in what way that shot differs in a critical way from a standard ground-
stroke.

When normal groundstrokes are played, your first concern is to get the ball over the net, and then you usually
want the ball to bounce deep in your opponent’s court. On the drop shot, your first concern is, again, to have
the ball clear the net, but now you want the ball to bounce as short as possible — hopefully having a second
bounce inside the service box. This difference in the resultant shot can be accomplished by visualizing the net
and court in a different manner.

For the normal groundstroke, you often visualize a window at the net, and your shot must pass through this
window in order to both clear the net and not bounce long. (Too much clearance above the net will often result
in a shot bouncing beyond the baseline.) Since, on a groundstroke, you are usually hitting the ball from your
own baseline at a height that is not too different from the height of the net @ feet), you must lift the ball up
with your stroke so that it will clear the net. Most of you probably visualize the ball’s trajectory having a maxi-
mum height (its apex) at the location of the net. This trajectory will produce a shot that bounces fairly deep into
your opponent’s court, which is what you want.
222 Chapter 23

For a drop shot, you hit the ball with less pace, and usually with considerable
backspin or underspin. Why is this? Both hitting with less pace and with con-
siderable backspin or underspin will produce a ball trajectory that will not go
very far after its bounce. This is what you want. The more speed a ball has,
the farther it will go between its first and second bounce, so you clearly want -
less pace. The more backspin on the ball, the more the ball will slow up when
it bounces, reducing the distance between first and second bounce.

However, the most important thing with a drop shot is not to have the peak of
the ball’s trajectory located at the net. A peak of the ball’s trajectory near the
net will result in a very poor drop shot since the ball will bounce fairly deep
into the court. For a good drop shot, you want the peak of the trajectory (the
highest point the ball reaches) approximately halfway between you and the
net. If you are on the baseline (which is a poor place to hit a drop shot from),
the peak of the trajectory should be about 20 feet from you. If you are 10 feet
inside the baseline when you hit the ball (a good place to hit a drop shot from),
then the trajectory peak should be 15 feet from you (and 15 feet from the net).

If you can visualize in your mind an imaginary wall or barrier 5 or feet high,
that is halfway to*the net, a good drop shot will just clear that obstruction.
However, if as your shot clears the imaginary
wall, it is still going up, your shot will bounce
well into the court rather than bounce close to
Figure 23.1
the net. If, as your shot clears the imaginary
wall, it is coming down, your shot will end in
Backspin Drop Shots the net. If as your shot clears the imaginary
From 10 and 20 Feet From Baselilne wall, it is at its peak (and moving essentially
horizontally at that point), your shot will clear
Peak the net and bounce very short.

— 10’— Figure 23.1 shows the trajectory of balls hit

(with backspin) from 10 feet and 20 feet into
Baselline Net the court. For each hitting location, one tra-
jectory shows the results when there is a peak
at the location of the net, the other shows the
Peak result when the peak is half way to the net.
Several things are apparent from these dia-
— 10’— grams.

Baselline Net
1. The closer you are to the net when you hit
Peak
the ball, the shorter your drop shot can be.

2. A ball trajectory that peaks half way to the

net produces a much better drop shot than a
20’ ———_
shot whose trajectory peaks at the net.
Baselline Net
An obvious trajectory that one might think of
would have the ball peak near the net and
Peak
also cross the net only inches above the tape.
This will result in a shot that bounces very
long (probably out of the court) because to
produce this trajectory you have to hit the ball
Baselline Net very hard. Where the typical shots shown in
the figures correspond to balls hit at 20 to 30
 The Drop Shot 223

mph, to produce a ball trajectory that peaks near the net and just clears the
tape would require an initial ball speed of over 80 miles/hour. That is not what
is wanted for a drop shot.

Your opponent has hit a nice drop shot and you charge in to get to it before
the second bounce. Hitting it on the dead run, you give the ball a gentle flick,
and your return sails long over the baseline. Why does this happen? You have
been practicing that gentle flick from the area around the service line for years,
and in practice it always clears the net and lands well inside the baseline. Now
in a match, it sails long. Did you choke in a pressure situation? No! You hit it
exactly the same way as you did when you were standing on the service line
and practicing all those times. Only this time you were not standing still, but
you were running forward, and your forward speed must be added to the rac-
quet head speed that you get from the flick.

When you are running to get to a ball, your speed is about 8 to 10 mph
(miles/hour), so your racquet will be moving at 8 to 10 mph faster than if you
hit the same shot while you have your feet planted. This will add about 10 to
14 mph to the ball speed, which may be enough to send it out over the base-
line. Why does adding 10 mph to your racquet speed add 14 mph to the recoil-
ing ball speed off your racquet? That is because the speed of a ball off your
racquet is given by the formula V,,,, = ACOR x V,, + (1+ ACOR)V.. A typical
value of ACOR is about 0.4, so the term (1+ACOR) has a value of 1.4. If you
add 10 mph to V,, then V,,,, will increase by 14 mph.

To analyze this matter further, look at Figure 23.2 of ball trajectories for a flat
shot hit from near the service line. There are three curves in this figure, the
lowest one showing the minimum angle (about 6 degrees) needed for a 40
mph ball to just clear the net. Any 40 mph shot leaving the racquet at a small-
er angle will hit the net. The upper curve shows the maximum angle (about 19
degrees) for a 40 mph ball to just land on the baseline. Any ball leaving the
racquet at a larger angle will go long over the baseline. For a 40 mph shot to
be good, it must leave the racquet at an angle between the angles shown by
those two lines.

Trajectories for Returning A Drop Shot Figure 23.2

A B Cc
Initial velocity 40.0 mph 40.0 mph 50.0 mph
Initial angle 6.3° 18.9° 12.0°
Initial spin 0 0 0

B
224 Chapter 23

Let us take a typical case. In practice, you hit your shots so that the ball leaves
the racquet at 40 mph. This means that the angle must be between 6 degrees
and 19 degrees in order for it to be good. To play safe and give yourself a mar-
gin of error over the net and inside the baseline, you try to hit the shot so that
the ball leaves the racquet somewhere between 11 and 14 degrees. In this
angular range, the ball easily clears the net and bounces well inside the base-
line. Now look at what happens when you add 10 mph to your ball speed. If
you again have the ball leave your racquet at the same 11 to 14 degrees, it will
sail well out because it is moving at 50 mph. Looking at the figure, you can
see that at a 50 mph ball speed (the middle trajectory), a shot leaving the rac-
quet at 12 degrees will go long.

What should you do about this when you opponent hits that short drop shot?
If you can get there early and stop before you hit the ball, there is no prob-
lem. Your years of practice will let you get the shot in. If you must hit the ball
while you are in full flight, either hit it much easier than you did in practice
(reduce your racquet head speed on your flick), or aim your shot lower so that
the ball clears the net by a much smaller margin than you did i practice. Either
of these will work, but it takes the mental discipline to remember to do it in
the middle of a hectic point when you are on a dead run.
 How Time Is Key
in Tennis Strategy

Daniela Hantuchove

By Howard Brody

IME is the crucial element in a tennis match and most players do not seem to realize it. Almost every-
thing you do when you are trying to win a point is based on an attempt to control time. You always try
, to reduce the time your opponent has to get to a ball or hit a ball. You usually attempt to give yourself
as much time as you need when you are going to hit the ball. In each of the cases listed below, you will see
how your attempt to control time has determined your choice of what strategy to use and how you play. You
do these things to win points, without ever realizing that what you are doing is attempting to control time.

Why do you try to hit the ball hard in tennis in order to win more points?

You know that the harder you hit the ball, the smaller are your chances of that shot landing in the court. This
is because of two distinct reasons. When you swing harder, you have less control over your racquet, and when
the ball speed off of your racquet is increased, your “acceptance window” for the shot being good is reduced.
If you hit a ball at low speed, it can go in Cand in the court), whether it passes over the net, close to the net
top, or very high over the net. If you hit the ball hard, it can only land in the court if it passes close to the top
of the net. But hitting the ball harder also reduces the time your opponent has to get to your shot and to return
the ball. If your opponent does not have enough time to get to your shot, the result is a winner for you. The
less time your opponent has to prepare for a shot, the more likely it is to result in an error or a weak shot. On
a groundstroke, if the ball leaves your racquet at 50 miles/hour, your opponent has almost 1.5 seconds to get
226 Chapter 24

to it and return it. If the ball leaves your racquet at 70 mph, your opponent has
about a full second to get to it. This difference of half a second is quite signif-
icant and it may not allow your opponent time to set up and hit a good shot.

Why do you pounce on short shots your opponent has hit? Why would
you come in and hit a ball on the rise, even though it is a much tougher
shot to bit and it does not produce an appreciably higher ball speed off
of your own racquet?

Moving in and taking the ball early reduces the time your opponent has —
both the ball round-trip time between his/her swing and the return and also
the time between your hit and the ball getting to the other baseline. The total
time is important, because if it is too short, your opponent may not have had
time to get back into position. Remember, when you hit a winner, it means
your opponent did not have sufficient time to get to it. If the ball leaves your
racquet at 60 mph, just moving your hitting point from 5 feet behind the base-
line to 5 feet inside the baseline will reduce the time your opponent has from
1.28 seconds to 1.04 seconds. This is a quarter of a second less that your oppo-
nent has to play the shot.

Why do you try to get to net to win points?

Getting to the net allows you a larger angular range for your returns and it
greatly reduces the amount of time your opponent has to get to your return.
Those are the reasons why you can hit so many more winners when you vol-
ley from inside the service line, as compared to hitting groundstrokes from the
baseline. When you are at net, it essentially reduces the time your opponent
has by half, compared to the time your opponent has for a groundstroke hit
from the baseline.

Why do you try to hit your shots deep in the court (bounce near the
baseline on groundstrokes, the service line on serves), when doing so
increases your chances of making an error (your ball going long)?

Hitting shots deep reduces the time your opponent has to hit the ball, since
the ball slows down when it bounces. When a ball bounces short, it takes that
ball an extra 0.1 second to get to the baseline, compared to a ball that has
bounced near the baseline. When your shot bounces near the baseline, unless
your opponent retreats, it gives your opponent less time to hit the ball after the
bounce. As an example, if the ball leaves your racquet at 60 mph, for every 4.5
feet closer to the baseline it bounces, your opponent has 1/10 of a second less
time after the bounce to react. If your opponent does retreat well behind the
baseline to obtain more time, you then get much more time to get to his/her
return. It is very difficult to hit a winner from well behind the baseline — the
ball just takes too much time in its flight. Hitting the ball deep also gives you
more time to get to, and to react to, your opponent’s return of that shot.

Why dont drop shots work well when you hit them from behind the
baseline?

Drop shots don’t work well when you are hitting from the baseline region
because your opponent has more time to get to them, compared to the time
your opponent has when you hit a drop volley or a drop shot from well with-
in the baseline. When hitting from behind the baseline, it is also harder to try
 Time and Strokes 227

to keep your ball's trajectory short, because when you are hitting from that
position, you must give the ball more forward velocity to make it go over the
net.

Why should your approach shot (and volley) be deep?

Approach shots and volleys should always be hit deep because it forces your
opponent to hit from behind the baseline, which gives you more time to react
and cover more of the net. If you can force your opponent well behind the
baseline, you will be able to crowd the net and not be passed — you will have
time to cover both down-the-line and crosscourt shots. If your opponent lobs,
it gives you more time to get back under the ball.

Why is a topspin lob more effective than a backspin or flat lob?

You can hit a topspin lob with much more forward speed (and still have it go
in) compared to a flat or backspin lob. This gives your opponent at net less
time to get back and cover it. On the bounce, a topspin lob tends to not lose
much speed, where a normal lob will lose 40% of its forward speed on the
bounce. Your opponent will not have the time to run down a topspin lob after
the bounce if it gets over his/her head.

Why is there a premium on getting the first serve in when playing dou-
bles?

Many people have a weak second serve that allows the receiver to move well
inside the baseline when returning it. This can turn the server’s partner, who
is at net, into a “sitting duck.” Hitting, even a mediocre shot, at the net player
when you are well inside the baseline gives that player so little time to react
that he/she is handcuffed and will often return it with a weak volley, if at all.
To counter this, the net player can retreat back to the service line to regain
some time, but then most of the advantage of being at the net is lost.
 Reducing
Unforced Errors

By Howard Brody

f you are a typical tennis player, you will lose many more points by making errors than you will lose because
of winners hit by your opponent. If your rate of errors can be reduced, it should lead to an increase in the
number of points that you win. There are several ways in which your error rate can be reduced, and some
of them will be discussed here.

An improvement in stroke mechanics, lots of practice, and better concentration are obvious ways to reduce your
error rate, but these are not the subject of this chapter. Appropriate shot selection and a different racquet or
stringing can also lead to a reduction in your error rate, if you are willing to follow advice based on scientific
analysis.

So
Lateral Errors

Errors come about because the ball has hit the net or gone long (an error of depth), or because the ball has
gone wide (a lateral error). If you aim your shots so as to have the ball go down the middle, you have an allow-
able error of almost ten degrees to the right or to the left before the ball goes wide, and this is a sizable mar-
gin for error. If your shot bounces well short of the baseline, then your margin for side to side error is even
greater. The problem is that most of you don’t want to play safe and hit every shot down the middle. You want
to go for the corners or go down the sideline. You may want to pass an opponent at the net. You want to go
230 Chapter 25

for an occasional winner or at least make your opponent run for the ball once
in a while. When you do that, you are inviting lateral errors to occur.

However, you can reduce the number of errors from shots that go wide, even
when you go for corners or side lines by remembering this advice. Don't
change the ball angle! If a shot is coming to you cross court, return It Cross
court. If a shot is hit down-the-line, return it down-the-line. When you change
the ball angle by attempting to return a cross court shot down-the-line or return
a down-the-line shot cross court, you are asking for error troubles. The reason
for this comes out of the physics of the ball-racquet interaction.

When your return does not change the ball angle, the ball’s impact direction is
perpendicular to the face of the racquet at contact. It will then leave your rac-
quet in a direction perpendicular to the face of the racquet. The ball will go
out in the direction of the racquet motion whether you swing hard or soft. This
is not the case when you attempt to change the ball angle. In this case, the
direction that the outgoing ball makes relative to your racquet face depends on
how hard you swing your racquet. This is illustrated in Figure 25.1. The high-
er the relative ball-racquet speed, the closer the ball will be to perpendicular
to the racquet face as it leaves the racquet. If your swing is slow, then the ball
will leave your racquet at a larger angle. Therefore, when your return changes
the ball angle, the chances of you making an error and the ball going wide,
depend on how hard you swing. When you do not change ball angle, since
the ball direction does not depend on how hard you swing, lateral errors
should be reduced. It is your choice.

This advice also holds when you volley. If

you punch out at your return, the ball will go
close to where you aim it. If, on the other
Ball Angle vs Racquet Head Speed hand, you just try to block the ball on your
25
volley, and you are changing angles, the ball
will slide off at a large angle, possibly going
ine) =) wide.

However, if you return every shot back to

are on
where it came from, your opponent will
quickly catch on, and will be there waiting
—s oO
for it. Knowing the facts about the ball-rac-
Ball
(degrees)
Angle quet interaction, you can reduce the errors
that might occur even when you change the
ball angle. If you are not going to hit the ball
hard, then aim the ball a little closer to the
10 20 30 40 50 60 70 80 center of the court. If you are really going to
swing out, then you can aim your shot a bit
Racquet Head Speet (ft/sec)
closer to the sideline or the corner with con-
fidence.

Figure 25.1 The famous statement that the angle of

reflection equals the angle of incidence holds
Ball angle leaving the racquet versus racquet head
speed. This is for a ball incident at 20 degrees and
for light reflecting from a plane mirror, not
60 feet/sec for a typical groundstroke. The harder the tennis balls rebounding from a racquet.
ball is hit, the closer is the resultant ball direction to
perpendicular to the string plane. Often, in a match, when you find yourself
well ahead of your opponent, you may ease
 Error Reduction 231

up a bit, and not hit your shots quite so hard. You do this to reduce the num-
ber of errors you are likely to make. This can lead to a problem. If you con-
tinue to aim the ball the same way, only not swing as hard, balls that previ-
ously went down-the-line will now end in the alley.

A similar problem can be the result of you changing your game plan in the
middle of a match. You become concerned about the final outcome, so instead
of hitting out and playing your regular game, you decide to play it safer and
ease up on your strokes, Again, balls that previously went down-the-line will
now end in the alley. You end up making more, not less errors. People will
claim that you “choked,” when actually all you did was not understand the
laws of physics.

A second cause of lateral errors is timing errors. If you strike the ball too early
or too late in your swing, the ball will not end up going in the direction you
want it to. One way to reduce this sort of error is to follow through your swing
in the direction that you want the ball to go. When you swing this way, it looks
as if you are trying to keep the ball on the strings longer. However, what you
are actually doing is keeping the racquet head in the proper lateral angular ori-
entation for a longer period of time (using less wrist in your swing), and there-
fore making the shot less sensitive to timing errors. In baseball, when you miss-
time a pitch, you get a foul ball, and you get another swing. In tennis, when
you miss-time a shot, you may lose the point.

Where the side-to-side errors in tennis are

basically a matter of court geometry, the ver- Figure 25.2
tical or depth errors depend on many factors,
rape 3 Vertical acceptance angle leaving the racquet versu
and your margin for error is often an order of p 9 g oa id
: : ; : - ball velocity. The lower curve is for a ball just clipping
magnitude smaller than the side-to-side error
the top of then net and the upper curve is for a ball
margin. This means, that on occasion, you landing on the baseline. The distance between the
will be trying to thread a vertical needle with two curves is the angular acceptance for the shot
your shot. In order for a shot to be good, it being good as a function of the ball speed.
must be hit through a window — either a vir-
tual window in vertical angle as the ball
leaves the racquet that will allow it to clear
the net and bounce in the court, or an effec- Acceptance Angle vs Ball Velocity
tive window just above the net, the upper (Groundstroke from baseline)
edge of which corresponds to a ball trajecto- 20
ry which bounces right on the baseline. This
is shown in Figure 25.2. For a given ball
speed, if your ball passes over the net above 15 Ball Long
the top of this window, it will bounce beyond
the baseline.
10
Window Size and Velocity. The size of this
window decreases markedly as the ball speed
increases, but increases as topspin is added.
Racquet
Off
Angle
Ball Ball Hits Net
Since a groundstroke is hit at a height of
about 3 feet, and the net is about 3 feet high,
it is gravity that pulls the ball down into the 0
court, otherwise the ball would sail over the 40 20 60 We ; oie 20 pg DOaeale
baseline. When the ball is not hit hard, after SAPS eT DN
232 Chapter 25

it crosses the net, gravity has more time to pull it into the court before it gets
to the baseline. Therefore, the vertical window increases in size if you don’t hit
the ball as hard. The larger the window, the less likely you are to make an error
because of the ball failing to go through it. If the window is large compared to
the variation in your strokes from shot-to-shot, you will be known as a steady
player who does not beat him- or her-self. If you select a small window (by
hitting the ball too hard, for example), then you will have great difficulty in
winning, and most points will end early with you making an error, even if you
have reasonable control when you stroke the ball hard. The choice of window
size is selected by you, and therefore, your error rate also is effectively select-
ed by you. If you want to hit the ball hard you will make more errors, both
because the window is smaller and because it is more difficult to control the
racquet on a harder swing. Practice and correct stroke mechanics will reduce
the variation of your shots from stroke to stroke, but they will not change the
vertical window size. Only shot selection can do that.

The fact that shots that have higher velocity have smaller windows (and lead
to more errors) may lead to an unanticipated source’of errors. When a ball is
coming at you with a very high velocity, if you stroke it back with your nor-
mal swing, your resulting shot will be a higher speed ball. This is because the
speed of the struck ball is given by the formula:

V(struck) = One, (incident) +(1 +€,) -V (racquet) (25.1)

where the parameter e, is about 0.40. This means that, if the ball coming
toward you is hit harder, unless you reduce your swing speed somewhat, your
return shot also will have a higher speed. The result will be a reduced window
for your shot to be good, and an increase in
your error rate. If you don’t want this to
Acceptance Window vs Hitting Height happen, you should slow down your swing
a little when your opponent is blasting the
(Groundstroke from baselline)
es ball at you. As a bonus, this also might give
you better racquet head control.

Window Size and Hitting Height. You also

can increase the size of the window on any
shot (which reduces the probability of mak-
ing an error) by hitting the ball when it is
high, as opposed to striking it when it is
down around your ankles. The incoming ball
oy
(degrees)
trajectory does not always give you that
Accpetance
Window
option, and many players do not have the
control or power to hit shots at shoulder
height. But if you can hit the ball when it is
{ 2 3 4 5 6 7 8 higher, vertical errors will be reduced. Figure
Hitting Height (tt) 25.3 is an illustration of the increase in the
vertical acceptance window with hitting
Figure 25.3 ie and the decrease of window with ball
speed.
Vertical angular acceptance versus hitting height for a ground-
stroke. The three lines represent three different ball velocities. Window Size and Spin.
The chances of a shot being good (size of the acceptance win-
dow) increase as the hitting height increases and decreases
Topspin. You i : in-=
as the ball speed increases. Eee can increase the small win
dow that comes from hard shots by putting
 Error Reduction 233
topspin on the ball. Topspin causes the air to provide a downward force, in
addition to gravity, that causes the ball to dive into the court. But topspin also
has some penalties associated with it. Many players have almost no control of
their shots when they attempt to use topspin, so even though the window size
increases, their shots spray all over (and often out of) the court. One of the
reasons for this may be the following. When a ball bounces, it normally comes
off of the court with considerable topspin due to the friction force between the
ball and the court surface. When you try to return such a shot with your own
topspin, not only must you turn the ball’s direction around, you also must turn
its spin around. In order to do this, you must swing your racquet 1.5 times
faster than if you want your return to be flat (with no spin on the ball). If you
are already swinging your racquet fairly hard (which is why the window is
small), the loss of control due to a 50% increase in swing speed may be a major
cause of the increase in your errors.

Backspin. Backspin (under spin or chop) shots have a reduced window

because the ball tends to float long after clearing the net. (The interaction of
backspin with the air produces an upward force on the ball that counteracts
gravity.) Yet, many intermediate level players seem to be able to play a steady
game chopping away at the ball and getting it to go through the small window
they have set for themselves. This is because they do not have to hit the ball
very hard to return it, due to the fact that the incoming ball has acquired top-
spin because of its bounce. All they have to do is turn the ball’s direction
around, since the ball is already spinning the way they want it to. Not having
to spank the ball really hard means that the player can use a compact stroke,
get away with poor racquet preparation, and
still have good racquet head control. This
often reduces their shot to shot variation suf-
Acceptance Window vs Angle at 45 mph
ficiently to more than compensate for the
(h it from corner)
reduced window size.

Window Size and Shot Selection. Another Over Baseline

option you have is the down-the-line versus
Beyond
cross court decision when you are hitting Sideline
from the corner. This has already been dis-
cussed with respect to change of angle, and Good Shot
on
now it will be analyzed with respect to errors
in depth. The vertical window for a success-
Window
Vertical
ful shot is always greater for a shot hit cross e Hits Net
court rather than down-the-line. This is 5
because the court is much deeper (the base-
line further away) when you hit at an angle, 0
along the hypotenuse of a triangle. You are ee 0 5 10 15 20 25 30
effectively hitting into a bigger court when Horizontal Angle
you go cross court, compared to down-the-
line. The harder you hit the ball, the greater
is the advantage of cross court over down- Figure 25.4
the-line. If your shot lacks pace, you will not Vertical angular acceptance versus horizontal angle for a 45
make many more depth errors going down mph groundstroke hit from the baseline at the corner. The dif-
the line. If, however, you really swing out at ference in angle between the upper lines and the lower line is
the ball, then the cross court shot will result the actual vertical angular acceptance of the window as a
in far fewer errors. Figures 25.4, 25.5, and function of the hitting angle (from down-the-line at zero
25.6 show how the window of acceptance degrees to crosscourt at 19 degrees) and is shown in Figure
depends on the angle at which the ball is hit 25,6.
and the ball velocity.
234 Chapter 25

You often hear TV commentators claim that

when hitting from the corner, you should go
crosscourt rather than down-the-line since
Window of Acceptance vs Angle the net is higher down-the-line, and there-
fore, you must lift the ball higher to clear the
net. This is not true. Since the distance from
the corner to the center strap is greater than
the distance to the net down-the line, you
must lift the ball to the same vertical angle to
clear the net for both cases. It is the greater
depth of court behind the net that increases
the window cross court, not the variation in
net height. The opposite effect occurs when
Window
Acceptance you are hitting from the center of the base-
line. Hitting down the middle has a lower
net, shorter distance to the net, and shorter
court behind ¢he net. This is almost exactly
0) 5 10 1S 20 25 30 balanced by a shot toward the corner which
Angle (degrees) goes over a slightly higher net, has a greater
distance to the net, and a bigger court to hit
into behind the net. The vertical acceptance
Figure 25.5
window for a shot down the middle is almost
Window of vertical acceptance for a groundstroke hit from the the same as one into the corner.
corner for three different ball velocities.

String Tension, Dwell Time and Control.

Often you hear the advice “keep the ball on
Ratio of Acceptance Window, your racquet longer for better control.” It’s
Cross Court/Down-the-Line versus Ball Speed easy to keep the ball on the strings longer —
5
just reduce the string tension and the ball
will spend more time in contact with the
3 strings. Ct is a well documented scientific
G4
£& result that looser strings lead to longer ball
= dwell times.) But will this lead to better con-
ae) trol? Probably not. You have often heard the
iy)
a
o
advice “string tighter for more control, loos-
op 2 er for power.” Both of these pieces of advice
ro) cannot be correct since they contradict each
2 other. What is correct?
os 1
or
The racquet frame and the stringing in that
0 frame can influence the number of errors
40 50 60 70 80 90 100
that you make. For a long time the literature
Ball Speed (mph)
of tennis has stated that tighter strings lead to
better control, and better control means
Figure 25.6
fewer errors. However, nowhere in the liter-
Ratio of vertical angular acceptance window for crosscourt ature can you find an experiment, data, or a
versus down-the-line as a function of ball speed for a ground- scientifically valid argument explaining why
stroke hit from the corner. There is a small advantage to hit- this is so. (An article appearing in the Winter
ting cross court at low ball speed, but at high ball speeds, it is 1999 edition of the International Sports
very important to hit cross court rather than down-the-line. Journal was the first to show a measurement
of the correlation between string tension and
control.) It has long been known that lower
 Error Reduction 235
string tension does produce greater ball speed (power), but the arguments
about the correlation between tension and control are usually anecdotal.

Spin Shots and Control. There is a second reason why tighter strings can
lead to fewer errors. Often when you are hitting a heavy spin shot, the ball will
clip the frame and you will get a complete miss-hit. By stringing the racquet at
a higher tension, the chances of such a miss-hit will be reduced. On a heavy
spin shot, the strings meet the ball at a large angle, as opposed to nearly nor-
mal on a flat shot. The ball then slides across the strings before rebounding off.
The distance the ball slides along the strings depends on the angle of the rac-
quet, the racquet head speed, and the dwell time of the ball on the strings. The
shorter the dwell time (the tighter the strings), the shorter is the distance that
the ball slides along the strings, and the less likely is the ball to run into the
frame. Therefore, increasing the string tension will reduce the probability of a
frame miss-hit and increase the acceptance window of ball impact locations on
the head of the racquet. As a general rule, when hitting shots that are designed
to have a lot of spin, you don’t want the initial ball impact location to be at the
center of the head (along the long axis). Since the ball is going to slide or roll
along the strings for a certain length, you want an initial impact point that is a
distance away from the axis, which will allow the ball to cross the axis while
rolling or sliding, and then depart the strings on the other side of the axis. This
will also reduce the probability of striking the frame.

Headsize, Polar Moment, Off-Center Hits and Control. In normal play, you
attempt to hit the ball at a specific location on the strings, possibly the center
of the head. If you miss that location, you may make an error. This is because
the ball comes off the racquet at an unanticipated angle or with a speed high-
er or lower than desired. The result of this slight miss-hit is a ball trajectory that
often ends in the net or over the baseline.

When the ball hits the strings at a location that is not along the long axis of the
racquet (toward the side of the head), the frame will twist and the rebound will
come off with lower speed than the same shot hitting on axis. Assuming the
long axis of the racquet is parallel to the ground, a ball impact above the axis
will result in the racquet twisting, a rebound angle greater than desired, and
also less ball speed. This combination of effects Carger vertical angle and lower
ball speed) will often lead to ball trajectories that still end landing in the court.
If, on the other hand, the ball strikes below the axis, the rebound ball angle
will be lower than desired and the ball speed again will be lower. This com-
bination often leads to shots that end in the net. (An error reducing strategy
would be to attempt to always hit the ball slightly above the axis. However, the
unanticipated consequence of such a strategy might result in tennis elbow from
the constant twisting of the racquet due to off-axis ball impacts.)

The fractional reduction of rebound ball speed depends on how far off axis the
impact is and a physical property of the racquet called the moment of inertia
about the long axis. The larger this moment of the racquet, the less the
rebound ball speed will be effected by off axis hits and the more stable the
racquet will be against twisting. Therefore, by going to a racquet with a larger
polar or roll moment of inertia (the moment about the long axis), you will
make fewer errors due to the miss-hits, some of which are almost inevitable.
As a general rule, racquets with a wider head (as opposed to widebody, which
is thicker) will have a larger moment of inertia and be more stable. A frame
that is 25% wider (10 inches instead of 8 inches) can result in a 50% greater
236 Chapter 25

moment. This means that using a racquet with a wider head usually will result
in fewer errors due to this type of miss-hit. Howard Head knew what he was
doing when he designed the original Prince oversize racquet.

Racquet Balance, Sweetspot and Control. If the ball hits the racquet along
the long axis of the racquet, but at different locations along that axis (closer to
the tip or to the throat), errors of depth can again occur. For many racquets
and classic ground stroke swings, the ball rebound speed decreases as the ball
impact point moves from the throat area toward the racquet tip, all other
parameters held constant. If you are attempting to strike the ball at the center
of the head and trying to hit a deep shot, ball impacts that occur closer to the
throat are likely to go long because the ball will come off of the racquet with
too great a speed. This can be compensated for by consciously hitting your
normal (central impact) shots so they bounce short. Now the balls that inad-
vertently impact closer to the throat will land in the court. This is a very con-
servative strategy, and is comparable to hitting every ball softly to insure a large
window and few errors. ‘

Some of the modern, very stiff, handle-light frames now available, have suc-
ceeded in moving the power region further up into the head of the racquet.
When one of these frames is combined with a flexible wrist swing (not the clas-
sic long, flowing stroke), the effective maximum power point of the racquet
can move up so that it is at or near the center of the head. This is because the
center of the head is moving faster than the throat on these type of swings.
Calculations show that when the power maximum is at the center of the head,
the variation of rebound ball speed as the ball impact point moves away from
the center, is greatly reduced, compared to the classic racquet used with a clas-
sic stroke. In other words, the “power spot” has been enlarged and made more
uniform. Clearly, this will reduce errors of depth.

In addition, there is a built-in error correcting mechanism that occurs when the
maximum power point is at the center of the racquet head. Any shot that miss-
es the center will come off the racquet with slightly less rebound speed than a
shot that hits at the center. If you then attempt to hit your shots so that balls
that strike the racquet at the center of the head land deep in the court, depth
errors (balls going over the baseline) due to variation in impact location will
automatically be reduced. Any shot that hits the strings slightly closer to the
throat or slightly closer to the tip than the center of the head, will come off of
the racquet with slightly less speed, and therefore will land a bit shorter. These
types of slight miss-hits will not result in balls that land long, and therefore
depth errors are reduced. Because this effect is working in your favor, you can
hit your shots deep, go for the baseline, and play a more aggressive game
S
 Introduction to
String Behavior

Justine Henin, Belgium

By Howard Brody

Introduction —
he strings in a tennis racquet perform a number of functions. Their principal purpose is to absorb the
kinetic energy from the relative ball-racquet velocity, turn it into potential energy (stored energy of elas-
tic deformation), and then give that energy back to the ball in the form of an outgoing ball velocity or
kinetic energy. Since the string plane is the only thing that touches the ball, it is also the agent for putting spin
on the ball and for controlling the outgoing direction and speed of the ball.

A racquet is strung with thin strings, rather than having a fabric or a membrane covering the head of the frame
because of air resistance. Even with the relatively thin strings taking up just a small fraction of the area, there is
some difference in air resistance between a strung and unstrung racquet. To get a feeling for what air resistance
does, perform the following experiment. Cover the face of your racquet with a sheet of paper and swing it, then
remove the paper and swing it again. Note the difference in the way the racquet feels. You might also see if you
can detect the difference in feeling between a racquet swung with the face perpendicular to the ground (a nor-
mal swing) and the same racquet swung with the face parallel to the ground (knifing through the air). It is inter-
esting to ask whether you as a player, can tell the difference in air resistance between a racquet strung with 1.4
240 Chapter 26

mm thick nylon and one strung with 1.3 mm thick nylon? Probably not, but a
scientist who had a well-instrumented wind tunnel might notice a difference.

While players and racquet stringers talk about the tension in the strings, it is
actually the string plane deformation that determines many of the playing char-
acteristics of the racquet. Racquets strung at the same tension with the same
type of string can play very differently, depending on the head size and the
string pattern or string density. A measurement of string plane deformation
takes all of these factors into account, yet tension (specified in Newtons,
pounds or kilograms) is still the standard measure for specifying to a racquet
stringer how you want the racquet strung (because that is what the stringing
machine is calibrated in). Don’t ask your stringer for a specific string plane
deformation, because the odds are that you will not get it, unless you have a
very experienced stringer who checks strirfg plane deformation on every job
he does. Tension is an acceptable parameter to work with if you always use
the same racquet model (same head size and string spacing or pattern); but to
compare the string performance between two different racquets, string plane
deformation should be used. For small deflections of the string plane, string
plane deformation does not depend on the type of string used, so nylon, gut,
poly, etc., when strung in the same racquet at the same tension, will give about
the same string plane deformation.

String plane deformation can be measured by at least two different types of

commercially available devices. The one most commonly used is an instrument
(such as an RDC) that pushes on the string plane with a known, static force in
a direction perpendicular to the face and measures the distance the string plane
deforms or deflects. A second type of instrument places a mass on the strings
at the center of the head and measures the resonant frequency of the mass-
string system. The frequency of this oscillation depends on the square root of
the string plane stiffness. Both of these types of devices are usually calibrated
so they produce a number that is closely related to the tension in the strings,
once the head size and string spacing are taken into account.

¢ String plane deformation determines most of the playing characteristics of the

racquet that are due to stringing.

¢ For a given racquet, an increase in string tension will produce a decrease in

string plane deformation. (The tighter the strings, the more you have to push
on the string plane to get a certain deformation.)

e As a general rule, if all else is held constant, a frame with a larger head will
have more string plane deformation.

e As a general rule, if all else is held constant, a frame with a larger spacing
between strings (a more open string pattern) will have more string plane defor-
mation.

° A racquet with a greater string plane deformation will play softer.

When the ball hits the strings, or when you pluck or strum the strings, there is
a sound, almost musical in character, if there is no vibration dampener installed
 Overview of Strings 241

in the string bed. For most racquets, the strings will vibrate at about 500 to 600
Hertz (cycles per second) when displaced from their equilibrium position, and
that is what you hear. Many people tap their racquet’s strings to assure them-
selves that they are still at the correct tension. This is because the frequency
that the string plane vibrates at is directly proportional to the square root of the
string tension. In fact, the frequency at which the entire string plane vibrates is
close to the frequency of vibration of a single string that makes up that plane
at that tension. You cannot compare tension between two different size rac-
quets by listening to the ping. At the same tension, the bigger head will have
a lower pitch. (The frequency goes as the inverse of the string length, so a
longer string will have a lower frequency than a shorter string at the same ten-
sion.) Between two identical racquets at the same tension, you will get a dif-
ferent sound if different gauge strings are used. If you do have two identical
frames, strung with the same string, then you can compare them, and the lower
pitch will be caused by a lower tension. For those of you with “perfect pitch,”
you will be able to tell just how much the tension in your racquet has
decreased in time.

If you place a string dampener in the string bed near the throat of your rac-
quet, it will damp out the vibrations of the string (not the frame) very quickly
and most of the pitch information will be lost. You will hear a thud, rather than
a ping. The Rules of Tennis do not allow you to place a string dampener with-
in the pattern of crossed strings.

e There is no evidence that string vibrations can cause any harm to the player.

e There is no evidence that string vibrations have any effect on power or control.

44
well Time

The string plane deformation also determines the dwell or contact time of the
ball on the strings. The more the string plane deforms for a given load, the
longer a ball will spend in contact with those strings. This means that lower
string tension produces longer dwell times and string planes with less deflec-
tion produce shorter dwell times.

It is counter-intuitive, but the harder you hit the ball, the shorter the dwell time.
You somehow have a picture in your mind of a harder hit ball sinking deeper
into the string bed, and therefore taking longer to come out. What actually hap-
pens, is that the harder you hit the ball, the stiffer the strings get, as does the
string plane. Even though the ball sinks in more, it also snaps back sooner due
to the increase in string plane stiffness.

Many types of strings get stiffer or less elastic as they are stretched, so when
you hit the ball harder, the impact may feel harsher, because of the shorter
dwell time and the increase in string stiffness. This will contribute to the
“boardy” feeling that some synthetic strings produce.

There is an old adage in tennis that says “Keep the ball on your strings longer
for better control.” As far as I know, there is no scientific justification for this
statement in the literature of tennis, or anywhere else. In fact, there is another
old adage that says, “Tight strings give better control.” When you go to tighter
strings, the time the ball spends on the strings DECREASES (shorter dwell time),
so these two old adages disagree with each other. There is nothing that you
can do with your swing that will increase the 0.004 seconds (4 milliseconds) of
242 Chapter 26

ball dwell time on the strings, except not hit the ball as hard, which will prob-
ably improve control by itself. What can be done, and should be done, is to
follow through the stroke in the direction you want the ball to go. Then slight
errors in timing will not translate into errors in ball rebound angle. This follow
through in the ball rebound direction will improve control and give the appear-
ance that you are keeping the ball on the strings longer.

When the new Type 3 ball (6% larger in diameter) was introduced, it was test-
ed extensively in both the lab and on the courts. Players claimed that the larg-
er ball felt heavier in play than the standard Type 2 ball, even though it was
actually slightly lighter. Lab tests showed that the larger ball was just a bit soft-
er, leading to a slightly longer dwell time. Since a ball that is actually heavier
will have a slightly longer dwell time, it is possible that this is the effect that
causes players to claim the bigger ball feels heavier.

A tennis ball is designed (and required by Rule 3 of the Rules of Tennis) to lose
energy when it deforms. When you drop a tennis ball from a height of 100
inches onto a hard surface, the Rules of Tennis state that it must rebound to a
height of at least 53 inches and not more than 58 inches. This means that 42
to 47 percent of the energy the ball had when it hit the ground is lost — dissi-
pated in heat, sound, or other mechanisms. Another way to characterize the
interaction is through velocity. The way to calculate the velocity return in a col-
lision is to take the square root of the rebound height to drop height ratio. This
is called the Coefficient of Restitution (COR), and it is the ratio of the rebound
velocity to the incident velocity. As an example, for a rebound height of 55
inches, the ratio is 55/100 = 0.55 and the COR = 0.74, which is the square root
of 0.55.

In collisions where the ball deforms more (a

higher velocity collision with the hard sur-
Bounce Height/Drop Height
face), the fraction of the energy lost increases,
0.65
which means the COR decreases. If the ball
does not deform as much (for example, it is
dropped from a lower height than 100 inch-
es), the percent of energy loss will decrease
and you have a higher COR. This is displayed
graphically in Figure 26.1, which shows the
ratio of rebound height to drop height (frac-
tional energy return) for various drop heights.
If you could keep the ball from deforming
Bounce
Height/Drop
Height
when it collides with something, then the
energy loss in the ball might be reduced and
the ball will rebound with a higher velocity
0 0.5 1 ipo 2 2s) 3 3.5 4
(or it will be “livelier”). If a ball is dropped
Drop Height (meters)
from 100 inches, not onto a hard surface, but
onto the strings of a racquet (with its head
Figure 26.1 firmly clamped) the ball will rebound to a
Bounce Height/Drop Height. The ratio of rebound height to pans ues 2 ang 20 vache Sls Sch
drop height versus drop height for a tennis ball on a hard collision, the strings deform, absorb some of
(concrete) surface. the ball’s incoming energy, and then return it
to the ball very efficiently (with essentially no
loss of energy in the strings). The strings act
 Overview of Strings 243

like a trampoline, absorbing some of the incom-

ing energy, and then returning it to the ball. The Energy Return vs
more energy the strings absorb, the less is left Ball Stiffness/String Stiffness
in the form of ball deformation. The strings give (clamped head)
back 95% of the energy they absorb while the
ball gives back only about half of the energy it
stores. Clearly, you will get more “power” if you
arrange it for the efficient strings to absorb
more of the incoming energy and have the dis-
sipative ball store less energy (in the form of
deformation). The looser the strings Cow ten-
sion), the larger is the fraction of energy stored Energy
Return
by the strings, and therefore, a larger fraction of
the incoming energy will be fed back to the
rebounding ball. A calculation of this effect is
shown in Figure 26.2. However, there is a limit
to this, since you cannot play tennis with a but-
terfly net. When the strings begin to move with- Ball Stiffness/String Stiffness
in the string plane, rubbing each other as they Figure 26.2
move, energy is lost.
Calculation of Fractional Energy Return versus Relative
Ball/String Stiffness. As the strings become softer, they
How do we know that strings are about 95% absorb then return more of the energy the ball is bringing in.
efficient in returning energy? If you clamp the
head of a racquet so the string plane is hori-
zontal and drop a non-deformable (hence non-
energy dissipating) ball on it, it will rebound to
about 95% of its drop height. This has been
done with a wood Bocce ball and also with a steel ball. The drop heights were
such that the energies involved were comparable to a typical ball-string impact.
These results were also fairly independent of whether the strings were old (in
the racquet for years) or new and also independent of the type of string. If you
were to do the same experiment with a tennis ball (which is deformable and
dissipates energy when deformed), the rebound height would be about 80% of
the drop height, compared to a 55 to 60% return when a tennis ball is dropped
on a hard surface. This does depend somewhat on the string tension.

¢ Tight strings (high tension, stiff string plane) will return a little less energy to
the outgoing ball.

¢ Loose strings Cow tension, softer string plane) will provide a little more
power, but strings that are too loose will dissipate energy.

e For a given tension in the strings, longer strings (a bigger racquet head) will
produce a softer string plane and more power.

e For a given tension, an open string pattern (more space between the strings)
will produce a softer string plane and more power.

Are there limits to how low you can go in tension and still get a bit more power
(rebounding ball speed) from your strings? As long as the deformation of the
strings is perpendicular to the string plane, the power should increase. If the
strings begin to move within the string plane, rubbing against each other, then
it is possible to not get back out of them the 95% of the energy you might oth-
erwise expect. (You can’t play tennis with a butterfly net).
244 Chapter 26

rhe Feel of the Strings

If you reduce the string tension (reducing the string plane stiffness) you should
produce a more comfortable, softer feeling racquet. Primarily, this is not
because with lower tension you get more power, so you don’t have to swing
as hard. This extra power is a small effect compared to the true reason. When
you reduce the tension in the strings, the dwell (contact) time of the ball on
the strings increases. The increase in dwell time leads to a reduction in the
peak force of the strings on the ball, which makes the impact feel more com-
fortable.

To return a shot, the ball’s momentum must be turned around, and a change
in momentum is caused by the impulse of the strings on the ball. An impulse
is a force times the length of time the force acts (it’s actually the integral of
force over time), so if the time increases (longer dwell time) the force decreas-
es when the same impulse is needed to retyrn the shot. Another way of look-
ing at this is to think about the shock being spread over a longer time, so its
peak value is reduced.

As a general rule, to have the strings play the same, the string tension should
be scaled with string length. When you go to a racquet with a larger head, you
should increase the tension in the strings in proportion to the increase in string
length, if you want a comparable feel — the racquets to play the same.

All tennis strings are elastic — they stretch, absorbing energy, then return it to
the ball. Some strings require more force to elongate a certain amount. The
ratio of the incremental force required to increase the string length a certain
percentage is called the modulus of elasticity. The smaller this number, the soft-
er are the strings and the softer they play. Steel strings (piano wire) or Kevlar
strings have a very high modulus of elasticity, and even though they give back
essentially all the energy they absorb, they hardly stretch at all when impacted
by a tennis ball. Nylon and gut have a low modulus of elasticity, so they
deform when a ball impacts and they play well. Nylon and gut differ in elastic
modulus when they get to high tension. Both may stretch about the same
amount in being brought to 60 pounds tension, but synthetics get stiffer Chigh-
er modulus) above that tension, where gut does not. Since when the ball hits,
the strings in the region around the impact area stretch a great deal, their ten-
sion increases a lot. The synthetic materials become stiffer, and you do not get
as much “pocketing” of the ball in the string plane or the same soft feel that
gut gives.

Chapter 30 explains how to measure the elastic modulus for various strings.

The thickness of a string (its diameter) is specified by a gauge number, where

a higher gauge number means a thinner string. Where almost all tennis meas-
urements and specifications are often given in English units (tension in pounds,
string length in feet, etc), the string diameter is traditionally given in millime-
ters (mm) rather than inches. So this means that a 16-gauge string can be about
1.30 mm in diameter and a 17-gauge string can be about 1.25 mm in diameter,
UNSTRETCHED. When under tension, the string stretches and becomes slight-
 Overview of Strings 245

ly thinner. (The amount the diameter changes for a change in length varies
according to the material. The ratio of how much the diameter changes for a
fractional change in length is called Poisson's ratio and it runs from about 0.1
to 0.3 for typical materials.)

As a general rule, the tensile strength of a given string is proportional to the

string’s cross sectional area or its diameter squared. You may think of the string
as made of fibers, the number of which goes as the area of the string and the
strength also goes as the number of fibers. Hence, the strength goes as the
diameter squared. However, since strings often have a unique construction, this
is just a rule of thumb.

The elasticity of a string is proportional to the inverse of the string’s cross sec-
tional area. The thinner the string, the more it will stretch under a given load
and the more elastic it will be. Since elasticity is a desirable trait in a string, as
another general rule, thinner strings play better. This again assumes the con-
struction is the same, and the thinner string has not been produced by just
stretching a somewhat thicker string down to a thinner diameter.

The better playing string (control, comfort, power) is usually the string which
is thinnest, so it may break sooner. If breaking strings is a problem, then a
thicker string will often be advantageous.

_Wear and Breaka oo

How come some recreational players seem to be able to play with the same
set of strings for years, while other players cannot get through a match with-
out breaking a string at least once every set or two? There are a number of fac-
tors here that depend as much on the style of play as on the inherent string
properties. If a player does not hit the ball hard and does not use spin on the
shots, a set of strings may not break for a long time. On the other hand, a play-
er who hits the ball hard can easily break strings, particularly mains near the
top of the frame if they are worn or notched.

When a ball impacts on the strings near the frame, the strings tend to stretch
asymmetrically, and the percentage of string elongation on the frame side of
the impact is much greater than a typical impact near the center of the head.
This increases the tension in that short section of string a great deal, possibly
well beyond the tensile strength limit of that string, so it breaks. A good exam-
ple of this is a very hard serve, struck near the tip of the racquet or on a miss-
hit near 3 or 9 o’clock. The strings on the short side of the impact location are
fixed in place by the frame. The ball may penetrate the string plane by a con-
siderable amount, stretching the short lengths of string beyond their limit.

Another cause of premature string breakage is due to a particular type of wear.

If a player hits with spin, the mains saw back and forth on the crosses, notch-
ing the mains. This notching reduces the cross sectional area of the string capa-
ble of carrying a load, making it effectively a thinner, weaker string at the
notch. When sufficient string has been sawed away, the remaining string can-
not support the incremental tension induced by the impact of balls, so it fails.
If the courts have loose granules that get picked up by the ball, they may get
into the strings and then they tend to act as an abrasive in the sawing process,
which accelerates the breakage. Testing of string in wear chambers (where
strung racquets are hit by balls repeatedly) show that higher string tension may
246 Chapter 26

increase string life, compared to lower string tension and this seems counter-
intuitive. This result is because at higher tension there is less string movement,
hence less sawing, less wear, and less notching.
4

Racquet frames come with many different types of stringing patterns, ranging
from 45-degree strings to three-way stringing to a rectangular grid or fan pat-
tern of strings with various spacings. In each case, the manufacturer (or inven-
tor) is attempting to accomplish something. We will analyze only the rectan-
gular grid patters, since they are by far, the most commonly found patterns.

If the head of the racquet were rectangular or square, as opposed to what it

actually is — oval, round, or teardrop in shape - then all the main strings and
all the cross strings would be the same length. Since the string behavior is a
function of string tension divided by string length, this would make the analy-
sis much easier. However, since for most racquets, the strings near the periph-
ery of the head are shorter than the ones that are in the middle, they play
somewhat stiffer and with a bit less power. In addition, a string tends to play
stiffer near its end, as opposed to in its middle. This is a second effect that
tends to stiffen the string plane around its edges. To compensate for these
effects, the racquet designer goes to a non-uniform string pattern, with more
space between the peripheral strings than between the central strings.

one
There is an old adage in tennis that says, “String tight for control, loose for
power.” Is this true, just a perception that players have, or another of the ten-
nis myths that are not correct when put to the test? If you look at some of the
tennis magazines of 20 to 30 years ago, you would find just the opposite
advice. The logic on power went something like this: most big hitters string
their racquets rather tightly, so it follows that tighter strings give more power.
The logic on control went something like this: to control the ball better you
want it to spend more time on your strings, and lower tension leads to longer
dwell times. Hence the prevailing (and erroneous) argument was: string loos-
er for control, tighter for power. However, we still see advice that tells the play-
er to keep the ball on the strings longer for better control.

We now think we know better. Looser strings give you just a little more zip on
the ball and this has been confirmed in both laboratory testing and in theoret-
ical models. Where it is relatively easy to measure rebound ball speed in the
lab, control is another matter, and most (but not’all) of the data is anecdotal.
One prevalent theory is that when a ball strikes the racquet off of the princi-
pal axis, the racquet tends to rotate about that axis, and the ball comes off of
the strings at an errant angle. The looser the strings, the longer the ball dwell
time, the more the racquet will rotate while the ball is in contact, and the larg-
er will be the ball angle relative to the desired ball direction. This scenario has
been verified in lab testing by firing balls off center at racquets with different
string tensions. There was a correlation between string tension and the errant
angle the ball came out at.

The fact that the ball speed will be in error for off-axis impacts tends to com-
pound the problem.
 Overview of Strings 247

There is a second effect that can cause a ball to come off of a racquet’s strings
at an angle.

If a ball impacts on the strings at some distance from the center of the head,
the deformation of the strings is asymmetric. This asymmetry in string defor-
mation should cause the ball to rebound at an angle that depends on a num-
ber of factors. The higher the tension (the less the string plane deformation)
the smaller is this effect and the farther the impact is from the center, the big-
ger is the effect. It is the opposite direction from the direction the ball would
go if the rotation of the racquet is considered. This effect tends to make the
ball rebound back toward the center (a focussing effect), while the rotation
effect tends to make the ball rebound away from the center.

Since there is good experimental evidence for the change in rebound angle
due to the racquet rotation, this asymmetry of string deformation effect must
be small compared to that effect.

Rebound Angle
When a ball impacts at the center of the head, normal (perpendicular) to the
string plane, and the racquet is also moving normal to the string plane, the
ball’s exit or rebound angle will be approximately normal to the string plane.
When a ball impacts at the center of the head at some angle to the normal to
the string plane, the ball will rebound at some angle relative to that normal. As
a general rule, the angle of rebound and angle of incidence are NOT equal to
each other, as they are when light hits a mirror. You can calculate the rebound
angle, to a good approximation, by breaking up the problem into a ball veloc-
ity normal to the string plane and parallel to the string plane. You then solve
for the normal and parallel velocities after the impact, and then combine them
to get the rebounding angle. When you actually carry out the calculation, for
most ball impact angles, the ball goes into the rolling mode while on the
strings, so the final parallel component of ball velocity is about 0.60 times the
initial parallel component of velocity. The normal or perpendicular component
of ball velocity is given by the formula:

Vout = Ga * Vin + (1 + e,):V, (26.1)

where:

e, (apparent coefficient of restitution) is the ratio of ball rebound speed to

incident rebound speed for a free racquet as obtained in a lab test.

V, is the speed of the racquet head at the point of ball impact.

V,,, is the incident ball speed (normal).

Vout is the rebounding ball speed (normal).

The value obtained for v,,,, and 0.60v (parallel) are added as vectors and the result-
ant ball velocity and angle are obtained.
248 Chapter 26

Typical values of ACOR run from 0.5 (near the throat) to 0.20 (near the tip),
and typical racquet head velocities are often similar to incident ball velocities.
However, when the ball is hit hard, the ball tends to come off at an angle close
to the normal to the string plane. When the ball is not hit hard, the exit or
rebounding angle is close in magnitude to the incident angle. If the racquet is
not moving, then the rebounding angle (relative to the normal) can be greater
than the incident angle. This effect can be a cause of errors in ball placement,
since the angle of rebound from the racquet depends on how hard the ball is
hit.

Figure 26.3 shows the angle of the ball off of the racquet as a function of the
racquet head speed for an incident ball of 60 feet/sec at an angle of 20 degrees.

When the ball is not incident normal to the

string plane, it will come off the strings with a
Set side-spin (spinhing about a vertical axis). It is
3 possible to calculate the amount of spin on the
2 ball by taking the final parallel ball velocity com-
@ ponent (about 60% of the incident parallel ball
& speed) and dividing it by the ball radius. This
s will give you the answer in radians/sec. To con-
vert this to revolutions/sec, divide by 2m or 6.28.

Grommet
{02 20" 8 280 50 60 70 80 ee =
Racquet Head Speed (ft/sec) Grommets and grommet strips are placed in the
frame of a racquet to protect both the strings
Figure 26.3 and the frame itself. The grommets and grommet
strips are made from a very tough material
Ball Angle off of the Strings for Different Racquet Head which prevents the strings from breaking
Velocities. The ball is incident at 60 feet/second, at an through the thin-walled composite frame and
a of 20 degrees to the direction of the racquet nor- also prevents string breakage by softening the
bends and sharp edges where the strings go
though the frame.

Recently, the size and shape of the grommets has been changed by many man-
ufacturers, to make them bigger, rather than have them fit closely around the
string. What is the purpose of having enlarged grommets with extra clearance,
particularly in the string plane deformation direction (perpendicular to the
plane)?

It is felt by many that longer strings play better, feel better, and give more
power than shorter strings. However, the ITF in Rule 4 says, “The frame of the
racquet shall not exceed 12.5 inches (31.75 cm) in overall width. The hitting
surface shall not exceed 15.5 inches (39.37 cm) in overall length and 11.5 inch-
es (29.21 cm) in overall width.” This seems to put a clear limit on string length
(15.5 and 11.5 inches) until the manufacturers and racquet designers realized
that they could get around it — but only slightly around it. If they could sus-
pend the string from the outside of the frame, rather than the inside of the
hoop, they could effectively get an extra inch of string length in the crosses
and at least that much extra in the mains. This meant a redesign of the grom-
mets so that the string was free to move on the inner part of the frame, hence
larger grommet holes are used.
 Overview of Strings 249

This also reduces the number of broken strings due to ball impacts very close
to the frame. With the old, smaller grommet holes, the string could not deform
or deflect beyond the inner part of the hoop, so there were large stresses in
the string when the ball hit near the frame. With the larger grommet holes, the
string is not constrained by the inner part of the hoop, but the outer part, so
the stress on the string is reduced and there is less breakage.
 a oe

aan Sieh Ley

ae
eee ‘
&

8
=
3=
<oc

Where Do Baby
Strings Come From?

Tim Henman, Great Britain

By Crawford Lindsey

tring is different than most of the solid objects you encounter in A Hexagonal Crystal
your day-to-day life — especially metals. The primary difference Structure of a Metal
= in their properties lies in how their atoms are arranged. In metal,
the atoms are arranged in a crystalline lattice —- every atom has the
same neighbors in the same location as every other atom. And every
atom is bonded with all its neighbors (Figure 27.1).

String is made of different stuff. Instead of a uniform geometric struc-

ture, string is made up of an aggregation of lots of molecular chains in
various degrees of entanglement or alignment with their neighbors. And
the bond between neighboring chains is not as formal as between lat-
tice members in metals (Figure 27.2). Figure 27.1

In a crystal structure, every atom

If you apply a force to the metal, it will be distributed instantaneously has the same arrangement of
to all the atoms which will deform their positions until an equal force is atoms around it as does every
built up to oppose the external force. In the string, the force will not be other.
252 Chapter 27

Stringy as efficiently transmitted. Note the different ways these molecule

Structure of String chains can respond to a force: they can straighten, stretch, rotate,
bend, slip past one another, and break.

These chains are known as polymer chains.

Figure 27.2 What is a polymer? “Poly” means many and “mer” means unit. So
polymer means many units. “Mer” is short for monomer, or single
String is made of long chain mole-
unit. These units are groups of atoms (molecules) arranged in a
cules known as polymers. These mol-
ecules may be twisted, straight,
pattern that is then repeated end-to-end, over and over. So that’s it,
entangled, oriented and unoriented. a polymer is a molecule arranged in a long chain of atoms. These
A more uniform orientation means a chains can be several thousand units long and are thus sometimes
stiffer and stronger string. referred to as macromolecules or giant molecules. Though even at
that, polymer chains are still microscopic.
8
String is a plastic. Like all plastics, it has its beginnings in the
ground as oil and its chemical raw materials are by-products of oil
refining. These chemicals go to resin manufacturers (these may be
the oil companies or the likes of Dupont, ICI) who convert them
into resins. This is where the chemicals become polymers. This is
where atoms and molecules combine to become nylon, PEEK, poly-
ester or Kevlar®. Polymer resin (usually in the form of flakes, gran-
ules or pellets) is then processed (heated into liquid) and shaped
(extruded) into the filaments that make up
POLYMER strings. This too is done by the Duponts of the
It's A String Too! world, though larger monofilaments are some-
times extruded by the string manufacturers.
Repeating Unit

So let’s meet a polymer. Figure 27.3 presents the

simplest shape for a polymer (polyethylene).
There are a few things to notice about this thing
to make us feel more comfortable and familiar
with it. First, it is a thing. Polymer molecules have
@ shape, length, and weight. They are put together
with parts, just like anything else in the world.
1. Start 2. Break the 3. Chain reaction begins
These parts, likewise, have different relative
with this double bond
shapes, sizes, and weights. Some ingredients are
H oH H H H H|H H
I
c=C
I ed]
-c-C-—
I I I
c-c+C-C
| bigger and stronger than others.
| i] ies I I I ]
HH lal an H H]|H H
Monomer There is also an arrangement of parts. The parts
have to be put together in an exact order or you
i Carbon © Hydrogen Bond no longer have the same thing. And just like any
other building materials, you can’t be trying to fit
square pegs in round holes in order to stick these
Figure 27.3 parts together. As we will see, only certain com-
“Polymerization” = Fancy word for making long chain binations of atoms can bond together.
molecules. When the double bond in the starting
monomer is broken, a bonding site becomes avail- It’s no different than building a jungle gym in the
able on each carbon, setting up a chain reaction of backyard, and we just have to think in the same
polymerization — stringing together units into molec- way. Don't get scared by formulas. They just tell
ular sized strings. They're called polymer chain mol-
us how many 2 x 4s and planks we need. And
ecules.
when two pieces fit together we nail them
 String Chemistry 253

(bond) together. To build the whole structure (all the mole-

Chain Tinkering
cules in the string), we build the molecules to specification
Factors affecting string properties
and then fit all the molecules together in the right arrange-
ment and relationship to each other and bond them to each
1. The chemical ingredients determine the polymer.
other.
2. Average length and weight of the polymer chains,
called the degree of polymerization. As the chain
The molecule has some parts. In the polyethylene diagram, gets longer, it also gets heavier. The molecules are
the circles are atoms. The chain of carbon atoms is the back- not all of the same length, so the average is impor-
bone (carbon is the backbone of virtually all polymers). The tant. The longer and more uniform the molecules
hydrogen atoms off to the side are known as pendant or are, the stiffer the string. _
side atoms. It’s easy to see that if the backbone is strong and 3. The amount of chain entanglement which in turn
stiff, so will be the molecule. It’s also evident that the num- depends on length and shape. More entanglement
ber, size, and shape of the side groups will determine how means less slippage.
4. Arrangement and location of side groups affects
closely you can fit molecules together, how flexible they will
how molecules can fit together.
be, the amount of entanglement and how easily they can
5. Regularity of pattern of molecules relative to each
move around relative to their neighbors. Add them all up
other (crystallinity). More crystallinity creates more
and you have an idea what the string will be like. A stiff “structure.”
string, for example, will have a stiff backbone or have mol- 6. Closeness of packing chains together for stronger
ecules that fit together and are oriented into a stiff structure, bonds.
or both. 7. Orientation of the chains in the direction of
applied stress. The more molecules that are aligned
The properties of a string come down to the size, shape, parallel to the axis of the string, the stiffer will be the
length, arrangement, and bonding of molecules. string. —
8. Types and strength of bonds between atoms with-
in a molecule string and between molecules.

The properties of the string depend on how all

the atoms and molecules are bonded which
influences and is influenced by size and fit of
atoms and molecules.

Bonding is electric. Chemistry is sort of electric

YANKING THE CHAIN
tinker toys — building by matching positive and Polymer Responses to Tension
negative attractions together. It’s all about attrac-

“oyedeg ste?
tion and repulsion and reaching an equilibrium
where the energy between all the parts is at the
lowest possible balance point. This is the most
Relaxed
stable condition. When everything is in its lowest
energy state, it takes the most energy to break
apart the bonds. If it’s not in its lowest energy Chain Straightening
state, there is either extra attraction or repulsion
forcing the atom to behave one way or another 06969096
e GES eS Qtg0ee
— i.e., unstable — and you only need to add a Chain Stretching
little extra to nudge it into doing something. hk hh ak ek ek ed ane
Here’s the key: bonds between atoms on the
te ek bk ok ek he
Chain Slipping
polymer chain are stronger than bonds between
two separate chains. It’s like a plate of spaghetti.
If you pull on one strand, first the loose end
straightens, then stretches slightly, and then it Figure 27.4
begins to slip its way out of the plate (Figure
27.4). This happens because the bonds along the Chain straightening and stretching are elastic phe-
nomenon. Chain slipping is plastic involving perma-
spaghetti are stronger than those between two
nent deformation.
separate pieces. At some point, the amount of
254 Chapter 27

entanglement may grab the spaghetti, and further pulling

BONDS stretches and finally breaks the piece. This is similar to
Making Baby Strings what happens in a string when it is stretched. The string is
the mound of spaghetti and the molecules each individual
noodle. It’s up to the chemist to control how much of what

OO |
re) (H> happens when.

Bonding ona Chain

Understanding why the bonds along the polymer chain are
stronger than those between chains is helpful in visualiz-
ing why and how a string behaves. The bonding behavior
of atoms depends on the behavior of the electrons in its
outer shell. Briefly, the atom is composed of a nucleus
made up of positively charged protons and an equal num-
Figure 27.5 ber of neutral neutrons, An equal number of negatively
charged electrons orbit around the nucleus in shells. There
In string, molecular chains are held together by
atoms sharing electrons. This is known as a cova-
are a maximum of 7 shells and a maximum of 32 electrons
lent bond. Carbon needs 4 electrons to complete
in a shell. The inner shell has a maximum of 2 and the
its outer shell, hydrogen one. Each bond repre- outer shell 8 electrons. The number of electrons in the
sents two electrons — one from each atom. By outer shell is what determines an atom’s bonding behav-
sharing, each atom completes the other, thus ior. The most stable, lowest energy state for an atom of any
forming a very strong bond. element (except hydrogen and helium) is to have 8 elec-
trons in the outer shell. The goal of an atom’s life is to
complete its outer shell. In a polymer, this is done by shar-
ing electrons (Figure 27.5).

This sharing bond is known as a covalent bond. “Co” means sharing, and
“valence” refers to the number of electrons an atom must share to complete its
outer shell. Carbon always has 4 sharing bonds, hydrogen 1, nitrogen 3, and
oxygen 2. So in diagrams, you'll always see that many lines (bonds) joining the
respective atoms. If you don’t, a shorthand notation is being used that leaves
out some hydrogen atoms.

Sometimes an atom (especially carbon) shares multiple bonds with another

atom. In that case, there are two lines between the atoms. In the instance of a
double bond, the atoms would be sharing 4 electrons — two from each atom.

So these molecular-sized strings are made up of atoms along the chain sharing
electrons. These are very strong bonds.

Bonding Between Polymer Chai

Attraction between two separate polymer chains is a different story. It’s not so
much bonding as it is a little hanky panky on the side. When atoms share elec-
trons, the sharing isn’t always equal. Depending on the relative sizes of the two
atoms and their respective tendencies to attract electrons, the electrons from
one atom may tend to spend more time around the other. So even though the
bonding arrangement as a whole is electrically neutral, it is not equally neutral
all over its surface. Sort of a miniature magnet is set up with a positive and
negative end. Figures 27.6 and 27.7 show how the positive and negative ends
of bonds on separate chains attract each other.
 String Chemistry 255

BONDS Extra Molecular Affairs

Unequal Sharing...

N —
i
C-N-C—-C-C
i
ed e tia haar
H H H_O O
cet
Im dnl ial © fF) ©
Nylon | | Th, lp
Chain #2 —-N-C-N-C-C-C
| |
H H

..Leads to... —>

Figure 27.6 Figure 27.7

Hydrogen's shared electrons spend most of their Polymer chain atoms are promiscuous. This
time hanging out in the space between its nucle- may not be moral, but it makes for good string.
us and the carbon atom. This leaves the hydro- These secondary forces hold chains in place.
gen nucleus with its positive proton sticking out
looking for a good time — the negative end of a
polar bond on another chain.

Hydrogen, for example, only has 1 electron, so when it is in a covalent bond

it becomes very polar as its electron hangs out closer to the atom it is bonded
with, leaving an “exposed” positive back-end. These kind of forces are known
as secondary forces and this type of polar hydrogen bond is the strongest of
these. So chains with a lot of hydrogen atoms as side groups tend to have
stronger secondary bonds keeping the chains in alignment. This is true of
nylon strings.

Okay, so we want to start bonding a

bunch of stuff together to make a string.
To do this, just visualize taking our simple KEY PLAYERS IN STRINGS
polymer in Figure 27.3 and imagine sub-
stituting different configurations of atoms Or rH 7 C=C \
tot =O) = C6 — =e C=
in place of some of the hydrogen atoms
or added into the backbone. There are mCoNG Sisiaue
several of these configurations called Amide Ether Aromatic
functional groups that appear in string
materials, and these typically dictate the O O
string’s properties. Consequently, their ll ll me ae
-C-O-C- -C-
group name usually appears in the mate-
rial’s name. For instance, if an amide Ester Ketone Aromatic (view #2)
group is in the repeating unit, it is a
polyamide. Some of these groups appear Figure 27.8
in Figure 27.8. Entire classes of materials are defined by these functional groups.
They may appear in formulas in different orders than above, but
Molecules that have simple linear back- they are all equivalent. Remember, carbon must have four bonds,
bones are called linear molecules and so where you don't see four lines, the hydrogen atoms are left out
those that have rings are called cyclic or for simplicity.
256 Chapter 27

aromatic. Generally speaking, a ring

STRING FORMULAS in the backbone will make a mole-
cule stiffer and stronger than a linear
H /H\ H O /H\ O molecule. PEEK and all the aramids
[ee ek aalcetimeetiony abl: 0) ait like Kevlar® have these rings. The
N+ C7N-CHCFE complete formulas for string materi-
4 :b als are presented in Figure 27.9.
: p These formulas represent the repeat-
Polyamide: Nylon ing units that are strung together to
make the molecule chains.
= = H
« cer S | oll é pene If you look at these formulas as if
ens ok oY hae os ee 7 they were physical structures made
on oo n of 2x4’s, and if you squint, you can
Aramid: Kevlar visualize how stiff or flexible that
stlucture will be just by its layout. A
rf ‘ ‘ / Ca \ fn few highlights are worth pointing
O C-C O-C-C C= out.
| | S Ce & “7
H H/m n
Nylon: Nylons as a group are
Polyester: Polyethylene terephthalate (PET) referred to as polyamides. The C—O
and N—H bonds in nylon are polar
C=C C=C O C=C with the O having a negative end
‘ a : ) ca e and the H having a positive end.
or a Gate ie Sa oreeS Ne pe Ss NS ee “4 e These polar side atoms attract their
opposites in neighboring molecules,
tending to increase stiffness and ten-
Bee el eee tele one sile a These secondary bond-
ing sites combined with the symme-
aiguieie’2 try of the polyamide chain allows
The subscripts a and b in nylon are changed to 6 and 12 accordingly these molecules to align themselves
for nylon 6/6 or 6/12, etc. Likewise with polyester. Changing the m side by side or neatly fold back upon
subscript results in a different polyester. The n subscript indicates the themselves in a very close-packed
number of times everything in the brackets will be repeated. fashion.

These regions of regular alignment are known as crystalline regions. The

degree of crystallinity also increases strength and stiffness. Because of the
length and different shapes of polymer molecules combined with only weak
secondary forces to hold molecules in position, you'll never get 100% crys-
tallinity in polymers. Sixty to seventy percent is high. So most materials have
crystalline and amorphous regions mixed together.

Nylon also comes in many varieties. There is nylon 6, nylon 6/6, nylon 12, and
nylon 6/12, among others. This differentiation is the number of carbon atoms
between (and including) the amide groups. Two numbers indicate that two
monomer groups came together with a different number of backbone atoms
and were linked with the amide group. This is indicated in the nylon formula
by the a and b subscripts.

Kevlar: Kevlar is a strange animal. You'll notice that it is both a polyamide and
an aromatic. Hence the name “aramid” which is a shortening of “aromatic
polyamide.” It is not considered a nylon, however. Its strength and stiffness
come from the aromatic rings. These rings are stiff not only because of the
shape but by the nature of the bond. All six of the carbon atoms share the elec-
 String Chemistry 257

trons of the double bonds. These electrons travel around the circle holding the
structure together in a very strong, stiff bond.

A copolymer refers to the structure of the chain itself. The chain of a copoly-
mer has more than one repeating unit within the polymer chain and is creat-
ed during the polymerization process. Depending on the design of the copoly-
mer chain, you will achieve an average of the properties of chains formed from
the individual repeating units or even a synergistic effect that can boost one or
more of the properties.

Sometimes, different polymers are blended together in the polymerization

process. These materials are known as polymer blends or alloys. Each chain of
the different polymers retains its unique configuration during the blending
process, and secondary bonds form between the different chains to create the
polymer blend material. Depending on the materials and how much of each
is blended, you will achieve an average of the properties of the different mate-
rials.

How does the behavior of these polymer chains show up in the whole tennis
string? String is schizophrenic. Sometimes it acts like an elastic solid, like when
it is being hit by a tennis ball. And sometimes it acts like a slow moving bot-
tle of catsup, as it does if you hang a weight from one end and it continually
elongates, or when it loses tension sitting in a tennis racquet. Hitting a tennis
ball is instantaneous, and sitting around in a tennis racquet is time dependent.
This behavior is known as viscoelasticity — behaving both as a viscous liquid
and an elastic solid.

This difference is important in string, but not in metal. If you hang a weight
from a metal string, it will stretch just enough to exactly counter the amount of
the force. It will maintain that length (with virtually no creep) until the force
is removed, whereupon it will instantly return to its starting length. A nylon
string, on the other hand, will continue to get longer indefinitely. When the
force is removed, it will recover all or most of the elongation, but it too will
take time.

But if you impact both with a tennis ball, they both instantaneously absorb and
return the energy of the impact force. They respond the same to a fast force
but differently to a slow force. Why?

Consider a ball of Silly Putty. If you throw it on the floor, it will bounce. It will
act elastically to the force of impact. But if you just lay in on the floor for a
couple of hours, it will flatten out under the force of its own weight.

The reason for these different behaviors goes back to the difference between
the uniform pattern of atoms in a metal crystal compared to the entangled mass
of polymer chains. Upon impact, polymer chains straighten in the amorphous
regions and stretch in the crystalline regions. These are instantaneous and
recoverable reactions, just like the behavior of atoms in a completely crystalline
metal structure. But in a string that is under constant tension over time, chain
slippage and disentanglement occur. That’s why the string constantly elongates
258 Chapter 27

with a steady force or the tension constantly relaxes in a strung racquet. The
chains have time to rearrange themselves and move relative to each other. The
secondary forces between chains are not strong enough to prevent this, where-
as in a metal, everything is linked by primary forces.

As you can see, there are a lot of variables that can be affected to change the
properties of string. String manufacturers must calculate how much work and
money is worth tinkering to get 3, 4 or 5% improvements. And typically the big
quantum leap improvements in materials occur when an application in an
industry much larger than tennis justifies a huge development expense. Then
that material can trickle down to other applications.

It's a mind-boggling task when you consider that you can take a formula and
organize the atoms in it into different structures called isomers (same formula,
different structure). A formula like C4gHg9 can be arranged in millions of dif-
ferent structures. What does it cost to tinker and fine tune in this manner? What
will it cost the consumer, and how much is each percent improvement worth
to the consumer and what percent is necessary for him to notice?
 Performance
Properties of
Synthetic Materials

Jennifer Capriati, United States

By Crawford Lindsey

hoosing a string can be a bewildering affair. There are bazillions of constructions and materials, but when
you get right down to it, to the untrained eye, most strings look alike. You can’t point to features on a
string and say, “See this thingamajingie...it increases control (or dampens vibration or whatever).” String
features are invisible. That’s why you can say almost anything about a string and sell it — once. But when an
irate customer comes back to hang you with that string, a fancy jitterbug is no substitute for basic knowledge
as to why certain materials actually behave as they do (see Note 1).

at Materials?

The glamour area of leading-edge materials technology is in the stiffer, lighter, and stronger materials.
Development for applications requiring these properties make headlines. This is great for making racquets, but
what about string? Most of the new post-nylon materials are only used for extreme circumstances like excessive
string breakage. The bulk of synthetic string material is still nylon. What are these other materials and how do
they compare to nylon?

The materials most commonly used in string are detailed in Table 28.1. Materials in the same family tend to
behave alike. The different names within a family are merely different licensed brand names for the same mate-
260 Chapter 28

rial or patent work-arounds for similar but propri-

eee. Material Families etary branded material. Each product may also
Material Company Family have several application-specific versions with dif-
Nylon Dupont/All Polyamide
ferent strength, stiffness, ultraviolet sensitivity,
abrasion resistance, temperature behavior, creep
Polyester DuponvaAll Polyester
resistance, etc. So, referring to a material just by its
Kevlar Dupont Aramid
name can do it an injustice. It depends on what
Twaron Akzo Aramid
variety you are working with in what situation. For
Tecnora Teijin Aramid
example, there might be better or worse formula-
Vectran Hoechst Celanese ECRs
tions of polyester for a specific player type.
Spectra Allied Signal UHMWP*
Dyneema DSM UHMWP
How are all these materials different from each
Zyex ICI PEEKS
other in strength, stiffness, and play?
Note:
1. LCP = liquid crystal polymer
UHMWP = ultrahigh molecular weight polyethelene
PEEK = polyetheretherketone
2. Each family is a different way of aligning and bonding Strength is the force necessary to break a square-
long strings of molecules in straighter, closer packed, and
longer parallel chains. The more this is done, the stiffer the
inch of a material while pulling along its length.
material. The chart at right ranks the materials in terms of
tensile strength.

The strength values are big numbers. What do they

Table 28.2 | Strength — mean relative to a benchmark that we can relate to?
Table 28.2 indicates each material’s strength com-
pared to nylon. Carbon and steel are also listed for
Material Strength Ib/in? (000s) Times Nylon
an additional frame of reference.
Carbon 518 3.6
Tecnora 498 35
Dyneema 497 3.5 What IsStiffness?
Kevlar 424 3.0
Spectra 422 3.0
We think of elasticity in terms of how much a mate-
Twaron 421 2.9
rial stretches. The force resisting stretch is the elas-
tic force. It competes against the stretching force to
Vectran 412 2.9
return the material to its original shape. The meas-
Steel 399 2.8
urement of this elastic restoring force is called
Nylon 143 1.0
modulus. This is the same “modulus” that we
Polyester 140 1.0
encounter in racquets. Modulus is defined as the
Zyex 115 0.8
tension (force per square inch) divided by the
strain (elongation divided by original lenght).
“Modulus” and “stiffness” are used interchange-
eee Stiffness— ably, so the higher the modulus, the stiffer the
material. The stiffness chart (Table 28.3) ranks the
Material Stiffness Ib/in? (000,000s) Times Nylon common string materials and compares each to
Carbon SOLS 48 nylon. The stretch chart (Table 28.4) indicates how
Steel 28.9 A much each material stretches at the breaking tension.
Spectra 16.4 23
Dyneema 16.2 23
Twaron 11.9 17
Technora VOxS) 15 Breaking strength tells us how a string material
Kevlar 10.3 15 dies; stiffness tells how it lives. Stiffness describes
Vectran 9.4 13 how a material behaves under tension. The tool
Polyester PX 3 most often used to show the relationship of
Nylon 0.7 1 strength and stiffness is the “tension/stretch curve”
Zyex 0.6 0.9 (Figure 28.1). This curve shows the “lifeline” of a
material by plotting tension against the percentage
 String Materials 261

of stretch at each tension when being pulled at a con- Table 28.4

stant rate. Sounds boring, but check it out...it has very
interesting implications. Stretch at Break

The slope of the curve is the stiffness. For materials like Material % Stretch
steel, the curve is virtually a straight line at normal Zyex 34.0
stresses, so the slope is the same everywhere (curve A Nylon iKe).7
in Figure 28.1). In that case you can say the material has Polyester 14.9
definite stiffness, and you know how it will behave in
Technora 4.6
all circumstances. But for most string materials, the
Kevlar 3:6
curve changes along its length (ike curve B or C), and
Twaron 3.6
so, therefore, does its stiffness. Consequently, stiffness is
not an absolute term when it comes to string materials. Dyneema 3:5
Vectran 3:3
Instead, string materials behave differently at different Spectra 2.9
tensions. When you say a string is stiff or elastic, it may Carbon Ue
depend on what tension you are talking about and what Steel 1.4
other string you are comparing it to. Look at the differ-
ence in the two fictional string materials displayed in
Figure 28.2. String materials very often have “S-shaped” Figure 28.1
curves. The slope, or stiffness, at any point along these
two curves would indicate something quite different vis- Tension/Stretch Curve
a-vis each other. Each would be considered stiff or soft
x Strength =
compared to the other depending on the tension.
Breaking Point
Strength and stiffness combine to create some very inter-
esting behaviors (Figure 28.3). Some materials are strong
and stretchy and some strong and stiff. The first are
Possible
tough, the latter brittle. “Brittle” means that it breaks eas-
Shapes
ily on impact. But how can it break if it is so strong? (Ib/in2)
Tension Slope indicates Galore
After all, we use Kevlar, which is strong and stiff (and
behavior between
brittle), to prevent breaking. What’s the deal? two tensions

An example of a boat tied to the dock by a long rope

Percent Stretch
helps to explain this paradox. The rope’s job is to bring
the boat’s velocity to zero as it floats away. The boat

Figure 28.3
Figure 28.2

Stiffness Changes w/Tension Strength and Stiffness

Strong & Strong &
Stiff Stretchy

Both at same tension but

material B has stretched
more getting up to that
tension. But at tensions
(Ib/in2)
Tension directly above the "Xs,"
the slopes change and Weak &
material A stretches more (Ib/in2)
Tension Stretchy
than B. Therefore, which
material is stiffer depends
on the tension.

Percent Stretch _ Percent Stretch

262 Chapter 28

drifts until the line becomes taut. If the rope is very stiff, it won’t stretch and
the boat comes to a screeching, instantaneous halt. If it is stretchier, it will ease
to a leisurely stop. The force exerted on the rope by the boat (and vice versa)
is equal to the boat’s mass times its acceleration (deceleration). Since the boat
weighs the same in each case, the force will be greater when the deceleration
is greatest — with the stiff rope. The stiffer rope has to be stronger because it
creates higher forces. If it isn’t, it will break while the “weaker” one does just
fine.

Sometimes stronger is stronger, and at other times, weaker is “stronger.” It

depends on the situation. Kevlar is used to prevent string breakage because its
strength is so high compared to nylon’s that it makes up for its lack of stretch
and the higher forces it incurs. This is a classic case of property tradeoffs when
comparing materials.

If the boats above used a long stiff rope and a short stiff rope of the same mate-
rial, the longer one would stretch more (because there’s more of it, but it
stretches the same percentage amount) and thus create less force. The short-
er one might break while the longer one holds firm. This applies to racquet
strings too. There is less impact force felt on longer strings than
Figure 28.4 shorter ones. Thus, you might be able to get away with a stiffer
material on a center main, but not on a shorter side main.
- Breaking Energy, -
Strength & Stiffness The principle is the same as catching a baseball. If you move your
hand backwards while catching, you lessen the force because the
Strong & Strong &
(o-ay| Stiff Stretchy time of the impact becomes longer as it is spread over a greater
N
= distance. A string stretches to do the same thing. The energy is
a
a
absorbed, the force lessened, and dwell time increases. Figure 28.4
2
c demonstrates how much energy it takes to break the four fiction-
”
al materials illustrated in the figure. The area under each curve is
- 5 the energy needed to break each material. It is apparent why the
strong brittle material might break on impact. If both the strong
materials have the same breaking strength, the stretchy one would
require more energy to achieve breaking force because more ener-
gy goes into stretching as opposed to increasing the force, as it
Percent Stretch does in the stiffer material. Depending on the impact energy, a
material may or may not build up enough force to break.

Strength and Stiffness byWeight |

Table 28.5

Density Ever hear the brag, “Strongest substance in the universe by

weight”? Being strong and stiff is one thing, but how much “stuff”
Density (Ib/in3) does it take to be that strong? That’s where a material’s density is
Material
important. Density is the material’s weight per cubic inch. The den-
Steel 0.282
sity table (Table 28.5) shows the densities of several materials.
Carbon 0.064
Some string materials are definitely heavier than others.
Kevlar 0.052
Vectran 0.051
If you are looking for the most strength or stiffness for the least
Technora 0.050
weight, you look at the “specific strength” and “specific stiffness,”
Polyester 0.050
which is strength or stiffness divided by density. These are your
Zyex 0.047
ultimate strength and stiffness values. Each material’s density and
Nylon 0.041
specific properties are presented in Tables 28.6 and 28.7. Using
Dyneema 0.035
steel as a value of 1, other materials compare quite favorably. All
Spectra 0.035
materials listed are stronger than steel by weight, although they are
not all stiffer.
 String Materials 263

Strength by Weight Stiffness by Weight

Material Strength (in.) (000,000s) Times Steel Material Strength (in.) (000,000s) Times Steel
Dyneema 14.2 WE5 Carbon 708.6 (of)
Spectra 12.0 6.3 Spectra 467.7 4.6
Technora O29 Ore Dyneema 460.6 4.5
Kevlar 8.2 4.3 Technora 209.0 2.0
Vectran 8.2 4.3 Kevlar 196.6 1.9
Carbon 8.2 4.3 Vectran 186.0 1.8
Nylon 3:5 1.8 Steel 102.7 1.0
Polyester 2.8 125 Polyester 41.8 0.4
Zyex 2.4 1.3 Nylon 16.1 0.2
Steel ew 1.0 Zyex 11.9 On

Table 28.6 ; Table 28.7

So, just throwing out a wild, random example, a string made with just a little
Dyneema (first on the strength by weight list) and a lot of Zyex Cast on the
list, not counting steel) might be as strong as one made of all Kevlar (middle
of the list). Looking at specific properties (properties by weight) like this gives
a good indication of why certain combinations of materials have been used in
a string to achieve a specific playing characteristic.

That brings us back to our original question — why is nylon still used despite
all the high tech materials? For racquet sports, it has the best combination of
stretch and strength which leads to the best combinations of power, control,
comfort, and durability. The higher tech materials are usually much stiffer at all
tensions. Stiffness works for racquets but not for most strings. That’s why so
many constructions are designed to provide for a softer feel, even if new mate-
rials aren’t coming along to do the same thing.

Notes

Note 1. Some caveats: There is a difference between material behavior and the
behavior of a structure made of a material. The structural properties of a rac-
quet or string depend both on the material used and how it is used. This chap-
ter deals only with material. Second, the data accumulated here are obtained
by static testing — constant rate pulling on material — and are the types sup-
plied by chemical companies. But string is used in dynamic impact situations.
These dynamic data are not readily available, so predictions have to be made
from the static data. Third, only the standard modulus of materials are com-
pared here. Some materials have dozens of varieties with different customized
’ properties. These could flip flop relative positions on a chart, but they aren’t
important to understanding the big picture.

Note 2. All data were obtained from material supplier specifications.

 as
&
=
c
S
as

How a Synthetic
or Gut String Is Made

Yevgeny Kafeinikov, R

By Crawford Lindsey

~ here’s good news and bad news. The good news is, of the hundreds of strings on the market, there real-
ly may be some difference between them all. And the bad news ... well, it’s the same as the good news.
™ How can that be? Aren’t there just a few basic constructions, like solid core with one, two, or three wraps,
or multifilament, or monofilament like a polyester? That’s true, but within each construction the combination of
materials, diameters, wraps, twist angles, resins, coatings, and sundry other considerations give every string dif-
ferent properties. How the string is constructed makes a big difference. A string is the sum of its parts. Whenever
you add one fiber to another, you always get more than you started with. More strength and more stiffness.
Whether you get “less more” or “more more” depends on the composition and orientation of the added mate-
rial. For example, adding an elastic wrap does not make up for the stiff core and make the string more elastic.
It actually adds more stiffness overall. It does, however, make the total string more elastic than if you added a
stiffer wrap. Either way, the sum will be stiffer and stronger than either part by itself. And whatever you end up
with, you are getting it with the qualification “All other things being equal.” Its important to always qualify
“more” and “less” with “compared to what.”

Core Princip

The Core Dictates. First, consider the core. In a string of a given material, the core will dictate performance
because the core is generally the fiber with the greatest diameter. The thicker a fiber, the stiffer, stronger, and
266 Chapter 29

Natural Gut less elastic it is, and the better it will maintain tension (not all the
Construction and Play time — see Chapter 35). It’s the work horse of the string. And by
being so dominant, it also limits the potential contribution of any
By Rod Cross other element’s properties. As the core goes, so goes the
string.The least elastic fiber in a string will also dictate perform-
Why Players Like Natural Gut
ance. The analogy of a stringbed is a perfect example. No matter
how long the main strings might be, if the crosses are very short,
atural gut is still the preferred string of
N many professional players since they like
stiff, tight, and close together, they will not allow the mains to do
the way it feels and plays. Exactly what their thing. It’s the same with string construction — a stiff fiber
that means is hard to pin down, but one out- wont allow the elastic one to stretch as far as it would by itself.
standing feature from laboratory tests is that nat-
ural gut has a smaller dynamic stiffness than any So, in a solid core single wrap string made of only one material
other type of string. All natural gut strings share like nylon 6/6, the least elastic part would be the thicker core. But
this property, regardless of which manufacturer if it is a composite string (made of two or more materials), the
makes it. Most nylon strings stretch further than stiffest part could be either the core or the wraps, regardless of
natural gut when they are installed in a racquet, diameter. The performance of a string with a stiff core and elastic
but the initial part of the elastic curve is irrele- wraps will be determined by the core. But a string with an elastic
vant. The relevant thing is how far the string
core and stiff wraps is more complicated. It depends on the mate-
stretches and how much the tension increases
rial, diameter, and orientation of the wrap. A very stiff material can
when it is used to hit a ball. For example, sup-
pose that a 12 inch length of nylon stretches by
be made thin enough to be as elastic (or nearly so) as a more elas-
1 inch when it is installed in a racquet and ten- tic core material. Or it can be wrapped at an angle to increase its
sioned to 60 Ib. When used to hit a ball at 120 elasticity (more on that later), or both. Thus, there are more
mph, the string will stretch further, by about 0.2 options starting with an elastic core.
inch, and the tension will increase by about 40 Ib
during the impact. The string returns back to its Material vs. Diameter. Even so, the fiber’s material composition
normal stretched length and normal 60 |b ten- is usually more dominant than its diameter. The elastic properties
sion when the ball leaves the strings. The extra
of two fibers must be pretty close for changing diameters to cause
0.2 inch stretch lasts only for the time the ball is
their properties to add together instead of one limiting the other.
on the strings, but that's the important part.

A 12 inch length of natural gut will stretch by Material vs. Construction. Materials are also more important
about 3/4 of an inch when it is tensioned to 60 Ib. than construction. What a string is made of is more important than
Gut is less elastic than nylon at tensions up to what it looks like. No matter how you combine nylon elements,
about 50 Ib. But it is more elastic than nylon at you're not going to get gut. Single wrap, double wrap, braided,
tensions above 50 Ib. When used to hit a ball at multifilament — nylon is still nylon..
120 mph, a 12 inch natural gut string will stretch
further, by about 0.22 inch, and the tension will Materials contribute more to playability, whereas construction has
increase by about 20 Ib during the impact. The a greater impact on durability. That’s why wrap “construction” was
extra stretch is only slightly more than with
invented in the first place — to protect the material in the per-
nylon, but the increase in tension is a lot less.
formance core. And that’s where the magic formula comes in —
This has a significant effect on the softness of
the impact. The string tension increases during to make all the elements additive and complementary while still
an impact from 60 Ib to a maximum of about 100 getting all the desired characteristics.
Ib in a nylon string (when a ball is hit at about
120 mph). With a natural gut string, the tension Strength vs. Elasticity. Strength and stiffness are always additive.
increases from 60 Ib to a maximum of only 80 Ib. Every element that contributes to the cross section is adding
Since the average tension during the impact is strength, regardless of its other properties. For example, if the core
smaller with gut, the impact is softer and the has a tensile strength of 100 pounds and the wraps all add up to
average stiffness of the string plane is lower, as 100 pounds, then the string’s total tensile strength will be 200
if the racquet was strung at a lower tension. pounds. But if string is as strong as the sum of its parts, it's only
as elastic as its least elastic part. And the least elastic is usually the
Another property of natural gut is that it holds its
strongest. So what limits one characteristic augments another.
tension better than most other strings. If one
racquet is strung with natural gut and another is
strung at the same tension with a nylon or poly-
ester string, then after several days or after sev-
eral sets, the natural gut string will still be close
Wraps were invented to protect the core — especially from notch-
ing (the biggest culprit of breakage). A thick monofilament mate-
 String Construction 267

rial with a notch can be weaker than a thinner fiber without one. to its original tension while nylon strings will
If an outer fiber of a wrapped string notches and breaks, the fail- have dropped a few Ib in tension and polyester
ure doesn’t propagate through the rest of the string like it would strings even more. For a professional player,
if it were a monofilament. The outer layer localizes or isolates the this is important. For an average social player,
crack. this is not so important. In fact, social players
may prefer nylon or polyester since a loss in ten-
sion results in a softer feel. A racquet strung with
Single wraps predominate, but double and triple also exist. For a
natural gut may feel too stiff for a social player,
given gauge, the more wraps you have, the smaller the core they
especially if it is strung at the same high tension
are protecting (unless you make the diameters of the wraps small- used by professionals.
er instead). But then you have a weaker core surrounded by indi-
vidually weaker protectors. Now the wrap strategy becomes Why Cows Hate Natural Gut
strength in numbers. “You can get some of us but not all of us.”
Natural gut strings are more expensive since
But with more elements you also get more space between fibers, they are not as easy to make as _ synthetic
so you have less fiber material. And with less fiber in the cross- strings. As the name implies, natural gut is made
section (the rest is filled with resin), you have less strength. So, the from a natural product, namely the intestines of
cows. The small intestine of an animal is essen-
strong core that we originally set out to protect has become weak-
tially a long flexible tube designed to help the
er, the protectors are weaker, and all of them added up are weak-
animal digest its food. It is naturally quite elastic
er than we started with. But a funny thing happened on the way
since it needs to expand and contract to function
to the string factory. As more and more layers and fibers are added properly. It takes about three cows to make a
in an attempt to protect the core, strength goes down but playa- single tennis string. Not because the intestine is
bility goes up. With more and more parts, the string becomes soft- too short but because the part used to make a
er, more flexible, and more playable. Taken to the extreme, we string is very thin.
end up with a multifilament. Multifilaments thus have less tensile
strength, but perhaps are more durable in terms of longevity. A Many people think that natural gut is made from
single fiber will break all at once, a multifiber a little at a time. cats. The small intestine of a cat is only 4 feet
Which one goes first depends on what kind of forces each long and is therefore too short to make a tennis
string. The word "catgut" seems to have evolved
endures.
from the use of natural gut in a musical instru-
ment called a "kit" or perhaps from the name of
Twists the town in Germany where the strings were
made. The intestine of a cow or sheep is about
Twists and Durability. This is where string construction gets fun, 120 feet long, enough to string three racquets.
and where a string can really be fine tuned. Wraps are twisted The part of the intestine used for making tennis
around the core, and multifilaments are all twisted around each strings (or for making sausage skins from sheep
other. What is the purpose of the twist? If the wrap fibers are all intestines) is the outermost layer called the
serosa.
oriented in a straight line along the axis of the core (and some-
times they are), they will contribute the most to tensile strength
The serosa is removed and cut into long ribbons
(the breaking strength when pulling along the string’s length) and
which must then be cleaned in a series of salt
stiffness. But if one of the fibers breaks, that fiber loses its contri- and other chemical baths over several days. A
bution to the string’s properties. If the fiber is twisted around the sufficient number of ribbons (about 18) is
core instead, its tensile strength will be less (because it’s no longer assembled and twisted as a long string and
oriented in a straight line), but if it breaks, the fibers on either side dried under tension over several days in a tem-
of the break are held in place by the friction and braiding effect perature and humidity controlled room. The
of neighboring fibers in the twist. A good percentage of the fiber’s string must be thoroughly dried and not just
strength will be lost, but some is retained due to the interaction dried on the outside. After drying, the string has
with fibers around it. the appearance of ordinary string or rope. The
string is then polished using a fine grit sanding
machine to form a smooth, round and clear
Twists and Elasticity. Twisting also builds elasticity into the con-
string. Polishing may involve rotating a long
struction. In effect, it makes the string more spring-like. For exam-
stretched string while the polishing machine
ple, a straight metal wire might have a stretching stiffness of thou- trevels along the length of the string or it may
sands of pounds per square inch. If it is coiled into a spring, the involve passing the string through a small pol-
force required to stretch it will be much less. It’s the same with ishing head. The final process is to treat the
string — twisted wraps take less force to stretch than straight line string with a protective coating such as
wraps. polyurethane to improve abrasion resistance
and to prevent moisture entering the string.
268 Chapter 29

Twist Angle and Elasticity. The angle of the twist is also important. The
tighter the wrap (the greater the angle compared to the core), the more elastic
it will be. But it will also have less tensile strength — the greater the angle, the
more of the load the core will have to bear. But as the string loses tensile
strength at increased angles, it gains in hoop (or bending) strength which
comes into play on knots and turns between grommet holes.

Twist Angle and Energy Return. A tighter wrap may also diminish the
string’s energy return. This is because the very spring-like action that makes the
string more elastic can rob it of resiliency. The force of ball impact is absorbed
spirally which makes it soft, but is given back the same way, which limits the
unified directionality available for the energy return. In a straight line of fibers,
all the fibers are moving in the same direction. With twisted fibers, each fiber
is moving in somewhat different directions relative to each other and to the
core. Not only does this disperse the energy but it also causes more friction,
which uses up even more energy.

Twist Angle and Fiber Count. The angle of the twist also determines the
number of fibers you can fit on any wrap layer. The higher the angle, the fewer
fibers of a given diameter that can fit. But whether it is twenty fibers or one
fiber, it is the same amount of material in either case. For example, imagine a
core with a six-fiber-wrap arranged in a straight line. As the angle of that wrap
is increased, the orientation of each fiber to the cross section becomes greater
and the cross section of those fibers become elliptical. The elliptical cross sec-
tion becomes wider with every increase in angle. As it gets wider, fibers must
be eliminated to fit the circumference of the string. Ultimately, as the angle
approaches 90 degrees, you are left with just one fiber wrapped perpendicu-
larly around the core. But though you lost fibers along the way, you had to
increase the length of the remaining fibers to account for the tighter wrap. So
you end up with the same total amount of material but fewer fibers and much
different properties. Basically, the volume is being filled in a different configu-
ration, but it is still the same volume and thus the same amount of material.

Bonding. Fibers get all the glory, but resin is the unsung hero. Without resin,
any fiber is too limp to string. The resin bonds all the fibers together and gives
the string its body and shape. Without the resin, a string would be just like a
rope or twine. Resin is typically applied to every fiber at the intersection where
they all come together from the bobbins to be twisted onto the core. If a fiber
is not “wetted,” premature fray, breakage, tension loss, and poor resiliency can
result. In general, the resin is more elastic than the fibers to keep it from break-
ing down when the string is stretched and to allow the fibers to do their thing.

Resin Types. Resins come in different flavors. There are different starting com-
pounds and additives that contribute to making strings different from each
other. Resin can be stiff or elastic, depending on how the pieces are intended
to work together. But not every resin bonds with every fiber. For that reason,
nylon fibers will usually have a nylon resin, Zyex a urethane, and Kevlar might
have either. So not all combinations are possible.

How Much Resin? The amount depends on the number of fibers. The more
fibers there are, the more resin is necessary to fill the gaps between them and
to bond every fiber to each other. For example, the “void volume” of a 16
 String Construction 269

gauge multifilament may be as high as 22-24%. That is the amount of space

that will be occupied by resin and coating.

It’s ironic that the more fibers there are, the less performance material there is.
Or more to the point, more fibers means less fiber. This also means less
strength and tension maintenance (remember, all other things being equal).
The properties of the wraps, core, and resin in a string don’t add up to what
they would be if the string had been a monofilament of the same total diame-
ter. Which brings us to another irony, the monofilament will be stronger, but
because of the catastrophic consequences of notching, it might not last as long.
Strength and durability are not the same thing.)

Resin and Performance. Resin itself isn’t a performance material, but instead
enables other materials to perform. First, it protects each component fiber and
the string in total. It acts somewhat like a windshield safety glass. As in the
glass, there are dissimilar layers that prevent crack propagation. Each crack is
isolated by the “glue line.” If one fiber gets nicked, the crack is not propagat-
ed throughout the string. This is why you can notch a multifilament and not
get a catastrophic break like you might in a monofilament.

Secondly, resin transfers the load from fiber to fiber by shear forces. Without
resin, you’d get more breakage because fibers couldn’t as easily transfer the
load to each other. Some would end up bearing a disproportionate amount and
break. Resin by itself is not a strength component because it has no direction-
al fiber alignment, but it helps strength by judiciously transferring the load.

And last, resin allows a multi-part string to be one and many at the same time.
In other words, though it bonds fibers together as a whole, each fiber can still
behave somewhat individually. The human muscle is an example. Depending
on how you flex a certain muscle, different fibers will do different work, as evi-
denced by the muscle’s different shapes. It’s the same with strings. Depending
on the speed, angle, and location of ball impact, the individual fibers will elon-
gate in different ways. Theoretically, individually responding fibers will dis-
tribute the load more efficiently than a solid core which will deform uniform-
ly. So, even though the resin holds the fibers together, if one wants to stretch
a littke more than another, the resin gives and takes to balance it all out.

The coating is for durability and stringability. Usually the materials are the same
as the resins because the same primary consideration is true — finding a coat-
ing that will bond with the fibers. But a coating can have many different addi-
tives and can have multiple layers, each with a different mixture and thickness.

Durability. A coating can be hard or soft. If it is hard, it will resist abrasion,

friction burn, and string wear. If it is soft, it will resist string movement and
wear because the strings stick together more. Each is a different approach to
the same thing — durability.

Lubrication. Lubricants are also added to the coating or as a last layer itself.
A lubricant serves the same purpose as waxing the string to facilitate weaving.
Lubricants are usually a silicone or light oil, though materials like teflon are
also used. The balancing act is to find a lubricant that will ease stringing, not
gum up the clamps, and will evaporate quickly to prevent excessive string
movement.
270 Chapter 29

Coatings also contain other additives such as coloring and plasticizers which
keep the outer glue line between fibers from becoming too stiff and brittle
when the string is bent during stringing.

Need you ask? Everything so far has answered that. But it’s so confusing
because everything seems to be a tradeoff. To gain in one property, you lose
somewhere else. Each time you change one factor, you change both durabili-
ty and playability. So, if there is just one difference between strings, then there
must be at least two ... but probably many more you may never know about.

Draw Ratio. For example, consider two strings with a nylon 6 core and outer
wrap. Though made of the same material, the individual fibers can be prepared
differently before putting them together in a string. One such difference is the
“draw ratio.” Nylon fibers originally get extruded at a larger diameter than the
final string. While they are still hot, they are pulled or “drawn” down in diam-
eter. The draw ratio is the starting diameter compared to the final diameter. The
draw ratio determines how stiff and strong a filament is. When the hot plastic
first comes out of the nozzle, it is very weak. Pulling it orients and aligns the
molecular chains. As it is pulled, stiffness and strength are added. But there is
an optimum point. If it’s pulled too much, it becomes stiff and brittle. So, the
same nylon 6 core can be less stiff by not drawing it as much, while the outer
wrap could be stiffer by drawing it more. The same nylon 6 construction can
have different cores and wraps, and you won't know until you play.

And, if that’s not enough, some manufacturers heat and draw the completed
string again for a final “orientation.” So, saying a string is a certain material like
nylon 6 and a certain construction like a single wrap does not tell the whole
story. Different draw ratios, wind angles, resins, coatings, and final orientation
treatments can all make a difference to what looks like the same material and
construction.

It looks like the same string! Why does one cost $10 at retail and the other $23?
Well, as we have seen, they probably aren’t the same string. Each is finely
tuned, and it takes knowledge, R&D and machinery to optimize the perform-
ance gain of each tune-up. That’s what you're paying for. And eeking out lit-
tle differences in the lab can result in big differences on the court. You are pay-
ing for performance, not materials, labor, or packaging. You are paying for the
intangibles that make one string play better than another. Sometimes that’s hard
to swallow in a black and white material world, especially with regard to some-
thing that seems as simple as a piece of nylon. You’re not paying for features,
you're paying for benefits. Sure, a company can charge anything it wants with
a rationale like that, but it can only get it if the product delivers. And if it does,
it’s worth it.
 String Properties
and What They Mean

Anna }

By Rod Cross and Crawford Lindsey

s simple and innocent as string appears, it is very difficult to test. It has many complicated properties,
and the results are sometimes contrary to expectation. Testing it is like opening a Pandora’s Box of new
concepts and ways of thinking about string.

We tested 89 strings. Strings were between 15 and 18 gauge and represented a cross-section of materials. Prior
to tension testing, each new piece of string was weighed, gauge measured, and friction tested. Each single piece
of string, measuring 320 mm (12.6 inches) long, was mounted into the tensioning apparatus and clamped. The
string was tensioned to 20 kg (44.1 Ibs) and the elongation was measured. Tension was raised to 28 kg (61.7
Ibs) and held there for 10 seconds. Elongation was measured again.

To test tension loss, the string was allowed to sit for 1000 seconds (16.7 minutes) while data signals from a strain
gauge were processed every one or two seconds to record tension. Many previous tests had shown that the rate
of tension loss becomes linear (on a logarithmic scale) before that time, and tension loss can be accurately
extrapolated for any time beyond 1000 seconds.

The string was then impacted by a hammer swinging like a pendulum. The hammer weight (one pound includ-
ing lead head and wooden shaft) and speed at impact (8.63 ft/sec) combined to simulate the force necessary to
hit a serve 120 mph. Ten separate impacts were recorded. For each impact, change in tension (before, during,
272 Chapter 30

after impact), string deflection,

50 a duration of hammer on. strings,
45 a0 force on strings and hammer, and
deflection = y =0 ension in-going and out-going hammer
BG 40 1.5 00 velocity were recorded.
=
eS
10 8 Deflection and hammer speed data
2 Ss
2)
(cS We were collected by digitizing the sig-
®
be
0.0 nals from a laser that passed
, through an optical grid on the ham-
-0.5 mer as it impacted the string. Figure
35 30.1 shows the oscilloscope pres-
entation of this information.
Figure 30.1 Stiffness was calculated using all of
this data.
Tension vs time during an impact with a polyester string and the corre-
sponding laser signal used to measure string displacement. Each new
maximum in the laser signal corresponds to an extra 2 mm displace-
There are two astounding facts that
ment. The tension starts to rise only after the string has moved a few stand out from all the data collect-
mm. ed. First, gut is simply amazing. It
ranks “highest” (.e., at the extreme
end — qualitative interpretation
being left to the player) in every category except friction. The other astonish-
ment is that out-going hammer speed (“power”) is about 95% of the in-going
speed for every string — no matter the tension, type of string or the number of
impacts. This later result is presented in greater detail in Chapter 32.

Other observations also jump out. Strings of similar material all group togeth-
er. (There were five main groups of material in our tests — gut, zyex, nylon,
polyester, and kevlar.) Material composition is the biggest differentiator
between strings, not the construction of those materials within the string. But
for a given material, construction can make a difference.

What Is Dynamic Stiffness? The most important physical property of a string

is its dynamic stiffness. The stiffness of a string tells us how far it will stretch
when the string tension is increased. The dynamic stiffness was measured after
the string was tensioned to 28 kg (61.7 lbs). The string was left for 1000 sec-
onds (16.7 minutes) to measure the loss in tension with time. It was then
stretched rapidly by hitting it sideways with a hammer so that it stretched an
extra 6 mm (1/4 inch) or so lengthwise. The string tension rose and fell with-
in a period of about 30 ms (milliseconds, or thousandths of a second), at which
time the hammer bounced off the string. We measured the increase in tension,
DT, and the increase in length, DL. The dynamic stiffness, k, is given by k =
DT/DL and it is measured in lb/in. For example, if the tension increases by 50
lb when an impact stretches the string lengthwise by 0.25 in., then k = 50/0.25
= 200 Ib/in. This is a typical result for a nylon string. With a very stiff string like
kevlar, the tension increased by about 120 Ib when it was stretched by 0.2 inch,
so k = 120/0.2= 600 Ib/in. The softest string was natural gut, with k = 108 Ib/in.

Measuring Dynamic Stiffness. We tested string samples of free length 320

mm (12.6 inch), held firmly in metal jaws. “Free length” is the distance between
the jaws at the entry point of the string. A one pound hammer was swung into
each string at a speed of 8.63 ft/sec to impact the middle of the string and to
 String Properties 273

simulate the effect of serving a ball at about 120 mph. (The hammer weight
and speed reproduced the impact energy calculated for one string’s share of
the impact of a 120 mph serve.) The hammer deflected the string sideways by
about 1 inch, resulting in a lengthwise stretch of about 1/4 inch (note: side-
ways deflection and lengthwise stretch are not the same thing). The resulting
measurement of k therefore refers to a string of length 12.6 inch. If we had test-
ed a string twice as long, it would have stretched twice as far and k would have
been half as stiff.

Quick Stretches vs Slow Stretches. Impacting a

string with a hammer simulates the effect of hitting
a tennis ball. It is not exactly the same, but the Stiffness: Elongation vs Tension
important thing is the rapid stretch rather than a
slow stretch. During a slow stretch (Figure 30.2),
First Pull
such as when you string a racquet or when it is ianonnnnanrnnnnnnenecnonnintennnionntibnnniiiin scrim

pulled slowly in a pull test machine, the string will

keep stretching even if the tension is held constant. 8 -
. es

This phenomenon is known as creep. If the string S pre 4 ‘Second Pull

has settled down at say 60 lb for a few days, then 5
as soon as you increase the tension above 60 lb, -
the same thing happens. For example, if you =
increase the tension from 60 lb to 80 Ib by pulling =
on the string, it will stretch immediately by a cer- a.
tain amount, but it keeps stretching forever at the
higher tension. Measuring DT/DL under those cir-
cumstances obviously does not give a consistent 0 20 40 60 80 100
measure of k. In order to get a reliable measure- Tension (Ibs)
ment of k without the effects of creep, it is neces-
sary to stretch the string quickly by impacting it Figure 30.2
instead of pulling it and allow it to return quickly
to its pre-stretch length before it has a chance to The graph shows percent elongation measured in an Instron
creep further (Figure 30.1). pull machine, stretching at 4 in./min. This is a lengthwise
static pull similar to a stringing machine. The first pull acts
On the first impact, a few molecular bonds get bro- to pretension the string. The second pull shows that the
ken and the string stretches a bit further than usual. string is stiffer after a prestretch. The stiffness (the slope of
After that, it settles down and it is then possible to the curve) varies with tension, but is roughly constant above
measure k reliably. That is why we impacted the 40 Ib.

string 10 times with the hammer. The dynamic stiff-

In the impact tests, the string was stretched at about 1000
ness varied a tiny bit from the first impact to the
in./min by hitting it with a hammer. Stiffness is defined as
last, but usually by no more than 3%, despite the the impact tension rise divided by the impact elongation
fact that each impact caused the string tension to (note: elongation is not the same as deflection).
drop immediately after each impact. To smooth out
this variation in k, we took the average of all 10
impacts for every string tested.

Stiffness and Materials. The dynamic stiffness of a string, for a lengthwise

stretch, depends mainly on the type of material and not the tension. However,
all strings were tested at the same starting tension of 28 kg. We did not test the
strings at other starting tensions, so we did not specifically look at a large range
of different string tensions. However, the tension dropped after each impact,
and that allowed us to measure k over a small range of string tensions. The 3%
variation in k over this small range of tensions for a given string was negligi-
ble compared with the 800% difference in k between gut and kevlar. We test-
ed two gut strings and both had k = 108 lb/in. Kevlar strings ranged from 500
274 Chapter 30

to 886 lb/in. Polyester strings ranged from k = 251 to 310 Ib/in. Most of the
strings we tested were nylon, and all had values of k between 161 and 225
Ib/in. Only one zyex string was available, with k = 145 Ib/in.

In terms of stiffness, this makes comparing one type of string with another rel-
atively easy. Despite the fact that we tested hundreds of different strings, there
were really only five catagories of string. In order of increasing stiffness they
are gut, zyex, nylon, polyester and kevlar (aramid). For a given material, dif-
ferent construction methods can make a small difference to the stiffness, but
the biggest difference in stiffness comes when you switch to a different mate-
rial. Comparing one nylon with another is not so easy. The differences are not
large, in which case other effects such as tension loss with time are then just
as important in making the comparison.

What Is It? When a player hits the ball, each string in the racquet stretches by
an amount that depends on its length and its dynamic stiffness. The player is
not actually aware of this and couldn’t care less whether the center main
stretches by 0.2 inch or 0.3 inch. The player is only aware of what is felt
through the handle and what happens to the ball in terms of its flight path and
spin.

The sound made by the strings and the ball is probably an important psycho-
logical effect, but it has no effect on the flight of the ball and only a tiny effect
on the feel. Some players like both the sound and the feel of the strings vibrat-
ing, but the force transmitted to the handle by this vibration is tiny compared
with the main impact force. The only reason players are aware of the string
vibration at all is that the strings vibrate for about one second after the impact,
while the frame vibrations die out much faster, in about 30 ms.

So, as far as the player is concerned, the most important effects of the strings
are felt in terms of the impact force and the duration of the force. Both these
effects depend on string tension and dynamic stiffness, which together, help
determine string plane stiffness. But that’s getting ahead of ourselves because
the actual string plane stiffness depends on the number of strings in the rac-
quet, the tension and the dynamic stiffness. But to begin to sort this out, we
first need to look at the behavior of a single string.

Measurements. The force on a single string and the impact duration both
depend on string tension and dynamic stiffness in a way that is entirely pre-
dictable but very surprising. The force varies with time during an impact. It
starts off at zero when the hammer (or a ball) first contacts the string, it increas-
es to a maximum value when the string has stretched as far as it will go, then
the force drops back to zero as the hammer bounces off the string. We meas-
ured only the maximum or the peak force (Figure 30.3) as well as the duration
of the force (Figure 30.7). (Actually, we measured the increase in tension in the
string (Figure 30.4 and 30.5) and the sideways displacement of the string
(Figure 30.6), but it is easy to calculate the sideways force from that data.)

The increase in tension and the peak force during an impact both increase as
stiffness (k) increases. For example, on a gut string, the tension increased by
22 |b and the peak force was 28 Ib. On a kevlar string, the tension increased
typically by about 100 lb and the peak force was typically about 44 Ib. The
 String Properties 275

Peak Impact Force vs Stiffness Tension Increase vs Stiffness

50 120

Kevlar
45 —s

40

35 D
a
j=)(=)
Oo
oO

(Ib)
Force
Peak
30 288S&S)

(Ib)
Increase
Tension

25.
0 200 400 600 800 41000
0 200 400 600 800 1000
Stiffness (Ib/in) Stiffness (Ib/in)

Figure 30.3 Figure 30.4

Peak force is the peak sideways (in the direction of the The biggest difference between types of string is the
deflection) force on the string, calculated from the increase in tension during an impact. The increase in
deflection and the increase in tension. In a fully strung, tension for a gut (20 Ib) is much smaller than any other
hand-held racquet, the peak force on the stringbed (i.e., string. The tension rise in a kevlar string is huge by com-
not a single string) is about 10 times larger and it is the parison. All nylon strings were about the same, rising
same as the peak force on the ball. Ball soeed depends about 40 Ib. Similarly, all poly strings were about the
on both the force and its duration. When force goes up, same, with a tension rise of about 50 Ib.
duration usually goes down. Peak force is largest on
Compared to the increase in tension of gut, the average
kevlars and smallest on gut. That's why gut feels soft
nylon rose about 1.5 times gut, polyester about 2.5
and kevlar feels stiff.
times gut, and kevlar over 4 times gut. There is over a 5
times difference from the lowest to the highest tension
rise in the graph.

Peak Impact Tension vs Stiffness String Deflection vs Stiffness

160 30

29
140
28
e Pf
€= 26
S [o)
a
vt
25
py Kevlar
80
(mm) 23
Deflection
String
(Ib)
Tension
Impact
Peak
60 22
0 200 400 600 800 1000 0 200 400 600 800 1000
Stiffness (Ib/in) Stiffness (Ib/in)

Figure 30.5 Figure 30.6

The graph shows all 89 strings. Peak Impact Tension is During each impact, the hammer pushes the string side-
the original tension plus the increase in tension during ways by the amount (on average) shown in the graph.
the hammer impact, averaged over 10 impacts. For The same thing will happen on each string in a fully
example, gut was at 55 lb on average before each strung racquet, depending on how hard you hit the ball.
impact. The tension increased by 22 lb from 55 Ib to 77 In general, the stiffer the string, the less deflection.
lb during each impact. However, some stiffer strings may have the same or
higher deflection than softer strings. This can happen
when a stiffer string loses so much tension during
impact that it will stretch more than expected during the
impact.
276 Chapter 30

impact duration was about 31 ms on gut and about 24

impact Duration vs Stiffness ms on kevlar.
Explanations. These numbers might look a bit
strange at first sight. Kevlar is about 6 times stiffer than
gut and the tension rises about 5 times more, but the
peak force is only 50% larger and the impact duration
is only about 20% shorter. Why is this? If kevlar is 6
times stiffer than gut, why isn’t tension rise and peak
force 6 times more and impact duration 6 times less?

Kevlar is very stiff, so it doesn’t deflect very far when

(ms)
Duration
Impact hit by the hammer. Being so stiff, only a small deflec-
tion of the string causes the tension to increase dra-
matically, more than doubling the initial tension. But
the tension is directed along the string, not perpendi-
0 200 400 600 800 1000 cular in the direction of impact. Consequently, a very
Stiffness (Ib/in) large increase in tension along the length of the kevlar
string, combined with a small sideways deflection,
Figure 30.7
results in a relatively small force acting in the direction
Impact duration is the time that the hammer spent in of the impact. The force on the hammer is only about
contact with the string before it bounced off, meas- 50% larger with a stiffer string than the force on a gut
ured in millisec (ms). 30 ms = 0.030 sec.
string. The net result is that though the peak force on
The impact duration of a ball on a strung racquet is the kevlar string is larger (50% larger), that force acts
about 5 ms. The hammer was heavier than a ball for a shorter time (20% shorter), and the hammer is
and a single string is not as stiff as the whole ejected at exactly the same speed for both strings.
stringbed, hence the impact duration was about 6
times longer. Dividing these results by 6 will give a
Impact force and duration will cause different strings
good estimate of the time spent by a ball ona
to feel differently with respect to shock and feel, but
strung racquet.
they both conspire to eject the hammer at the same
speed from each of the different strings.

:
q

Strings Do Not Lose Much Energy. All of these effects agree with detailed
calculations of what should happen (See Chapter 13). What we didn’t know
before we did the tests was how well the strings would survive after repeated
impacts of the hammer. We expected that some would survive better than oth-
ers and that some would lose their elasticity or resilience after a sufficient
number of impacts. After 200 impacts on a nylon string, we gave up. None of
the strings lost any of their resilience or elasticity. They all lost tension after 10
impacts, but the hammer kept bouncing off every string at between 95% and
98% of the incoming speed for all of the strings. Each hammer impact lasted
about 30 ms, which is six times longer than the normal 5 ms impact of a ten-
nis ball on the strings. Two hundred hammer impacts is, therefore, equivalent
in effect to 1200 normal impacts, each equivalent to a 120 mph serve.

If a string loses no energy during an impact, then the hammer will bounce off
the string with exactly the same speed as the incoming speed. In fact, we found
that the energy loss was not zero, but it was close enough to zero to make no
significant difference in the calculations.

We did not test strings at tensions other than the starting tension of 62 pounds,
but we did do some calculations to predict the results. (Actually, our impact
 String Properties 277

tests tested strings at tensions ranging from 34 Static Tension Loss

to 62 pounds because each string was at a dif- 64
ferent tension at impact depending on how
much static, Figure 30.8, and impact tension, 62
Figure 30.9, it had lost.) The calculations are
shown in Figures 30.10-30.14 and are explained 60
below. 2 5 ee
(ee Sh oS Ne ee Pee) oe
Deflection: Perpendicular vs_ Linear 2 56
Stiffness (or Deflection Stiffness vs Pulling 5
Stiffness). As the graphs demonstrate, the r 54
sideways deflection of the string is smallest for
stiff strings and largest for soft strings, but it 2. SSSSSESS=<=4-—~ ;
does not vary by as much as one might expect 50 ‘ ae Kevlar2
for different strings, nor does it vary much by 0 200 400 600 800 1000 1200 1400
changing tensions. For example, if tension is 55 Time (seconds)
lb, then the string deflects by 29 mm for gut
and about 22 mm for a kevlar string. The side- :
. Figure 30.8
ways deflection increases at lower string ten-
sions, but only by 1 or 2 mm (Figure 30.10). Tension vs time after clamping at 28.0 kg (61.7 Ib).
Intuitively, one might expect the deflection to All strings lose tension rapidly during the first 100
sec, and then more slowly after that. All strings were
double when the tension is halved, but this is
held at 61.7 lb for 10 seconds before clamping.
not the case. The reason is that the string gets
Kevlars had a wide range of tension loss with time,
stiffer, in a direction perpendicular to the string,
but most lost tension severely in the impact tests.
the more it deflects.

Strings act a bit like ordinary springs in that the

harder you pull, the more they will stretch. But Impact Tension Loss
there is a difference. The stiffness of an ordi-
nary spring doesn’t change when you stretch it.
If you pull a spring twice as hard it will stretch
twice as far. If you pull or push sideways twice
as as hard on a string, the string doesn’t move
twice as far. It moves less than this. The side- Polyester
ways stiffness is proportional to the string ten- met is* SEN ¥Ha yee

sion, so when the string is pushed sideways,

the string stretches and the tension increases so (Ib)
Tension
Kevlar
the stiffness also increases. The lengthwise stiff-
ness might also increase a bit, but this is a rel-
atively small effect. In our impact tests, the
lengthwise stiffness did not change significant- cneeensnenmne ne
ly when we stretched each string. 0 50 100 150 200 250 300 350 400
Time (secs)
10 Impacts

Figure 30.9
The graph shows tension vs time while each string was subject to 10 impacts. The tension rise during each impact was
not properly recorded.For this test, data was only recorded every 2 seconds and whether or not the impact itself (the
spikes in the graph) was captured was just a matter of chance. It was not relevant to the test at hand and should be
ignored.The loss in tension after each impact is clear, especially after the first impact.
One impact with the hammer is roughly equivalent in effect to 6 very fast serves since one hammer impact lasted about
6 times longer than a normal impact of the ball on the strings.
278 Chapter 30

Deflection vs Stiffness At Different Tensions Tension Rise vs Stiffness At Different Tensions

140

120
= @
E £2 100
E 3
grf E
x

2
80

60
© 40
i=
=

A
20

0)
0 200 400 600 800 1000
Stiffness (Ib/in)
Stiffness (Ib/in)
Impact tension rises more for a string at lower tension
Deflection increases when tension is lowered for a compared to higher tension. The stiffer the string, the
given string. The stiffer the string the less extra greater that difference.
deflection you will get by lowering tension.

; igure 30.11
Figu
Figure 30.10

Force vs Stiffness at Different Tensions Impact Duration vs Stiffness At Different Tensions

45

@ 40
~ =
2 =
= 2 35
8
i}
85
rg 4 30
x)
o
£ 25

0 200 400 600 800 1000 20

Stiffness (Ib/in) 0 200 400 600 800 1000
For a string of any given stiffness, peak force does not Stiffness (Ib/in)
change much by varying starting tension. A higher Impact duration is more significantly increased by
tension does not significantly increase the force on the lowering tension for soft strings than for stiffer strings.
string, nor does a lower one lessen it much.

Figure 30.12 Figure 30.13

Relationship of Deflection, Force, and Tension. If a string is subject to a

sideways impact, the increase in tension, when the tension is low, is larger than
that when the string is initially at a high tension (Figure 30.11). The reason is
that when a string starts off at high tension, it is already fairly stiff in a side-
ways direction. The hammer comes to rest fairly quickly with only a relatively
small stretch and a small increase in tension. If a string starts off at a low ten-
sion, the hammer pushes the string a relatively long way before it comes to
rest. As a result, the peak tension is only slightly smaller if you start with a
string at low tension.

Furthermore, the peak force on any given string hardly varies at all if you
change the initial tension. The peak force increases if you use a stiff string such
as kevlar, but it doesn’t really matter if you string kevlar at 33 Ibs or 55 lbs. The
peak force on the strings will be about the same. The same is true of nylons,
polys and gut. Regardless of the initial tension, or the subsequent loss in ten-
 String Properties 279

sion over time, the peak force on the string will be about the same. New and
old strings should therefore feel much the same when measured in terms of
the peak force on the strings. The main difference in feel between new and
old strings will probably be due to the impact duration, which is longer on old
strings at low tension. Is that the only difference? What about power?

ose Power or Resilience?

Difference Between New and Old Strings. Most players report that new
strings perform better than old strings. New strings are crisper and give better
ball control, while old strings are less lively and are often described as “dead.”
By dead, most players seem to mean that old strings lack the power of new
strings. However, some players say the opposite, that they get more power out
of the strings as they get older. New strings are at high tension and old strings
are at a lower tension. Most players will tell you that strings lose their resilience
with age or use. But our tests show that they don’t. It is possible that strings
will lose resilience if they are stretched almost to the breaking point, but we
didn’t stretch the strings that far. We stretched them only as far as they stretch
in a 120 mph serve. So we are left with some really tough questions here. Do
old strings actually lose power or don’t they? If they do, then how come the
hammer keeps bouncing off at the same speed after lots of impacts? If they
don't, then how come most players think they do?

Steel Ball Drop Test. Further tests were done to look at this. A two pound
steel boule was dropped from a height of 100 inches onto the strings of sev-
eral racquets, with old and new strings, with the racquet clamped to the floor.
The ball was held by an electromagnet so that it could be more easily dropped
in the middle of the strings. The ball bounced to a height of between 89.7 and
90.3 inches off all the strings. This variation is too small to make any practical
difference. So, that settles it. Old strings DON’T lose power and they DON’T
lose resilience. They absorb all of the impact energy of a hammer or a steel
ball and give most of it back, regardless of how old they are. The ball bounced
off the strings at 95% of the incoming speed, exactly the same as when the
hammer bounced off a single string.

Player Percep ions of Power —

When a player says that new strings give more power than old strings, it is not
a comment based on any accurate measurement, but rather a perception of
power. The player might genuinely think the ball is coming off the strings
faster, when in fact, it is not. How can this be possible? That is something we
can’t answer and we are not even certain it is true, but there is clearly a dif-
ference in what players experience and what the above measurements and cal-
culations are telling us. Some suggestions follow.

Old Strings Are Looser and Have MORE Power. We know that strings lose
tension with time and with use. We also know that if you want more power
you should string at a lower tension. That’s because the strings will absorb
more of the impact energy and the ball will absorb less, so less energy will be
wasted in the ball. At least, that’s the case with a tennis ball. A steel ball does-
n't compress as much and loses no energy. That’s why the steel ball was used
to test the strings. But this tells us that old strings will have MORE power than
new strings when they are used to hit a tennis ball. That’s a bit embarrassing.
We can’t use that argument to explain why old strings seem to have LESS
power.
280 Chapter 30

Effect of Impact Duration on Power and Control. Perhaps there is a clue

in the difference in impact duration between old and new strings. A ball or a
hammer will spend more time on a soft string than a stiff string. So, as a string
ages and the tension drops, the ball will spend more time on the strings.
Though it will still come off at the same speed (maybe even a slightly higher
speed if we believe that low tension = more power) it will come off at a dif-
ferent angle. Why?

On strings at high tension, the ball will come off the strings after about 4.5 ms.
With strings at low tension, the ball comes off after about 5 ms. For a serve or
a smash, the racquet head swings through an angle of about 8 degrees in 4.5
ms or about 9 degrees in 5 ms, depending on how fast the racquet is swung.
For a groundstroke, where the racquet is swung at a slower speed, the racquet
will swing through an angle of only 3 or 4 degrees during the time the ball is
on the strings. A one degree change in the rebound angle means that the ball
will land in a different spot, but a good player should be able to compensate
by hitting the ball a bit earlier, one degree before the usual impact point.
Nevertheless, if a player finds that the ball lands in a different spot, the ques-
tion is whether the player will perceive this as a change in ball speed or a
change in rebound angle. However, we don’t think this is the effect we’re look-
ing for.

Psychological Effects: Is Power All In Your Head? Perhaps there are psy-
chological effects involved here as well. If a ball stays on the strings for only
a short time before it bounces off, then players might assume that it also comes
off at a higher speed because it came off faster. The word faster is ambiguous
itself. Faster can mean sooner or it can mean higher speed. Players sometimes
complain that pressureless balls are like bullets since they come off the strings
so fast. They might spend a shorter time on the strings, but that doesn’t mean
they come off at a higher speed. Suppose a ball stays on string A for 4.5 ms
and bounces off at 60 mph. On string B, it might stay on the string for 5.0 ms
and still bounce off at 60 mph. But the player might think it came off at a high-
er speed from string A because it came off sooner and felt crisper. We are only
guessing here. No one has done an experiment with a radar gun to check this
out, but it might actually be true. There is definitely nothing wrong with the
physics of this. The speed off the strings depends on the force acting on the
ball as well as the duration of the force. A small force acting for a long time
can generate the same speed as a big force acting for a short time. That’s exact-
ly what we measured with the hammer bouncing off different types of string.

Sound and Power. The sound of the impact might also play psychological
tricks. If you hit a drum at high tension, you get a crisper sound than at low
tension. If you use a string dampener, the shot sounds different (better for
some and worse for others), but it can’t make any difference to the speed of
the ball since only a microscopic amount of energy is involved in the string
vibrations.

Strings Don’t Lose Power, They Lose Trajectory. There is another possible
reason why old strings might seem to lose power when in fact they don’t.
Suppose you pick up a racquet with new strings and hit the ball consistently
over the net to land on the baseline. Then you pick up the same model rac-
quet but with older or well-used strings. If you play the same shot and the ball
keeps landing short or in the net, what will you conclude? Will you say the old
strings have lost power? What if you have to hit the ball a lot harder to get it
 String Properties 281

over the net or onto the baseline? Will you still say that the old strings have
lost power? You will probably answer yes to both questions. But suppose the
ball comes off both sets of strings with exactly the same speed when you swing
the racquet with the same speed. Then the strings have the same power, but
the ball could come off the strings at a different vertical angle. The ball might
bounce off new strings at a higher angle than off old strings, possibly because
new strings give a better grip on the ball than old strings. It might have some-
thing to do with the way the strings slide or don’t slide across each other side-
ways in the string plane. That would explain why the ball clears the net with
the new strings and lands short or in the net with old strings.

The question now is, can a player judge the speed better than the vertical
angle, or is it the other way around? Again, no one we know of has tested this
with a radar gun, so we can’t say for sure. Until that is done, we cannot give
a definite answer to the question why old strings seem to lose power even
though there is no loss of resilience.

In the mean time, have a look at the ball trajectories in Figure 30.14. These
were calculated on a computer. The ball slows down through the air due to air
resistance and lands on the court at a slower speed than when it was first hit.
Can you guess which ball started off with the highest speed? Which one start-
ed off with the lowest speed?

Trajectories, Spin and Shot Placement

ae)
= A
Baseline
—_ 6
[40]

i 4
2S
s
x
DeNs
Yo
F - in 60 80 100
Horizontal (ft)
1 = 70 mph, no spin, 10°
2 =70 mph, 16 revolutions/second, 10°
3 = 70 mph, 28 revolutions/second, 10°
4 =70 mph, no spin, 6°
5 = 70 mph, 16 revolutions/second, 6°
Note: the angles of 6 and 10 degrees are the launch angles above the horizontal

Figure 30.14

The answers are that they all started at the same speed, but the spin or the
angle was different in each case. The trajectories are shown side-on. The play-
er’s view is from above and behind, so the angles won’t be as easy to pick. A
player might say, “I’m not getting any power out of these strings,” but a spec-
tator might say, “The ball is coming off the strings at a low angle.” Power
describes ball speed, and control describes a consistent or expected ball angle.
If the trajectories (4) and (5) are consistently low with old strings, and trajec-
tories (1) - (3) are consistently high with new strings, then control is good, but
282 Chapter 30

it might seem that power is different. If the trajectories are not consistent, then
the problem is with control, and that is probably a sign that the string tension
is too low. At low tension, the ball stays on the strings a bit longer, giving the
racquet more time to rotate about the long axis through the handle. If the ball
strikes the strings off-axis then the racquet will rotate about the long axis and
the ball will come off the strings at an angle that is either too low or too high
depending on whether the ball strikes below or above the axis.
 An Introduction
to Elastic Energy

Goran {ivanisevic, Croatia

By Crawford Lindsey

tring Tensionand Power

-ennis players live by the cliche “string loose for power and tight for control.” But though many live by it,
many don’t understand it. In fact, many courtside arguments on this topic have been more vituperative
than on a disputed match point line call. So, is this dictum a law of the universe or a convenient myth?
The crux of the matter is this — somewhere inside of us is the little kid looking at the world for the first time
saying, “Daddy, if I pull the racquet string very, very tight, then it will have more power in it, and it will send
the ball away faster and farther.” The reasoning, of course, is that there is more energy in the string, so there
must be more power. A little thought experiment will help to expose the fallacy of this statement.

Steel Balls and Bounce Height. Where does this energy that’s all couped up in the tighter string go? Imagine
two racquets infinitely secured in all directions so only the string face will deflect upon impact. One is strung at
50 and one at 70 pounds. Since we want to know what is happening in the strings only, we will use nonde-
formable steel balls for our test. This eliminates the dynamics of a deformable rubber ball interacting with the
strings. We will thus be able to see which strings have the most power.
284 Chapter 31

(Figure
Drop a steel ball weighing the same as a tennis ball on each racquet
happen? The possibili ties for
31.1). Which ball will rebound higher? What will
ball rebound are shown in Figures 31.1-31.3.

Follow The Bouncing Balls

Tension & Rebound

Does Lower Tension Rebound More?

Figure 31.1
Guess #1:

The 50 Ib stringbed will deflect more and rebound the ball higher because of a tram-
poline effect.

Does Each Tension Rebound The Same?

Oo Oo

EP LLL EKA AAAS

ia a laa ay aa,
ESETITSUT GLP LTIT EToy
LEG TITTIESTITS EF LST
TUTE il

Figure 31.2
Guess #2: It will bounce the same height on both racquets.

You might say that since no energy is getting lost in ball deformation because they
are made of steel, all the kinetic energy of the balls is going into the strings. Since
each ball has the same energy at impact and each stringbed will receive the same
energy, then each ball might bounce the same height with each tension. But then you
rethink. The 50 pound stringbed will deflect more than the one at 70 pounds. Does
energy get used up in stretching the string? If so, then the 70 Ib stringbed will lose
less energy stretching.
 Power and Strings — Part 1 285

Does Higher Tension Rebound More?

PRN ONAN

Figure 31.3
Guess #3: So now you're convinced that the bounce will be higher on the 70 pound
racquet because more energy is “available” since it took more energy to stretch the
70 |b strings tighter. And round and round you go.

Now you are more confused than ever because all three scenarios seem rea-
sonable. What happens is this. Strings don’t actually “have” power. The ener-
gy in the strung racquet, whether at 50 or 70 pounds, is not available to the
steel ball. That energy is all distributed in the equilibrium balancing of all the
stresses and strains as the racquet and strings pull on each other. The only
energy available to return to the ball is the energy added by the impact. The
preexisting tension in the string is irrelevant. That’s not extra energy, it’s equi-
librium energy, or by definition, zero energy. The racquet starts in equilibrium
and ends in equilibrium. In between, energy in equals energy out. There is
nothing extra in either racquet to give. The energy is just passed around a lit-
tle differently in each racquet and with the steel ball.

Since there is no extra “power” or energy obtained in tensioning the strings,

and since there is no energy going into deforming the ball, the only place ener-
gy can be lost is in stretching the strings. As will be explained in more detail
in Chapter 32, all strings lose only about 5% of the incoming energy in stretch-
ing. And in the typical ranges of playing tensions, there is little difference in
energy loss at different tensions. The answer to our conundrum is therefore
clear. The steel balls will bounce to the same height.

The reason this is an important conclusion is that we have determined that the
“intrinsic power” of all strings at all tensions is about the same. We did this by
eliminating energy loss in the racquets and balls by clamping the racquets so
they could not recoil and using steel balls so the ball could not deform. But
we know that tennis balls don’t bounce to the same height when dropped on
a clamped, stationary racquet (the looser stringbed bounces the ball higher).
And since different strings have the same intrinsic power, the answer must lie
in the ball or the dynamics of the interaction between the ball and strings at
different tensions. Let’s take a look.

Tennis Balls and Bounce Height. Now we are back in the real world of rub-
ber tennis balls. But our little trip to Steel Ball Land has taught us one very
important thing. Power is all about energy conservation in colliding objects. A
tennis ball’s impact energy is the same (because the steel balls weighed the
286 Chapter 31

same as a tennis ball), and thus, the amount available to give back to the balls
is the same. But the tennis ball is much more compressible, so the stiffer the
strings are in comparison to the ball, the more the ball compresses. The key
point here is that a tennis ball is designed to lose about 45% of its share of the
stored impact energy due to internal friction and heat during its deformation.
(The energy isn’t actually “lost.” It is converted to other forms that are not of
any use to propelling the ball.) Thus, that energy is lost for accelerating the
ball. The relative deformability of the two objects is going to determine where
the incoming energy will go — more into string or ball deformation (see
Chapter 32 for a detailed analysis). And where that energy goes determines
how much help it will be in rebounding the ball.

Because so much energy goes into ball deformation, kinetic energy conserva-
tion is of the utmost importance in creating tennis power. As we have seen,
one way to do this is to make the ball harder. That would be illegal. So for ten-
nis, the equation becomes very simple. For the range of typical playing ten-
sions, the looser the strings, the more energy will be stored in the strings and
returned to the ball. The tighter the strings, the more the ball is compressed,
and the more energy is lost.

There is a point, however, where more energy is lost in stretching and mov-
ing the string than the energy saved by not compressing the ball. The tension
where that occurs is somewhere in the 40-pound range, depending on swing
and incoming velocity. At that point, lowering tension will produce less power.

So, lower tensions are not by their nature more powerful than higher ones. It
depends on the nature of the colliding objects, the dynamics of the collision,
and how low you go. But if one of the objects is a tennis ball, then for the
range of common playing tensions, lower tensions mean more power (that’s
the good news, but the bad news is that it is not a “significant,” nor, to most
Figure 31.4 a a perceptible
: A ;
players, amount more, as we will see in
Chapter 32). But it’s not that lower tensions are more pow-
erful, but that they are more capable of storing and return-
What Goes Up... ing a ball’s impact kinetic energy without turning some of it
Must Come Down on the ball itself and crushing it.
(depending on input energy)

Pe ae The Bow and Arrow Analogy. But you still can’t get rid of
F 7 ; that feeling that tighter strings somehow have more power,
AQ can you? Still, you’re secretly thinking, “But what about all
eo aees that extra energy I put into the string by pulling it to 70
iv > pounds? Isn’t it just like a bow and arrow? Won’t the tighter
\ bow send the arrow farther?”
+ Equal deflection” 50 Ibs : ; ae an
- Unequal energy input The bow and arrow is a little different because it is not a
* Unequal power collision and the input energy is provided solely by the
wo operator. As shown in Figure 31.4, the archer fully extends
/ \e 70 Ibs his arm holding the bow and pulls the string back to his
Yd cheek with the other. The tighter bow (70 Ibs) is not more
i es i powerful because of the starting tension, but because it
/ takes more input energy to pull the string back an equal dis-
\ tance to the cheek than with the 50 lb bow. Unequal ener-
= Eaualinpueneray gy went into each pull, so unequal energy comes out. What
- Unequal deflection 50 Ibs is equal here is the string deflection back to the cheek, not
*- Equal power the energy needed to do the deflecting.
 Power and Strings — Part 1 287

And if you still won't give up and are thinking that pulling tension on the
stringing machine is like pulling on a bow and arrow, you're right. There is
more energy locked up in that 70 pound racquet, but it is just that — locked.
To release it, you have to cut out the strings. Imagine an ant crawling along
your center main Gmagine you're stringing and haven't installed the crosses
yet). If you cut the string just behind its path, the ant will go flying further with
the tighter racquet. But ant hurling is another sport and a different application
of the tensioned string that doesn’t help one’s forehand much.

We conclude that looser does result in more power, but not because the strings
are more powerful, but because they crush the ball less and more of the ener-
gy is thus available to be stored in the strings and then given back. It’s about
energy conservation and transfer during impact, not inherent power or pro-
duction. It’s not that loose strings have more energy, it’s that up to a point, they
absorb more and give back more when interacting with a tennis ball. They are
better transformers; they are not the energy plant.

The same sort of reasoning applies to long vs short strings, all other things
being equal (Figure 31.5). In other words, with steel balls, both a 13-inch and
a 15-inch string of the same gauge and material and at ther same tension are
potentially equally as powerful depending on the nature of the collision. The
longer string will deflect more, however. The reason is simple. When a string
is pulled from the end or impacted in the center, it stretches the
same everywhere. So, if you stretch a string 10%, every inch along
the string is stretched 10%. If one string is 15 inches and the other Figure 31.5
13, then the 15-inch string will stretch 2” x 10%, or 0.2 inches
more than the 13-inch string. A long string isn’t more stretchy; it
just has more inches to stretch and thus deflects more on impact. Long and Short Tension

Of course, for a tennis ball, the results are the same as in the loose
70 Ibs — = Long
vs tight situation. More deflection means more cushioning, less
ball deformation, and more energy return. Again, it’s about ener-
gy conservation and location of conservation, not energy creation.
All things being equal, longer strings convert more incoming ener- 70 Ibs
gy into outgoing energy. Longer strings have a better exchange * Equal tension
rate than shorter strings at the same tension. * Equal impact energy
* Unequal length
Different deflections for different lengths is, of course, why pro- * Unequal deflection
portional tensioning exists. For the above reasons, different length * Unequal power
strings at the same tension will be saving different amounts of
incoming energy. If you’re looking for power, the shorter strings
will give energy away. If it’s control you want, longer strings at the 70 Ibs
same tension might introduce too much power or variability into
your hit.
F 50 Ibs
The basic stringing formula for proportional tensioning is to
reduce tension in the same proportion as the difference in length
+ Unequal tension
of two adjacent strings. This is a good starting point, though it is
* Unequal length
essentially arbitrary for the following three reasons.
+ Equal impact energy
+ Equal deflection
(1) First, it assumes that for every x percentage increase or + Equal power
decrease in tension, there will be a constant x percentage increase
288 Chapter 31

or decrease in elongation. Instead, elongation changes proportionately more at

lower tensions than higher tensions. So, as you lower the tension on the short-
er string, you may gain more than a proportional amount of deflection.

(2) So, using the basic formula may actually over-compensate a little. But that’s
okay because each string going outward from the impact is exposed to less
impact. That’s because the ball is round and the impact is absorbed outward
from the first impact point. So each outward string needs to be even looser to
stretch the same amount as the one next to it because it is enduring slightly
less impact energy.

(3) And last, longer strings at higher tensions will also lose tension at a faster
rate. Even as you’re lowering the tension on the shorter string, the longer,
tighter string is losing creep tension faster than the shorter one. Your tension
differential will therefore end up less than the difference in your reference ten-
sions for the two strings. :

So, if anything, you might want to err on the side of lowering tension on the
shorter, looser string by more than the same proportion as the difference in
string lengths. Obviously, unless you’re a mathematician, trial and error will be
your guide.

If you tension a 15 gauge monofilament string to 50 pounds, it is being ten-

sioned equally from the outside to the inside (at least after the tension has had
time to equilibrate inside the string). The entire cross-section is being pulled to
50 pounds and is elongating the same amount in the process. By this reason-
ing, you can think of a 17 gauge string as embedded in the middle of the 15
gauge string. Now imagine that you are pulling this string on a continuous ten-
sion machine. What if we could get to this buried 17
gauge string by melting away the outside layers? What
GAUGE AND TENSION would happen? As the layers melt away, the string will get
longer and longer with each disappearing layer. Why? It
nirewrh 50 Ibs is exactly the same piece of nylon that was there to begin
= with. It’s not that the nylon suddenly became more elas-
tic because it is thinner. It’s that the 50 pounds seems like
0 more to the string now than it did before. It’s still the
= same 50 pounds, but now the number of pounds per
Two strings one is two atoms thick and the other one square inch of cross-section is much greater. Each poly-
atom thick. At 50 pounds, the thicker string has twice mer chain is not sharing the 50 pound weight with as
as many atoms to share the load. It stretches less and
SEES, : '
Se many other chains as before, so it must stretch further to
build up enough tension to resist the load. So it’s not that
thinner is stretchier, but that thinner carries more load per
Figure 31.6 square inch than thicker, so it stretches more (Figure
31.6):

If we now put all the layers back on and then increase the tension until the
stress in pounds per square inch equals its 17 gauge counterpart, we'll find that
they have both elongated the same amount. The nylon elongates the same
amount when subjected to the same stress in pounds per square inch, no mat-
ter how thin or thick it is.

So, different gauge strings may be at the same tension, but they are not at the
same stress. It is the stress that determines how tight or loose the string feels
upon impact. Also, if you pull a 16 and a 17 gauge to equal tensions, not only
 Power and Strings — Part 1 289

will the 17 gauge elongate more to get to that tension, but the higher the ten-
sion, the greater the difference in their diameters will become compared to the
difference in their starting diameters. This in turn means the stress increases
faster as tensions increase for a thinner string compared to a thicker one.

And then the same reasoning as above applies: the tennis ball will bounce
higher off the 17 gauge because it stretches more, absorbs more energy of
impact, and then gives it back. The pounds per square inch are greater in the
thin string but the amount of energy being added to it is the same. (Naturally,
all is not always so simple. Chapter 35 describes some exceptions to this prin-
ciple.)

So, a long, thin, loose string has as much power in it to start with as a short,
fat, tight one. If you play tennis with steel balls, it won’t make any difference.
But if you use tennis balls, the first one will (usually) give you more energy
return than the latter.
 How Much Do
Strings Affect Power?

-lena Dementieva, Russia

By Rod Cross

he most surprising results to come from the string tests described in Chapter 30 are that

e strings don’t lose power with use

¢ tension makes almost no difference to the energy loss in a string

e there is no significant difference in energy loss between different string types.

This is so contrary to common belief that it seems to be wrong. Most players at an elite level replace their strings
many times each year since they say that the strings feel dead after five or six sets, as if the strings have lost
power or resiliance (see Note 1). Most players also prefer the “crisp” or “lively” feel of a new set of strings.
Consequently we will look more closely in this chapter at the effect of different string tensions and string types
on the energy loss in the ball. Impacting a string with a metal hammer or a steel ball is the best way to meas-
ure the energy lost in the strings, but it does not tell us how much energy is lost in the ball when the string type
or tension is varied.

The conventional wisdom is to string a racquet at low tension for more power. The reasoning is that the ball
will squash less in a soft impact, so it will lose less energy and come off the strings at a higher speed.
292 Chapter 32

Furthermore, the racquet will vibrate less in a soft impact, since the force on
the strings is smaller, so the racquet bends by a smaller amount and less ener-
gy is wasted in racquet vibrations. If less energy is wasted in both the ball and
the racquet, then the laws of physics guarantee that the ball will come off the
strings at a higher speed. However, the question that needs to be asked is, by
how much?

Suppose the string tension is reduced from 60 to 50 pounds. If you can serve
a ball at 120 mph with strings at 60 pounds, then you can serve it even faster
if you drop the tension to 50 pounds. But as you will see in this chapter, meas-
urements and calculations indicate that you will serve the ball at about 120.8
mph if you drop the tension to 50 pounds. The increase is ridiculously small
and probably not noticeable.

Old strings do NOT lose power. They will feel softer since the string tension
drops over time and with use, but if anything, the ball will come off the strings
a tiny bit faster as the strings get older. The sensation that old strings feel dead
and have reduced power is probably due to the fact that the force of the ball
on the strings is reduced at low string tension, with the result that the force act-
ing on the hand is also reduced. If the force on the hand is reduced, players
may interpret that as a dead feel, and they may equate that with the feel of hit-
ting the ball at a lower speed. The racquet will vibrate less when the tension
is reduced, so players will also notice a reduction in the vibration level, as if
the ball was being hit at a lower speed. Conversely, many elite players tend to
prefer relatively stiff strings such as polyester or kevlar, possibly because the
force on the hand is greater, and that will give them the sensation of greater
power.

When a player serves a ball, the racquet is swung at a speed that varies from
one end of the racquet to the other. The tip travels the fastest, and the handle
travels the slowest. If the speed at the impact point is V, then the ball will be
served at a speed v given by

v=(1+e,)V (32.1)

where e, is the apparent coefficient of restitution (ACOR) at the impact point.

Equation (32.1) is derived in Chapter 11. The value of e, can be measured by
firing a ball at right angles to the string plane and measuring the rebound and
incident speeds. e, is the ratio of these two speeds, so it a number that meas-
ures how well the ball bounces off the strings. For example, if the ball
rebounds at 40% of the incident speed, then e az 0.4. If it bounces at 10% of
the incident speed, then e, = 0.1. If e, is 0.4, then the serve speed will be 1.4V
or 40% greater than the racquet speed. If e, = 0.1, then the serve speed will
be 1.1V or 10% greater than the racquet speed.

Changing the string type or tension has no effect on the racquet speed.
Consequently, the only effect that we need to consider is the effect on e A ile
value of e, increases when the string tension is lowered, but only by a small
amount. Similarly, changing from nylon to natural gut will reduce the string
plane stiffness, but e, will increase by only a small amount.
 Strings and Power — Part 2 293

Figure 32.1

ACOR vs Distance from Tip of Racquet

A typical graph of ACOR vs impact position
on the strings is shown in Figure 32.1. All
Tip Middle of strings Throat
racquets and all strings give a similar result 0.5
to this in that the ACOR is largest near the
throat and smallest near the tip. If you
think about this for a second, it means that 0.4
racquet power is largest near the throat and
smallest near the tip. Racquet manufactur- 0.3
ers don’t tell you this. They tell you that
racquet power is biggest in the sweetspot ACOR
region near the middle of the strings. It 0.2
turns out that we are both correct. If the
racquet is at rest when the ball hits it, then 0.1
the power is a maximum near the throat.
However, if you swing the racquet to hit
the ball, then the point of maximum power
shifts away from the throat to a point near e S 0 Is
the middle of the strings. That’s because Distance from Tip (inch)
the middle of the strings is traveling at a
higher speed than the throat, so the ball
then comes off faster from the middle of the strings. The ACOR in the middle
of the strings is usually about 0.3 for a light racquet, about 0.4 for medium
weight racquets, and about 0.5 for a heavy racquet. The ball, therefore,
bounces faster or higher off a heavy racquet, but the disadvantage is that you
can’t swing a heavy racquet as fast as a light racquet.

You can measure the increase in ACOR with racquet mass yourself very easi-
ly. Place the racquet on the floor and put your foot on the handle. Then the
racquet behaves as if it is as heavy as the floor. Then drop a ball onto the mid-
dle of the strings. The ball bounces much higher since it can’t transfer any of
its energy to the racquet. The energy is taken up partly by the strings and part-
ly by the fact that the ball squashes a bit, but almost all of the energy stored
in the strings is then given back to the ball. If you drop the ball directly on the
floor, then the floor is too stiff to absorb any of the impact energy so it won't
have any energy to give back to the ball.

Equation 32.1 tells us something that is perhaps very obvious. If you swing the
racquet twice as fast, so that V is doubled, then the ball is served twice as fast
since v is also doubled. It also tells us something that is not quite so obvious.
If you can somehow double the ACOR, then the ball speed won't be twice as
fast. It will only be about 20% faster. Doubling swing speed doubles ball speed;
doubling bounce speed from the strings adds only 1/5 to the total ball speed.
For example, suppose that you somehow increase the ACOR from 0.25 to 0.5.
One way to do this is to serve from a point near the throat instead of a point
near the tip. But then, the racquet speed V is different, since the tip will trav-
el faster than the throat. But suppose you use a super-dooper ball and a super-
dooper racquet with super-dooper strings so that the ACOR is doubled from
0.25 to 0.5 for an impact on a particular spot of the strings. Then V is the same,
and v will increase from 1.25V to 1.5V. That’s an increase in serve speed of
0.25V, or an increase of 20%.
294 Chapter 32

More realistically, you might get a 10% increase in ACOR if you use a heavy
racquet with strings at low tension, in which case ACOR could increase from
say 0.40 to 0.44. That gives an increase in serve speed from 1.40V to 1.44V,
which is an increase of 0.04V, or an increase of 3%. But if you go for a heav-
ier racquet with a bigger ACOR, then you won't be able to swing it quite as
fast, in which case you might actually end up serving the ball a bit slower. For
an impact in the center of the strings, the ACOR is typically about 0.4. The ball
stiffness is usually about equal to the string plane stiffness, in which case, about
half of the elastic energy is taken up by the strings and the other half is taken
up by the ball. As the ball comes off the strings, about 95% of the energy stored
in the strings is given back to the ball, and about 5% of the energy stored in
the strings is lost. We know that from the hammer and the steel ball impact
tests.

We also know that 45% of the energy stored in a ball is lost during a collision.
If you drop a ball from a height of 100 inches onto a concrete surface, then all
of the impact energy is taken up by the ball when it squashes. In that case, the
ball stores all the impact energy, but it loses a lot of this energy as it bounces
off the concrete. The ball bounces to a height of about 55 inches, which means
that it loses 45% of its stored energy. This is specified by Rule 3 of tennis. A
ball dropped from a height of 100 inches onto a concrete slab must bounce to
a height of between 53 and 58 inches.

_ Increasing Bounce Speed by Lowering Tension __

What happens if we change string tension, keeping racquet mass and speed
fixed, while hitting in the same stringbed location? If you decrease the string
tension, then the ball will absorb less of the impact energy by squashing a bit
less, and the strings will absorb more since they will stretch a bit more. The
ball still loses 45% of whatever energy it stores, and the strings still lose 5% of
whatever energy they store. Since the strings store more energy than the ball
when they are at low tension, and they give more of it back, then the ball will
bounce a bit faster or a bit higher.

A simple calculation (which is done in the next section) indicates that if you
drop the tension by half from 60 to 30 Ib, then the ACOR will increase by 7%
from 0.40 to 0.43. That’s a handy increase, but no one would want to drop the
tension that far. A drop from 60 to 50 lb is more reasonable, in which case the
ACOR will increase by 2.3% from 0.40 to 0.41 and the serve speed will increase
from 1.40V to 1.41V. That’s an increase of only 0.7%, not worth bothering
about. Similarly, if the strings lose 10% of their stored energy instead of 5%,
then the effect on the serve speed is also insignificant.

So why does reducing tension have so little affect on bounce speed and, ulti-
mately, serve speed? The ball and the strings both act like springs, as described
in Chapter 13. If you apply a force F to a spring, it will compress by an amount
x given by F = kx, where k is called the spring constant or the stiffness of the
pune The spring will then store an amount of elastic energy given by
Eaaakxe 2.

The ball and the strings both have a spring constant about 200 Ib/in., but of
course this can vary from one string or ball to another. Now suppose the ball
 Strings and Power — Part 2 295

has a spring constant k, and the strings have a spring constant k.. If they push
against each other, then the force on the ball is the same as the force on the
strings, sO

in k Xp = k,x,

where x, = compression of the ball and x, = distance moved by the strings

with respect to the frame.

The energy stored in the ball is E, = k,x,°/2 and the energy stored in the
strings is E= k.x.*/2. The total elastic energy stored in both the ball and the
strings is

1 gece oO gaa cr dee

so the fraction of E, stored in the strings is

Fg ghee) (Kk, tix”).

But x, = k.x./k, so (with a little algebra that we will skip here):

(32.3)

We will now apply the last formula to a case where the ball and the strings are
equally stiff and where the ball is incident with 200 units of kinetic energy (the
actual units are not important here). We first consider a case where the racquet
head is clamped so that all the kinetic energy of the incident ball is converted
to elastic energy in the ball and the strings. None of this energy will be wast-
ed in pushing the racquet head backwards, and none is wasted in causing the
racquet frame to vibrate. Similarly, no extra energy is given to the ball by
swinging the racquet towards the ball.

Forexample.if ke-= ko =,200 Ib/in., then E,/E, = 200/400 = 0.5, so half the
total elastic energy is stored in the ball and half is stored in the strings (see
Note 2). Since the total energy available is 200 units, then 100 units are stored
in the ball and 100 units are stored in the strings. Since 5% of the energy in the
strings is lost and 45% of the energy in the ball is lost, the total energy lost is
5 + 45 = 50 units or 25% of the total elastic energy. The ball will therefore be
ejected out of the strings with 200 - 50 = 150 units of kinetic energy. Since it
was incident with 200 units of kinetic energy, the ratio of the rebound speed
of the ball to its incident speed will be given by e = COR = (square root of
150/200) = 0.866. Note that the COR and the ACOR are equal when the rac-
quet head is clamped, as explained in Chapter 11. If the racquet is hand held,
and if the ball impacts in the middle of the strings so that the frame does not
vibrate, then the COR is unaffected, but the ACOR is less than the COR.
296 Chapter 32

Now suppose we reduce the string tension by half, so k, = 100 Ib/in. Then
E,/E. = 200/300 = 2/3 so 2/3 of the total elastic energy is stored in the strings.
Of the initial 200 units of energy, 133 units are stored in the strings and 67 units
are stored in the ball. The string loses 0.05 x 133 = 6.6 units and the ball loses
0.45 x 67 = 30.1 units. The total energy lost is 6.6 + 30.1 = 36.7 units = 18.4%
of the total. The energy lost is less than before, so the ball will come off the
strings a bit faster. In this case, the ball will be ejected with 200 - 36.7 = 163.3
units of energy, so e = COR = (square root of 163.3/200) = 0.904.

We have just shown that when the string plane stiffness is halved, the COR
increases by only 4.3%, from 0.866 to 0.904. This is when the racquet head is
clamped. What we are really interested in is the increase in ball speed when
the racquet is swung towards the ball. To calculate that, we can use the results
found in Chapter 12 to calculate the ACOR, e,.

For a hand-held racquet, the ACOR is given by

(eM, - m)
e,= (32.4)
(M, + m)

where M, is the effective mass of the racquet, m is the mass of the ball and e
is the COR. For an impact in the middle of the strings, M, is about half the actu-
al mass of the racquet. A typical value for M, is about 170 gm, m = 57 gm, so
e, = 0.40 when e = 0.866, and e, = 0.43 when e = 0.904. This represents a 7%
increase in e,, even though there is only a 4.3% increase in e. Halving the
string tension from say 60 lb to 30 lb therefore increases the serve speed from
1.40V to 1.43V or by 2.1%. Consequently, decreasing the string tension from 60
Ib to 50 lb will increase the serve speed by 1/3 of this, or by a miserable 0.7%.

We are assuming here that string plane stiffness is proportional to string ten-
sion. This is valid for small deflections of the string plane, but in a very fast
serve, the string plane will deflect by an inch or more, and then the string plane
stiffness increases considerably during the impact. The increase in stiffness is
typically about 30% for natural gut strings, but it can be a factor of three for
kevlar strings. The string tension can increase from 60 lb to 180 Ib during a fast
serve with a kevlar string. The tension drops back to 60 Ib when the ball leaves
the strings, but during the time the ball is on the strings, the string plane stiff-
ness is up to 3 times larger than the 60 lb stiffness. Furthermore, strings at low
tension will deflect more than strings at high tension, with the result that the
string plane stiffness at low tension will be almost the same as string plane stiff-
ness at high tension. As a result, the actual increase in serve speed will be less
than 0.7% when the tension is reduced from 60 to 50 lb.

The best one could possibly do to increase the serve speed is to change the
rules of tennis so that no energy is lost in the ball or the strings at all, and then
get the manufacturers to come up with the appropriate lossless balls and
strings. Then e = 1, and for the the above racquet, e, = 0.50. One could serve
this magic ball at a speed v = (1+ eV = 1.5V rather than at speed v = 1.4V
 Strings and Power — Part 2 297

using a conventional ball. A 140 mph serve would then become a 150 mph
serve. At the opposite extreme, suppose that we use a totally bounceless plas-
ticene ball with e = 0. Then e, = -0.25 for the above racquet, and the ball
would be served at speed v = 0.75V. The ball would stick to the strings when
it was served (or perhaps get sliced up into several pieces), and the racquet
would slow down from speed V to speed 0.75 when it hit the ball. Even if the
ball was ejected out of the strings, it would be impossible to return the ball
since it would roll along the court instead of bouncing.

We have estimated the effect on serve speed of changing string tension by con-
sidering the effect on the ACOR. The same answer is obtained if we examine
the way in which energy is shared between the racquet and the ball. When you
serve a tennis ball, only about 1/4 of the initial kinetic energy in the racquet is
converted to elastic energy in the ball and the strings. This is explained in detail
in Chapter 12. If the strings are as stiff as the ball, then half of the elastic ener-
gy is stored in the strings and the other half is stored in the ball. Essentially
none of the energy in the strings is lost and almost half of the energy in the
ball is lost. As a result, about 25% of the stored elastic energy is lost, and hence,
1/4 of that, or 6.25%, of the initial racquet energy is lost.

If the strings are only half as stiff as the ball, then 2/3 of the elastic energy is
stored in the strings and 1/3 is stored in the ball. If the ball loses half its stored
energy and the strings lose none, then 1/6, or 16.7% ,of the elastic energy is
lost, and hence, 1/4 of that, or 4.2%, of the initial racquet energy is lost.
Dropping the string tension by half therefore increases the total energy avail-
able by only 2%. Some of this is taken up by the ball and some is retained in
the racquet.

Suppose that the ball is served at speed v = 1.41V, where V is the speed of the
racquet at the impact point (1.41 looks like a strange number to pick, but it is
as good as any other number, and it happens to be the square root of 2, which
simplifies the following calculation). The ball speed will be 1.41V at a point
near the middle of the strings where the effective mass is about 1/2 the actual
racquet mass, M. Consequently, the kinetic energy of the racquet before the
collision is about MV’/4, and the kinetic energy of the ball after the collision is
5 mv? = 3 m(1.41V)* = mV. The energy given to the ball divided by the initial
racquet energy is therefore about 4m/M. This ratio is about 4 x 57/340 = 2/3
for most racquets. We should really consider rotation and translation of the rac-
quet to get an accurate value of the initial energy in the racquet (see Note 3),
but a rough estimate is good enough for our purpose here. For a serve, the ball
ends up with roughly 2/3 of the initial racquet energy, meaning that about 1/3
of the energy is retained in the racquet or lost. If the racquet came to a com-
plete stop, then all of the racquet energy that is not lost would be given to the
ball, but that doesn't happen when a ball is served.

Suppose that at a string tension of 60 Ib, the ball ends up with 67% of the ini-
tial racquet energy. If we decrease the string tension from 60 Ib to 30 Ib and
all of the 2% extra energy available goes into the ball, then the ball will end
up with 69% of the initial racquet energy. That represents a 3% increase in the
energy in the ball or a 1.5% increase in ball speed (see Note 4). Consequently,
a decrease in string tension from 60 Ib to 50 lb will result in an increase in ball
speed of less than 1%.
298 Chapter 32

Notes

1. If a 12-inch length of string is stretched until the tension is say 60 Ib, it will
stretch by about 1 inch, and then it will be about 13 inches long. If the pull
force is removed, then the string will shrink back towards its original length,
but it will remain slightly stretched, about 1/4 inch longer than its original
length. In this respect, strings are not perfectly resiliant, meaning that they do
not return to their original shape or length when they are stretched and when
the stretching force is removed. However, this has no relevance to the behav-
iour of a string in a racquet. When a string is installed in a racquet then each
12-inch length will stretch by about one inch, but if the string is tied to the
frame, then it will remain at that stretched length for the rest of its life. Every
time a ball is hit, the strings stretch a bit further, but they always return to their
strung length when the ball flies off the strings. In that respect strings are 100%
resiliant. However, the tension drops a tiny amount every time a ball is hit.
Strings never lose resiliance in a racquet, but they dolose tension.

2. The stiffness of a tennis ball depends on how fast it is compressed. The rules
of tennis (Appendix C) specify that a ball must compress by about 1/4 inch
under a load of 18 lb so the ball stiffness is about 18/0.25 = 72 lb/inch when
the ball is compressed slowly. In SI units, this corresponds to a compression
of 6.25 mm under a load of 8.2 kg (80 N) and a stiffness of 80/0.00625 = 12.8
kN/m. When a ball is tested according to the rules it is compressed much more
slowly than during a normal bounce. A ball incident at 13.5 m/s on concrete
compresses by about 4 mm when the force on the ball is 240 N and the stiff-
ness is then 240/0.004 = 60 kN/m (337 Ib/in). The ball is much stiffer during a
fast compression since it takes about 0.2 ms for the wall to buckle. After the
wall buckles, the stiffness drops to about 17 kN/m (96 lb/in). Further details
are given in Chapter 37.

3. The kinetic energy of a racquet is

KE = $MV_2
Nile cm.
+ cm
41?

where M is the actual racquet mass, V. cm. is the speed of its center of mass, Icm
is the moment of inertia about its center of mass, and @ is the angular speed
of the racquet. If the racquet is rotating about an axis that is a distance h from
its center of mass, then the moment of inertia, 1, about that axis is given by the
parallel axis theorem

T=]. cm
+ Mh?

The velocity of the center of mass is Nee TO)

KE ==e 2)3Mh’o? + 31.cm w? = 1

2

Alternatively we can write this as

cE
KE SMV 2 ewhere Vo Dihe = (g 2 ew2
Nie

V is the speed of a point on the racquet a few cm past the balance point of
the racquet. For example, suppose that a uniform racquet of length L = 70 cm
 Strings and Power — Part 2 299

is rotating at @ = 80 rad/s about an axis that is 10 cm beyond the butt end of

the handle. Since the balance point is in the middle of the racquet (at the cen-
ter of mass), then h = 45 cm, V_, = hw = 36 m/s , 7
I,em = MI#/12 and

V2 = 362 + (0.49/12) x 802 so V = 39.5 m/s.

If V = dw then d = 0.493 m, meaning that V is the speed of a point 4.3 cm past

the balance point (towards the middle of the strings).

The kinetic energy of a racquet can also be expressed as +M.V?, where V is

the speed of the impact point, and M, is the equivalent mass at the impact
point. This relation is exact during and immediately after a racquet and ball col-
lision, but not before the collision. Prior to a collision with the ball, the racquet
can be swung about any axis that the player chooses. It turns out that in this
situation, +M.V? is typically about 10% less than the actual kinetic energy of
the racquet.

4, There are many mathematical tricks that physicists use to simplify calcula-
tions so that the calculations can be done in their head rather than on a cal-
culator. Students these days are trained to use calculators rather than having to
rely on “mental” arithmetic, so these tricks are not as well known as they
should be. Most of the tricks give an answer that is close enough for practical
purposes, but they are not exactly correct. For example, if I buy 20 items from
the supermarket, then I expect to pay around $80 since the average price of
each item is typically about $4. Similarly x? is 9.87, but that is close enough to
10 if I want to figure out the answer to a calculation in my head. If I want the
square of 6.5, then that's about 6 x 7 = 42, but I know I should add 0.25 to the
answer, so the answer is actually 42.25. I can add up all the numbers from 1
to 100 in my head since 1 + 99 = 100, 2 + 98 = 100, 3 + 97 = 100. There are
about 50 such pairs of 100 so the answer is about 5000.

If I increase x by 10%, then x? will increase by about 20%. For example, if x =

10 then an increase of 10% gives x = 11, and x? increases from 100 to 121.
That's an increase of 21%, but it's close enough to 20% not to worry about. If
the kinetic energy of a tennis ball increases by 3% then the increase in speed
is actually 1.489%, so my estimate of 1.5% was pretty close. The general rule is
that if y = x", then an increase in x of p% will give an increase in y of about
np%. The answer improves as p gets smaller. Mathematically, dy/dx = nx“") =
ny/x so dy/y = n(dx/x), where dy is a small increase in y and dx is a small
increase in X.

One of my favorite tricks (admittedly a bit esoteric) is to get the quality factor,
Q, of a damped vibration such as the damped sine wave in Figure 7.2. The
quality factor of a vibrating system is formally defined as Q = 2m x (energy
stored)/(energy lost each cycle). A high Q oscillation is one that damps out
very slowly since very little energy is lost during each cycle. A low Q oscilla-
tion is one that damps quickly. I can tell by looking at Figure 7.2 that the Q is
about 15 since that's the number of cycles I can count. It can be shown math-
ematically that the amplitude of an oscillation decreases to 1/27 th of its origi-
nal amplitude after Q cycles.
 Tension
Factors in Tension
and Tension Loss

Mikael Pernfors, Sweden

By Crawford Lindsey

nce upon a time, string tensioning was done by hand. The string was twisted around a wooden dowel
to apply tension. The stringer plucked the string, listening for just the right “reference” tone, and then
“clamped off” with an awl.

The crosses were done differently. Since weaving damps tonal pitch, stringers would try to match the humps in
the weave instead. This didn’t necessarily match the mains, but an experienced stringer could judge what ten-
sion (hump) he was pulling compared to the mains and how it would affect the overall stringbed stiffness.

In each case, stringers were trying to achieve the same string stiffness on each string — not the same tension.
The mains were tonally set such that as you go out from the center mains, the pitch gets lower. Shorter strings
are stiffer at the same tension than longer strings, so stringers would pull each shorter main with less tension,
thus giving it a lower pitch but close to the same stiffness as the string on each side of it. This technique was
based on the principle that a string’s frequency varies depending on the length, gauge, and tension of the string.

Matching humps, though not as scientific, did the same thing. Hump size varied in stiffness depending on the
pattern density and the tension at each point. Experience dictated the “reference hump.” This was the art of the
stringbed. Even though the technique seems prehistoric compared to our age of continuous pull digital readout
machines, there was a lot of science there also.
302 Chapter 33

We have come a long way since those days, however, the key both then and
now is stringbed stiffness, not string tension. Tension is just a means of achiev-
ing different degrees of stiffness at different locations. It is not the end in itself.

WhatIs Tensi
Newton’s Third Law of Motion says that every action must have an equal and
opposite reaction. Newton wasn’t a stringer, but he knew that if the reference
stringing tension were 50 pounds, but the string pulled back at more than 50
pounds, then the stringing machine or its operator would go flying out the win-
dow. It’s the same with ball impact. The string deflects just enough to exactly
match the combined energy of the incoming ball and the swing of the racquet.
Your racquet is an oddly shaped atomic computer that instantaneously figures
out how much to deflect to bring the ball to a complete stop and then turn it
around. This is really kind of magical. What is happening?

It’s all about tension. My computer does not fall through my desk because the
force of its weight is exactly matched by the desk. The desk does this by
becoming tensioned (actually, compressed in this case). It deforms just like
strings do. Deformation is necessary to create tension or compression.
Deformation is all about pushing and pulling on atoms. All objects are made
of atoms in relative equilibrium, and atoms are made of a balance of positive-
ly and negatively charged particles. When atoms are pushed closer together,
past their equilibrium positions (compression), same charge particles build up
repulsive forces to each other preventing them from being pushed any closer
Figure 33.1 by an external force of that magnitude. If
atoms are pulled apart, the attractive forces
of opposite charges try to pull them back
into equilibrium. The further you pull them
apart, the greater the force required to do
so (up to breaking tensile strength). This is
what we feel as tension.

But not all tensions are created equal. If

you pull every string in a racquet at the
same tension, which is typically the case,
then every string will deflect a different
amount when hit by a tennis ball. Figure
33.1 shows a graphic of a representative
string plane deflection at different impact
points.

The variances from point to point are a

result of pulling different length strings at
(mm) the same tension. So, when you ask a
Strings
of
Deflection
stringer to string a racquet at 50 pounds,
you are really saying, “I want every point
on my racquet to play differently from
every other point by the amount caused by
a 50 pound tension.” Well, that’s not what
you mean at all, though that is what you’re
getting. Most players probably think that 50
pounds will make their racquets play the
same everywhere. In general, and all things
 String Tension 303

being equal except string length, identical tensions will mean different deflec-
tions and stiffnesses. Strings with the same stiffness will have equal deflections
but different tensions. The stiffness resulting from any given reference tension
is totally relative to the length, gauge, materials, construction, and properties
of a string. If you hang a weight from two unequal length strings at the same
tension, the longer string deflects more. For this reason, some players have
their racquets strung by the “proportional stringing method.” With this method,
the string tension is lowered on each string in direct proportion to its percent-
age length of the longest main strings.

Oh, if this were the only complication in determining how tension will affect
the feel of the stringbed. The rest of this chapter details the real bugaboo —
string tension is always changing.

String Tension Is a Moving Target

As soon as you pull and clamp the string, it is a different tension. Ten seconds
later, it’s different again. Two weeks later, different still. Every day your racquet
is at a different dower) tension than the day before. The reference tension is
the setting on the stringing machine for the brief instant of the pull on a spring
tension machine or until it is clamped on a continuous pull machine. So, if the
tension is never the same anyway, what’s the big deal about tension? Players
and stringers obsess about a pound here and a pound there. Does any of this
make a difference if it is just going to go away?

String loses tension by a process called “creep.” It works in two ways — creep
elongation and creep tension relaxation. A string pulled at a constant force will
elongate forever. It does so at a decreasing rate as time goes by, however.
Tension a string in a continuous pull machine and come back a few years later,
and the tension head will be at the end of the track. On the other hand, a string
held at a constant length (as in a racquet) can’t get longer, so, instead, it loses
tension at a continuously decreasing rate. It’s easy to picture elongation during
pulling, but tension relaxation is more difficult because you haven’t actually let
go of the string. Where does this tension go and why? After all, we pulled it to
50, tied it off at 50, and now the racquet is holding it there at 50. Right? Wrong.
We will see that a string loses tension for the same reason it continues to elon-
gate. It all has to do with the alignment and stretching of molecules. But they
don’t all stretch the same amount at the same rate. This results in a tension
equalization between them as time goes by.

How much tension are we talking about? Figure 33.2 shows the two faces of
creep.

_Why Does
As you can see, most of the creep happens in the first 20 to 30 seconds,
whether through elongation or tension loss. This happens because string has a
split personality. It is viscoelastic which means it has both viscous (which takes
time) and elastic (which happens instantly) behavior. If you were to stand on
top of a vat of honey, the force of your weight would cause you to slowly sink.
If instead you stood on top of a big pogo stick, you would immediately depress
the spring to equal your weight. In both cases, the weight is the same, but in
304 Chapter 33

Pulling Tension Clamping Off

AS.
Creep Elongation Creep Tension Loss

= 0.25

o i)is)

jo)Bes oa
(Elongation

(>)=e oO

ane 0 10 20 30 40 50 60 70 80 90 100 110120 0 1 2 3 4 10 20 30 40 50 60

Time (seconds) Time (seconds) after clamping

Data: Dr. Carl Love Data: Dr. Carl Love

Figure 33.2

If you leave a string at 50 pounds in the tension head of a con- Typical behavior of a nylon string when clamping
tinuous pull machine, the string will continue to elongate. The immediately after pulling tension. The majority of the
string is at 50 pounds no matter when you clamp off, but the tension loss occurs in the first 10 to 30 seconds (as
internal configuration of the string depends on when you do it. with elongation in graph at left). String is always
It's as if early on the string is just faking it, never really getting moving — it elongates forever as long as you are
to 50. As you keep pulling, it gives up the fight, stabilizes, and pulling on it. The second you clamp it off and con-
says, "Ok, you win, | will stay here if you just stop pulling on strain its length, it begins losing tension. It is never
me." On the steep part of the curve, all the polymer molecule both a constant length and tension at the same
chains (the molecular-sized parts) haven't had time to share the time. This is the stringer's uncertainty principle: if
load equally, so when you clamp off, there will be a substantial you know either the tension or the length, you can't
tension equalization between molecule chains. The overall ten- exactly know the other. One or the other is always
sion will then go down. See right. changing.

Figure 33.3 one, the result of the force is time dependent, and in
the other, it is instantaneous. String is a combination
of the two.
How String Loses Tension: An Analogy
To visualize how this split personality contributes to
0 Ibs tension
* spring and honey jar
elongation and tension loss in a string, let’s take our
|<< Dipper honey and spring analogy further. String elongates as
|
Ibs_|
60 if it were composed of a coiled steel spring attached
* spring elongates to a dipper in a jar of honey (Figure 33.3). When the
* dipper remains stationary string is tensioned, the spring immediately elongates
60 Ibs plus1 minute
but the dipper stays put. As the continuous pull
honey dipper moving adding machine keeps pulling, the spring remains stretched
to longer overall length
* same spring length
at the same length, but the dipper starts to move in
the honey. As it does so, the string continues to elon-
60 Ibs clamp after-1 minute
* same overall length gate. Now, if we clamp off, the time dependent
* dipper pulled out more pulling on the dipper continues because it is. still
* spring decreases
* tension declines being pulled by the force of the spring. But since the
length of the string is now fixed, the spring must
decrease its elongation to make up for the movement
 String Tension 305

(“elongation”) of the dipper. And the tension necessary to maintain the spring’s
new length is less, so its tension goes down. This in turn results in less pull on
the honey dipper, so it pulls out less and the tension in turn declines, but less
than before. This creep circle continues indefinitely. This is a simplified analo-
gy of the actual process, but it allows us to visualize how this creepy behavior
is possible.

The real life counterpart is this: the force of tensioning or impact is immedi-
ately reacted to by the atomic pogo sticks (the bonds between atoms) in the
straight sections of the polymer chains, but it takes time for the other less ori-
ented chain molecules to unfold, straighten, untangle, and move relative to
each other. As a result, some molecule chains are bearing more of the load
than others when the string is first pulled. The faster and harder it is pulled,
the more of the load these chains will bear in proportion to the rest. When the
string is clamped, time takes over and the tensions on the various chains begin
to equalize. As they do, the overall tension goes down, though the length stays
the same, as it must. It’s the same as happens in a racquet when strings are
pulled at different tensions. As soon as you release the clamp between the two
strings, they equalize their tension, yet keep their length.

What does this mean to a stringer, if anything? Figure 33.4 demonstrates what
happens to a 70-pound string throughout the stringing process. With a force

CHANGE IN STRING TENSION DURING STRINGING

70
| * Reference tension: 70 lbs
68 * Final tension: 58.43 Ibs
| * 10X @ tension head: 70 Ibs
66 * Tension 10X opposite side of racquet: 59.89 Ibs 1M ony
= | * 10X final: 51.56 Ibs
oD - After 12 hours: 1M: 57.91; 10X: 48.54 Ibs
a oe | Creep Loss:
a Tension on 1M
5 pe | measured after
g@ 60 | pee aca’ Weave Gain: Tension on 1M
= measured after each cross
S string was installed
3 |
=

ide

1M 2M 3M 4M 5M 6M 7M 8M 1X 2X 3X 4X 5X 6X 7X 8X 9X 10X11X12X13X
14X 15X 16X 17X 18X 19X
Pattern: 16 mains (M, 8 on each
. . Data: Dr. Carl Love
String Location side) and 19 crosses (X).
(Tension on 1M measured after installation of each string)
Figure 33.4
String 1RM (first right main) was monitored throughout the stringing process. The tensions above are
the tension on string 1RM immediately after each string in the racquet was installed. While the mains
are being installed, tension goes way down due to creep. During cross string installation, tension goes
back up because each weave elongates the main a little and frame deformation occurs pulling in the
sides, thus increasing the tension 0.2—-0.5 pounds per string. Creep is still occurring at this point, but it
has stabilized at a much slower rate.
306 Chapter 33

gauge placed on the first main and the center cross, tension was measured
throughout the stringing procedure.

As you can see, during stringing, the main string tension drops way below ref-
erence tension (the same is happening for every other string also). But then
the weaving of the cross strings brings it back up. Each hump in the weave
elongates the main string a little, thus increasing its tension and the frame dis-
torts, pulling the sides in and lengthening the head from tip to throat.

And, of course, the same applies to the cross strings, but with a catch. Not only
does the same tension loss occur after clamping, but it does so by a different
amount on different parts of the string. That’s because, though the string is 70
pounds at the tension head, it is only a percentage of that as you travel across
the racquet. The furthest part of the string from the tension head is not being
pulled as tightly because the tension is gobbled up by friction in the weaves.
The difference can be as much as 15 pounds depending on the texture of the
string and the reference tension. When you clamp off, the end with the high-
er tension will creep more and faster than the end at lower tension. And the
cross string going back in the opposite direction will do the same in the oppo-
site order. So it cancels itself out. But again, 70 pounds isn’t always 70 pounds.

Tension Loss and“~

Every string has a different structure, material, and degree of polymer chain
alignment, so each string will creep differently than the next. So, is there any-
thing you can do about it? Are you destined to be a victim of creep?

Well, yes and no. It can’t be stopped, but it can be reduced. The question is,
should it be reduced? This is where the argument for and against prestretch-
ing comes in. There are three kinds of prestretching. The first takes place in
manufacturing. Here a string is stretched at high temperature to get the desired
alignment of the polymer chains. When cooled, this alignment is set, deter-
mining many of the string’s properties.

The second, and most familiar, prestretching is the old door knob method.
Here you tie the string to the door knob and pull the other end with the force
of your weight. This is to straighten (get the coils out) and/or take some of the
stretch out before it is installed.

The last method (used separately or in combination with the second) is to

allow the tensioned string to dwell in the tension head of a continuous pull
machine for an extra length of time (e.g., 10-30 seconds). This further elongates
the string and stabilizes the creep. We will concentrate on this method here
because it is easiest to work with for illustrative purposes, and it actually comes
into play every time you string, whether you know it or not.

Figure 33.5 shows three different illustrations of how different combinations

of tensions and pulling techniques can result in many different stringbeds.
Two racquets strung at 60 pounds with and without extra tension head dwell
time or prestretching will be very different stringbeds. The prestretched racquet
will be much stiffer. But if you’re not looking for a stiffer feel, that’s not good.
To compensate, you could still prestretch, but string at a lower tension to get
a softer stringbed. But then you ask, why bother with the prestretch if I could
just end up at the same final tension by using a higher tension to begin with
and without all the extra effort?
 String Tension 307

Case #1: Different Reference Tensions Case #2: Same Reference Tension Case #3: Different Reference Tensions,
but Same Method (Pull & Clamp) but Different Methods Different Methods, Same Result?
65 65 65
60 second dwell
(or go answer phone)
D oO
62 62

oloOo
30 second dwell
65 Ib ref. tension 59 (or answer a customer's question) 59

oO(=)

55 Ib ref. tension 56 56

Pull and clamp

53
45 lb ref. tension
Tension
machine)
(any
Ibs BSoO
pull)
(continuous
Ibs
Tension Pull and clamp Tension
(continuous
Ibs
pull)
30 second dwell
50 ——
(ge)ol
OR Sea OMAIS
I 20Re25res0
——
O° 256i OF go 4 506 0 5 10 15 20 25 30
Time (seconds) Time (min.) Time (hours) Data: Dr. Carl Love

Figure 33.5
Higher reference tensions lose more Three stringers starting at the same Two stringers starting at different
tension faster during stringing. The 65 reference tension does not mean that reference tensions but using differ-
Ib string lost 23.4%, the 55 Ib string they will all end at the same final ten- ent methods end up at the same
lost 20.2%, and the 45 Ib string lost sion. Stringing technique, specifically, final tension. (Actually, the graphs
15.3% tension in 30 seconds. The dif- tension head dwell time or prestretch- just intersect and are still losing ten-
ference between final and reference ing, can make a big difference. These sion at different rates.) The question
tension thus depends on reference are completely different stringbeds. is, are these similar stringbeds
tension. Each string started 10 pounds The more you elongate, the less ten- when they intersect? Even though
apart and ended up only 5 or 6 sion you lose and the more stable the you end up in the same place, does
pounds apart. tension over time, but it's also much it matter how you got there? See
stiffer. text for explanation.

We begin to see that there is a whole series of combinations that end up with
the same internal molecular configuration. In principle, you could create a
chart that describes the final tensions arrived at by using different combinations
of tension and pull time. There should be several combinations that arrive at
the same equilibrium tension after a suitable time period. For example, after
sitting for 24 hours, a string pulled to 70 pounds for 3 seconds might be the
same as one pulled to 50 pounds for 60 seconds or 55 for 30, or 60 for 7. If
that’s true, then does it matter how you get there? Table 33.1 explores this.

The figure shows some surprising results. First, let's make some generaliza-
tions.
1. The longer you pull on the string at any tension, the more it elon-
gates and the smaller the diameter becomes.
Table 33.1

TIME AND TENSION

Tension Pull Time Diameter Stress @ Pull Elongation % 12 Hr-Tension 12 Hr-Stress

70 lbs OES -4.57% 35,508 Ibs/in2 13.64% 48.10 Ibs 24,399 Ibs/in@
65 lbs 60s -6.90% 35,039 Ibs/in2 14.87% 58.20 Ibs 31,373 Ibs/in@
60 Ibs 60s -7.12% 32,747 Ibs/in2 13.78% 55.98 Ibs 30,553 Ibs/in2
55 Ibs 60s -5.84% 29,894 Ibs/in2 13.71% 51.57 lbs 28,030 Ibs/in2
50 Ibs 60s -5.62% 26,842 Ibs/in2 12.09% 46.65 lbs 25,044 Ibs/in2
Data: Dr. Carl Love
308 Chapter 33

2. A string at a lower tension pulled for a longer time may elongate

more and become thinner than a string pulled at a higher tension for
a shorter time interval.
3. If time of pull is constant (e.g., 60 seconds in this case), then change
in diameter, length, final tension, and final stress declines as reference
tension declines.

Look at the string tensioned to 70 pounds and clamped immediately. Its ten-
sion and stress 12 hours later are closest to the 50 pound string that dwelled
in the continuous pull tension head for 60 seconds. That’s remarkable. The
string pulled to the highest tension equals the one pulled to the lowest tension,
but for a longer amount of time. Yet, though close to the same 12-hour ten-
sion, the elongation and change in diameters were quite different, which indi-
cates that the internal configuration of the molecular alignment is probably
somewhat dissimilar. If so, they still will play a little differently, but not as if
they were strung 20 pounds apart.

ttention to Detail

So, can you get to the same place starting at different tensions? Maybe. The key
is understanding where “there” is. If a player likes the tension and string you
used last time, that’s easy. The key, however, is in reproducing your method.
As the graphs show, the difference between clamping off 4, 7 or 10 seconds
on different strings will result in a big difference on where those strings settle
out compared to each other. The key is consistency. Consistency from string to
string and from string job to string job for each customer. You want to clamp
off at the same point on the elongation curve every time.

On a spring tension machine, this happens automatically because when the

tension head locks out, the string stops elongating and tension relaxation sets
in immediately. So clamp time is not as important on this type of machine. But
even with this machine, if you prestretched the string prior to stringing, then
time once again enters the equation, but in a different way. How long did you
prestretch, at what tension (hard to determine on a door knob), and how long
has it been since you prestretched? The last is the most important. Over time,
Figure 33.6 a good percentage of the creep elongation and molecular
alignment caused by prestretching is lost. So, if the pre-
stretch duration, tension, and time since the prestretch are
Pull Angle and Tension Loss not all exactly the same each time, you put the string in
the racquet (for the same string model), then you are
installing a different string each time. Each string job will
then have a different stiffness at the same tension.

andTension __
Another pesky little variable contributes to final tension
not being the same as reference tension — the angle of
To tension head pull. This refers to the angle the tension head makes with
"40°
When pulling at an angle, not all of the force is
the pulled string. Depending on which string you are
concentrated along the line of the string. Some of the pulling, as the string comes out of the grommet, it bends
force is lost in pulling sideways. Tension is also lost in at a certain angle as it enters the tension head (Figure
the added friction caused by the angle.
33.6). Pulling at this angle will result in less tension than
Source: Dr. Carl Love if the tension head were in a direct line with the string as
it emerges from the grommet. Table 33.2 shows the result-
 String Tension 309

ing tension as measured at two different angles

with three different tensions. In point of fact, then, Pull Angle and Tension Loss
you are pulling each string at a different tension if
it is pulled at a different angle. And because each String Tension (Ibs)
string is at a different tension, when you release the
50
clamp between them, they will equalize some-
where in between. 46.1 37.0
40° m0.5 44.9 34.8
This is not bad either. This is the way stringing
machines are designed (except True Tension Source: Dr. Carl Love
machines). Still, some stringers have been known
to compensate by locking the tray table in a straight Table 33.2
line position for each string pull. This will make the
resulting stringbed much stiffer. But you don’t have
to do this because matching reference tension in
the final product is not your goal. It is to duplicate
the stringbed stiffness that the player likes.

3 Tension Hea

When you pull tension, the length of string

between the grommet and the tension head is
essentially being prestretched. If the tension head Figure 33.7
is in a different position on the track each time you
pull, then some lengths might be pretensioned and
others not. This might not seem important if you Temperature and Tension Loss
pull, clamp, and release very quickly, but in fact, it
is. It amounts to a double pull which can be shown iS MM 45 degrees
to increase stringbed stiffness significantly. As long ™ 105 degrees
as you pull from the same spot on the track, you
will pretension the same amount of string and adja-
cent strings will have the same degree of stiffness.
If you don’t, the further back on the track you pull,
the more you prestretch and the tighter will be
your stringbed.

What if you string one racquet in your garage in the

Tension
morning at 45 degrees ambient temperature and
one in the afternoon at 105 degrees? If you string
them both at 70 pounds, will they result in the
same stringbed? The graph in Figure 33.7 shows
what happens. At higher temperatures, molecules
move more easily relative to each other, so the
creep curve is accelerated. At lower temperatures,
it slows down.

Tension loss while stringing occurs all over the 1

place. It’s like a big leak in the dike. But you don’t
Data: Dr. Carl Love
necessarily have to stick your finger in it to plug it
310 Chapter 33

up. Tension loss is not a bad thing (it’s natural and unavoidable) as long as you
understand it and know how your actions with respect to it can influence the
final stringbed stiffness. The goal of stringing is not to be at the reference ten-
sion when you finish the job. That would be an unbearably stiff stringbed. The
goal is to be at the desired stringbed stiffness produced by setting reference
tension at a certain value for a certain string. Stiffness is not tension. Stiffness
is combination of string type plus the tension and the time at tension before
clamping or the pretstretching procedure used, if any. On a continuous pull
machine, time at tension means more elongation and molecular alignment and
results in a stiffer stringbed. Same tension, different feel. On a spring tension
machine, the issue only arises if you prestretched ahead of time or double
pulled. Either way, time is the key element. Any procedure related to elongat-
ing a string must follow an exact procedure with each task exactly timed to
reproduce the same results, string job after string job. And the same applies to
the pull angle, tension head location when pulling, and room temperature.
Fancy techniques don’t mean anything unless everything else is also kept con-
sistent. ;
 D
ot
@
=
co
oO
ta

How To Measure It
and What It Means

Lindsay Davenport, United States

By Rod Cross

tis easy to string a racquet at, say, 60 pounds or 65 pounds and to know that it is strung at that tension, but
how do you check the tension after the racquet is strung, say one week later or even one minute later? There
are several ways, as experienced stringers will already know. The most common way is to hit the strings with
one hand or with another racquet and listen to the sound they make. A high pitch "ping" means the strings are
tight and a low pitch means the strings are loose. A better way is to use a racquet diagnostic machine, or a
Stringmeter, or maybe a fancy electronic gadget such as the ERT700. But how do you know if these devices give
the correct answer? The problem is how can they be tested? What else can they be tested against?

When you string a racquet, one end of the string is tied down and the other end can be connected to a spring
or a strain gauge to measure the tension. But when both ends of the string are tied to the frame, this method
of measuring the tension is no longer available.

The next best method is to push the string sideways with a known force to see how far the stringbed deflects.
That is the method used in Pacific and Babolat racquet diagnostic machines (RDC) (Figure 34.1).
312 Chapter 34

The answer depends on the string tension as well as

several other factors such as the length of each string,
the number of strings, the stiffness of each string, and
the diameter of the object (usually cylindrical) that is
used to push the strings sideways. For that reason, a
racquet diagnostic machine can’t give a reliable meas-
ure of the string tension, but it does give a reliable
measure of the stringbed stiffness. In fact, the
stringbed stiffness is more important than the tension
itself, since the stringbed stiffness determines how far
the strings will deflect, and how much the ball will
squash when you hit a tennis ball. Even so, stringbed
stiffness is usually quoted in DA or RA units ("made
Figure 34.1 up" units indexed against the machine’s zero poinD,
RDC stringbed stiffness measuremen.t
which is just another way of admitting that the actu-
al stringbed stiffness is not known. (Another prob-
lem, of course, is that the actual stringbed stiffness
depends on how hard you hit the ball. The stringbed
gets stiffer the more it stretches, but that is another
story.)

Figures 34.2 and 34.3

Top and underside views of Stringmeter recording main

string tension. A racquet diagnostic machine is useful in measuring
whether the string tension or the stringbed stiffness
has decreased over time, and it will indicate roughly
by how much, but it still doesn’t give an accurate or
absolute measurement of the actual string tension
because all the other factors are not properly
accounted for. That is why the Stringmeter was
invented.

The Stringmeter contains two small cylindrical rods

1/3 of an inch apart that can be inserted anywhere in
the stringbed to measure the tension of any desired
string. The idea is to rotate the rods so that a partic-
ular string is pushed on one side and pushed in the
other direction on the other side to form a little kink
in the string (Figures 34.2 and 34.3). The rods are
connected to a spring-loaded dial that is marked with
graduations to read the tension in Ib or kg. There is
also a small adjustment needed for string diameter,
which one does by rotating the position of the scales
slightly. The only problem with this is that it is diffi-
cult to know whether the answer is correct. One can
test the device on a single long string that is pulled to
60 pounds, and it will correctly read 60 pounds, but
that is not the same as testing it on a fully strung rac-
quet. In fact, the Stringmeter has three separate scales
— one for long straight strings, another in pounds for
a strung racquet, and also a kg scale for a strung rac-
quet. The difference is due to the fact that the kink in
a single long string has a different shape to the kink
when the strings overlap.
 Measuring Tension 313

Friction between the mains and the cross-strings prevents the string kinking
over the full length of the string, with the result that the string kinks only over
a short section of the string. A bigger force is needed to kink a short string than
a long string, by an amount that is difficult to estimate with any accuracy.
Consequently, the pounds and kilogram scales on the Stringmeter are difficult
to check to determine if they are correct. Nevertheless, a Stringmeter can be
used to determine if the tension has dropped with time, and roughly by how
much, and it can be used to measure the tension in every single string if you
have the patience. A surprising result is that the tension in the cross-strings is
often a lot less than the tension in the mains. The Stringmeter is the only instru-
ment I know of that can give this particular information. In principle, one could
use a Strain gauge located on every string to measure the tension in each string,
but the Stringmeter is a lot more convenient.

Every other method of measuring string tension measures all the strings at
once, so these methods give only the average string tension in all the strings
— not the tension in any given string. A high-tech method is used in the
ERT700 (Figure 34.4).

It is high-tech since it contains two boxes of elec-

tronics and a digital display, but the basic method is
fairly simple. The ERT700 forces the strings to vibrate
back and forth, in much the same way that they
vibrate when you hit them with your hand. The point
of doing this is that the vibration frequency depends
Figure 34.4
on the string tension. It is the same as in a piano or
guitar or violin. Tuning up a guitar involves adjust- ERT stringbed stiffness measurement.
ment of the tension in each string so that it vibrates at
just the right frequency. If you increase the tension it
will vibrate faster and the pitch rises. If you drop the
tension, the string vibrates more slowly and the pitch
drops. The ERT700 detects the frequency of vibration
of the whole string plane and assigns a number to the
tension. However, the tension indicated by the
ERT700 is not the actual tension since the vibration
frequency depends on string tension, string length,
the area of the stringbed, the mass of each string, the
number of strings, and the actual string pattern. The
formula for the actual frequency is too complicated to
figure out the actual tension.

The ERT700 uses a small plastic box (which they call

a ball simulator) that is clipped onto the strings, w
Ww

preferably in the middle of the string plane. The box

io
Q

has a mass of 34 gm, which is a bit less than a stan- 4

dard 57 gm tennis ball, but bigger than the mass of Zz
g
wu

the strings themselves. The strings have a mass typi-

cally about 15 or 16 gm. Without the box in place, the
strings themselves will vibrate at about 500 Hz or 600
Hz when you hit them, depending on the string ten-
sion. When the box is added, the extra mass of the
box slows down the vibration frequency to something
314 Chapter 34

between 140 and 180 Hz, again depending on the string tension. Inside the box
is a small brass weight attached to a piezo disk. When a voltage is applied to
the disk, the piezo causes the brass weight to vibrate in the same way that a
piezo buzzer works in a watch, or a computer, or an alarm, to make a buzzing
sound. By varying the vibration frequency of the piezo until it matches the fre-
quency of the strings plus the simulator box, a resonance frequency is estab-
lished and it is used to give readings of string tension, dynamic tension, and
percent power vs control. All these numbers are based on one measurement
of the frequency, so they are all just different ways of describing the same num-
ber.

Provided that one is willing to accept that all the numbers are artificial, then
the ERT provides a useful and easy-to-use guide as to whether string tension
has dropped over time and by roughly how much. The power reading varies
from 20% to 80%, even though a decrease in 10 pounds in string tension will
increase the ball speed by less than 1%. In other words, the readings need to
be taken with a grain of salt. :

_Artificial Tension Calibrati

The manufacturers of racquet diagnostic machines and the ERT700 have decid-
ed that the simplest thing to do is to make sure that the tension indicated by
their device is nearly the same as the pull tension, at least immediately after a
racquet is strung. As time goes by, the tension drops, and these devices will
indicate that the tension drops, and they agree fairly closely with each other,
but they all give a reading that is higher than the actual tension. There are four
good reasons why the tension immediately after stringing is, in fact, a lot lower
than the pull tension, as described below. However, stringers and customers
are unlikely to be happy about this. If you string a racquet at a pull tension of,
say, 60 pounds, and the actual tension immediately after stringing is only 40
pounds, then what will you tell your customer? If the customer asked for 60
pounds and you give him 40 pounds, then he is not going to be very
impressed. That’s why these tension measuring devices are calibrated to read
the pull tension rather than the actual tension. It keeps everyone happy and
no one knows the difference, including the stringer.

Bond Slipping. If a string is stretched to a tension of 60 pounds, and then

clamped at a fixed length, the tension will immediately start to drop since the
bonds in the string keep breaking whenever the string is under tension. The
weakest bonds break first, but it takes a longer time to break the stronger
bonds. The tension will drop by about 8 pounds in the first 15 minutes. After
that, the tension drops at a much slower rate. However, the main point is that
the tension drops dramatically while a racquet is being strung. After that, the
loss in tension is hardly noticeable. Polyester strings drop the most in the first
15 minutes.

Frame Distortion. Another problem is that all racquet frames are light and
flexible. Even if you clamp the head at six or more points, the bits that aren't
clamped can move during the stringing process. Graphite racquets are made to
be stiff in a direction perpendicular to the string plane, but that is done at the
expense of making them soft in a direction parallel to the string plane. Every
time you add another string, it pulls the frame inwards in the direction of the
 Measuring Tension 315

string, and stretches the frame outwards in a direction perpendicular to the

string. If you add another cross string, the tension in the previous cross-strings
drops, and the tension in the mains increases. That is why the tension in the
cross-strings is usually a fair bit less than the tension in the mains. The total
pulling force of say 10 strings each at 60 pounds is 600 pounds. That is a huge
force on the frame, and it is no wonder that the frame distorts even when it is
clamped at six or more points. A poorly mounted racquet can expand in width
by an inch or more when the mains are strung. That is a terrible thing to do to
the frame, but people do it, and then say it doesn’t matter since the frame pulls
back into shape when the crosses are added.

Friction. Friction is another reason why the string tension after stringing is a
lot less than the pull tension. The tension at one end of a string in a racquet is
not necessarily the same as the tension at the other end. When the mains are
strung, then the tension at each end is the same, since there are no cross-strings
to interfere along the length of any given string. But when the crosses are
installed, they are woven through the mains. Friction between the mains and
the crosses then results in the tension at the pulling end of a cross-string being
higher than the tension at the other end. This is easy to demonstrate. Take a
three-foot-long string and weave it across the main strings in a fully strung rac-
quet. Clamp a 2-pound weight on the bottom end of the string. Then, holding
the racquet so it can’t move, lift the other end of the string. In order to lift the
weight at the bottom end, you will need to pull with a force of about 4 pounds.
That means the tension in the top end is 4 pounds and the tension in the bot-
tom end is only 2 pounds. Or, if the tension at the pull end is 60 pounds, the
tension at the other end will be about 30 pounds.

Pull Angle. The fourth reason has to do with the pull angle. The pulling head
on some machines has to be a bit lower than the racquet in order to rotate the
racquet. The string makes an angle of about 10 degrees with the horizontal
between the pulling head and the edge of the racquet frame. It then passes
through a grommet hole. Friction in the hole ensures that the tension after the
string passes through hole is less than the pull tension, typically by about 2
pounds.

This all adds up to a total tension loss of at least 10 pounds, and maybe a lot
more by the time the racquet is strung, depending on the stringing technique.

In 2001, Rob Bower (Lecturer in Sport Science at the University of Technology,

Sydney, Australia) and I strung five identical Volkl Pro Comp racquets all with
the same string but at five different tensions. We wanted to see how well play-
ers could pick the differences in tension just by hitting a ball a few times with
each racquet. They were instructed not to hit the strings with their hand or with
another racquet since the differences in the sound are easy to identify. A string
dampener was inserted in each racquet to deaden the sound, since we want-
ed the players to pick the differences in the feel or performance, not the dif-
ferences in the sound. After testing 41 different players it became obvious that
they were not very good at picking differences. More than half of them could
not pick a difference of 15 lb. We then tested 18 professional satellite players
at a tournament in Sydney. For them, we used four identical Topspin 630 Pro
racquets, all a bit heavier than the Volkl racquets and more suited to their
game. They were not much better than anyone else at picking the tensions.
316 Chapter 34

Eleven of the 18 satellite players could not pick a difference of 11 Ib, even
though most of them specify the tension to the nearest lb when stringing their
racquets.

Though many accomplished players can’t pick small differences in string ten-
sion, a small minority can. The above players were not asked to identify spe-
cific tension, just to choose which was strung looser or tighter than the other.
In the process of doing the experiment, we had to measure tension in many
different ways — reference tension, pull tension during stringing, stringbed stiff-
ness (not tension) after stringing, actual average stringbed tension (not stiff
ness) after stringing, actual individual string tensions after stringing, and per-
ceptual tension, which was what the experiment was all about. These are all
different concepts but very frequently used interchangeably, lumped into one,
and referred to simply as "string tension." To add to the confusion, the avail-
able methods to determine these things use different calibrations and measure
different parameters, yet often use the same vocabulary. Figuring out what we
were measuring was important to our test, but it is also important to the
stringer. It is important to know what each device is actually measuring and
how it is useful to you. The main thing we discovered is that all the devices
do a good job measuring relative changes in racquet tension, but actual ten-
sion was much harder to pin down. Fortunately, once the racquet is strung,
stringers and players are primarily interested in relative tension changes, not
absolute measurements. Unfortunately, many are not aware that is what they
are doing.

Determining ed Tension: Method |

The main thing we needed to be sure about in our experiment was the string
tension. The stringing machines were accurately calibrated against a load cell
to make sure the pull tension was correct, and exactly the same procedure was
used to string each racquet. At the time, we didn’t have a Stringmeter or an
ERT700 or a portable racquet diagnostic machine, so I invented my own way
of measuring the tension at regular intervals before the playtests and after-
wards, to make sure the differences in tension stayed about the same as when
the racquets were first strung. For that purpose, a small piezo was attached to
the strings to measure the vibration frequency of the strings. That was enough
to show me that the strings at high tension vibrated at a higher frequency (as
I could hear with my own ears) and that the relative differences in the tensions
were correctly maintained throughout the tests. The tensions all dropped slight-
ly over a few days, but not enough to worry about.

Bees soa sites

The average tension in all the strings can be calculated directly from the vibra-
tion frequency of the strings. The key to this measurement is the fact that the
vibration frequency depends on the total area, A, of the string plane and not
on its exact shape. In fact, the vibration frequency of the whole string plane is
almost exactly the same as that of a single string of length L, where A = L?
(Cross and Bower, 2001). This result indicated that the string tension in each
racquet was a lot lower than the pull tensions, as one might expect from the
fact that all strings drop rapidly in tension after clamping (see Chapter 33).
Nylon strings drop by about 9 lb in just 17 minutes, which is about the time it
takes to string a racquet. My tension results were even lower than that, so I
 Measuring Tension 317

checked the measurements using a microphone to record the sound of the

vibrating strings. This time the frequencies and hence the tensions were slight-
ly higher than I found with the piezo. The problem with the piezo was that it
weighed 2.2 gm, a lot less than the 15 or 16 gm weight of the strings, but
enough to lower the vibration frequency by about 15%. Even so, the tensions
were all about 30% lower than the pull tensions in all of the racquets, imme-
diately after stringing them. For example, the racquet strung at 60 Ib was actu-
ally at a tension of about 42 lb. But still, was the frequency determined tension
correct? The frequency calculation gives an average stringbed tension. How
does that compare to the tensions in individual strings?

The arrangement shown in Figure 34.5 was

used to investigate further.

A metal bracket was made up so that it could

be attached at the 3 o’clock position togeth-
er with a load cell to measure the tension in Load cell
two of the cross-strings. For this experiment reads 112
I used a 1.30 mm diameter polyester string.
After stringing the mains, the first seven
cross-strings were installed and the 8th and Full
; at
9th strings were threaded through a small Goh
bracket attached to the load cell. Each string
was tensioned to 62 lb. However, the reading
on the load cell was then 112 Ib. It should
have been 2 x 62 = 124 lb since two strings
each at 62 lb were pulling on the load cell.
The alignment of the pulling head, together
with friction in the grommet holes and fric-
tion between the mains and crosses all acted
to reduce the tension from 124 to 112 lb. As
each additional cross-string was added, the
reading on the load cell dropped further due
to contraction of the frame. The frame was Figure 34.5
actually supported at 10 points in a Pacific
stringing machine, but that was still not
enough to prevent the frame contracting slightly. By the time the job was com-
pleted, and on removal of the racquet from the stringing machine, the load cell
reading had dropped to 76 Ib, or an average of 38 lb in each cross-string. The
tension in the main strings was not measured due to restricted access near the
12 o’clock position. So here was a baseline actual tension against which to
compare both theoretical calculations and tension testing equipment.

These results seem to be at odds with the folklore that the tension immediate-
ly after stringing a racquet might be a few pounds lower than the pull tension,
but not much more than that. So I compared Stringmeter and ERT700 meas-
urements with my calculations using frequencies determined with a micro-
phone. The results of the various tests are given in Tables 34.1 and 34.2. Some
of the tests were done immediately after stringing and some were done a few
318 Chapter 34

Table 34.1. Volkl racquets immediately after stringing with 1.42 mm nylon.

Racquet Pull T (Ib) TP (Ib) f(Hz) TVib (Ib)

Volkl 1 39.7 46.3 504 26.7
Volkl 2 46.3 50.7 549 S4i7.
Volkl 3 50.7 52.9 565 33.5
Volkl 4 55.6 55.1 599 37.7
Volkl 5 62.6 58.4 645 43.6
Table 34.2. Topspin racquets immediately after stringing with 1.27 mm nylon.

Racquet Pull T (Ib) ERT (Ib) f(Hz) TVib (Ib)

Topspin 1 46.3 50.7 632 32.8
Topspin 2 52.9 52.9 653 35.0
Topspin 3 57.3 57.3 695. 39.7
Topspin 4 63.9 63.9 724 43.2

months later on the same racquets, which were put aside and not used during
that time. Pull T is the pull tension, TP is the tension indicated in a Pacific rac-
quet diagnostic machine, f (Hz) is the vibration frequency of the strings meas-
ured with a microphone, TVib is the tension calculated from the vibration fre-
quency of the strings, ERT is the tension indicated by the ERT700.

The results in Tables 34.1 and 34.2 show that the Pacific diagnostic machine
and the ERT700 give tensions that are similar to the pull tensions, but the ten-
sions calculated from the vibration frequencies of the strings are much lower
and are more consistent with the measurements made using the arrangement
shown in Figure 34.5. The vibration frequencies for the Topspin racquets were
generally higher than those for the Volkl racquets, even though both models
had exactly the same head size at 630 cm” (98 in’). The difference in frequen-
cy is due to the smaller diameter and lighter string used in the Topspin rac-
quets.

The Volkl racquets were strung on 30th Nov 2000. The

Topspin racquets were strung on 8th Feb 2001. The rac-
Topgpin mains quets were put aside after testing the players, and on
VoRrare 11th March 2001, the tension in every string was meas-
ured using a Stringmeter. I also re-measured the vibra-
<2 2 tion frequencies of the strings. The results for racquets
(1b)
Tension
~s-e : \ Volkl 4 and Topspin 4 are shown in Figure 34.6.
Volkl crosses

PS an eT In both cases, the tension in the cross-strings was much

—— Topspin 4 strung at 63.9 Ib less than the tension in the mains, a result that can be
explained by the fact that the tensions in the cross-
ae strings drop and the tensions in the mains increase due
to contraction of the frame when the crosses are
Figure 34.6 installed. The Topspin 4 frequency had dropped to 655
Comparing Stringmeter results with calculations. Hz, giving an average stringbed tension of 35.3 Ib. The
 Measuring Tension 319

Volkl 4 frequency had dropped to 558 Hz, giving an average stringbed tension
of 32.6 lb. These tensions correspond roughly to the average tension in the
cross-strings rather than the average tension in the mains if one can believe the
Stringmeter readings shown in Figure 34.6. Theoretically, one would expect
that the vibration frequency of the strings should be determined by the aver-
age tension of all the strings, not just the cross-strings. Consequently, I decid-
ed that either the Stringmeter readings were too high or the tensions deduced
from the string vibration frequencies were too low.

The ERT700 also uses a measurement of the vibration frequency of the strings
to determine the average string tension. On 11 March, the ERT700 readings
were 50.7 lb for the Topspin 4 racquet and 48.5 lb for the Volkl 4 racquet.
These tensions are close to the average Stringmeter tension in the main strings
rather than the cross-strings, and appear to be too high since the results in
Table 34.2 indicate that the ERT700 is calibrated to read the pull tension when
a racquet is freshly strung. All three methods of measuring the string tension
indicated that the tension dropped in all racquets during the period after they
were strung, but the actual tensions were proving hard to pin down.

5 ri. ve (gee

When a Stringmeter is used to measure the tension in a string, two prongs are
positioned on either side of the string then the prongs are rotated to twist the
string sideways (Figure 34.3).

Most of the length of the string is locked into position by the other strings, but
a small section of the string about 1-inch long gets stretched and twisted, espe-
cially in the section about 1/2-inch long surrounding the two prongs. If a sin-
gle long string is tested this way, the Stringmeter reading is lower than the actu-
al tension since a smaller force is needed to twist a long string than a short
string. That is why a separate scale is provided on the Stringmeter to test sin-
gle strings. Furthermore, a much larger force is needed to twist a stiff string
than a soft string. If one tests a steel wire under tension with a Stringmeter, the
reading on the Stringmeter is much higher than the actual tension.

Apparatus was set up to test the

Stringmeter as shown in Figure 34.7.

A single string about 14-inches long was

tensioned to either 40 or 60 lb, as indicat- <d—~>
ed on an accurate load cell. Two short
parallel bars, each perpendicular to the
string, were positioned a distance d apart
so that the string passed through a small
hole in each bar. A Stringmeter was then
used to measure the tension in the string
between the bars as the distance d was
varied from 150 mm down to 15 mm. The Stringmeter
sideways movement of the string was
thereby restricted to only a short section Figure 34.7
of the string of length d.
Apparatus to test Stringmeter measurements
320 Chapter 34

75 As shown in Figure 34.8, the Stringmeter reading

increased as d decreased. At large d the reading
70 Stringmeter
for a nylon string was smaller than the actual ten-
Readings sion, but for a kevlar string, the reading was larg-
2 65
oD) er than the actual tension. At a distance of about
=
> 60 Nylon 60 Ib 40 to 50 mm, corresponding to the actual distance
oO
o over which the string is twisted in a racquet, the
f 655
i) Stringmeter reading was almost spot-on for the
® 50 Kevlar 40 Ib nylon string, but was way too high for a kevlar
Ee
string.
£ 45
Le@))
Polyester 40 Ib
ve 40

35
20 40 60 80 100
My conclusion is that the old-fashioned
Figure 34.8 d (mm)
Stringmeter is still the best instrument available to
measure actual string tension, provided that it is
only used on relatively soft strings such as nylon and natural gut, and provid-
ed that that is what is useful to you. But at the same time, it is more difficult
to determine what has happened to the entire stringbed.

The other methods also indicate whether the tension in a racquet has dropped,
and roughly by how much, but the actual values are not correct. The ERT700
and racquet diagnostic devices read too high.

All the devices are good at telling relative differences in a racquet over time or
between two racquets. As long as you understand what each device is telling
you, and how each differs from the other, then you will be able to make intel-
ligent decisions when it comes to stringing tensions.

References

R. Cross and R. Bower, Measurements of String Tension in a Tennis Racket,

Sports Engineering, 165-175 (2001).
 Are Thin Strings
Softer or Stiffer?

Nicolas Kief

By Rod Cross

Summary |

ny given tennis string is normally available in several different diameters. The thinner versions might be
expected to be more elastic and to lose tension faster over time. Of 18 different string pairs tested, 8 of
® the thin versions were dynamically stiffer and 8 maintained tension better than the corresponding thick
versions. These effects were investigated by a controlled experiment where a thin string was compared with a
thick version of the same string made by placing two pieces of the thin string next to each other.

Introduction
Tennis strings are available in a range of diameters from about 1.20 mm (18 gauge) to about 1.40 mm (15 gauge).
Manufacturers have devised their own gauge system where 17 gauge is about 1.25 mm, 16 gauge is about 1.30
mm and 15L gauge is about 1.35 mm. The nominal diameters vary slightly from one manufacturer to the next.
It is commonly assumed that thinner strings are softer or more elastic, but this is not necessarily the case.

A sample of 18 different pairs of 16 and 17 gauge strings were tested. As demonstrated in Table 35.1, eight of
the thinner strings were found to be dynamically stiffer than the thicker versions.
322 Chapter 35

The parameter of most relevance in a tennis string is the increase in elongation

for a given increase in tension, starting at the tension at which a racquet is nor-
mally strung. This parameter can be termed the dynamic stiffness, and it is this
Table 35.1

Dynamic Stiffness, Elongation, and Tension Loss for 16-17 String Gauge Pairs
Numbers shown in bold are measurements that are contrary to conventional wisdom.
Dynamic
Brand String Name Diameter Stiffness (k) Stretch (mm) Tension Loss*
mm Ib/in at 62 Ib Ibs

Babolat Super Fine Play 16 1.32 208 35.0 11.3

Babolat Super Fine Play 17 1.26 202 34.5 10.5
Babolat VS Power 16 1.29 108 26.0 US
Babolat VS Power 17 1.24 100 — 29.2 7.8
Gamma Live Wire 16 1.31 77 SCHS) ae
Gamma Live Wire 17 27 180 34.0 WS
Gamma Live Wire XP 16 leoz 161 S210) 12.8
Gamma Live Wire XP 17 121 183 29.8 13.0
Gamma Synthetic Gut 16 1.30 214 Shas) 320
Gamma Synthetic Gut 17 1.24 211 31.0 9.1
Gamma TNT 16 ESS 188 33.0 8.9
Gamma TNT 47 12s 207 34.3 10.2
Gamma TNT 18 1.16 197 43.0 WS
Gamma TNT Pro Plus 16 lei 187 38.5 10.8
Gamma NNiseroPiusk Ze 123 163 40.5 10.9
Gosen OG-Sheep Micro 16 1.33 209 35.0 Ae O
Gosen OG-Sheep Micro 17 1.30 207 31.0 9.3
Prince DNA Helix 16 1.30 192 SHI) 9.4
Prince DNA Helix 17 W225 192 32.8 9.9
Prince Perfection 16 1.29 206 PIS 11.4
Prince Perfection 17 25 204 Pip ts) 11.3
Prince Sweet Perfection 16 1-295 183 34.0 12.2
Prince Sweet Perfection 17 129 178 37.0 11.7
Prince Synthetic Gut 16 w/Duraflex 1.29 202 SRLS 8.9
Prince Synthetic Gut 17 w/Dura flex 1.25 198 32.0 10.1
Prince Synthetic Gut Original 16 1.32 194 41.5 10.8
Prince Synthetic Gut Original 17 1.24 212 30.0 9.1
Prince Synthetic Gut Soft 16 1.29 213 30.0 8.7
Prince Synthetic Gut Soft 17 128 207 32.8 9.3
Tecnifibre NRG2 SPL 16 1.34 170 31.5 8.4
Tecnifibre NRG2 SPL 17 1.23 184 30.0 8.3
Wilson Extreme Control 16 1.29 210 29.0 8.8
Wilson Extreme Control 17 1-28 196 29.5 9.2
Wilson Sensation 16 1.32 183 29.0 9.0
Wilson Sensation 17 4.25 213 29.8 10.1
Wilson Sensation Ice 16 11:82 194 34.0 11.4
Wilson Sensation ice 17 1.25 204 33.0 11.3
Wilson Sensation NXT 16 1.32 208 29.0 10.5
Wilson Sensation NXT 17 1.24 202 29.8 10.4
Wilson Syn Gut Extreme Control 16 1.30 198 29:5 9.3
Wilson Synth Gut Extreme Control 17 ees 218 29.4 8.3
“Tension loss is total loss in string 1000 seconds after being pulled to 62 Ibs and then impacted 10 times
with a force
equivalent to 120 mph serve.
 String Gauges 323

parameter that was observed to be larger for 8 of the 18 thin strings (a string
with a higher dynamic stiffness means that the string elongates less for a given
tension rise than does another). Neither the elongation nor the dynamic stiff-
ness, at tensions lower than that at which the racquet is strung, are of any con-
sequence regarding the performance of tennis strings. So don’t go by the
stretchiness of a string during stringing. This occurs in a tension range below
the strung tension and is irrelevant. When you hit the ball, the tension rises
rapidly above the strung tension, and the string stretches by an amount that
depends on the dynamic stiffness. Typically, a nylon string that is soft when
you string it is about twice as stiff when you hit the ball. The reason that it is
twice as stiff is only partly because it is stretched faster. The main reason is that
the string does not stretch as easily at tensions above the normal strung ten-
sion, even if you stretch it slowly.

The stiffness of the string plane in a racquet refers to the deflection in a direc-
tion perpendicular to the string plane when a force is applied perpendicular to
the string plane. For small deflections, this is independent of the dynamic stiff-
ness of the strings and would be the same for any type of string provided the
string tension, string length and number of strings remained the same. For the
larger deflections encountered in practice, the string plane stiffness does
depend on the dynamic stiffness since the string tension increases significant-
ly during an impact with a ball, typically by a factor of about two for a fast
serve. The increase in tension is smaller for soft strings such as natural gut, and
larger for stiff strings such as kevlar.

A graph of tension, T, vs elongation, x, for small and large diameter strings is

shown in Figure 35.la. For any given string, a graph of T vs x will be a smooth
curve, where the slope (i.e., steepness of the curve) at low T is typically about
a factor of two smaller than at high T, at least for nylon or polyester strings.
Strings made from natural gut are more linear.

For purposes of illustration, the curve for a nylon or polyester string can be
approximated by two linear segments, as in Figure 35.1b, showing the simpli-
fied graph of a thin and thick string (the normal shape of the curves is illus-
trated in Figure 35.1a). If one assumes that the material properties of the thin
and thick strings are the same, and if the strings are the same length, then for
any given elongation x, T will be proportional to the cross-sectional area of the
string. At any given tension, the extension of a thin string will therefore be larg-
er than that of a thick string. (Note: a tension vs elongation curve is different
than a stress vs strain curve, though they look the same. The former plots how
tension changes with a given elongation in any given piece of string — thick
or thin. The stress vs strain curve looks similar but it takes the size of the string
into account — factoring in the cross-sectional area and length of the string. As
a result, the stress vs strain curve will be the same for any string length or diam-
eter of a given material. It measures the property of the material of a stan-
dardized dimension. The tension vs elongation curve measures the behavior of
string of a particular dimension. Therefore, the same level of stress (tension
divided by cross-sectional area, i.e., pounds per square inch) will occur at dif-
ferent tensions for different gauges. Both gauges behave the same at the same
stress; it’s just that the thinner string achieves that stress at a much lower ten-
sion. The two terms are incorrectly used interchangably in common discourse.)

In practice, racquets are strung at a tension of about 55 pounds, regardless of

the string diameter. As a result, thick strings can operate in a region where x
324 Chapter 35

Figure 35.1

Tension vs Elongation

100 ue

75
2 2
6 5 Operating Tension
@ 50 3 50
ed 2

0 0
0 10 20 30 0 10 18 20 24 25 28 30
Elongation (mm) Elongation (mm)

Figure 35.1a Figure 35.1b

A typical tension vs elongation curve. The A simplified version of the tension vs elonga-
string gets stiffer the more it is stretched, indi- tion curves for two nylon strings of different
cated by the steeper slope of the line. diameter.

As shown in Figure 35.1b, when the thick string is strung at 50 Ib, it stretches by 18 mm (point A). When
the thin string is strung at 50 lb, it stretches by 25 mm (point B). Up to that point, the thin string is softer
than the thick string, since it stretches futher, which is what everyone would expect.

Now, suppose a ball hits the string and the tension suddenly rises from 50 Ib to 70 Ib and then back to 50
lb, all during the 5 ms impact with the ball. If it hits the thick string, the string stretches from point A to
point C, or an extra 6 mm. If the ball hits the thin string, it stretches from point B to point D, or an extra 3
mm. This time, the thick string stretches more than the thin string, so the thin string is stiffer.

In this example, the dynamic stiffness of the thin string is twice as large as the thick string since it stretch-
es by an extra 3 mm, while the thick string stretches an extra 6 mm (for the same increase in tension from
50 to 70 Ib).

=
and the dynamic stiffness k = dT/dx (i.e., change in tension divided by the
change in length) are both small, while thin strings can operate in a region
where x and dT/dx are both large. In other words, the dynamic stiffness of a
thin string may be larger than that of a thicker string of the same material,
depending on the shape of the stress vs strain curve and the operating tension.
In practice, the string tension can rise by a large factor during an impact, in
which case an average dynamic stiffness can be defined as DT/Dx, where DT
is the increase in tension and Dx is the increase in elongation.
 String Gauges 325

Most strings are constructed as composite materials with an inner core and an
outer protective coating to enhance durability. When the string diameter is var-
ied, the relative dimensions of the core and the coating may not remain fixed.
Nevertheless, a string that is marketed under the same name would be expect-
ed to have similar proportions of core to coating thickness, and the thinner
gauge would normally be expected to be more elastic and stretch further at a
given static tension. Indeed, this was the case for most of the 18 string pairs
tested. The exceptions may be due to a change in coating thickness or core
composition, despite each string pair having the same name. It might even be
possible to manufacture a 17 gauge string simply by stretching a 16 gauge ver-
sion.

Since the precise composition and manufacturing techniques used to construct

different gauge strings is unknown, a controlled experiment was devised to
investigate the effects of varying the string diameter. For this purpose, the prop-
erties of a single 18 gauge nylon string (string A) were compared with a string
(B) of twice the cross-sectional area mounted in the test device as two lengths
of string A in parallel. Strings A and B were both tensioned, in separate exper-
iments, to 62 pounds and then clamped at a fixed length (320 mm). The ten-
sion in each string decreased with time due to stress relaxation. The tension in
string A dropped by 9.2 lb to 52.8 lb after 1000 seconds. The tension in string
B dropped by 11.9 Ib to 50.1 Ib after 1000 seconds. This result is consistent
with the common observation that thin strings often hold tension better than
thick strings, despite the fact that the stress in a thin string is larger. It is also
consistent with the observation that when string A was tensioned to 31 lb, the
tension dropped by 6.1 Ib to 24.9 Ib after 1000 seconds. When string B was
tensioned to 62 lb, each “half” of the string was tensioned to 31 lb. If the ten-
sion in each half drops by 6.1 lb after 1000 seconds, then the total tension in
string B would drop by 12.2 lb, essentially as observed.

The dynamic stiffness of strings A and B was measured immediately after the
1000 second delay, by impacting ten times with a hammer as described in
Chapter 30. In theory, one should be able to determine the dynamic stiffness
of a string simply by measuring its elongation curve under quasi-static condi-
tions using an Instron or similar materials testing device (these devices are like
stringing machines — they pull relatively slowly and at a constant rate of elon-
gation. That’s why the rate of pulling is always specified in testing. These rates
are typically 10 mm to 500 mm per minute.). In practice, this is not as accurate
or as relevant as an impact method since the elongation curve depends on the
rate of stretch. For example, if a string is stretched by an additional 1 mm and
then clamped, the tension may increase by say 2 lb while the string is being
stretched, but it immediately starts to decrease as soon as the string is clamped,
typically by about 0.4 lb in the first two seconds. This problem can be avoid-
ed if the string is stretched sufficiently rapidly (a hammer impact stretches the
string at a rate of about 1000 inches per minute).

For the impact tests, the impact duration was about 29 milliseconds for each
string, and the impact energy was 1.63 Joules per impact (equivalent to the
energy on one string in a 120 mph serve). The increase in tension during each
impact was 37.8 lb on average for string A and 43.1 lb for string B. Calculations
revealed that string A had a dynamic stiffness k = 196.6 lb/in. averaged over
10 impacts, at an average tension of 49.7 lb. For string B, k = 235.5 Ib/in aver-
326 Chapter 35

aged over 10 impacts, at an average tension of 48.2 lb. An average tension is quoted since the
tension decreased slightly following each impact. In this case, the thinner string was softer, but
not by the factor of two that one would expect if the stress vs strain curve were linear.

To understand this, consider a simple spring. Suppose it stretches one inch when you pull it
with a force of 25 lb. Suppose also that it stretches by 2 inches when you pull it with a force of
50 Ib. Such a spring is said to be linear, since when you double the force, you double the
amount it stretches. Now suppose you put two such springs in parallel, and pull with a force of
50 lb. Then there is a force of 25 Ib acting on each spring, so they each stretch by 1 inch. The
two springs in parallel are twice as stiff as only spring since they stretch only one inch at 50 Ib,
whereas one spring alone stretches by 2 inches at 50 lb. Similarly, two identical tennis strings
in parallel would be twice as stiff as one string alone, if the strings behaved as linear springs.
In fact, it turned out that strings A and B (i.e., two A strings in parallel) had about the same stiff
ness, which means that the strings were not linear. For string A, one can infer that the dynam-
ic stiffness at a tension of about 48.4 lb is a factor of 1.7 larger than the dynamic stiffness at a
tension of 24.2 lb. If it were a factor of two higher, then the dynamic stiffness of strings A and
B would have been the same. If the stiffness at 48.4 lb was the same as the stiffness at 24.2 Ib
(i.e., the strings were linear) then string B would be twice as stiff as string A.

A surprising result of testing thin and thick versions of commercially available tennis strings is
that thin strings can be dynamically stiffer and they can maintain tension better than the corre-
sponding thick versions. In this chapter, it has been shown how this can arise. Two linear strings
connected in parallel will be twice as stiff as one. Tennis strings are nonlinear and have the
property that the stiffness increases with tension. Consequently, the stiffness of two parallel
strings can be larger, smaller, or the same as that of a single string depending on the stress vs
strain curve and the operating tension. Similarly, the rate at which a string loses tension is a non-
linear function of tension. The rate of tension loss increases with tension, but the loss rate may
not increase in proportion to the tension. As a result, two strings in parallel can lose tension at
a faster rate than a single string.
 ‘
‘
/
|
:

ct
 Rules and Types

By Howard Brody

hy should you care about the tennis ball? Both you and your opponent are going to be using the
same ball. How can it possibly give you an advantage over your opponent? How can a different ball
change your enjoyment of the game?

Let’s start off by assuming that the balls you are using meet all the International Tennis Federation (ITF) and
USTA specifications as to rebound, compression, size, weight, etc as spelled out in Rule 3 and Appendix 1 of
the Rules of Tennis. Then what choices do you have? Each manufacturer has several lines of balls of different
quality, with names like “Championship,” “ATP,” “Grand Prix,” etc. Within each brand and line there may be sev-
eral types of balls, such as “Heavy Duty,” “Regular Duty,” “All Court,” “Hard Court,” etc. The extra duty or heavy-
duty balls are designed to last longer on abrasive courts, while the regular duty balls are usually designed for
softer courts.

For recreational use you may use any type of ball on any type of court and be within the rules of tennis if the
balls are ITF approved (it will say ITF or USTA Approved on the can or box).
330 Chapter 36

d and Pressureless
Balls _
Most balls are fabricated with an internal air pressure that gives them their
bounce and feel when you hit them. Unfortunately, the rubber walls of most
balls are slightly permeable to air, so balls are sold in pressure cans, to keep
the internal air from leaking away and the ball going dead while the can might
is sitting for months on the shelf in the store. Once the can is opened and the
balls are exposed to normal atmospheric pressure, the air begins to seep out
through the walls. The balls then begin to lose their bounce and after a peri-
od of time become soft and “dead.”

There are also some balls that have an internal air pressure that is the same as
the ambient air pressure (they are called pressureless or non-pressurized).
Therefore, the air cannot leak out and cause the ball to go dead. This latter cat-
egory gets its bounce, stiffness and resiliency predominantly from its rubber
shell. Pressureless balls are popular in Europe where. balls, because of their
high cost, are expected to last for many weeks.

Note that all balls that meet ITF and USTA standards are comparably hard when
you slowly compress them. At present, there are no rules on the dynamic stiff-
nesss or hardness of a ball. (in a dynamic situation, the ball is compressed
quickly, as if it were being hit.) You can easily tell which balls are pressure-
less, since they are usually sold in a cardboard box and not in a pressure can.
Tretorn has a hybrid ball that has material other than air inside of the ball to
give it some of its dynamic properties.

In addition, there are high altitude balls. There are two separate and distinct
reasons why pressurized tennis balls play differently at high altitude, compared
to the way they play at or near sea level.

1.) The coefficient of restitution (COR, or how lively the ball is when hit or
when it bounces) of a pressurized tennis ball depends on the internal pressure
relative to the outside (atmospheric or ambient) pressure. At high altitude, the
outside, atmospheric pressure is reduced. This results in an increase in the
pressure difference between the internal pressure of a normal tennis ball and
atmospheric pressure. The larger this difference in pressure, the greater will be
the COR (the ball will be livelier). This increase in COR will affect both the
bounce of the ball on the court Git will bounce higher) and the speed of the
ball as it comes off of the strings (it will give you more power).

2.) The flight of the ball through the air (its trajectory) is influenced by the force
of air resistance. The more air molecules the ball runs into, the more the ball
will slow down. At high altitude, the ball will run into less air molecules (the
air is thinner), so it will not slow down as much (the ball will “fly” a bit more).

Reducing the internal pressure in a ball (so that it does not bounce as high
when dropped at sea level) will produce a ball that, at high altitude, has a COR
that is similar to the COR of a regular ball at sea level. This reduction in bounce
height is presently part of Rule 3 of the Rules of Tennis.
 The Tennis Ball 331

Using a larger ball at high altitude will result in an air resistance force
closer to the force encountered by a standard ball at sea level and
should lead to ball trajectories that are closer to the ones observed at sea
Air Resistance
level.
And The Big Ball
It then makes sense that the Rules of Tennis (Rule 3 specifically) be he force of air resistance is deter-
modified to recommend that a larger pressurized ball (Ball Type 3) with | mined by the number of air mole-
reduced rebound height be used when tennis is played at high altitude. cules that the ball runs into,
If such a ball is not available, then a larger ball (Type 3) with normal which is proportional to the density of
internal pressure can be used. If non-pressurized balls are to be used, it air (the number of air molecules/cubic
is recommended that they be Type 3 balls, if available. centimeter). This density of molecules is
linearly dependent on the air pressure,
The official ITF Rules of Tennis specify many of the ball’s parameters, and therefore both the density and pres-
sure are the same function of altitude.
including the diameter, rebound height (when dropped from 100 inch-
The pressure and the density are pro-
es) and the static compressibility. Some players who play on slow, soft
portional to the sea level value multi-
courts (such as red clay) prefer harder tennis balls (less compressible) plied by a function of the altitude. This
because they tend to play faster (not lose as much speed on the function is:
bounce). Starting in January of 2002, the ITF is allowing a ball that is
harder than the standard ball to be used in tennis. These balls are called e to the negative mg/kT multiplied by
Type 1 balls. the altitude h, where e is the base of the
natural logarithm and equal to 2.8.

eon (36.1)
Many players and spectators feel that the game of tennis has gotten too When the values of m (the mass of a
fast and the serve is becoming too dominant, particularly on fast courts, molecule of air), g (9.8), k (Boltzmann’s
such as grass. The proposed solution to this problem is to allow the use constant) and T (degrees Kelvin) are
of a larger tennis ball (6 percent greater diameter) on fast courts. The put in, the result is e to the negative
larger ball will hit more air molecules in its flight, increasing its air resist- height, h, above sea level (in kilome-
ance, and slowing the larger ball down more than a standard size ball. ters) divided by 9.
These larger balls can be used on any court in recreational tennis and
are intended for men’s professional tournaments on fast surface courts e("9) (36.2)
such as grass. These lager balls are called Type 3 balls.
This means that at a height of one (1.0)
kilometer = 3281 feet, the density (pres-
There is an interesting sidelight to this larger ball idea. When you watch
sure) will be e ’° = 0.895 times the den-
some of the top players (such as Pete Sampras) prepare to serve, they sity (pressure) of the sea level value. At
ask for several balls, then reject all but one. On what basis are they a height of 2.0 km (6562 feet) the den-
selecting that ball? They know that a ball whose cover is fluffed up acts sity will be e *° = 0.80 times the densi-
as if it had a greater cross section. It will have more air resistance and ty of air molecules at sea level.
will slow down more in flight, giving their opponent more time to get
to and return their serve. A ball with a felt cover that is somewhat mat- A ball that has a diameter 6.5% greater
ted down will get to (or past) their opponent faster, which they feel than a standard ball (13.4% greater
gives them a competitive advantage. In addition, all tennis balls are not cross sectional area) will have the same
air resistance force on it at a height of
exactly the same size to begin with. The Rules of Tennis allow a little
1.13 km (3700 feet) as a standard ball
over a tenth of an inch variation or latitude in the diameter of an
at sea level.
approved tennis ball. So the players select the ball, which to them,
appears to be smallest, and will play the fastest on their serve. a’ 8) _.4/1.134

When tennis balls are placed in a wind tunnel and their air resistance A 6.5% larger ball will then have the
measured, the results depend very much on the fuzziness and condition same trajectory at 3700 feet as a stan-
of the felt cover on the ball. (in fact, before one test, the ball was shaved dard ball would have at sea level.
with an electric razor.) As the ball fluffs up, then begins to wear down,
clear changes in the ball’s performance can be noted. That is why, in (All of the above assumes that the tem-
perature is the constant with altitude)
major tennis matches, the balls are changed every nine games. However,
332 Chapter 36

there are no rules in the Rules of Tennis concerning the aerodynamic proper-
ties of the felt cover on a ball, other than it be a fabric.

When the bounce of these different type of balls is examined in actual play
using high speed photography, it is discovered that the harder balls (Type 1)
tend to bounce a bit lower than the standard (Type 2) balls and the larger balls
(Type 3) tend to bounce a bit higher than the standard (Type 2) balls.
 Testing Ball Properties
for Perfect Bounce

Mats Wilander, Sweden

By Rod Cross

he rules of tennis are very specific about the properties of a tennis ball. In addition to having the correct
weight and diameter and color (white or yellow), a ball must bounce to the correct height when dropped
onto a concrete slab from a height of 100 inches. Furthermore, an approved ball must squash by 1/4 of
an inch + 0.04 inch (6.5 mm + 1 mm) when subject to a load of 18 lb (8.2 kg). If the ball squashes more than
the specified amount, it will be too soft and it won't bounce very well. If it squashes less than the approved
amount it will be too stiff and it will bounce too high. Given all these restrictions, one might expect that a ball
will behave in a very predictable manner. In fact, a tennis ball is full of interesting surprises as described below.

The modern way to test the stiffness of a ball is to place it between two parallel metal disks and to push the
disks together using a materials testing machine. These days, such machines are computer controlled and cost
about $10,000 or more. The machine compresses the ball at a controlled rate and gives a plot of force vs com-
pression. A typical result is shown in Figure 37.1. The plot is not a straight line since a tennis ball gets stiffer the
more it compresses. Furthermore, when the force applied to the ball is reduced to zero, the decompression curve
lies below the compression curve. The machine used to compress the ball was instructed to pause at 10 kg for
5 seconds on the way up (10 kg is the specified 8.2 kg load plus a specified preload of 1.8 kg) and to pause at
10 kg for 10 seconds on the way down. This is written into the rules to give the person using the machine suf-
334 Chapter 37

ficient time to take a reading. The rule is no longer strictly necessary since a
computer can take readings 100 times a second if it is instructed to do so. In
the old days, the reading was taken with a manually operated Stevens machine
(and is still done so in some countries), but the rules haven't yet caught up
with modern technology.

Flour eer! The curve shown in Figure 37.1 is a typical hystere-

sis Curve, meaning that the curve for an increasing
i force is different from the curve for a decreasing
force. At a load of 10 kg, the compression is 8 mm
a on the way up and 10.5 mm on the way down
= (after the specified waiting times). This ball passed
Zz oe the test since it compressed by the approved
S amount (6.5 mm + 1 mm) when the load increased
ee 20 by 8.2 kg (from 1.8 kg to 10 kg). It also passed the
specified test on the way down.
10 :
The area between the two curves represents the
energy lost in the ball. The area between the top
curve and the horizontal axis represents the work
required to compress the ball, and it therefore rep-
resents the energy stored in the ball at maximum
compression. The area between the bottom curve
and the horizontal axis represents the energy given back by the ball. The dif-
ference is the energy lost. The area under the top curve in Figure 37.1 is 5.03
Joules. The area under the bottom curve is 3.87 Joules. The energy lost is there-
fore 5.03 - 3.87 =1.16 Joules, which represents 23% of the stored energy (5.03
Jp

There is an interesting problem with this result, since it is not consistent with
the 100 inch bounce test. When a tennis ball is dropped from a height of 100
inches it must bounce to a height between 53 and 58 inches. The height of a
ball above the ground is a measure of its potential energy. Consequently, if the
ball bounces to 55% of its drop height it loses 45% of its energy during the
bounce. Therein lies the problem. The compression test indicates that a ball
should lose only 23% of its energy when it squashes. If it lost only 23% of its
energy during a bounce it would bounce a lot higher than 58 inches.

One difference between the drop test and the squash test is the time taken to
squash the ball. When a ball is dropped on a hard surface, it spends only about
0.0045 seconds in contact with the surface. The ball is squashed over a period
of about 30 seconds when measuring its compression. Another difference is
that the ball contacts only one surface when it is dropped, but it is squashed
between two surfaces in a compresion test. A third difference is that the com-
pression test squashes the ball by just over one inch, but a ball dropped from
100 inches compresses by about 3/4 inch.

The behaviour of a tennis ball during a rapid compression is not the same as
that for a slow compression. The force and compression can be measured dur-
ing an actual ball bounce by dropping the ball on a pressure sensitive plate
(Cross, 1999). Typical results are shown in Figure 37.2 (force vs time) and
Figure 37.3 (force vs compression) for a ball incident vertically downwards at
 Ball Testing for Bounce 335

5.2 m/s or 13.5 m/s onto a hard surface. The sur-

Figure 37.2a
face was a ceramic piezoelectric block which gen-
erates a voltage proportional to the force on the
block. The hysteresis curve for a rapid compression
is much fatter than one for a slow compression,
meaning that the ball loses more of its stored elas-
tic energy when it compresses and decompresses
rapidly.
(N)
Force
The results in Figures 37.2 and 37.3 are typical of
all tennis balls. There are only very slight differ-
ences between different balls. At low ball speeds,
Ball speed 5.2 m/s
the area enclosed by the hysteresis curve is about
35% of the maximum stored energy and the ball
bounces to about 65% of its drop height. As the ball
speed increases, the peak force on the ball increas- Time (ms)
es, the impact duration decreases and the ball loses
a greater fraction of its impact energy.

Figure 37.2b
The main reason for the increased energy loss is
that a tennis ball is much stiffer when it is com-
pressed rapidly and hence more work is needed to
compress it, and more energy is stored in the ball
for any given compression. However, the wall of
the ball buckles after a short time delay and the ball
then becomes softer. When the wall buckles it loses
a lot of its stored elastic energy. The wall also buck-
les during a slow compression but it buckles at a (N)
Force
much earlier stage when there is very little elastic
energy stored in the ball.

During a rapid compression a small region of the

ball in contact with the surface compresses locally,
with the result that the force on the ball increases Time (ms)
by up to half its maximum value before the wall
starts to buckle. It is much harder to compress a
piece of rubber than to bend it, and the more it is

Figure 37.3a Figure 37.3b

200

Ball speed 5.2 m/s Ball speed = 13.5 m/s

150
vag
100
(N)
Force
(N)
Force
50

Omene? 4 6 8 10 5 10 1S 20 25 30 35
Compression (mm) Compression (mm)
336 Chapter 37

compressed the stiffer it gets. As a result, the force on the ball increases expo-
nentially with time during the first 0.2 ms of the bounce. The ball is about 6
times stiffer than normal just before the wall buckles. The initial contact with
the surface sends a transverse bending wave around the circumference of the
ball, but the wave travels relatively slowly around the ball. The wave can be
seen travelling around the ball using a small piezo disk glued to the ball, or by
high speed photography.

As a result of the slow wave speed, the rubber wall in the contact area can
compress by a few mm before the wave travels any significant distance away
from the initial contact point. As shown in Figs. 37.2a and 37.2b, there is a sud-
den change in ball stiffness after about 0.2 ms, which is the time at which the
wall buckles. The ball incident at 13.5 m/s travelled 2.7 mm before the wall
buckled, compressing the rubber and cloth to about half its original thickness
in the contact region. During a slow compression, this time delay is insignifi-
cant and the ball buckles at a much lower force and at a much smaller value
of the ball compression. This effect can be seen by comparing Figs 37.2a and
37.2b. At the lower ball speed, the wall buckles when the force on the ball is
50N. At the higher ball speed, the wall buckles when the force on the ball is
230N. Buckling occurs not at a fixed value of the force but after the bending
wave travels about 1 cm away from the initial contact point.

Another interesting effect is that the bottom of the ball can bounce upwards
while the rest of the ball is still coming down. As a result, the bottom section
of the ball pops up inside the ball like a small bubble. This effect can be
observed by cutting a tennis ball in half with a hacksaw. If one half of the ball
is then pressed down on a hard surface by hand, the contact region pops up
inside the ball. If the ball is pushed hard enough, it turns itself inside-out.

High speed photography shows that the top of the ball also forms a dimple
since the top of the ball continues downwards while the sides are bouncing
up. Even after the ball has bounced up off the surface, the top continues to
bounce up and down for a short while, as the dimple pops out of the ball and
pops back in again at a rate of about 500 cycles per second (two cycles dur-
ing the 4 ms impact period). This generates a short ping which is part of the
sound generated by the ball when it bounces. Some balls vibrate more than
others and sound different when they bounce, but all approved tennis balls
bounce to essentially the same height and take essentially the same time to
bounce off the surface, regardless of whether they are the pressurised or pres-
sureless variety.

The vertical bounce test shows that a tennis ball incident vertically on a hard
surface, at a speed of about 7 m/s, bounces upwards at about 75% of its inci-
dent speed. Under normal playing conditions a tennis ball does not fall verti-
cally downwards onto the court surface. It approaches the surface at an angle
of around 20° to the horizontal and at a speed of around 30 m/s for a fast serve.
Such a ball drops vertically by about 3 m in 0.4 seconds or at a vertical speed
of about 7 m/s for a fast serve. Measurements of the bounce speed for a typi-
cal serve have been made for many different surfaces and they reveal a sur-
prising result. A ball incident at about 20° bounces off the surface at around
20° O to the horizontal
. .
but the vertical speed after the bounce can be as high as
 Ball Testing for Bounce 337

90% of the vertical speed before the bounce. In other words, the ball behaves
more like a superball than a tennis ball in terms of its vertical bounce proper-
ties. The ball bounces much higher than one would expect from the standard
100 inch drop test. At 7 m/s, a ball should bounce to a height of about 55 inch-
es or about waist height. In fact, some balls incident vertically at around 7 m/s
bounce to head height. This is a typical result for a kick serve with lots of
topsin.

There is as yet no generally accepted explanation for the higher than expect-
ed bounce. One possibility is that a ball may sweep some sand ahead of it to
form a small ramp and is then deflected upwards by the ramp. However, high
bounces are also observed on hardcourts without any sand. In that case, ball
spin may supply an explanation, as follows.

A ball bouncing at an oblique angle bounces with topspin even if it has no spin
before it bounces. If the ball is served with topspin then the amount of topspin
increases after the ball bounces. For example, a ball can be served so that it
spins at about 60 revolutions/sec before it hits the court, and it will then spin
at around 100 revolutions/sec after it bounces. The spin increases from 60 to
100 rev/sec during the 4.5 ms time that the ball is in contact with the court sur-
face. If we assume that the average spin during this time is 80 rev/sec then the
ball rotates by 80 x 0.0045 = 0.36 revolutions during the bounce.

The point on the bottom of the ball that first contacts the surface will rotate
during the bounce so that it is nearer the top of the ball when the ball lifts off
the surface. That point will start to squash when the ball first contacts the sur-
face but it will not squash very far because it rotates off the surface before the
wall has had time to collapse. As shown above, the wall starts to collapse after
0.2 ms. As the ball rotates, each new section of the ball coming into contact
with the surface will also take 0.2 ms to collapse. Consequently it might be
possible for the ball to keep rotating and not collapse at all if the ball is spin-
ning fast enough. The ball is very stiff before the wall collapses. If the wall
does not collapse at all, or if it does not have time to fully collapse, then the
ball will compress by only a small amount and it will lose very little energy
during the bounce. A spinning ball might therefore bounce much higher than
a non-spinning ball. Detailed measurements would need to be made to see if
this explanation is correct and whether it applies to all court surfaces or only
some surfaces. An alternative explanation for the high COR is described in
Chapter 39.

Reference

R. Cross, Dynamic properties of tennis balls, Sports Engineering, 2, 23-33

190):
 J» paian a

. “6 ae PS BAD Bi
» eine’
: e oe @ewsenn
“ rr or)
7 sont Laat

ea? erY)
 @=
6
=
ic
cS
Cc

An Introduction

Monica Seles, United States

By Howard Brody

he bounce of a tennis ball on a tennis court is a complicated physical situation, but by making some sim-
plifying assumptions, the problem can be easily analyzed. These assumptions include keeping the ball a
* hollow sphere (no deformation), having the court hard (no court deformation), and having the friction
force (the force on the ball parallel to the surface) obey the simple law of friction, as typically taught in an intro-
ductory physics course.

When a ball interacts with a court surface by bouncing, there are three forces acting on the ball, a normal force
(perpendicular to the surface), a friction force (parallel to the surface), and the force of gravity. When the forces
involved are examined, it turns out that the normal and friction forces are at least an order of magnitude greater
than the force of gravity on the ball, so the gravity force can be neglected during the few milliseconds that the
ball is in contact with the court’s surface. The ball’s motion is then broken up into a normal or vertical velocity
and a parallel or horizontal velocity. The problem is then solved for each motion separately, and then the veloc-
ity components are recombined to get a final result.

When a tennis ball bounces on a court, it does not return to its original height. In fact the Rules of Tennis state
that when a tennis ball is dropped from a height of 100 inches onto a hard (concrete) surface it must bounce
340 Chapter 38

to a height between 53 and 58 inches. This ratio of heights (between 0.53 and
0.58) is the ratio of energy returned to the ball after the bounce to the energy
the ball had just before it hit the ground. There is a term called the coefficient
of restitution (COR), first used by Newton and Wren in the sixteen hundreds,
that can also be used to describe the collision of the ball with the ground. The
COR is the ratio of the ball’s vertical velocity just after hitting the ground to its
vertical velocity just before its impact with the ground. The COR is the square
root of the rebound height ratio, so for a ball to be approved, the COR must
be between .728 and .762 (which is the square root of 0.53 and 0.58). Note,
this 53 to 58 inch return is onto a hard surface and for a vertical drop of 100
inches only. There are other surfaces that may produce a different rebound
height for a “legal” ball, since some surfaces are dead, and absorb energy. If
the ball is dropped from less than 100 inches, the COR will be slightly higher.
In the analysis that follows, we will make the assumption that the COR of a
ball is the same whether it is just dropped or whether it is rebounding at an
angle as part of its trajectory and it has a horizontal velocity as well. In other
words, we are assuming the rebound height ratio and the COR of a ball are
independent of the horizontal velocity of the ball.

When a tennis ball hits the ground and rebounds back upward, the direction
of the ball’s velocity reverses, and a force is required to do this. The actual
change in velocity is the difference between the incident and rebounding
velocities, and it is equal to the incident velocity times (1 + COR). The ball has
a momentum equal to its mass times its velocity, and Newton’s Second Law
says that an impulse (force x time) is required to change the momentum of an
object. The magnitude and direction of the impulse is equal to the change in
momentum of the object. For a bouncing ball, the ground produces a vertical
force (called the NORMAL force) equal to the change in the ball’s vertical
momentum (which is mv(1 + COR)) divided by the time of the ball’s contact
with the ground.

Horizontal Velocity
When a ball bounces, there is a frictional force parallel to the court on the bot-
tom of the ball. The rougher the court surface, the larger is the friction force.
The roughness can be characterized by a parameter called the coefficient of
friction (COF). The magnitude of the friction force is also proportional to the
magnitude of the normal force. The bigger the normal force, the larger is the
frictional force. Combining these two effects gives the actual friction force on
the bottom of the ball, and that friction force is equal to the COF times the nor-
mal force (COF x N). The change in the ball’s horizontal momentum will be
equal to the frictional force multiplied by the contact time of the ball on the
surface (the horizontal impulse). When you put this together with the formula
for the Normal force in the previous section, you get that the change in the
horizontal velocity of the ball, due to its bounce, is COF(1 + COR) times the
vertical velocity component of the ball before the bounce.

Since there is only one horizontal force on the ball when it is in contact with
the court surface, and that force acts on the bottom of the ball, there is a net
torque on the ball. This torque caused by the ball skidding or sliding on the
surface, produces a topspin on the ball. The result is that most shots, after the
bounce are found to have a fairly large amount of topspin on them. If the ball
strikes the ground at a large angle (the ball has cleared the net by a huge mar-
gin) and the COF is high, it is possible for the ball to begin rolling before it
 Bounce of the Ball 341

leaves the surface (see Note 1). Once the ball begins to roll instead of sliding
or skidding, the frictional force is greatly reduced. When a ball goes into a bit-
ing or rolling mode, it will have lost 40% of its original horizontal velocity com-
ponent. On a typical serve, the ball will lose about 30% of its original velocity
component parallel to the ground.

Think about the slow clay of Roland Garos or the fast grass at Wimbledon.
What is it that makes these courts fast or slow? How does this “speed” change
the way you play your game?

It is quite clear that the only way a court can influence the speed of the game
(the speed of the ball) is for the ball to bounce. There are several very differ-
ent effects that influence the ball when it bounces. The ball may pick up bits
of clay and become heavier. This will slow down the game because a heavier
ball comes off of the racquet with less speed, but the more massive ball will
slow down less due to air resistance. These two effects tend to cancel each
other out. The bounce may fluff up the ball a bit and therefore it will have
more air resistance. This will slow down the speed of the ball in its flight.
However, these are not the effects that are normally considered when the caus-
es of fast or slow courts are discussed.

It is the friction between the ball and the court that determines the speed (or
PACE) of a court. If the friction force is small, then the ball will lose little of its
horizontal velocity component when it bounces, and you will have a fast court.
When the friction force is large, the ball will lose a larger fraction of its hori-
zontal velocity component, and the court will be considered slow. If the COR
(rebound height) is the same for both a fast and slow court, because the hor-
izontal components of the velocity will be different, on a fast court the ball will
rebound with a lower angle than the angle it comes off at on a slow court. It
is also clear that on a fast court, the player will have less time to get to the ball
and prepare for a shot than on a slow court.

Surprisingly, on ground strokes leaving your opponent’s racquet at 60 miles an

hour, the ball will arrive at your baseline only 0.07 seconds earlier on a fast
court compared to a slow court. On a 110 mile an hour serve, the difference
in time between a ball coming to you on a clay court and a grass court is just
a few hundredths of a second! This seems to fly in the face of all that you
thought you knew about fast and slow courts. How can these sort of numbers
agree with our idea that there is a perceptible slowdown in play on clay courts?
Why is it that when you work out the mathematics, the ball may be moving
18% slower when it reaches the baseline on a slow court than on a fast court,
yet it takes only 5% longer to get there?

In going from baseline to baseline, the ball spends most of its transit time in
the air before it bounces. Any slowdown due to the bounce can only affect the
ball during the very short distance that the ball moves between the bounce and
the baseline. Therefore, an appreciable loss of speed at the bounce will not
lead to an appreciable increase in the time it takes the ball to get to the play-
ee

This still does not seem to answer the question of why the entire game seems
to slow down on clay courts? To understand this, we must investigate the ball-
342 Chapter 38

racquet dynamics. For a groundstroke, the speed at which a ball leaves the rac-
quet is determined not only by the racquet head speed but also by the speed
of the incoming ball. When you work out the physics of the interaction, about
25% of the outgoing ball speed is determined by the speed of the ball coming
toward the racquet. When a ball bounces on clay, there is an appreciable loss
of speed, so the return of that shot will be reduced in speed (unless the play-
er makes a very conscious effort to swing the racquet much harder). The result
is that on a clay court, with the players swinging with their normal strokes, the
ball speed BEFORE the bounce is reduced, compared to a fast court. This leads
to a general slowdown of the game. The shots are leaving the racquets at lower
speed and then slowing down more when they bounce.

Notes

1. Balls do not roll when they bounce since they are not perfectly rigid.
Instead, the bottom of the ball bites the surface. However, biting and rolling
are almost the same for a tennis ball. The differences are explained in Chapter
39:
 Friction, Sliding, Rolling,
Biting and Spinning Balls

Todd Woodbridge, Australia

By Rod Cross

Summary

" he main object of the game of tennis is to hit the ball over the net and into the opponent's court in such
a way that the opponent finds it difficult or is unable to return the ball. The ball needs to bounce at the
right speed, angle and spin, not only off the racquet, but also off the court. Players learn how to do this
by trial and error and years of practice and coaching. In this chapter, we examine the main factors that deter-
mine the way a ball bounces off the court from a physics point of view.

Tennis players rarely stop to think about what happens during the brief instant when a ball bounces off the court
or off the strings. They just let it happen, and they know from experience what happens when they hit the ball
a certain way. The ball spends only 0.005 seconds on the court or on the strings while it is bouncing, so there
is nothing the player can do during or after that time to alter the fate of the ball. In that respect, the bounce is
predetermined. During a bounce, the ball changes speed, it changes direction and it changes the rate at which
it spins. All these things change very rapidly. For example, a ball served at 100 mph accelerates from 0 to 100
mph in 0.005 seconds or about two thousand times faster than a sports car. Interesting things happen under such
344 Chapter 39

extreme conditions. The force on a ball served at 120 mph is enough to lift two
people off the ground; it is enough to squash the ball in half; and it is enough
to stretch the string plane backwards by an inch or more. It happens so fast
that it is impossible to see by eye, but it can be captured and measured using
the right instruments.

When a ball bounces off a surface, the ball always reverses its direction of
motion perpendicular to the surface, and it usually slows down in a direction
parallel to the surface. For example, a ball headed downwards at an angle to
the court bounces up and away from the surface. If it is traveling from left to
right before it bounces, it usually keeps traveling left to right after the bounce
at a reduced speed. A ball with lots of backspin can sometimes bounce back-
wards, which is an extreme example of slowing down in a direction parallel to
the surface. A ball with lots of topspin can sometimes shoot forwards at a
greater speed than it had before the bounce, but this is a rare event. It can hap-
pen in a topspin lob but it never happens with a serve or a normal ground-
stroke.

Motion of the ball in a direction perpendicular to a surface is easy to explain.

On the court, the ball just squashes and pushes itself back off the surface as it
expands back to its original shape. It is actually the surface that pushes on the
ball, but the ball pushes with an equal and opposite force on the surface. It is
also easy to explain why the ball slows down in a direction parallel to the sur-
face. It is due to the fact that friction acts backwards on the ball while the ball
is moving forwards in contact with the surface. All objects moving along a sur-
face experience a friction force that acts backwards on the object. It is for this
reason that any object that slides or rolls along a horizontal surface eventually
comes to rest if it is moving under its own steam (without an onboard motor
to keep it going).

The most difficult part of explaining the bounce of a tennis ball involves esti-
mating or calculating the magnitude of the friction force. Most things are easy
to explain in words. But scientific explanations involve more than just words.
In order to determine if the explanation is correct, it has to be tested against
calculations and measurements. The problem here is that there are three dif-
ferent types of friction:

° Sliding friction acts only when an object is sliding.

* Rolling friction is much smaller and acts only when an object is rolling.
e Static friction acts only while an object is at rest on a surface and when some
other force is applied in an attempt to push or pull the object along the sur-
face.

If you try to push an elephant along the ground, the elephant will remain at
rest due to static friction between its feet and the ground. But you can push a
car along a smooth horizontal surface since the friction force on a rolling wheel
is relatively small.

It turns out that when a ball bounces, static friction is sometimes more impor-
tant than the other types of friction, since the bottom of a ball can actually
come to rest while it is bouncing. The bottom of the ball always comes to rest
in the vertical direction when the ball bounces, but it can also come to rest in
the horizontal direction. In that case, we say that the ball grips or bites the sur-
face. Even though the bottom of the ball comes to a complete stop, the whole
 Ball Spin and Bounce Off Courts 345

ball doesn't come to a complete stop. If that is hard to visualize, consider what
happens to your foot when you are walking. You move one leg forward, plant
it on the ground and move the rest of your body forward. Your front foot
comes to a complete stop until you lift first your heel and then your toes off
the ground. The horizontal force that allows you to walk forwards arises from
static friction on your foot. Sliding friction arises when you step on a banana
skin.

A ball incident at a low angle on a surface usually slides throughout the

bounce. An example is a fast serve without much topspin. At higher angles or
when the ball is incident with a lot of topspin, it usually slides at the start and
then bites before it bounces off the surface. To complicate matters, one part of
the ball can bite the surface while other parts are stretching or compressing
and sliding along the surface. If tennis balls bounced in a really simple man-
ner, this chapter would be very short. But balls can do surprising things when
they bounce, which gives them some personality and it also makes them hard-
er to explain. An example is a basketball. When a basketball bounces, the fric-
tion force at the bottom of the ball changes direction six times during the
bounce. When a tennis ball bounces, the friction force sometimes reverses
direction once, sometimes twice, and sometimes not at all, depending on if and
when it bites the surface.

If a ball bounces off the court or off the strings of a racquet, it almost always
changes its rate of spin. The only case where it doesn't change its spin is when
it bounces at right angles to the surface. For example, if a ball is dropped with-
out spin onto the court or onto the strings of a racquet, then the ball will
bounce without spin. If the ball is thrown or hit onto the court without spin,
but at some angle to the surface, then it bounces with topspin. The faster the
ball is incident on the court, the higher it bounces and the faster it spins. An
interesting question is how one can make the ball spin as fast as possible. One
way is to throw or hit the ball as fast as possible, but what is the best angle to
throw it, and does the angle depend on whether the surface is fast (like grass)
or slow (like clay)? And what angle is needed to get the most spin off the
strings? Topspin is imparted when the racquet head rises upwards to contact
the ball. An insect on the strings sees the ball coming in at an angle and going
out with a different spin than it had coming in. The question is, can one get
more spin when the head rises at say 30 degrees to the horizontal, or is 45
degrees better and does it make much difference if the head is tilted forwards?
These questions are answered in Chapter 41. In this chapter, we concentrate
mainly on the bounce off a court.

A ball incident at some angle on the court will usually bounce with reduced
speed. The speed reduction is typically between 30% and 60%. For example, a
ball served at 100 mph will slow down through the air and hit the court at
about 75 mph. It then bounces off the court at about 50 mph. If it had no spin
when it hit the court, then it will bounce with topspin, spinning at about 6,000
rpm or 100 revolutions per second. The speed reduction and the extra spin
both arise from the same force acting on the ball. The force is due to friction
between the ball and the court and it acts on the bottom of the ball. The fric-
tion force acts in a direction parallel to the surface, and it acts backwards on
346 Chapter 39

the ball if the ball is moving forwards, slowing it down. The friction force also
exerts a torque on the ball, causing it to rotate or to spin. The bigger the back-
wards force, the more the ball slows down and the bigger the spin.
Consequently, the way to get the ball to spin faster is to make sure the friction
force is as large as possible. The same is true for spin off the strings. However,
the price to pay for extra spin is that the ball does not come off the strings as
fast as a ball hit without spin.

The friction force, F, depends on three factors:

e The force, N, exerted by the ball pushing downwards onto the court. That
depends on how fast the ball hits the court and it also depends on the angle
of incidence. If the ball slams vertically into the court, then N will be large. If
the ball is incident at the same speed but at a small angle to the court, then N
will be smaller, since the vertical speed of the ball will be smaller.

¢ The smoothness or roughness of the court surface. That is described by the

coefficient of sliding friction COF. For a fast court like grass, COF is about 0.6.
For a slow court like clay, COF is about 0.8. For all other courts, COF is about
0.7. The friction force is given by F = COF x N. At least, that is the case if the
ball slides along the court while it is bouncing. It comes as a surprise to most
people that the horizontal friction force does not depend on the horizontal
speed of the ball but it does depend on the vertical speed.

e¢ Whether the ball slides throughout the bounce or whether it bites the court
during the bounce.

Sliding means simply that the bottom of of the ball slides along the court. It
can do that throughout the whole bounce period or just at the start. If the inci-
dent ball has sufficient topspin, or if it acquires sufficient spin during the
bounce, then the bottom of the ball can actually come to rest on the court. That
occurs if the bottom of the ball spins backwards at the same speed that the
whole ball moves forwards. In that case we say that the ball “bites” or grips or
sticks to the court. If the ball starts to bite during the bounce, then F drops
gradually to zero and then it reverses direction as explained in Chapter 40. It
can even reverse direction twice if it starts to bite early in the bounce. This
complicates the situation and makes it difficult to calculate exactly what will
happen. However, measurements and calculations of bouncing balls have
helped to clarify the situation and reasonably good predictions can be made,
as described later.

The critical factors that determine whether the ball slides or bites are:

¢ the angle of incidence

¢ the coefficient of sliding friction, and
¢ the amount of spin on the incident ball.

A ball incident on the court at an angle of less than about 16 degrees to the
horizontal and without any spin will slide throughout the bounce unless the
court is really rough or covered in coarse sandpaper. A ball served at high
speed is incident on the court at an angle around 12 degrees to the horizon-
tal, and it slides throughout the bounce. A ball incident at an angle larger than
about 20 degrees usually bites the court, especially if it has a lot of topspin.
Typical cases where the ball bites are a slow topspin serve or a groundstroke
hit high over the net or with a lot of topspin.
 Ball Spin and Bounce Off Courts 347

When a player hits a ball, the angle of incidence of the ball on the strings is
usually around 45 degrees or more. The coefficient of sliding friction with the
strings is a lot smaller than that with the court. The ball will bite if the angle
of incidence (between the path of the ball and the string plane) is greater than
about 45 degrees, provided the ball is not spinning when it hits the strings. In
fact, a ball bouncing off the court with topspin is usually incident on the strings
with backspin. That makes it harder to hit the ball back with topspin and it
means that the ball is less likely to bite the strings. Most players prefer to say
that the strings bite the ball rather than the other way around, but the effect is
exactly the same in that the region of the ball in contact with the strings stops
sliding across the string plane if it bites the strings.

When a ball is pushed along the ground it doesn't slide and it doesn't bite, but
it rolls. When a ball bounces it can slide or it can bite, but it doesn't roll.
However, the effect of biting on a tennis ball is similar to rolling. The effect is
much the same when the friction force is small or when the friction force
reverses direction towards the end of the bounce, since the average friction
force during the latter part of the bounce is close to zero. When Howard Brody
first calculated in 1984 how a ball would bounce, he assumed that the ball can
either slide or roll, but it wouldn't bite. Howard was correct in assuming that
the ball can slide, but he was not correct in assuming that the ball would roll.
Nevertheless, it is sometimes a good approximation to assume that a tennis ball
rolls, and it makes calculations much easier. In fact, no one has ever calculat-
ed what happens when a tennis ball bites. Calculations have been done for a
solid ball that bites, but a tennis ball is hollow and that makes an important
difference. We therefore have only three choices at present. We can use
Howard's bounce model, or we can use the model developed for a solid ball
or we can be guided by experimental measurements of tennis balls.

An object that rolls along a surface travels a lot further than an object that slides
along the surface, since the friction force for rolling is much smaller. You can
roll a tennis ball from one end of the court to the other easily, but you can't
slide anything that far. If you slide your racquet along the court, it will proba-
bly stop before it gets to the net. You can't slide a tennis ball the whole length
of a court either. It might start to slide when you push it, but it will start rolling
almost immediately. The same thing happens to a bowling ball in a bowling
alley. A bowling ball starts sliding down the alley but it will start rolling before
it gets to the end. The reason that a ball can roll so far is that the bottom of
the ball does not push horizontally on the surface and the surface does not
push back horizontally on the ball. Any point on the circumference of the ball
comes to rest when it gets to the bottom because it rotates backwards around
the axis of rotation at the same speed as the axis moves forwards. A ball that
is completely at rest on a horizontal surface remains at rest because there is no
horizontal force on the ball. Similarly, if a ball is rolling along a surface, there
is almost no horizontal force acting on the ball since the bottom of the ball is
at rest.

When a ball is rolling, the horizontal speed of the ball is the same as the rota-
tion speed of any point on the circumference around the axis. Suppose that a
point on the circumference takes a time T to complete one revolution. The
rotation speed around the axis = circumference/time = 2mR/T = 2afR where f
= 1/T is the rotation frequency. For example, if the ball spins at f = 100 revo-
lutions/sec, then T = 1/f = 0.01 sec. A tennis ball has a radius R = 33 mm and
348 Chapter 39

a circumference 2nR = 207 mm. If a ball is spinning at 100 revolutions/sec, then

any point on the circumference spins around the axis at a speed of 207 x 100
= 20,700 mm/s = 20.7 m/s. If the ball is rolling along a horizontal surface, then
it travels along the surface at a horizontal speed v = 2xfR = 20.7 m/s. The num-
bers here are typical of the speeds at which a ball spins and slides during a
bounce. Even though a ball does not roll when it bounces, it always comes
close to rolling by the end of the bounce. If you know the horizontal speed of
the ball when it bounces, you can estimate how fast it spins to within about
30% by assuming that it was rolling when it bounced.

Suppose that a ball without spin is incident at some angle on the court, and it
starts sliding along the court at, say, 20 m/s. Friction at the bottom of the ball
will slow it down and it will also make the ball start to spin. The further the
ball slides the slower it slides and the faster it spins. If the ball speed drops to
a point where v = 2nfR, then the bottom of the ball will come to rest. If the
COF is small enough, or if the angle of incidence is small enough, then the ball
will bounce off the court before the bottom of the ball comes to rest.
Conversely, if the COF is large enough, or if the angle of incidence is large
enough, then the bottom of the ball will come to rest before it bounces. If that
is the case, then any increase in the COF or roughness of the surface will have
only a small effect on the final speed or the spin of the bouncing ball. That
would also explain why thin strings and thick strings make very little difference
or no difference at all to the ball spin off the strings (see Chapter 41).

A simple rule of thumb is that when a ball is incident on the court at a low
angle, it slides throughout the bounce and the spin is about 10 or 20% lower
than one would expect for a rolling ball. If the ball is incident at angles greater
than about 20 degrees, it will bite the surface and the spin will be about 10 to
20% higher than one would expect for a rolling ball. At least, this is the case
for a tennis ball. For example, a tennis ball that bounces with a horizontal
speed of 20 m/s will be spinning at about 100 revolutions/sec after it bounces.
If the ball bounces with a horizontal speed of 10 m/s, then it will spin at about
50 revolutions/sec after it bounces. We can subtract 10 or 20% from those fig-
ures if the ball was incident at a low angle, and we can add 10 or 20% if the
ball was incident at a high angle.

When a ball bites, the bottom of a ball comes to rest during the bounce but
the top of the ball keeps going. An elastic ball will tend to lean forwards and
then spring back as the ball bounces off the surface. Elastic energy is always
stored during the bounce due to compression in the vertical direction, which
allows the ball to spring back vertically and bounce off the surface. If elastic
energy is also stored in the bottom of the ball as a result of stretching in the
horizontal direction, then the bottom of the ball can spring back horizontally
as the ball lifts off the surface. This extra kick in the horizontal direction means
that the ball can spin a lot faster when it bounces. This effect is particularly
noticeable with a superball. A superball bounces higher than a tennis ball, and
it also spins much faster when it bounces, since it is almost perfectly elastic in
both the vertical and the horizontal directions.

In the mid 1970's, a gardener in Germany named Werner Fischer came up with
a bright idea to make a tennis ball spin as fast as a superball (US Patent
4,273,331). The idea was to make the strings almost perfectly elastic in a direc-
 Ball Spin and Bounce Off Courts 349

tion parallel to the surface of the stringbed. That way, elastic energy was stored
in the strings as well as the ball. The ball was given extra spin when the strings
sprung back to their original position. One way to do that is to string the mains
and crosses separately so that they are not woven or interlaced. The mains can
then slide freely over the cross strings. Fischer used plastic spaghetti tubing
over the strings to reduce friction, and this system was therefore called
spaghetti stringing, or double stringing, since the mains were tied together and
slid over the crosses. It was quickly banned by the International Tennis
Federation since it allowed a player to put excessive spin on the ball.
Nevertheless, some movement of the strings occurs even when the strings are
interlaced. It is unlikely that this will put any extra spin on the ball since the
strings don't move very far and since they don't spring back freely. It is com-
mon to see top players spend half their time between points pulling their
strings back into position. If the main strings don't spring back freely, then
energy is wasted in pushing them sideways and the energy is not recovered.
An interesting question concerns strings that don't move within the string
plane. The question is whether they don't move at all or whether they do move
and then spring back to where they were while the ball bounces off the strings.
High speed photography shows that the strings do move.

We mentioned above that a ball bouncing off the court or off the strings
changes speed in a direction parallel to the surface. It also changes speed in a
direction perpendicular to the surface. The ball itself does not necessarily trav-
el parallel or perpendicular to the surface, and is usually incident at an angle
of around 20 degrees on the court or around 70 degrees on the strings.
Nevertheless, the ball will have a certain speed in a direction parallel to the
surface and a different speed perpendicular to the surface. For example, sup-
pose that a ball is served from a height of 9 feet and it lands 0.5 sec later on
the service line 60 feet from the server. The ball drops in a vertical direction at
an average speed of 9/0.5 = 18 feet/sec (12.3 mph) and it travels in a horizontal
direction at an average speed 60/0.5 = 120 feet/sec (81.8 mph). If the ball trav-
eled in a straight line from the racquet to the court, it would travel a distance
of 60.7 feet. It actually travels in a curved path of length about 62 feet, in which
case the average speed along that path is 62/0.5 = 124 feet/sec (84.5 mph).

In practice, the ball accelerates in the vertical direction due to the pull of grav-
ity. In the above example, it would hit the court at a vertical speed of about
15 mph. The ball slows down through the air and would hit the court at a hor-
izontal speed of around 70 mph. Suppose the ball bounces off the court with
a vertical speed of 12 mph and a horizontal speed of 42 mph. The ratio of the
vertical speed after the bounce to the vertical speed before the bounce is called
the Coefficient of Restitution or COR. In this case COR = 12/15 = 0.80.

The COR is reasonably constant for a given court, regardless of the ball speed
or the angle of incidence. However, it can vary on an uneven or patchy sur-
face like grass. On grass courts, the ball doesn't bounce as well, and the COR
is about 0.7, but it depends on which patch of grass the ball hits and whether
the grass is new or worn. Values of the COR as high as 0.9 have been observed
on some grass surfaces and also on hard court surfaces, particularly at low
angles of incidence. The court speed in the vertical direction can be just as
important as the court speed in the horizontal direction. When people say that
grass courts are fast and clay courts are slow, they are referring to the speed
350 Chapter 39

in the horizontal direction. But grass courts are usually slow in the vertical
direction and clay courts are fast in the vertical direction. In other words, a ball
bouncing on clay will bounce faster in the vertical direction so it bounces high-
er. This can be a big effect. For example, if one court has a COR = 0.9, and
another has a COR = 0.6, then the vertical speed off the COR = 0.9 surface is
50% larger and the ball will bounce 2.25 times higher. The bounce height is
proportional to the COR squared.

The ratio of the horizontal speed after the bounce to the horizontal speed
before the bounce doesn't have a special name because it depends on the
angle of incidence. In the above example, the ratio is 42/70 = 0.60. On a fast
court the ratio is usually a bit higher and on a slow court the ratio can be lower,
but in practice, the ratio is usually between 0.40 and 0.75 on all courts and at
all angles of incidence of practical interest. In theory the ratio could be 0.9 or
more at very low angles of incidence, but a ball never hits the court at angles
less than about 10 degrees. The ratio is sufficiently important that it ought to
have a special name. In this chapter, we will call it “Court Speed,” since that is
what it describes, but it does vary with the angle of incidence.

The International Tennis Federation (ITF) uses a different word and a different
definition to describe the speed of a court, called Pace. The formula is slightly
complicated, but if the ball slides throughout the bounce, then Pace = 100(1 -
COF) where COF is the coefficient of sliding friction. For a grass court, COF is
about 0.6 and hence Pace = 100 x 0.4 = 40. For a clay court COF is about 0.8
and hence Pace will be about 100 x 0.2 = 20. The advantage of this definition
is that it doesn't depend on the angle of incidence. A disadvantage is that it
does not give a direct indication of how much the ball slows down.
Furthermore, two courts can have the same Pace but the ball will slow down
by different amounts if the COR on each court is different.

The “Court Speed” defined above depends on four factors, not just one. It
depends on

e the COF
¢ the COR
e the angle of incidence
e the spin of the incident ball.

The amount of spin on a ball that bounces off the court depends on all these
factors and it also depends on the ball speed. The spin of a ball bouncing off
the strings of a racquet depends on all of the above and it also depends on the
speed of the racquet, the direction of motion of the racquet and the angle of
the racquet face.

Every tennis player knows that the grass courts at Wimbledon are fast and that
the clay courts used for the French Open are slow. Just how fast or just how
slow is hard to measure. Not impossible, but not as easy as you might think.
The ITF has sponsored the development of a device called the Sestee to meas-
ure court speed. It works by projecting a ball at 30 m/s and at 16 degrees to
the court, and it uses infrared beams to measure the incident and bounce
 Ball Spin and Bounce Off Courts 351

speeds and angles. It is very accurate, it costs around $50,000 and it needs a
trained operator to use it. Tennis clubs can't afford to buy one. Others have
tried to measure court speed by dragging a tennis ball or a piece of rubber
across the court to measure the friction force. This is not as accurate but it is a
lot cheaper.

There is another way to measure court speed that is both relatively cheap and
accurate, and that is to film the bounce of a ball with a video camera. In fact,
it is something that you could easily do yourself if you wanted to. It could also
help resolve expensive legal issues that sometimes crop up when there is a
dispute between a court builder and a club that feels that the builder let them
down. The best sort of camera is a digital video camera since the images can
be downloaded to a computer with a video card in order to measure the speed
of the ball and the corresponding incident and rebound angles. If you focus
close up to the ball, you can also measure how fast it is spinning, provided
there is a mark or line drawn on the ball to measure the rotation angle from
one frame to the next. However, if you focus up close, then you can measure
the ball speed only if the ball is traveling slowly.

Video cameras capture images at 25 or 30 frames/sec which means that the

time between each image is 1/25 or 1/30 sec (40 ms or 33.3 ms). If a ball is
incident at 30 m/s, then it travels 1.2 m in 40 ms so the camera needs to view
several meters of the court in
order to measure the ball speed.
The ball will then be too small Figure 39.1
and too blurry to measure its Illustration rendered from video film of a ball bouncing at low speed ona
spin. But, if you throw the ball at Rebound Ace surface.
low speed, you can focus up
closer and measure both the
spin and the speed of the ball. A
fast shutter speed is needed to Bounce on Rebound Ace
avoid blurry images. A video
camera with a sports mode set-
ting will have a suitably fast
shutter speed. An example is
shown in Figure 39.1. This is a
composite drawing copied from
actual video film so that the
images are a bit easier to inter-
pret.

The drawing shows the position

of the ball in each frame, which
is 40 ms later than the previous
frame. The first frame is labeled 0.181)/0.04
as time zero and the next as 0.04 s 0.669)/0.04
etc. The ball was filmed as it passed
in front of a large screen with 0.080)/0.04
vertical and horizontal lines to
0.031)/0.04
calibrate the distance scales. A
straight line was drawn with a
felt pen across a diameter of the
ball, with a dot at one end to Rotation speed = 300° every 0.04 s = 20.9 revolutions/sec

distinguish one end of the line Incident angle = 29.5° Rebound angle = 43.4°
352 Chapter 39

from the other. The ball was thrown without spin so that it bounced on a fac-
tory sample of Rebound Ace glued to a slab of concrete. Rebound Ace is the
surface used for the Australian Open and is similar to the Deco Turf used for
the US Open in that it has a 1 mm thick acrylic and sand mixture on top of a
spongy rubber backing.

The position of the center of the ball is given in Figure 39.1, together with cal-
culations of the ball speed before and after the bounce.

The calculations in the figure are not highly accurate since they are based on
only two positions of the ball. A more accurate result would be obtained by
plotting a few more images and then using a best fit line drawn through all the
data points. The speeds in the horizontal direction are calculated from the x
coordinates of the ball. For example, the ball moved from x = 0.181 m at time
zero to 0.446 m at 0.04 s so it moved in the horizontal direction at a speed v,,
= 6.62 m/s before the bounce. After the bounce it traveled at a horizontal speed
V,, = 3.22 m/s. The speeds in the vertical direction are calculated from the y
coordinates of the ball. The ball was incident in the vertical direction at Vets
3.75 m/s and it bounced at speed ven 3.05 m/s in the vertical direction.

The court speed can be defined by the ratio v,, /v,, which for this bounce is
3.22/6.62 = 0.49. In other words, the horizontal speed after the bounce is a frac-
tion under half the horizontal speed before the bounce. At other angles of inci-
dence this ratio might be different, but at least we know the court speed for
an angle of incidence 0, = 29.5°. The angle of incidence is given by tan(®,) =
Weg Vg 7 0: 101 002" = 0.506.

The coefficient of restitution for this bounce is COR = Vy2 ee = 3.05/3.75 =

0.81. This is a bit higher than one gets for a purely vertical bounce, partly
because the vertical speed of the incident ball is slower than the vertical speed
in a standard 100 inch drop test.

The ball was found to spin at 20.9 rev/sec after the bounce. It looks like the
ball is spinning backwards slowly in Figure 39.1 but it was actually spinning
forwards rapidly, by 300 degrees every 0.04 sec. This was checked by switch-
ing the camera to a high speed mode to record at 100 frames/sec. Not many
cameras have this feature. An interesting result is found if one compares the
spin of the ball to its horizontal speed. Any point on the circumference rotates
at speed v = 2nRf = 4.33 m/s where R = 0.033 m and f = 20.9 rev/sec. This is
greater than the horizontal speed v,, = 3.22 m/s, which means that any point
on the ball in contact with the surface is sliding backwards on the surface by
the time the ball leaves the surface. The ball started sliding forwards on the sur-
face at the start of the bounce since it had no spin to start with, but then it
acquired enough spin during the bounce to slide backwards. The ball did not
roll. Instead, it gripped the surface for a while and then it started sliding back-
wards.

Results obtained at other angles of incidence are shown in Figure 39.2, togeth-
er with results obtained on three other surfaces. One surface was a smooth slab
of concrete. The other two were made by bonding sheets of P800 and P150
emery paper to two other smooth concrete slabs. P800 is a fine grade and P150
 Ball Spin and Bounce Off Courts

is relatively coarse, but one can Rebound Ace Smooth Concrete Slab
rub a hand firmly across the
surface without cutting the skin.
P150 means that there are about
150 particles per inch on the
emery paper. Results on real
courts are described below. The
advantage of these “artificial”
court surfaces is that the sur-
faces are more uniform than a
0
real court, so there is less vari- 0 10 BO si) 40 50 0 10 20 30 40 50
ability to contend with. In addi- 0, (degrees) 8, (degrees)
tion, the hard concrete surfaces
are not deformed by the ball, so
P800 on Concrete Slab P150 on Concrete Slab
that is one less variable to
worry about. It is not easy fig-
uring out how a ball bounces. 0.8 0.8
Having only a few variables MeV > ees a
p56 2 0.6
makes the task a bit easier. 0.6

Each circle in Figure 39.2 repre-

? 0.4 0.4

sents a single bounce, filmed by 0.2 0.2

throwing a ball at a speed of
0
about 7 m/s at various angles 10 15 20 25 30 35 40 45 50° 10 15205 25 2 SO some O
onto each surface so that the
0, (degrees) 9, (degrees)
ball spin could also be meas-
ured. The dashed lines’ in
Pioure) 155).20 represent: “the Figure 39.2
expected result at very low
Court speed vs angle of incidence 0, for four different surfaces.
angles of incidence, and the
solid lines are included to show
the transition from sliding at
low angles to biting at high
angles. 0, is the angle of incidence (between the path of the ball and the hor-
izontal). All the results shown in Figure 39.2 were obtained by using at least
three images of the ball before and after the bounce.

The most obvious result is that the court speed is large at low angles of inci-
dence, and it is smaller at high angles of incidence. At low angles, the speed
decreases as the angle increases, but at high angles the speed remains constant
as the angle increases. The difference is due to the fact that the ball slides
throughout the bounce at low angles, and it bites at high angles. On the
rougher P150 surface, the speed remained constant at all angles which means
that the transition from sliding to biting occurred at an angle of incidence less
that 13°. On the smooth concrete slab, the ball started to bite only when the
. . . oO
angle of incidence was increased to 35°.

Have a good look at Figure 39.2 and see if you can decide which surface was
the fastest and which was the slowest. The speed of each surface is summa-
rized in Table 39.1, showing the speed (ie., v,, /V,,) at angles of incidence of

16 degrees and 30 degrees. At 16°, the fastest surface was the smooth concrete
slab and the slowest was the rough P150 surface, as one might expect. But at
30°, the fastest surface was P150 and the slowest was the smooth concrete slab.
354 Chapter 39

The ball spin on each of the four surfaces is shown in Figure 39.3. This is plot-
ted in terms of the ratio S = v,..,/Vxa where Vin = 2nkf is the speed of a point
on the circumference of the ball, R = 0.033 m is the ball radius, and f is the ball
spin in rev/sec. If S is less than 1, then the ball
Table 39.1 slides throughout the bounce. If S$ = 1, then the
ball started to roll during the bounce, but if S >
Summary of results in Figure 39.2
1, then the ball bit the surface when it bounced.
Surface Speed at 16° Speed at30° Bite Angle S=1 The angles at which S = 1 are shown in Table
39.1. These angles are all a fraction lower than
Smooth slab 0.74 0.49
the bite angles shown in the table, which means
Rebound Ace 0.70 0.54
that the effect on the court speed due to biting
P800 0.65 0.51
must be similar to the effect due to sliding at the
P150 0.56 0.56
point where S = 1, and the difference is impor-
tant only when S is greater than about 1.15.

Figure 39.3
Ball spin vs angle of
incidence on four dif-
ferent surfaces. Rebound

Smooth
concrete slab

10 20 30 40 50 60 10 20 30 40 50
8, (degrees) 9, (degrees)

Measurements of the COR for bounces on the smoothest and roughest of the
four surfaces are shown in Figure 39.4. A correction was made when analyz-
ing the data to account for the acceleration due to gravity in the vertical direc-
tion. This affects calculations of ball speed in the vertical direction but no such
correction is needed to calculate the ball speed in the horizontal direction. In
Figure 39.4, the COR is plotted as a function of Vyy Since it is well known that
the COR decreases as Ma increases, at least for a ball incident vertically on a
surface. That is easy to understand since the ball squashes more in a high
speed impact and it loses a greater fraction of its stored elastic energy. The ball
bounces higher as the incident speed increases, but the ratio of the rebound
speed to the incident speed decreases as the incident speed increases.

When the ball was dropped vertically from various heights onto each of the
four surfaces, the COR was found to decrease slightly as the drop height
increased, but the effect was relatively small, and the COR was about 0.79 for
all drop heights on all surfaces. Figure 39.4 shows the results for a vertical drop
on the P150 surface. A larger drop in the COR would occur if the ball was inci-
 Ball Spin and Bounce Off Courts 355

Figure 39.4
Smooth Concrete Slab P150 Surface
COR vs vertical
speed on two dif-
Oblique bounce ferent surfaces.
oblique bounce

COR ®e—— eo
vertical drop

V4 (m/s) er (m/s)

dent vertically at much higher speeds, but this is not particularly relevant in
tennis. Even for a very fast serve, the ball hits the court at a vertical speed of
only about 8 m/s. An approved ball should have a COR of close to 0.75 for a
100 inch vertical drop, but the balls tested were a bit bouncier than this.

Of greater interest and significance is the fact that the COR increases when the
ball is incident at an oblique angle, as shown in Figure 39.4. This result is sur-
prising, and no one knows for sure why it happens. There are several possi-
ble explanations, such as:

e The ball forms a depression in the surface and is deflected upwards by the
ramping slope of the depression.

e The ball spins so fast that each new section of the ball contacting the sur-
face moves off the surface before it has time to compress very far.

e¢ Some of the kinetic energy due to horizontal motion of the ball gets chan-
neled into the vertical direction.

While each of these effects might make a contribution in some cases, the first
two effects cannot be important in the present case, since concrete is too rigid
for the ball to form a depression and since the COR was enhanced even at low
ball speeds. The third explanation looks more promising. Suppose that you tie
a string on each side of a ball and pull horizontally. The ball will stretch in a
horizontal direction. Now place the stretched ball on a horizontal surface and
cut both strings at the same time. The ball will spring back to its original spher-
ical shape and bounce up off the surface. That is an interesting effect since it
means that a force can be applied in a horizontal direction and the result is
motion in the vertical direction. A similar effect can be engineered under dif-
ferent circumstances using levers or pulleys but a tennis ball does it by spring-
ing back into shape.

A ball can be expected to stretch horizontally when it bounces, since the bot-
tom of the ball will slow down faster than the top of the ball. The horizontal
stretch would be bigger on a rough surface than on a smooth surface and
would be more noticeable if the horizontal speed of the ball is a lot larger than
the vertical speed. This is consistent with the results in Figure 39.4, since the
COR is larger on the rough P150 surface than on the smooth concrete surface.
356 Chapter 39

Furthermore the COR is largest at low angles of incidence where v,, is small.
The same effect is seen on real courts, as described in the following section.

Measurements of court speed on a Rebound Ace court and on a clay court are
shown in Figure 39.5. In this case the ball was served down the center line to
land on the court at about 30 m/s and at an angle of incidence of about 12.
Bounces were also filmed by hitting or throwing a ball from near the net at
about 15 m/s to land at higher angles of incidence. It was found that the ratio
Vv, /V,, does not depend significantly on the incident ball speed, so all the
results for the different ball speeds can be combined into the one graph with-
out affecting the speed vs incident angle graphs.

Figure 39.5 Rebound Ace (v, = 15 - 30 m/s) Clay Court (v, = 15 - 35 m/s)
1
Court speed of real courts
at real ball speeds 0.8
Vyo I Vy
, 0.6

Ae; 0.4

0.2

0
1@ is 20) 2s co) S 4@ 0 10 20 30 40
0, (degrees) 8, (degrees)

The ball was filmed as it passed in front of two vertical ball cans 1.00 m apart
in order to calibrate the horizontal and vertical scales on the film. The cans
were placed along the center line, the
camera axis was at right angles to the
center line, and the ball was projected
parallel to the center line so that an
accurate measurement could be made
of the height of the ball in each frame
(Figure 39.6).

These results show that clay is indeed

a slow surface but the Rebound Ace
court tested was not much faster. The
result for Rebound Ace in Figure 39.2
differs from the result in Figure 39.5
since the court had a rougher finish
=e - 1.00 mf due to the extra sand content. The
speed of Rebound Ace can be varied
Figure 39.6 by changing the amount and size of
the sand in the acrylic and by sweep-
Video film of a 90 mph serve on Rebound Ace. The three frames on the left
ing the acrylic with a broom when it is
show the ball approaching. The right frames show the ball after it bounced.
applied. The broom leaves small
Filmed at 100 frames/sec with 2 ms exposure each frame. The shadow of
the ball provides some depth information, but this was not needed.
ridges, with the result that sweeping
across the court generates a slower
surface than sweeping along the court.
 Ball Spin and Bounce Off Courts 357

Rebound Ace Court Clay Court Figure 39.7

COR vs angle of inci-
v= 15 to 30 m/s dence on real courts.

COR COR

10 15 20 25 30
0, (degrees) 0, (degrees)

Measurements of the COR for these two courts are shown in Figure 39.7. The
COR varies quite a bit from one bounce to the next, so grass courts are not the
only courts with a variable bounce. Both courts have COR values greater than
the value 0.79 observed for a purely vertical bounce on these courts, especial-
ly at low angles of incidence. A COR greater than 0.9 is normally observed only
with a superball. If the COR was 0.79, a fast serve would be expected to
bounce just above waist height. But since the COR is sometimes as large as 0.9,
the ball will sometimes bounce around shoulder or head height even without
any deliberate topspin. A head-high bounce is usually the result of a kick serve
with lots of topspin, but it can also happen in a “flat” serve if the receiver
stands far enough behind the baseline.

Sse een ts ote Seen eee

The coefficient of sliding friction is usually represented by the Greek letter u
and it can be calculated from the speed and COR measurements obtained at
low angles of incidence. The relevant formula is

Vyo /V,,= 1- u(1 + e) tan(8,) (39.1)

where e is the COR. A value for u can be calculated tor

‘ ; Table 39.2
each and every bounce provided the ball slides through-
out the bounce. The average value of uw and e for each of Summary of Court Speed Results
j, wae: Ps : ;
the above surfaces is given in Table 39.2. A reliable value Surtace ‘
for u could not be found for the P150 surface since the ball
did not slide throughout the bounce even when the angle Pec ni on
8 Smooth Rebound Ace 0.62 0.80
of incidence was only 13°. If the ball bites, then the above P800 on concrete 0,73 0.82
formula underestimates the true value of uw. This table P150 on concrete 20 0.85
highlights the fact that surfaces which are slow in the hor-
izontal direction are usually fast in the vertical direction Rebound Ace court 0.70 0.83
and vice-versa, but grass is a notable exception (depend- Clay court 0.80 0.85
ing on the length of the grass). Artificial grass courts have
values of uw and e typically about 0.68 and 0.82 respec-
tively.
 How A Ball Bites the
Court When It Bounces

Venus Willams, United States

By Rod Cross

Introduction

or a long time, people involved in the physics and engineering of sport assumed that when a ball bounces,
it either slides throughout the bounce, or it starts to roll when it acquires enough spin. The assumption
about sliding is correct, but the assumption about rolling has recently been found to be incorrect. When
pall bounce they bite the surface instead of rolling. As a result, balls that bite spin faster than expected, they
slow down more than expected and the friction force at the bottom of the ball does not suddenly drop to zero.
Instead, the friction force decreases slowly to zero and then reverses direction due to the fact that the ball starts
vibrating in a horizontal direction during the bounce. In this section we describe experimental results showing
that real balls bite when they bounce.

The arrangement used to measure the friction force on a bouncing ball is shown in Figure 40.1. A wood block
was supported on two cylindrical rollers so that the block could move freely in the horizontal direction with
almost no frictional resistance. A ball incident obliquely on the block bounced off the block at an oblique angle,
exerting a force on the block only during the brief period of the impact. The horizontal component of the force
360 Chapter 40

on the block is equal in magnitude and opposite in

direction to the friction force on the ball. The hor-
izontal force caused the block to accelerate in the
horizontal direction (from left to right) and it
Tennis ball
caused the ball to slow down. A 19 mm diameter,
0.3 mm thick ceramic piezo disk was fastened to
Smooth or rough surface one end of the block as a simple accelerometer.
The output voltage from the piezo is directly pro-
portional to the acceleration of the block so it pro-
=I
eZ vided a direct measurement of the time-varying
Piezos to measure N
Piezo disk friction force acting on the bottom of the ball. For
Wood block to measure F the low speed bounces studied in this experiment,
a block mass of 340 gm (about six times the mass
& Cylindrical rollers & of a tennis ball) was used to obtain an adequate
output signal from the accelerometer.

Figure 40.1 Two large area ceramic piezo blocks were mount-
Apparatus used to measure the friction
ed on top of the wood block to measure the verti-
force on a bouncing ball. cal force on the block. Each of the large piezos was
51 mm square and 4 mm thick and were connect-
ed electrically in parallel to act as a single, large
surface area (51 x 102 mm) force plate. The upper surface of the plate was
mechanically protected using 0.3 mm thick circuit board attached directly to
the plate with double-sided adhesive tape. This surface was quite smooth and
is referred to below as the low friction surface. For some experiments, fine
grade (P800) emery paper was taped firmly to the circuit board to study the
bounce off a surface with a higher coefficient of friction than the smooth cir-
cuit board surface. It is referred to below as the high friction surface.

Bounces at several different angles on both surfaces were filmed at 100

frames/sec with a digital video camera to determine the ball speed, spin and
angle before and after each bounce. Video clips were transferred to a computer
for analysis. The horizontal and vertical distance scales were calibrated simply
by measuring the actual and image diameter of each ball.

Simultaneous measurements of the vertical force, N, and the horizontal friction

force, F, are shown in Figures 40.2-40.5. Figure 40.2 shows the bounce of a new
tennis ball on the low friction surface, and Figure 40.3 shows the bounce on
the high friction surface. Figure 40.4 shows the bounce of a superball and
Figure 40.5 shows the bounce of a basketball, both balls being incident on the
low friction surface. In all cases, the ball was incident at low speed, typically
about 3 m/s. The properties of each ball are listed in Table 40.1.

The results in Figure 40.2 demonstrate that a tennis ball

Table 40.1 Ball Properties
incident on a low friction surface bounces without sig-
Ball Mass m
nificant reversal of the friction force during the bounce.
Radius R ae
ae At low angles of incidence the ball slides throughout the
Tennis 57.4 gm Be Ona bounce and the ratio F/N remains approximately constant
during the bounce. In Figure 40.2a, F/N = 0.27 + 0.03,
Superball 46.4 gm 23.0 mm which gives the coefficient of sliding friction. At higher
angles of incidence, the friction force dropped to zero
Basketball 589 gm 120 mm during the bounce, and it dropped to zero earlier as the
 Ball Bite and Bounce 361

Figure 40.2
Results for a tennis ball incident on the smooth, low friction surface.

2a 2b

50 100

40 =, tie
5 5
=) 30 Ss 6O
2 2
ae — 40
LL LL
2 10 iS 20
2)

= 0 Zz (6)

416) -20

t (ms)

angle of incidence was increased. These results are qualitatively consistent with
the bounce model described by Brody, who predicted that the ball would start
to roll at a progressively earlier stage as the angle of incidence was increased.
However, there is no sudden transition from sliding to rolling during the
bounce, and the ball was observed to spin faster than allowed by the rolling
condition. The fact that the friction force dropped to zero during the bounce
can be interpreted to mean either that the ball rolled or that the friction force
was positive in some areas of contact and negative in others. For reasons
described in more detail below, it can be inferred that the ball did not roll dur-
ing the bounce but it first gripped the surface and then commenced to slide
backwards.

In practice, the coefficient of sliding friction between a tennis ball and a court
surface is usually about 0.5 or larger. On such a surface, the ball is predicted
to slide throughout the bounce if the angle of incidence is less than about 16°.
At higher angles of incidence, Brody predicted that the ball would enter a
rolling mode, and hence, the friction force would drop to zero. Figure 40.3
shows that at these higher angles of incidence the bounce of a tennis ball on
a high friction surface is characterized by a significant reversal of F during the

Figure 40.3

Results for a tennis ball incident on the rough, high friction surface.

3a 3b

40 100

80
= 30 a

g £ 60
% 20 i
= = 40
z 2 20

0 ey
2 Zz
E10 -20
-2 0 2 4 6 8
362 Chapter 40

bounce. The reversal in the direction of F occurs earlier in time as the angle of
incidence is increased, allowing two reversals to occur in Figure 40.3b. In
Brody's model, the ball rolls at an earlier stage as the angle of incidence is
increased, but F does not reverse direction.

The bounce of a superball incident without spin at 36° on a low friction sur-
face is shown in Figure 40.4. The behavior of the friction force is qualitatively
similar to that for a tennis ball bouncing on a high friction surface in that F is
almost equal to N during the early stage of the bounce, and F reverses direc-
tion towards the end of the bounce period. However, the effect on the ball is
different in that a superball spins much faster than a tennis ball after the
bounce. The peripheral speed of the superball after it bounced, Rw,, was 1.75
times larger than its horizontal speed v,,.aon For the tennis ball in Figure 40.3, Ro,
is about 1.1 times larger than v,,.

Superball Basketball

F(Newton)
Nand

Figure 40.4 Figure 40.5

Superball on smooth surface. Basketball on smooth surface.

Figure 40.5 shows the bounce of a basketball incident with almost zero spin at
66° on the low friction surface. It is especially obvious in this case that the ball
vibrates horizontally during the bounce, causing the friction force to reverse
direction six times. The half period of oscillation is 15 ms in the vertical direc-
tion, and the full period of oscillation in the horizontal direction is about 4 ms.
There is no simultaneous 250 Hz oscillation in the N waveform, indicating that
this mode of oscillation involves horizontal displacements in the wall of the
ball rather than vertical displacements.

The difference between a ball that bites and a ball that slides is illustrated in
Figure 40.6. Part (a) shows a ball that is sliding along a surface while it rotates.
Itis assumed that every point on the ball is traveling from left to right at a hor-
izontal speed of 8 m/s and that every point on the circumference is rotating
about the axis at a speed of 6 m/s. The axis itself is traveling from left to right
at 8 m/s. The combined effect of these separate motions is shown in Figure
40.6b, A point at the top of the ball is traveling from left to right at 8 etbut
it is also rotating about the axis at 6 m/s so the combined speed is 14 ai A
point at the bottom of the ball travels left to right at 8 m/s, but it is also rotat-
 Ball Bite and Bounce 363

ing right to left at 6 m/s, so the Figure 40.6

combined effect is motion left Motion of a ball when it slides (a and b) or bites (c - f).
to right at 2 m/s. Every point at
the bottom of the ball therefore
slides along the surface at 2 14 m/s

m/s, and the friction force F acts

backwards as shown in Figure
40.6b.
10 m/s 10 m/s
The point at the front end of the
ball travels left to right at 8 m/s
and it also rotates downwards at
6 m/s. The combined effect is
that this point moves down-
wards and to the right at 10 m/s
(vector addition of two velocity
components at right angles
requires that we add_ the
squares of each component; in
this case 10” = 8” + 6”) 9.9 m/s 9.9 m/s

The effect of the friction force is

to reduce the horizontal speed
of the whole ball and to make it | ©)
rotate faster. Suppose that the
horizontal speed decreases to 7
m/s and the rotational speed
14 m/s
increases to 7 m/s as shown in
Figure 40.6c. The combined
effect of these separate motions
is shown in Figure 40.6d. In this
case, the bottom of the ball
comes to rest on the surface,
since the horizontal motion for-
wards is cancelled by backward
rotation. When an object is at
rest on a surface and remains at
rest, or when it is rolling at con-
stant speed, the friction force is
zero. In this case the friction
Rotation about axis and motion Combined effect of rotation about
force? isnot zero “since” the
to the right shown separately axis and motion to the right
whole ball is not at rest, and
since the ball arrived at this
point by stretching in the hori-
zontal direction while it was
sliding. The ball shown in
Figure 40.6b is stretched horizontally by a small amount. It stretches because
there is a horizontal force on the bottom acting to the left and because the
upper part of the ball is moving to the right. It might be easier to visualize this
effect if you imagine that the ball has a top half and a bottom half connected
by a spring. If the ball travels from left to right and you grab hold of the bot-
tom half, then the top half will tend to keep moving to the right, causing the
spring to stretch.
364 Chapter 40

The upper part of the ball in Figure 40.6d is therefore in a stretched state, and
it tends to pull to the right on the bottom of the ball. There is also a large force
on the ball in a direction perpendicular to the surface. If the coefficient of stat-
ic friction is large enough, the bottom of the ball will remain stuck to the sur-
face (like a book pushed against a vertical wall) and the static friction force on
the ball will act to the left. This stops the ball sliding along the surface and it
opposes the stretching force acting to the right.

The fact that F is not zero in Figure 40.6d means that the ball will continue to
reduce speed in the horizontal direction and it will continue to rotate faster.
Suppose that the horizontal speed decreases to 6 m/s and the rotational speed
increases to 8 m/s as shown in Figure 40.6e. The combined effect of these sep-
arate motions is shown in Figure 40.6f. The situation here is complicated by
the fact that the bottom of the ball tends to remain stuck. It eventually becomes
unstuck, but it does so progressively, not suddenly. Figure 40.6e shows one
instant of time where the middle part of the contact area is still stuck, and the
outer part of the contact area is starting to slide. As the ball rotates about its
axis, it slows from 8 m/s to 7 m/s when it first contacts the surface, and it slows
to 6 m/s at the center of the contact area. It then speeds back up to 8 m/s as
it rotates around to the back of the ball. This means that bits of the ball start
piling up at the bottom of the ball. The front edge of the ball gets compressed
and the wall gets thicker. The back edge of the ball is rotating faster than the
central stuck region so it is under tension and the wall thickness decreases
back to its normal size. The effect is the same as water flowing through a pipe
of varying diameter. Where the pipe gets fatter, the water slows down and
where the pipe gets thinner the water speeds up. Rubber does this by itself
without needing a pipe to stop it leaking out.

Figure 40.6f shows a central shaded area that is still stuck and where the fric-
tion force is still acting to the left. The annular region outside this area is slid-
ing Gn compression at the front or tension at the back) along the surface to the
left, and the friction force on this annulus therefore acts to the right. The total
friction force on the bottom of the ball can therefore be positive, zero or neg-
ative, depending on the size of the area that is stuck. The size of that area
decreases with time until it is zero, at which point the whole contact area slides
backwards on the surface.

The friction force acting on a bouncing ball slows it down in a horizontal direc-
tion and causes it to rotate. It also has another effect that we haven't mentioned
so far, and it is essentially the same effect that occurs when you apply the
brakes of a vehicle. That is, the weight of the vehicle shifts forwards. Every
time you apply the brakes, the front end of the vehicle moves down closer to
the road and then it bounces back up again when you come to a complete
stop. This is especially noticeable if you slam on the brakes to stop quickly.
The friction force on the wheels slows the vehicle in a horizontal direction and
Causes it to rotate due to the torque acting about the center of mass. As the
front end moves down, the vertical force on the front wheels increase. This
results in a torque that acts in the opposite direction and prevents the vehicle
spinning like a ball.

When a vehicle is at rest or moving at constant speed, the vertical force on the
front wheels is roughly the same as the vertical force on the back wheels.
 Ball Bite and Bounce 365

Together, the total vertical force on all four wheels is equal to the weight of
the vehicle. When the brakes are applied, the vertical force on the front wheels
increases and the vertical force on the rear wheels decreases as if the weight
of the vehicle shifted forwards. The engine and the passengers and everything
else stays roughly where they were so nothing actually moves towards the
front wheels, but the effect is the same so we say that the weight shifts for-
wards.

A similar thing happens to a tennis ball when it bounces, since the vertical
force at the front of the ball is larger than the vertical force at the back of the
ball. The distance between the front and the back of the ball is too small to
stop it rotating, but there is a shift in the “weight” of the ball, and there is a
torque acting backwards on the ball that stops it spinning as fast as it other-
wise would. The vertical force on a ball when it is bouncing has nothing to do
with its actual weight since its actual weight is tiny compared with the vertical
force due to compression of the ball. However, the front of the ball will com-
press more than the back of the ball due to the fact that the ball is rotating.
The amount of compression depends on the vertical speed at which the ball
hits the surface.

If you drop a ball from a height of, say, 3 feet, the ball will compress by about
1/4-inch before it bounces back up. If you drop it from a height of 12 feet, the
ball will hit the surface at twice the speed and compress by about 1/2 inch.
When a ball is spinning, as in Figure 40.6, the front of the ball strikes the sur-
face at a higher speed than the back of the ball so it will compress more. This
was not shown in Figure 40.6 since we con-
sidered only the combined effects of hori-
zontal motion and rotation. If we add in the
vertical motion, then we would get the
complete picture, including a shift in the
“weight” forwards. The result is that the
normal reaction force, N, acts vertically
through a point ahead of the center of the
ball as shown in Figure 40.7.

There are more symbols than necessary on

this diagram, but they are needed for a
complete mathematical description of the
bounce as described by Cross (2003a, b).
The important feature here is that if N is
shifted by a distance D, then there is a
torque ND that opposes the torque FR,
Figure 40.7
where R is the ball radius. Measurements of
ball spin and speed show that D is typical- Geometry of a bouncing ball.
ly about 4 mm, but it can be as large as 11
mm on a clay court. The large value of D on
clay is probably due to an accumulation of clay under and towards the front
of the ball, since the ball sweeps clay particles along and leaves a definite
depression in the court.

A consequence of all this is that the speed of a court depends not only on the
COF and the COR but also on D. The value of D is not important if the ball
slides throughout the bounce but it is important if the ball bites. It is for this
reason that, in Table 39.2, the smooth concrete slab was the fastest surface at
low angles, but it became the slowest surface at high angles.
366 Chapter 40

An interesting question is whether players base their estimate of court speed

on low angle bounces or high angle bounces or both. I suspect that it is main-
ly on medium pace, high angle bounces since the ball travels too fast in a fast
serve for players to make a reliable estimate of the court speed. Furthermore,
the percentage change in ball speed is larger at high angles than at low angles,
meaning that differences in court speed are magnified at high angles of inci-
dence, not only in a relative sense but also in absolute terms. A medium pace
groundstroke takes about 0.1 seconds longer to cross the baseline on a clay
court than on a grass court. A 0.1 sec difference is much easier to pick than
say a 0.02 sec difference. A fast serve at 120 mph that lands on a clay court
takes only about 0.02 seconds longer to cross the baseline than on a grass
court, assuming that it is served at the same angle and the same speed. But
how can players judge if the 0.02 s difference was due to a change in court
speed rather than a change in serve speed or angle? I suspect they can't. It
would be interesting to do an experiment using a radar gun to find out if they
can pick the difference between a 120 mph and a 124 mph serve, since the ball
would then take 0.02 s longer to cross the baseline for the slower serve.

Calculations of ball trajectories show that the ball can bite the court for a medi-
um pace groundstroke even if the ball is hit without any topspin. For example,
suppose that a ball is hit without spin from a height of 1 meter above one base-
line and it lands 2 meters inside the other baseline. If it is hit at 20 m/s, it would
need to be hit upwards at 20 degrees to the horizontal. It will pass 2.0 meters
above the top of the net and it will land 2 meters inside the baseline at an angle
of incidence of 31 degrees. Such a ball will bite on essentially all courts. If the
ball is hit at 25 m/s, it needs to be hit upwards at an angle of 11 degrees to
the horizontal in order to land 2 m inside the baseline. It will pass over the net
at a height of 1.0 metre above the net, and it will land at an angle of incidence
of 20 degrees. This ball is unlikely to bite if it has no topspin, but a small
amount of topspin will make it bite since the angle of incidence will then be
greater than 20 degrees.

A ball hit without spin at 30 m/s is relatively fast for a groundstroke and is like-
ly to slide without biting when it bounces. Such a ball would need to be hit
upwards at an angle of 6.7 degrees to the horizontal in order to land 2 m inside
the baseline, it will pass 0.5 m above the net and it will land at an angle of
incidence of 14.6 degrees. At a speed of 30 m/s or more, the speed of a court
will therefore depend primarily on the coefficient of sliding friction. At a speed
of 25 m/s or less, the speed of a court will depend on the coefficient of fric-
tion but it will also depend on D.

References

Brody, H. (1984) That's how the ball bounces, The Physics Teacher, 22,
494-497,

Cross, R. (2002) Measurements of the horizontal coefficient of restitution fora

superball and a tennis ball, American Journal of Physics, 70, 482-489.

Cross, R. (2003a) Grip-slip behavior of a bouncing ball, American Journal of

Physics, in press.
 Ball Bite and Bounce 367

Cross, R. (2003b) Measurements of the horizontal and vertical speeds of tennis

courts, Sports Engineering, in press.

Hierrezuelo, J, Catolicos R. and Carnero C. (1995) Sliding and rolling: the

physics of a rolling ball, Physics Education, 30, 177-182.

Tabor, D. (1994) The rolling and skidding of automobile tyres, Physics

Education, 29, 301-306.

Thorpe, J.D. and Canaway, P.M. (1986) Performance of tennis court surfaces II.
Photometric methods for measuring pace and bounce under playing condi-
tions. Journal of Sports Turf Research Institute, 62, 101-117.
 Racquet
Spin, Bounce and Strings

John McEnroe, United States

By Rod Cross

Introduction

» p layers these days like to put as much topspin or backspin on the ball as possible, consistent with reason-
able ball control. There are many different opinions as to how this is best achieved. The most common are:

e thin strings dig into the ball, allowing players to put more spin on the ball.
e thick strings, widely spaced, are better so that the ball squeezes between the strings
¢ Strings with a rough texture will get a better grip on the ball.
¢ Soft strings like natural gut act to cup the ball, which gives a better grip on the ball.
hit the ball on the rise
hit the ball when it is dropping
hit the ball towards one edge of the ball instead of in the middle
grip the racquet so that the racquet face is inclined forwards
for a right-handed serve, toss the ball over the left shoulder

Obviously, all of these methods can't be correct, but some of them are. From a physics point of view, there are
only three simple rules regarding spin:
370 Chapter 41

Spin Rule 1: The faster you hit the ball the faster it will spin.
Spin Rule 2: The strings don’t make much difference.
Spin Rule 3: Maximum spin occurs at an angle of incidence near 40 degrees.

Rule 2 will surprise many players and is explained in more detail below. Rule
3 means that an insect sitting on the strings would see the ball coming in at an
angle of about 40 degrees to the string plane. Most tennis players can’t hit the
ball comfortably at that angle, but a lot of table tennis players do (see Note 1).
The least amount of spin occurs if the ball strikes the strings almost at right
angles (depending on the spin of the incident ball) or if the ball hits the strings
at a very small grazing angle and just brushes the strings lightly. The reason
that about 40 degrees gives the maximum spin is that, at that angle, the ball
stops sliding and starts biting.

Players tend to use more spin on slow courts such as clay, in part to get the
ball to kick up at a steeper angle. On grass courts, the ball slides and tends to
stay low after bouncing. Players can take advantage of the low bounce on grass
by hitting the ball with less topspin, or with backspin, or by hitting the ball
faster and lower over the net, so that the ball does not bounce up into the com-
fortable waist high hitting zone.

If you watch a ball spinning, it is impossible to tell how fast it is spinning apart
from “pretty fast” or “pretty slow”. So how do players know if the ball spins
faster or not? The easiest way is to observe how much the ball curves through
the air. The bigger the spin the more the ball curves. If one is comparing, say,
thin strings with thick strings, it might be hard to pick the difference in spin or
the curve if the difference is only small. Are the players talking about 20% extra
spin with thin strings or is it only 2%? The best way to check this out is to
measure the spin, as well as the speed and angle of the racquet head, to make
sure that everything else stays the same when the string gauge is varied. Up
until about 1996 there were very few measurements of a spinning tennis ball.
High speed cameras are now available to capture the spin, but they are expen-
sive. As a result, measurements of a spinning tennis ball can still be counted
on one hand.

The amount of spin off a racquet or off the court depends on a whole bunch
of factors, some of which are important, some are not so important, and some
have never been measured, so it is not actually known for certain if they are
important or not. Recently, measurements of spin off the strings were made by
Simon Goodwill at the University of Sheffield. The ball was fired at the strings
at speeds up to 40 m/s, typical of a high speed groundstroke. These measure-
ments are the first to our knowledge on how the spin varies with string gauge.
The main result is that it doesn’t. It doesn’t vary much with string tension either.
Rob Bower obtained a similar result in 1996 for his Master’s thesis at Sydney
University. Rob fired balls at 45 degrees to the string plane and measured the
rebound spin for three different string tensions. The spin increased by only 5%
when the tension was increased by 50% (from 18 kg to 27 kg).

The basic physics of a ball that bounces obliquely off a surface is described in
Chapters 39 and 40. In general, the speed, spin and angle all change when a
ball bounces, by amounts that depend on whether the ball slides throughout
the bounce or whether it starts to bite during the bounce. We considered only
 Bounce Off A Racquet 371

a bounce off the court in Chapter 39, but a bounce off the strings is essential-
ly the same. The main difference is that the racquet head is pushed away from
the ball if the racquet is initially at rest. Alternatively, the racquet head slows
down if it collides with a ball. Another difference is that strings are relatively
slippery compared with most court surfaces so the angle of incidence at which
a ball stops sliding and starts to bite is typically about 40 degrees on the strings,
compared with about 20 degrees on a court. Most balls are incident at angles
less than 20 degrees on a court so most balls slide throughout the bounce on
the court. By contrast, most balls are incident on the strings at angles greater
than 40 degrees and they bite the strings (or equivalently, the strings bite the
ball).

The difference between sliding and

biting is illustrated in Figure 41.1. A t=1ms t=2ms t=3 ms
ball incident obliquely on a surface
without any spin will initially start to
slide on the surface. As it does so, it 20m/s ismnre 12m/s
slows down in a direction parallel to
the surface and it starts rotating. If it
acquires enough spin, then the bot-
tom of the ball can rotate backwards ee ate ceca
at the same speed that the whole — SLIDES ———-}
— BITES —}>
ball moves forwards, in which case
the bottom of the ball comes to rest Figure 41.1
and it bites the surface. The ball can Sliding and biting. A point on the bottom of the ball slides from A to B and
still advance forwards even though then remains stuck at point B when the ball bites.
some points remain stuck, since
points at the back of the ball lift up
off the surface as the ball rotates
while other points at the front of the ball rotate down onto the surface. This is
shown in Figure 41.1. The point on the ball in contact with point B on the sur-
face remains stuck when the ball bites, while the point on the ball that was
originally stuck on point A lifts up off the surface.

As explained in Chapter 40, biting and rolling are similar for a tennis ball. If a
ball rolls during the bounce, then the rebound spin and speed do not depend
on the coefficient of friction of the surface, but they do depend on the angle
of incidence (see Note 2). In essence, this explains the experimentally
observed fact that the ball spin off the strings does not depend on whether the
strings are thin or thick, rough or smooth or soft or stiff. These differences
would be important if a ball slides along the strings throughout the whole
bounce period, but this would require the ball to be incident on the strings at
an angle less than 40° to the string plane. More commonly, the ball is incident
closer to a right angle to the string plane, in which case the ball will bite before
it bounces off the strings.

When a ball bounces off the court, it acquires even more topspin than it had
when the player at the other end of the court hit the ball. A moderately fast
ball arrives at the receiver spinning between 50 and 100 revolutions/sec (314
to 628 radians/sec), but as far as the receiver is concerned, the ball could be
incident on the strings either with topspin or with backspin, as shown in Figure
41.2. Figure 41.2a shows a situation where the ball is incident on the court with
3/2 Chapter 41

Racquet Racquet
A = OC Racquet

= O=
iia
See
O S a
™e We Court
(c) hitting ball at the top
(a) hitting ball on the rise (b) hitting ball as it drops of its bounce

Figure 41.2
Hitting a ball on the rise or as it drops or at the top of its bounce.

topspin (meaning that the top of the ball is spinning forwards), it acquires even
more topspin when it bounces, and it is incident on the racquet with topspin
(turning the page through 90 degrees will show this more clearly). If the play-
er hits the ball as it is falling downwards, then the ball will be incident on the
strings with backspin. At least, it certainly looks that way in Figure 41.2b. In
fact, it might be the other way around depending on what the racquet is doing.

If the racquet is at rest when the ball hits the strings, then the ball is incident
on the strings with topspin if it is rising, and it is incident with backspin if it is
dropping, exactly as shown in Figure 41.2. But racquets are rarely at rest when
they hit a ball. If the racquet is moving horizontally at 20 m/s when the ball
hits the strings, and if the ball is traveling towards the racquet with a horizon-
tal speed of 20 m/s, then the ball approaches the racquet at a horizontal speed
of 40 m/s relative to the racquet. That will change the angle of incidence on
the strings (since the relative horizontal speed changes, but the relative verti-
cal speed is unaffected), but it won’t change a ball incident with topspin into
one incident with backspin (or vice versa).

Now consider a situation where the racquet is rising upwards at around 30

degrees to the court when the racquet hits the ball. If the ball is also rising
upwards at 30 degrees to the court, and if the speed of the ball is the same as
the speed of the racquet, then the height of the ball above the court will remain
the same as the height of the racquet, so the ball is heading at right angles onto
the strings. An insect sitting on the strings would see the ball coming in at right
angles. In that case, we can’t say that the ball is incident with topspin or back-
spin. It is incident with spin, but it doesn’t make any sense to say it is incident
with topspin or backspin. Think about Figure 41.2a, b and c. In (a) the ball is
incident on the racquet with topspin. In (b) it is incident on the racquet with
backspin. In (c) it is incident on the racquet with spin but it is neither topspin
nor backspin.

If a spinning ball is incident at right angles to the strings, then friction between
the ball and the strings will act to reduce the rate at which the ball spins. In
most cases of interest, there will be no reversal in the spin direction, unless the
ball happens to be a superball. If you happen to have a superball, then try it.
A superball that is dropped vertically onto a hard floor with spin will bounce
with a small amount of spin in the opposite direction. Suppose the racquet and
the ball are both traveling horizontally towards each other in Figure 41.2c. The
ball is incident from left to right with topspin with respect to the court surface,
but it is neither topspin nor backspin with respect to the strings. The ball will
 Bounce Off A Racquet 373

still spin in the same direction after it is hit, but :

it then travels right to left so it has backspin Bounce Off A Moving Racquet
with respect to the court. In order to hit a ball
so that it has topspin after it bounces off the
strings, the direction of the spin has to be =~ Racquet head
reversed when the racquet hits the ball, as indi- =-0
cated in Figure 41.3. That requires a bit of effort
on the part of the player, and it is necessary
that the racquet rise upwards faster than the G
ball. Even greater topspin can be imparted to
the ball if the ball is falling rather than rising Bounce off court
when it is hit (assuming that the ball is hit with
the same speed), since the ball will be incident ;
on the strings at a smaller angle to the strings. BT ed
Hitting a rising ball so that it rebounds with topspin.
If the ball is rising at 20 degrees and the rac-
quet is rising at 30 degrees, and if the ball and
the racquet have the same speed, then the racquet is rising up faster than the
ball as if the ball was falling downwards to meet the racquet. In that case the
ball would be incident on the strings with backspin, as in Figure 41.2b. But
what if the racquet is travelling at a different speed to the ball? That is harder
to think through logically but the situation can be analyzed with a bit of sim-
ple mathematics as shown at the end of this Chapter.

The performance of the strings in a racquet can be measured by firing a ball

at the strings and measuring the rebound speed and angle of the ball
(Knudson, 1997; Bower and Sinclair, 2000). Manufacturers do this routinely, but
in most cases, the ball is fired perpendicular to the string plane in order to
measure the performance of the racquet. Some manufacturers fire balls at an
oblique angle in order to measure the durability of the strings when subject to
repeated impacts. It is only very recently that the spin has been measured using
a high speed video camera. This was done in 2001 by Simon Goodwill at the
University of Sheffield as part of his PhD thesis, by firing balls at a racquet at
speeds between 22 m/s and 39 m/s and an angle of incidence of 50 degrees
to the string plane — i.e., 40 degrees to the normal (Goodwill and Haake, 2002).
The angle of incidence was not varied, but the amount of spin on the incident
ball was varied from -600 rad/s (the minus sign meaning backspin) to +100
rad/sec.

The racquet head was firmly clamped to stop it flying away every bounce, but
this does not detract from the relevance of the experiment. It simplifies the
experiment. It shows directly the difference between the strings and a court,
and it is relatively easy to calculate from these results the effects of NOT clamp-
ing the head. This is a similar situation to that described in Chapter 14 where
the COR can be measured by clamping the racquet head and bouncing a ball
at right angles off the strings. The experiment could also have been done by
clamping the handle instead of the head, which would be one step closer to
the real thing, but the results are still very interesting. Some of these results are
shown in Figure 41.4, and the main conclusions are:

(a) the rebound spin of the ball does not depend significantly on string ten-
sion, string type, string gauge, or spacing between the strings;
374 Chapter 41

(b) the rebound speed of the ball increases slightly as

300 the tension is lowered, and

250 +
(c) the rebound angle depends on string tension, string
type and the spacing between strings. The ball
® 200 -
72) rebounds at an angle closer to the normal as the string
ee (SOl > meee tension is reduced.
ras Gut 40lb
S 10) | All three parameters depend on the spin, ,, of the inci-
50 L 29m/s data (a) | dent ball as shown in Figure 41.4. Even though the ball
was incident with backspin, it bounced off the strings
@) i i | | ! | with topspin in all cases. As one might expect, the spin
-600 -500 -400 -300 -200 -100 0 100 of the ball is reduced when the incident ball has a lot
of backspin, but the effect is not very strong. For exam-
o, (rad/s)
ple, if w, is decreased from 0 to -600 rad/s, then @,
decreases from about 250 rad/s to about 150 rad/s.
Brody’s (1984) bounce model indicates that if w, is
{
decreased by 600 rad/s, then w, will also decrease by
600 rad/s if the ball slides throughout the bounce. If the
0.8 | Gut 40Ib ene ball starts rolling during the bounce, then Brody’s
model indicates that w, should decrease by 213 rad/s.
aa Gut 70lb >
=a a Syn’ 70Ib (same as The actual decrease in , from 250 to 150 rad/s can be
3 [ Syn 70 Ib) explained by the fact that the ball bites the strings.
SS OL) IE

oan V = 29 m/s (b) i Clues to the behavior of the strings can be found in
measurements of the v,, /v,, ratio, the COR, and the

(0) [ | | | ! !
ratio S = Rw, /v,,, as shown in Figure 41.5. On a tennis
-600 -500 -400 -300 -200 -100 O 100 court, v,,/V,, is typically about 0.4 to 0.5 at high angles
,(rad/s) of incidence, as described in Chapter 39. In Goodwill’s
experiment, the ball was incident at an angle of 50° to
the string plane, and the v,,/v,, ratio was about 0.42
on all strings tested when the ball was incident without
spin. This ratio dropped to much lower values when
Vie = 29 Ms the ball was incident with backspin, since it took longer
n
for the ball to stop sliding before it started to bite. The
results in Figure 41.5c show that the ball gripped all
@ Syn 70 lb foe strings during the bounce, since Rm, /v,, was larger
> Oe Gut 40 Ib
than 1.0 in all cases. When the ball first hit the strings,
S ‘ oS
it started sliding on the strings, but the friction force
= Syn 40 Ib caused it to slow down, reverse its direction of spin and
G ut 70 Ib (c) then bounce with topspin.

The ball slowed down most on natural gut at low ten-

-600 -500 -400 -300 -200 -100 0 100 sion and least on the nylon strings at high tension, as if
®, (rad/s) the ball dug deeper into the softer strings than the
stiffer strings. Alternatively, the softer strings might
Figure 41.4 stretch further in a direction parallel to the string plane,
acting in a manner similar to the spaghetti strings
Rebound spin, speed and angle for a ball incident at 29
m/s and at 40 degrees to the normal vs incident ball spin. described in Chapter 39. The latter effect appears to be
The racquet head was clamped to prevent it rotating. the most significant one. Measurements of ball spin and
Results are shown for natural gut and nylon (Syn) strung speed indicate that the normal reaction force acts ahead
at 40 Ib or 70 Ib. 8, is the rebound angle to the normal. of the center of the ball by a distance of about 3 or 4
mm on each of the strings tested and at all values of the
 Bounce Off A Racquet 375

incident spin. This is similar to the behavior on a court

(a)
surface. Even if the ball does dig into the strings, it does
not shift the normal reaction force any more than it does Gut = thick lines
on a court. On the other hand, the results in Figure 41.5¢ Syn = thin lines
show that Ra, /v,, is largest for gut at low tension, partic-
ularly when the ball is incident with a lot of backspin.
Si as Syn 70 Ib
Values of Rw, / v,, greater than about 1.4 are normally Sen
observed only with a superball or with spaghetti strings.
ays

Superballs spin much faster than other balls when they

bounce, and so do tennis balls when they bounce on
spaghetti strings. In both of these cases, there is a highly
elastic deformation in a direction parallel to the surface, -600 -400 -200 0
resulting in extra spin and a larger than usual slowing ®, (radians/sec)
down of the ball in a direction parallel to the surface.
Figure 41.4a shows that there is no significant effect on
spin, but Figure 41.5a shows that there is a big effect in
slowing the ball in a direction parallel to the surface. Part
of the reason for the low values of v,, /v,, with large 1

backspin is that the ball takes longer to stop sliding so the

friction force acts for a longer time, but the ball is espe- 0.95 _ Gut 40 Ib
cially slow at low string tensions. This indicates that loose Syn 40 Ib
strings move further in the string plane and take more
ge Syn 70 Ib
sideways energy out of the ball, but the strings do not
spring back, so there is not the same increase in ball spin
© 0.85 Gut 70 Ib
as there is with spaghetti strings. In other words, the large
values of Rw, /v,, shown in Fig. 41.5c are not due to 0.8 Gut = thick lines
excessive spin, but to a big slowing down of the ball on Syn = thin lines
the looser strings. 0.75
-600 -400 -200 0) 200 40C
An additional effect on soft strings is that the COR is ®, (radians/sec)
enhanced at all values of the incident spin (when com-
pared with stiff strings). The combined effect of the
increase in the COR and the decrease in v,, /v,, on soft
strings is that the ball bounces off the strings at a steeper
angle (closer to a line perpendicular to the string plane).
If you examine the diagrams below, you will see that this 3.5
will have the effect of sending the ball higher over the 3 Gut = thick lines
net. When players break a string during a groundstroke, Syn = thin lines
the ball usually flies out rather than landing in the net. woe
However, a more serious problem would arise if string >
movement at low tension was more variable than string on 2

movement at high tension, as one might expect. In that a

case the player would lose some control of the ball, espe-
cially if the player spins the racquet around between shots
and hits from both sides of the racquet. Some players 1 Syn 70 Ib

(especially Lleyton Hewitt) like to push the strings back

0.5
into position between each point. It seems to work for -600 -500 -400 -300 -200 -100 O 100
him! , (radians/sec)
In order to determine the behavior of a ball bouncing off
a hand-held racquet, the data shown in Figure 41.4 needs Figure 41.5
to be incorporated into a model that takes into account
376 Chapter 41

the speed and direction of the racquet, as well as the speed and direction of
the ball (Cross, 2000).

Figure 41.6 Suppose that the ball has topspin off the court and is rising upwards
as in Figure 41.6a. Suppose also that the racquet head is tilted
Method used to calculate the speed, spin
fowards and is rising upwards to meet the ball. The speed and
and angle of a ball hit by a racquet.
direction of the ball and racquet are represented by the length and
direction of the two arrows. It is easiest to think about this situation
if we first imagine that the racquet is at rest. To do that, we need to
change to the racquet frame of reference and then change back to
the court frame of reference, as follows:

If we subtract V, from the speed of both the racquet and the ball,
then the racquet is at rest, and the ball approaches the racquet at
speed v, and with backspin, as shown in Figure 41.4b. It bounces
off with speed v, and with topspin (if the ball is hit hard enough).
Thee vs arrow is obtained by drawing the parallelogram (dashed
lines), and v, needs to be calculated or measured separately. The
racquet will also change speed when it is struck by the ball.

Now add V, back to the speed of the ball (by drawing the paral-
lelogram) as shown in Figure 41.6c. Then the ball bounces off with
speed v,,,, and with topspin.

Note that changing the frame of reference has no effect on the spin
of the ball. It just changes the speed and angle of incidence or
reflection of the ball onto or off the racquet. This is an important
effect that is sometimes overlooked. If a ball is fired at a stationary
racquet to measure the angles of incidence and reflection, then
those angles will be completely different when the racquet is actu-
ally swung towards the ball.

The situation shown in Figure 41.6 is typical of a topspin forehand.

A ball hit on the rise is usually incident with backspin on the strings.
It is also incident with backspin if the ball is hit at the top of its
bounce or if it is hit as it falls towards the court, but the angle of
incidence changes in each case. This is shown in the following
example where the ball is traveling horizontally before it hits the
strings.

The data in Figure 41.7a are typical of a topspin forehand where the racquet
head is vertical and rising upwards at 30° to the meet the ball at the top of its
trajectory. From the geometry (see Figure 41.7d), we find that the ball
approaches the racquet at speed v, = 40 m/s with backspin and at an angle of
72° to the string plane.

Calculations show that the ball bounces at an angle of 72°, with speed Pinel te:
m/s and with topspin at 15 rev/sec, as shown in Figure 41,7b,
 Bounce Off A Racquet 377

Figure 41.7
Va Vin = 16.5 m/s va -248m/s E xample where racquet rises
i at t 30 30° to
meet the ball.
50 rev/s

(b) Vo =18m/s

(d) Incident ball geometry

sin(30) = 12.4/24.8 = 0.5 cos (30) = 21.5/24.8 = 0.866
165
tan(0) = 12.4/(21.5 + 16.5) =0.326 so @= 190
124 V,2 = 12.474 (21.5 + 16.5)= 1600
165 So. v,=40 and sin(18) = 12.4/40 = 0.31

Adding V, back to the speed of the ball (Figure 41.7c) shows that the ball
bounces with speed v._,out = 39 m/s, at 10° to the horizontal and with topspin at
15 rev/sec. Hit from the baseline, the ball would actually land beyond the
opposite baseline. A solution is not to hit so hard, or tilt the head forwards a
fraction.

We have seen how a groundstroke can be hit with topspin, but how about
serving with topspin? Does the same method work here? In order to hit a
groundstroke with topspin, the racquet head needs to rise upwards to meet the
ball. That way, the strings brush upwards against the back of the ball. The fric-
tion force between the ball and the strings then acts vertically upwards and
causes the ball to rotate with topspin. The ball is ejected forwards and upwards
out of the strings in order to pass over the net. That’s because the racquet head
is moving forwards and upwards when it contacts the ball, and hence the force
on the ball is forwards and upwards.

A serve is different since it is hit from a point well above the player's head.
That presents two problems. First, how can you get the racquet head to rise
378 Chapter 41

upwards as it contacts the ball when the racquet head is almost as high as you
can reach? Even if you can, how do you get the ball to project downwards so
it lands in the service court rather than fly upwards and over the back fence?
I have not seen any high speed film showing what happens when a ball is
served with topspin, but it is likely that (a) the racquet first makes contact with
the ball while the racquet head is moving upwards, and (b) the ball leaves the
strings 5 ms later when the racquet head is moving almost horizontally. In addi-
tion, a high ball toss causes the ball to move downwards along the strings dur-
ing the serve. That is equivalent in effect to the racquet head moving upwards
along the ball even if the racquet head is actually moving horizontally.

In a moderately high speed serve, the racquet rotates at about 80 rad/s just
before impact and slows to about 27 rad/s just after impact, as shown in
Chapter 20. At this speed, the racquet rotates 23 degrees in the 5 ms period just
before impact and 8 degrees in the 5 ms period just after impact, so it must
rotate by about 15 degrees during the impact itself. That is a large angle, given
that the ball must be projected downwards to pass through a narrow window
only 2 or 3 degrees wide in order to clear the net and land in the service box.
When hitting a groundstroke with topspin, the racquet face can remain verti-
cal immediately before, during and immediately after impact. That is not the
case when serving the ball.

The position of the racquet face during a serve is shown in Figure 41.8. The
racquet face is rising upwards at the instant of contact, it is moving forwards,
and it may also be moving sideways, i.e., moving up and across the ball from
say the 8 o'clock point on the ball towards the 2 o’clock point. Sideways
motion will cause the ball to spin about a vertical axis. As a result, the ball will
swerve sideways like a hook or a slice shot in golf. The situation just before
impact is shown in Figure 41.8a. The ball is falling downwards at speed v,,, and
the racquet is rotating in an arc at speed v,. Subtracting v, from both the rac-
quet and the ball (Figure 41.8b) shows the situation in a reference frame where
the racquet is rotating about an axis that remains at rest. In that frame, the ball
is incident at speed v,, and it bounces off the strings at speed v,. At least, that’s
Figure 41.8 what would happen if the racquet
Topspin serve
was originally traveling in a straight
line. In fact, the racquet rotates by
about 15 degrees during the impact.

If we rotate the racquet and the v,

arrow by 15 degrees, then we have
V the situation shown in Figure 41.8c.
2 This is not a strictly valid procedure
but it is roughly correct. The force
(a) Racquet in motion before impact (6b) Racquet at rest I EN ELE SSS I
perpendicular to the string plane
throughout the impact, meaning that
it varies in direction by up to 15
degrees during the impact. It varies
Vout by the full 15 degrees only at the very
WS = end of the impact. The effect on the
V2 Vo ball is that its trajectory will change
by something less than 15 degrees,
(d) Racquet in motion after impact and probably by about half of that.
(c) Racquet rotated 15°
 Bounce Off A Racquet 379

If we assume that the situation in Figure 41.8c is roughly correct, then the rota-
tion axis is still at rest but the racquet has rotated through 15 degrees. To
determine the situation from the player’s reference frame we need to add v, to
the speeds in Figure 41.8c, as shown in Figure 41.8d. Note that we have also
rotated the v, arrow through 15 degrees since the player sees that the racquet
is travelling horizontally at the instant the ball leaves the strings.

The result is that the ball is served with topspin and it is projected downwards
at a small angle to the horizontal. From the player’s point of view, the racquet
is traveling upwards (and forwards) immediately before impact, and _ this
imparts topspin. Since the ball is dropping vertically just before and at impact,
this acts to deflect the ball downwards. During the impact the racquet swings
around to a more horizontal path, but since the ball is traveling downwards
along the string plane, the racquet is still rising upwards with respect to the
ball. The friction force on the ball therefore imparts topspin throughout the
impact, even if the racquet is travelling horizontally near the end of the impact.

A rough estimate of the amount of topspin imparted to the ball is as follows.

Suppose that v,, = 3.7 m/s and v, = 35 m/s. Then the ball is incident in Figure
41.8b at v, = 35.2 m/s and at an angle of 84 degrees to the string plane. If the
ACOR is 0.4 then the ball will bounce in Figure 41.8b at v, = 14 m/s and at
about 84 degrees to the string plane. The bounce speed in a direction parallel
to the string plane is v,, = 14 x cos(84) = 1.46 m/s. Assuming that the ball rolls,
it will spin at v,, = Rw and hence w = v,,/R = 44 rad/s = 7 rev/s. This is much
smaller than the subsequent spin when the ball bounces off the court.
However it is enough to alter the trajectory of the ball significantly, as shown
in Figures 30.14 and 42.3. High speed video film of a topspin serve would be
needed to determine if this rough estimate is reliable.

Greater topspin can be imparted to the ball during a serve if the racquet head
first makes contact with the ball at an earlier stage during the swing. The head
will then be rising faster at impact, and the ball will be incident on the strings
(in the racquet frame of reference) at an angle that is not as close to the nor-
mal as shown in Figure 41.8b. The player might perceive that he or she is then
hitting the ball at a significant angle upwards from the horizontal, but the ball
will be (and needs to be) ejected from the strings at a small angle downwards
from the horizontal if it is served at moderately high speed.

References

R. Bower & P. Sinclair (1999). Tennis racquet stiffness and string tension effects
on rebound velocity and angle for and oblique impact. Journal of Human
Movement Studies. 37, 271-286.

H. Brody (1984) That’s how the ball bounces, The Physics Teacher, Nov 1984,
494-497.

R. Cross (2000) Effects of friction between the ball and strings in tennis, Sports
Engineering, 3, 85-98.

S.R. Goodwill and SJ. Haake, (2002) Why were spaghetti string racquets
banned in the game of tennis? 4th International Conference on the Engineering
of Sport, Kyoto, 2002.
380 Chapter 41

D. Knudson (1997), The effect of string tension on rebound accuracy in tennis

impacts, International Sports Journal, 1, 108-112.

Notes

1. It is easy to make a ball spin by dropping the ball vertically onto the strings
while the string plane is horizontal, and by “chopping” the bottom of the ball
with a sharp, horizontal swing. If the string plane is inclined at a few degrees
to the horizontal, the ball can be hit vertically upwards so it spins rapidly, and
one can repeat this several times so the ball spins faster and faster after each
chop. It might appear that the ball is incident vertically on the strings (since the
ball is dropping vertically and the string plane is moving horizontally), but an
insect sitting on the strings would see the ball coming in at an angle of only
about 40 degrees to the string plane, depending on the speed of the racquet.
We say that “in the racquet frame of reference” where the racquet is at rest, the
ball is incident at about 40 degrees and it bounces with topspin. A player serv-
ing a ball with heavy spin uses a similar chopping or brushing action, but in
the racquet frame of reference, the ball is incident typically at around 50
degrees. The object here is not to get maximum possible spin but to get as
much spin as possible consistent with a reasonably high serve speed. One has
to compromise by trading off spin against speed. Slice, kick and topspin serves
all involve spin where the ball rotation axis varies from about 45 to about zero
degrees to the horizontal and where the server brushes the strings upwards and
across the back of the ball in the direction that the server wants the ball to spin.

2. The simplest way to describe a bouncing ball is to assume that it is rigid and
that the surface on which it bounces is also rigid. The ball will start to roll dur-
ing the bounce at high enough angles of incidence. The horizontal speed and
the spin after the bounce are then related by v,, = Rw,, where R is the radius
of the ball and where v,, /v,, is 0.645 regardless of the coefficient of friction
and the angle of incidence and the initial ball spin. But for a given incident ball
speed, v,, decreases as the angle of incidence increases so V,. and , also
decrease, all the way to zero when the ball is incident at right angles to the
surface. A similar situation arises if the ball or the surface is flexible but the ball
bites the surface instead of rolling.
 Factors Influencing
the Flight of the Ball

Nathalie Tauziat, France

By Rod Cross

good tennis player knows instinctively how hard to hit a ball and at what angle to get the ball over the
. net and into the court so that the ball lands at just the right spot. This comes from years of practice.
™ Players might be surprised to know just what they are doing in terms of the actual ball trajectories. For
example suppose you hit a forehand at the baseline so that the ball lands on your opponent's baseline 78 feet
away. If you hit the same ball but 1% faster it will land 18 inches beyond the baseline. If you hit the ball one
degree higher it will land about 6 feet beyond the baseline, depending on the initial speed and angle. If you hit
the ball one degree further to the left it will land 16 inches further to the left.

When serving a ball at moderate to high speed, the ball must be served a few degrees down from the horizon-
tal. Too far down and the ball will hit the net. Not enough and the ball will be long. The range of possible angles
is only about two degrees, and it gets smaller if the ball is served faster or by a shorter player (Brody, 1987).
The range of angles increases if the ball is served with topspin. Typical ball trajectories are shown below, togeth-
er with an explanation of the trajectories.
382 Chapter 42

100 inch
(254 cm) drop
300 (a) Dropping a Ball (Free Fall)

Suppose that a ball is dropped from a height of 1.0 m (about

time speed waist height) and is allowed to fall to the ground. Provided
250 / 0 .00 Ss ‘ 0) that it is not hit or thrown then it starts out with zero speed.
It will accelerate downwards due to the force of gravity. If we
I 0.20s @® 2.0m/s ignore air resistance, then the vertical distance (y) travelled in
I @ time t is given by y = gt’/2 where g = 9.8 m/s’ is the acceler-
030s @ 2.9m/s ation due to gravity. After = 0.1 s, y = 0.049 m and after 0.2 s,
200 i . y = 0.196 m. The ball will hit the ground at t = 0.452 s. On
the way down, the speed increases and is given by v = gt. At
t=0,v=0. Att = 0.1, v = 0.98 m/s. Just before it hits the
| 0.40s @ 3.9m/s
y(em) Bae ground, at t = 0.452 s, v= 4.43 m/s. The corresponding results
for a ball dropped from a height of 100 inches (2.54 m) is
@
150) |= : shown in Figure 42.1. Air resistance is very small in this case
since the ball speed is small. If allowance is made for air
050s @ 49m/s
resistance, a ball dropped from a height of 100 in would hit
the ground at 0.725 s instead of 0.720 s.

r 060s @ 5.9 m/s

| (b) Horizontal hit with free fall

Now suppose that the ball is hit at 30 m/s (67 mph) from a
height of 1.0 m so that it starts moving in a horizontal direc-
tion, parallel to the ground, as shown in Figure 42.2. It is easy
to calculate where the ball will land provided we ignore the
force on the ball due to air resistance. Then the only force act-
ing on the ball is the gravitational force pulling the ball verti-
cally towards the ground. The ball will accelerate downwards
and change its vertical speed, but there is no change in the
Figure 42.1 horizontal speed since there is no force in the horizontal
direction. In other words, the ball will keep moving horizon-
tally at 30 m/s until it hits the ground. It starts out with zero
speed in the vertical direction, and moves down-
wards a vertical distance y = gt”/ 2, as in our first
example. It therefore takes 0.452 s to hit the
ground. During that time, it travels a horizontal
distance x = 30 x 0.452 = 13.56 m.

In this case air resistance does make a difference

since the ball is moving fairly fast and since the
air resistance increases as the ball speed increas-
es (Table 42.1).

If the ball has no spin then it will actually hit the

no spin ae‘ court after 0.470 s after traveling a horizontal dis-
tance of 12.46 m. The average horizontal speed is
12.46/0.47 = 26.5 m/s. The horizontal speed
drops slightly as the ball moves through the air,
and so does the final vertical speed since air
Figure 42.2 resistance exerts a force backwards and upwards
 Ball Trajectories 383

along the path of the ball. If the ball

has topspin, there is an additional | Table 42.1. Trajectories With Same Launch Speed and
force that acts at right angles to the | Angle Under Various Conditions
path of the ball, in a downwards and
backwards direction in this case. This | Conditions Velocity Time to Bounce __ Travel Distance
force is called the Magnus force. For No air 30 m/s 0.452 secs 13.56 m
example, if the ball is spinning at 20 | Air no spin 30 m/s 0.470 secs 12.46 m
revolutions/sec, the ball will hit the | jr w/ topspin (20rev/s) 30 m/s 0.410 secs 10.85 m
court after 0.410 s after traveling a
horizontal distance of 10.85 m.

The distance from the baseline to the net is 39 feet (11.887 m) and the height
of the net in the middle is exactly 3 feet (0.9144 m). A ball hit horizontally at
30 m/s will not clear the net even if it starts at a point 1.0 m off the ground.
To get the ball over the net it has to be hit upwards at a certain angle to the
horizontal. It can't be hit too steeply or it will land past the baseline at the other
end of the court. There is a certain range of angles that will get the ball over
the net and inside the baseline. The range of angles is relatively small.

(c) Hitting at Various Launch Angles

Figure 42.3 shows the trajectories of

a ball hit from a height of 1.0 m, at 2.5
30 m/s and at various angles to the weornsne a ;
horizontal, including the effects of air 2.0 — no spin
resistance. In order to just clear the a 0 eso 20 rev/s
net, a ball hit without spin must be
hit upwards at 4.0° to the horizontal.
If the ball has topspin at 20 revolu- _ baseline
tions/sec, it must be hit upwards at ‘at 23.77m
5.5° to the horizontal. If a ball is hit
without spin at 8.1° to the horizontal,
it lands on the far baseline. If the ball
has topspin at 20 revolutions/sec, it
must be hit upwards at 11.9° to land
on the baseline. The range of avail-
able angles is 4.1° without spin or
6.4° with topspin. Figure 42.3

(a) Drag force pushes ball backwards and slows it down.

A ball traveling through the air experiences a backwards force due to the fact
that the air pressure on the front of the ball is larger than the force at the back
of the ball. This force is called the drag force. The force is large because air is
heavy. Not as heavy as water, but a room full of air contains about 90 kg or
200 lb of air. When a tennis ball is at rest, air pressure exerts a force of 150 lb
on the front of the ball and 150 Ib on the back of the ball. The ball doesn't col-
lapse since the air inside the ball exerts a similar force outwards. When a ball
is traveling through the air, the force on the rear side of the ball is typically
about 0.2 lb lower than the force on the front side, depending on the ball
speed.
384 Chapter 42

(b) Formula for drag force.

The flow of air around a tennis ball is turbulent at all ball speeds of interest in
tennis, due to the relatively rough surface of the ball. This actually simplifies
the analysis of ball motion since the drag coefficient remains constant for all
speeds of interest. Other balls used in sport are generally smoother, but this
complicates the analysis since the drag coefficient usually decreases suddenly
at a sufficiently high ball speed (an effect known as the drag crisis). The drag
force is proportional to the ball speed squared and is given by the formula

F=C,Adv/2 (42.1)

where d = 1.21 kg/m? is the density of air, A = mR? is the cross-sectional area
of the ball, R is the radius of the ball, v is the ball speed and C, is the drag
coefficient. The only factor here that can be influenced by the player is the ball
velocity. If the ball was a flat, circular disk, the drag coefficient would be 1.0.
But a ball has a rounded nose which streamlines the air flow and reduces the
drag. For a relatively new tennis ball, C, is about 0.55. A used ball experiences
a slightly lower drag force since it has a smoother surface. A standard new ball
of radius 3.3 cm therefore experiences a backwards force

F = 0.55 x 0.00342 x 1.21 x v’/2 = 0.00114 x v’ Newton (N).

For example, if v = 10 m/s then F = 0.114 N. If v = 20 m/s then F = 0.450 N.

If v = 30 m/s then F = 1.026 N. We can compare this with the force of gravity
F = mg acting downwards on the ball. Since g = 9.8 m/s’ and m = 0.057 kg for
a standard tennis ball, mg = 0.057 x 9.8 = 0.559 N. The drag force is as large
as the gravitational force even at moderate ball speeds. At a very fast serve
speed of 60 m/s, the drag force is 7.3 times bigger than the gravitational force.
As a result, a ball served at high speed slows down rapidly through the air and
lands on the court at about 75% of the serve speed. In fact, a ball served at any
speed lands on the court at about 75% of the serve speed. A ball served at a
low speed experiences a small drag force, but it spends a relatively long time
slowing down.

(a) Force due to spin

An additional force arises if the ball is spinning. A spinning ball sets the air
around it in motion in a thin layer near the surface of the ball. The flow of air
around the ball is altered in such a way that the air pressure on top of a spin-
ning ball is decreased if the ball has topspin and is increased if the ball has
backspin. As a result, a ball that travels horizontally experiences a force down-
wards if it has topspin or upwards if it has backspin. If the ball spins about a
vertical axis then the force causes the ball to swerve sideways. The force due
to spin is called called the Magnus force and it always acts at right angles to
the drag force and to the spin axis. (see Figure 42.4). Consequently, if a ball
with topspin is rising upwards at an angle to the court, the Magnus force tends
to push the ball down onto the court and it pushes it forwards in a direction
paraliel to the surface. If a ball with topspin is falling towards the court sur-
face, the Magnus force pushes the ball downwards and backwards. A ball with
topspin therefore falls onto the court at a steeper angle than a ball without top-
 Ball Trajectories 385

spin. Conversely, a ball with backspin

Fq
tends to float through the air and falls to la (Gon
the court at a shallow angle.

(b) Formula for the Magnus force.

FL Fis
The magnitude of the Magnus force on a mg mg
tennis ball is given by

F=C, Adv/2 (42.2)

where C, is called the lift coefficient since

the ball experiences a vertical lift force if FL Fq Pr
it has backspin and is given by

CL = 12+ (Wein) (42.3) If

where Ve Sin = Ro is the peripheral speed

of the ball, R is its radius and w is the mg mg
angular speed about a horizontal axis per-
pendicular to the path of the ball. The
Figure 42.4
angular speed varies typically from about 100 to about 500 radians/sec which
translates to a spin between 16 and 80 revolutions per second (1 rev/sec = 6.28
radians/sec). Suppose that v = 30 m/s, w = 300 rad/s and R = 0.033 m. Then
a A 9.9 m/s, V/V oe = 3.03 and C, = 0.20. The lift coefficient in this case is
smaller than the drag coefficient, but not a lot smaller. The lift coefficient never
gets to be as large as the drag coefficient. At most, C, = 0.5, which is the limit
when v,,,,
pi
is much larger than v.

Ball spin also affects the drag coefficient

(Stepanek, 1985). A good fit to the experi-
mental data is given by

Cy = 0.55 + 1/22.5 + 4.2(VIV. spini.) 1

which indicates that C, increases from 0.55 to no spin
0.84 as v spin increases from zero to a value AQ TEV/Sia-==--=
much larger than v.

Armed with the relevant values of the lift and

drag coefficients, one can calculate the tra-
jectory of a tennis ball through the air for any
initial set of conditions. The equations
describing the trajectory are given in the
Appendix to this chapter. Some typical serve
trajectories are shown in Figure 42.5 for a ball
served at 110 mph (177 kph) from a height of 20
2.8 m (9 feet 2 inches). A person of height h
serves the ball from a height typically about
1.5h. For example, a person of height 6 feet Figure 42.5
386 Chapter 42

usually serves the ball from a point about 9 feet above the court, even if he or
she tosses it a lot higher. A high ball toss allows the server enough time to sight
the ball and judge its position accurately. A high ball toss also allows the serv-
er to maintain good balance while the tossing arm is lifted upwards and the
serving arm is moved backwards. If the toss is too high, the ball might be
thrown too far forwards or backwards.

The trajectories in Figure 42.5 were calculated for a serve down the center line
so that the ball just cleared the net or so that it landed on the service line (point
B in Figure 42.6). Results are given for a perfectly flat serve with no spin and
for a serve with topspin at 40 rev/sec. One of the advantages of serving with
topspin is that the available range of serve angles is increased (from 1.4° to 2.5°
in this case). The ball also approaches the court at a steeper angle and kicks
up at a higher angle. The results in Figure 42.5 are typical of a first serve. If the
serve speed is reduced to 85 mph and the spin increased to say 50 rev/s for a
second serve, the range of available angles increases to 3.8°.

A surprising result is that the range of available angles for a serve is almost the
same for a serve in the far corner of the service court (point D in Figure 42.6).
Even though the net is 3.2 inches (81 mm) higher along that path, the distance
from the server to the far corner is 18 inches (457 mm) longer. As a result, the
range of available angles is reduced by only about 0.2°. It is therefore almost
as easy to serve into the far corner
as to serve down the center line.
The, hard part. is’ tomserve even
wider than the far corner. There is,
however, one advantage of serving
towards the far corner. If you aim
for the center line and miss in the
horizontal direction, then you
have only a 50-50 chance that the
ball will be in. If you aim towards
the far corner and miss in the hor-
izontal direction then the ball can
still land in, provided you serve
the ball short.

To compensate for possible hori-

zontal direction errors, you should
never aim exactly down the center
Figure 42.6 line, but a bit towards the receiver.
Just how far is something that you
need to determine for yourself. Try serving down the line a few times and see
how far you miss (on average). That's how far you should aim away from the
line.
 Ball Trajectories 387

ix |:Trajectory Eq
Suppose that a tennis ball of mass m and radius R is traveling at speed v and
at an angle 8 upwards from the horizontal. The velocity in the horizontal direc-
tion is then v, = v cos(@) and the velocity in the vertical direction is v, = v
sin(9). The forces acting on the ball are mg downwards, F, = Cy Adv’/2 back-
wards and F, = C, Adv*/2 upwards and at right angles to the path of the ball
(assuming the ball has backspin). The equations of motion are then

mdv,/dt = -F cos (6) - F, sin (6)

mdv,/dt = F, cos(9) - F, sin(0) - mg

which can written as

dv,/dt = -kv (C, v, +C, v,) (43.a1)

dv /dt = kv (C, v, -C, vy)-g (43.a2)

where k = d.a.R’/(2m). If the ball has topspin, the sign in front of C, must be
changed in each of these equations. These equations can be solved numeri-
cally, but care is needed to avoid numerical errors. A good check is to solve
for a purely vertical drop without spin. In that case the equation of motion is

dv/dt = -kC,v?-g (42.a3)

which can be solved analytically to check the numerical result. For example, if
m = 57 gm, g = 9.8, d = 1.21 kg/m’, diameter = 2R = 65 mm, C, = 0.50, initial
speed = 15 m/s downwards starting at y = 30 m, then the ball hits the ground
at y = 0 at speed v = 21.006 m/s after 1.6199 sec.

The essential features regarding the horizontal motion of a tennis ball can be
described analytically if one ignores the small vertical component of the ball
speed. Consider a case where a ball is traveling horizontally at speed v with-
out spin, and where the equation of motion in the horizontal direction has the
form

dv/dt =-kC, v’ (42.a4)

The ball will subsequently develop a velocity component in the vertical direc-
tion due to the gravitational force, but this component is typically much small-
er than the horizontal component and can be neglected as a first approxima-
tion. Since the drag force increases with the velocity squared, one might expect
that the drag force on a ball would have a significantly greater effect at higher
ball speeds. However, the kinetic energy of the ball also increases with the
velocity squared. As a result, the percentage change in ball speed as a result
of the drag force, over a given distance, does not depend on the ball speed.
Equation (42.a4) can be integrated directly to show that
388 Chapter 42

viv, = 1(1+k Cys) (42.45)

where v. Oo is the initial speed and s = v,t is the distance the ball would travel at
speed v, in a time t. For a standard 66 mm diameter, 57 gm tennis ball, k C, =
0.020 m’' when C, = 0.55. Over the distance s = 17.888 m from a point 40 cm
in front of the baseline to the opposite service court line, the ball speed drops
to v = 0.737v,, regardless of the initial speed. For example, if v, = 160 kph, then
v = 118 kph, which agrees well with the numerical solution (113.6 kph) . The
numerical solution takes into account both the horizontal and vertical motion of
the ball.

Equation (42.a5) can also be integrated directly to show that the time taken to
travel a horizontal distance x is given by

t = [exp(k C, x) - 1/(k C, v,) (42.a6)

For a 66 mm diameter tennis ball, the time taken to travel from a point 40 cm
in front of the baseline to the opposite service line is therefore t = 21.5/v,. If v,
= 44.44 ms (160 kph), then t = 0.484 s, which also agrees closely with the
numerical solution (t = 0.487 s).

Suppose you hit a forehand or launch a ball from the middle of the baseline so
it passes over the middle of the net and lands on the service line tee. The ball
speed off the racquet or out of the ball launcher can be determined with the aid
of Figure 42.7. This shows the height at which the ball passes over the top of
the net when it is hit without spin from a height of either 0.7 m or 1.0 m above
the baseline. All you need to do is measure the height above the net and then
read the ball speed off the graph. For example, if the ball passes 2.0 m above
the top of the net, it started out at 16.2 m/s, and it lands on the service line at
an angle of incidence of 36°. If it passes 0.2 m over the net, it started out at
about 26 m/s and it lands on the service line at an angle of incidence of about
15%,

Ball launched at baseline Ae Ball launched at baseline

OD
G ball lands on service line ® ball lands on service line
Oo
[S
8 at angle of incidence
{od} Ss shown here
> ®
(e)
fe}
© launched 2
i= \ ath=1.0m = launched
oD)
‘oO launched “*S\ s launched > — Ne ieealD
iS ath=0.7m — 9 ath
= 0.7m
Sia @
its) 20 25 30 125 20 PAS) 30 $5}
v__ (m/s) v, (m/s)

Figure 42.7
Figure 42.8
Height over net vs ball speed
Angle of incidence on the court
 Ball Trajectories 389

References

H. Brody (1987) Tennis Science for Tennis Players, University of Pennsylvania

Press.

A. Stepanek (1988) The aerodynamics of tennis balls - The topspin lob,

American Journal of Physics, 56, 138-142.

R.D. Mehta and J.M. Pallis (2001) The aerodynamics of a tennis ball, Sports
Engineering, 4, 177-189.
Acceleration with sand to form a slightly rough surface as used on
the Rebound Ace and DecoTurf courts at the Australian
The rate of change in velocity. It is measured in ineters
and US Open. The amount of sand controls the speed
per second per second (m/s?): a = Av/t. If a vehicle of the court to a small extent. Some areas of a court
accelerates from 20 m/s to 30 m/s over 5 seconds, a = can wear more than others, typically leaving smooth
10/5 = 2 m/s?. If the speed stays constant at 20 m/s, patches near each baseline.
then a = 0.
Amplitude
Acceleration due to gravity (g)
The maximum displacement on either side of the equi-
The rate at which velocity increases for a freely falling librium position of a racquet that is vibrating.
body. Near the earth’s surface g = 9.8 meters per sec-
ond_ per second (m/s?). Angular momentum

Angular momentum is the momentum of a rotating

Air resistance object, but it is defined in a different way to linear
moment, since different parts of a rotating object trav-
Friction of the air upon a ball in flight. It generates a
el at different speeds. If a racquet of mass M is mov-
drag force acting backwards on the ball.
ing in a straight line at speed v, then its linear momen-
tum is Mv. If the racquet is rotating at angular speed a,
Acrylic
then its angular momentum is Iw, where I is the
An acrylic paint is a spreadable plastic compound moment of inertia about the axis of rotation. In a col-
mixed with pigment of a particular color used to coat lision between a racquet and a ball, the total linear
tennis courts. It can be thinned with water and dries and the total angular momentum (of the racquet plus
much faster than oil-based paints. It can also be mixed the ball) are both conserved.
392 Appendix A

Anisotropic Center of mass

A material that exhibits different properties when test- In order to describe the motion of a racquet or a ball,
ed in different directions. Graphite is an example. It is it helps to imagine that the whole mass of the racquet
stronger, for example, along its length than perpendi- or the ball is located at a particular point called the
cular to it. Titanium, on the other hand has the same center of mass (CM). The CM of a ball is inside the ball,
strength in all directions and is called isotropic. in the middle, despite the fact that there is nothing
there apart from some air. The CM of a racquet may
Atom also be located in the air space around the racquet,
and is typically somewhere near the throat. A racquet
Smallest unit of an element. It consists of a positively can be balanced on a narrow edge if the edge is direct-
charged nucleus surrounded by electrons with a total ly below the center of mass. If you toss a racquet in
negative charge that is equal and opposite the positive the air, the racquet will tumble and turn, but the cen-
charge of the nucleus. ter of mass traces out a perfectly smooth parabolic
path.
Axis of rotation

When a racquet or a ball rotates, it rotates about a par- Center of percussion

ticular axis in such a way that points furthest from the Located near the center of the strings, but a few cm
axis rotate at the greatest speed, and points on the axis closer to the throat. If a ball is struck at the center of
rotate at zero speed. Each point on the racquet or the percussion, there is no sudden movement of the han-
ball follows an arc (or possibly a full circle), the cen- dle towards or away from the net, and hence, there is
ter of which coincides with the axis of rotation. The no sudden shock felt by the hand or arm.
axis of rotation of a racquet, just before the racquet
collides with the ball, is typically about 6 inches Clamps
beyond the butt end of the handle.
Used by stringers to hold each string in position while
Balance point the next one is being installed. A good clamp proper-
ly installed should stay in place and not move very far
The balance point of a racquet is near the middle of
when the string is tensioned, but all clamps move a
the racquet, directly below its center of mass. It is the
small amount.
point where a racquet will balance on one finger or a
straight edge. It is usually quoted as the distance in cm
Clay
from the butt end of the handle to the balance point
and is typically between 32 and 39 cm. Also called cen- Clay courts are common in Europe, but not elsewhere.
ter of mass. The clay is made from finely crushed bricks or tiles
and has the texture of a fine powder at the French
Bite Open. It is called clay since bricks and tiles are made
by baking clay, but clay itself is soft, like plasticene.
Players and commentators often say that the ball
The court is relatively slow compared with most other
“bites” the court or that the strings “bite” the ball. They
courts, although players can slide into and out of shots
are correct in that the bottom of the ball can grip the
with ease since the fine particles make the surface
surface and come to a complete stop in a direction
slightly slippery under foot.
parallel to the surface. Even though the bottom of the
ball stops, the top and middle of the ball do not come
Closed face
to a stop. The result is that the ball bounces off the
court at lower speed in a direction parallel to the sur- Describes racquet head orientation at time of ball con-
face than when it slides across the surface, so it kicks tact. A closed face has the string plane tilted forward
up at a steeper angle off the court. and downward when the ball contact is made.

Bumper guard Coefficient of friction (COF)

The plastic strip around the tip of the frame used to When an object slides on a surface there is resistance
prevent the frame and strings getting scratched or to its motion caused by friction between the object and
damaged. the surface. The friction force F acts in a direction par-
allel to the surface, opposite the direction of motion of
the object, and is proportional to the force N on the
 Glossary 393

object acting perpendicular to the surface. The COF is the ball before the collision equals their sums after the
defined as COF = F/N and is a number typically collision: (MV), fore ri (MV) ster
between 0.1 and 0.9. The COF between a ball and
grass is about 0.6. Between the ball and a clay court it Court speed
is about 0.8. Between the ball and the strings of a rac-
quet it is about 0.4. If the ball rolls along a surface, The ratio of the horizontal speed after the bounce to
then COF decreases to about 0.05. the horizontal speed before the bounce doesn't have a
special name because it depends on the angle of inci-
Coefficient of restitution (COR) dence. In practice, the ratio is usually between 0.40
and 0.75 on all courts and at all angles of incidence of
When a ball is dropped vertically onto the ground or practical interest. In theory the ratio could be 0.9 or
onto a head-clamped racquet, the COR is defined as more at very low angles of incidence, but a ball never
the ratio of the rebound speed to the incident speed. hits the court at angles less than about 10 degrees. The
The COR is a number between 0 and 1 that describes ratio is sufficiently important that it ought to have a
how well the ball bounces. A plasticene ball does not special name. In this book, we will call it “Court
bounce at all off a hard surface and the COR in that Speed,” since that is what it describes, but it does vary
case is zero. A more general definition of the COR is with the angle of incidence. The International Tennis
used to describe the impact of a ball on a hand-held Federation has a different definition to describe the
racquet. If the ball impacts and rebounds at right speed of a court — Pace (see Pace).
angles to the string plane, then the COR is defined as
the ratio of the relative speed of the ball and the Creep
impact point on the racquet after the impact to the rel-
ative speed before the impact. For example, if the ball An object that is subject to a tension force will stretch
is incident at 10 m/s and the racquet approaches the by an amount that is roughly proportional to the force.
ball at 20 m/s, then the relative speed is 30 m/s. The Even though the force may be held constant, the object
ratio of the ball speed after the impact to the ball may keep stretching at a slow rate for hours or years.
speed before the impact is called the ACOR (apparent This effect is called creep and it is especially noticeable
coefficient of restitution), and it is usually measured in tennis strings and other plastic materials.
under conditions where the racquet is initially at rest. Alternatively, if a string is under tension, and it is
If the ball is incident obliquely on the court or on a clamped at each end so its length stays constant, then
racquet, then the COR is defined in terms of the per- the tension will slowly drop. Both effects are due to
pendicular components of the velocity of the ball and the gradual breakage and slippage of bonds in a mate-
the racquet. rial that is subject to a tension force.

Composite Constant pull stringing machine

A combination of two or more materials into a struc- When a string is pulled to a given tension, whereupon
ture. Usually one material contributes most to the the pulling is stopped, it will immediately begin to lose
strength (reinforcement) and the other holds the mate- tension. A constant pull stringing machine recognizes
rial together (the matrix). Racquets are usually a the decline in tension and pulls back up to the desired
“graphite reinforced matrix” and string a “nylon (or tension. This is a “continuous” or “constant” process
other) reinforced matrix” where the matrix is a resin. until the stringer clamps the string. It usually takes 2-5
seconds from pulling to clamping, so a continuous pull
Conservation of energy machine will stretch the string for a longer time than
will a lockout machine which stops pulling as soon as
In a closed system, the amount of energy remains con- the string reaches the desired (reference) tension. As a
stant. It may be converted to other forms, but the result, a racquet strung on a continuous pull machine
amount is always the same. When we say kinetic ener- will have a stingbed that is about 5-10% stiffer than
gy is lost in a collision of a racquet and ball, we mean one strung on a lockout machine. Many stringers know
it has been converted to another form that cannot be this and compensate accordingly.
used in propelling the ball off the racquet.

Conservation of momentum Copolymer

The total momentum before a racquet and ball collide Copolymers involve making long chain molecules
is equal to the total momentum after they collide. The (polymers) by using more than one type of basic
sum of the mass times velocity of the racquet and of repeating unit (mer). See polymer.
394 Appendix A

Covalent bonding Deadspot

Bonding between atoms by sharing electrons in the If a ball is dropped on a stationary, hand-held racquet
outer shell of each atom. This is how atoms bond to with the strings in a horizontal plane, then the ball will
one another in polymer chains that compose strings. bounce best near the throat and worst near the tip.
These bonds are very strong. The bonds between adja- There is usually a spot near the tip, especially with
cent chains within strings are not so strong and are light racquets, where the bounce height is zero. That is
called secondary or polar bonds. These are more called the dead spot. If the racquet is swung to serve
bonds of attraction rather than interaction and sharing, a ball at the dead spot, then the speed of the ball off
and as such, are weaker and are the reason string the strings is equal to the speed of the dead spot.
creeps — secondary bonds between adjacent chains
can “break” and the chains slide or slip relative to each DecoTurf
other.
The court surface used for the US Open. Players need
to wear thick-soled shoes on this court to get a good
Customization
grip, which is why shoes squeek more on this surface.
This typically refers to altering the weight, balance, Players at Wimbledon wear shoes with much thinner
swingweight, stability and sweetspot of a racquet by soles. The court is classified as a hardcourt and is built
adding lead tape to strategic areas of the racquet. up from an asphalt base using several layers of acrylic
mixed with rubber and sand to form a hard surface
Cyclops cushioned slightly by the rubber content.

A device called a Cyclops machine is used on some Density

courts to determine if a serve is in or out. Several par-
allel infrared beams pass over and beside the service Mass per unit volume of a substance.
line about 6 mm above the court. A ball that intercepts
one or more of the beams and not another can be Drag force
judged by the machine to be in or out. The machine
A ball served at 100 mph slows down through the air
can't be used on the baseline since a player's foot
and lands on the court at about 75 mph. The force
might interrupt the beam.
exerted by the air on the ball is called the drag force.
It acts backwards on the ball. The drag force is small
DA, RA and RDC units
at low ball speeds and increases at high ball speeds.
The stiffness or “flex” of a racquet frame can be meas- The drag force increases with the ball diameter and
ured by locating it between two bars and pushing on increases if the cloth fluffs up.
a third point to measure how far the racquet bends. It's
like bending it across your knee, but with a dial or Drawing
readout to give you a number. There are several rac-
quet diagnostic machines available to do this and to The process of stretching a string under heat during its
measure stringbed stiffness as well. Unfortunately, manufacture to orient all the polymer chains compos-
there is no industry standard and the different ing it in the same direction. The more unidirectional
machines don't use the same measurement units. The the polymer chains, the stronger and stiffer the string.
USRSA uses a Babolat machine calibrated in RDC
(Racquet Diagnostic Center) or RA units. Other Dwell time
machines such as the Pacific diagnostic machine use The amount of time that the collision between the ball
DA units which are similar, but not exactly the same.
and the strings takes. Typically this is about 0.005 sec-
Readings in both units vary typically from 55 (soft onds (5 milliseconds). It can be increased a millisec-
frame) to 85 (very stiff frame).
ond or two by lowering string tension and decreased
by increasing tension.
The stiffness of a racquet frame decreases as soon as it
is strung since the strings act to bend the frame further Effective mass
when the frame is deliberately bent to measure its stiff-
ness. Ideally, one should measure frame stiffness with- When a ball strikes a stationary racquet anywhere on
out the strings since the higher the string tension the the strings, the racquet recoils and it also rotates. The
lower is the frame stiffness. Stiffness readings are also combined recoil and rotation speed at the point of
affected slightly by the relative softness of different impact is greater than that for an impact at the rac-
grips. quet’s center of mass. As a result, the racquet behaves
 Glossary 395

at the impact point as if its mass is less than its actual tension drops slowly over time and every time a ball is
mass. The effective mass of a racquet near the tip is hit.
about the same as the mass of a tennis ball. In the cen-
ter of the strings, the effective racquet mass is about Elastic limit
half the actual mass.
The amount a material can be stretched, bent, etc., and
Elastic collision still return to its original form. Beyond that point, the
material remains permanently deformed.
A collision where all the kinetic energy is conserved.
So, the sum of the ball’s and racquet’s kinetic energy
Elastomer
before the collision would equal that after the collsion.
Of course, this does not occur in a racquet and ball Rubber materials that can stretch several times their
collision. Some of the kinetic energy is transformed length and return to the original length when the force
into vibrational energy in the racquet, ball and strings, is released. Elastomers are used in vibration damping
as well as into heat and sound. In this case, we say the in racquets. The elastomer stretches or compresses and
collision was inelastic. stores energy elastically. However, upon release of the
force, it returns the energy more slowly than it
Elasticity received it. Because the energy is transferred over a
longer period of time, the peak force experienced by
If a certain material is elastic, it can be stretched and it
the hand will be less.
will then return to its original shape when the stretch-
ing force is removed. The latter property means that
Flex
the material is also resiliant. The words elastic and
resiliant mean the same thing scientifically. In common The word flex means bend. When describing a rac-
usage, something that is very elastic can be stretched a quet, it refers to the DA or RA reading of racquet stiff-
long way, like a rubber band. Plasticene can also be ness.
stretched a long way, but it is not elastic since it does-
n't resume its original shape after stretching. Steel is Force
elastic provided it is not stretched too far since it will
return to its original length when the stretching force A force is a push or pull which can cause an object to
is removed. A tennis string is elastic if the stretching accelerate or decelerate or change its shape by stretch-
force is removed soon enough, otherwise it will not ing or compressing. A force is always exerted by some-
thing on something else. The earth pulls downwards
return to its original length when the stretching force
on an object of mass M with a force F = Mg where g
drops to zero.
is the acceleration due to gravity. In the SI system of
The amount of stretch under a certain tension force is units M is measured in kg, g = 9.8 m/s? and F is meas-
described by the stiffness or softness of the material. A ured in Newton. The earth pulls down on a 1 kg object
steel wire is a lot stiffer than a tennis string. It is also with a force F = 9.8 N. The weight of an object (as
less flexible meaning that it is harder to bend. If a steel opposed to its mass) is the force exerted by the earth
wire is stretched too far, it can be taken past its elastic on the object. The weight of an object of mass 1 kg is
limit and then it remains permanently stretched when therefore 9.8 N. In common usage it is often stated that
the tension force is removed. Some materials stretch the weight of a 1 kg mass is 1 kg or that it weighs 1
easily, but they don't return to their original length kg. In the USA it is more common to say that the earth
after stretching due to breakage of bonds in the mate- pulls down on an object of mass 1 Ib with a force of 1
rial. Such materials are said to be visco-elastic. Tennis lb and it weighs 1 lb. This avoids the problem of mul-
strings stretch or bend easily compared with say steel tiplying by g, but a more careful distinction between
wire of the same diameter so they are softer or less stiff mass and weight or force must be made when doing
and more flexible. Tennis strings are also visco-elastic. calculations. An object of mass M subject to a force F
If you take a given length of string, stretch it for a will have an acceleration a given by F = Ma or a = F/M.
while and release the tension, the string will be a bit If a string is stretched by applying a force F at each
longer than its original length. The longer it is under end, it will not accelerate since the total force on the
tension the more it stretches and the longer it will be string is zero (since equal and opposite forces are
when the tension force is completely removed. A applied at each end), but it will stretch by an amount
string in a racquet stays the same length for many x given by F = kx or x = F/k, where k is the stiffness
years since it remains under tension and since it is of the string. In this case, the force F is more com-
anchored at both ends at a fixed length. However, the monly called the string tension.
396 Appendix A

Frame of reference thick strings. Thin strings do not impart more spin or
less spin to the ball than thick strings.
In order to describe the speed and location of an
object that is moving, we need a coordinate system. Graphite
That system is our frame of reference. The frame need
not be at rest. It might itself be moving, but we usual- This is one of the materials used to construct almost all
ly prefer to use a reference frame that does remain at modern tennis racquets. It is one of the two main
rest. Your reference frame might then be different to forms of pure carbon (diamond being the other) and
someone else's reference frame. If you are sitting in a is used to make pencils. It can also be manufactured
train and throw a ball vertically into the air, the ball as a long and very strong filament which can be
will rise straight up and fall straight back into your woven as a cloth. Many layers of cloth are glued
hand, exactly as if you were on a tennis court. People together with epoxy resin to manufacture a tennis rac-
outside the train will see the train and the ball coming quet, often with other layers of kevlar or fiberglass.
towards them at high speed. We can describe any par- The result is a very light and very strong composite
ticular event in whatever reference frame we like, but material, especially when it is constructed as a hollow
the event will appear to be different in different frames tube. Graphite tubes are also used to construct modern
of reference, even though Newton's laws of tennis do arrows, kites, fishing rods, etc. Spider webs are actual-
not depend on which reference frame we choose. At ly stronger than graphite filaments (for any given
least, that is the case when two reference frames move diameter), but no one has figured out how to make a
at constant speed with respect to each other. Rotating tennis racquet from spider webs.
frames or accelerating frames or frames moving near
the speed of light need to be treated with more cau- Grommet
tion since Newton's laws don't apply in those reference
Plastic grommets are inserted through the holes in the
frames.
frame to protect the string from the sharp edges of the
holes in the frame.
Frequency

If any part of a racquet (or a guitar string) vibrates Gut

back and forth 100 times every second we say that the
Natural gut is manufactured from the intestines of beef
frequency is 100 Hz. Frequency is measured in Hertz
cattle. It is the preferred string of many professionals,
(Hz), which is the same as cycles per second. If the fre-
quency is 200 Hz, that actually means your hand gets but it is relatively expensive. Intestines are designed to
hit 400 times per second, since it is surrounding the stretch and shrink to accommodate a large meal and
handle on both the “back” side and “forth” side of the have excellent elastic properties. To manufacture a
vibration. string, the intestines are cut into several long ribbons.
It takes two or three animals to produce one string.
Gauge The ribbons are first washed and then twisted togeth-
er to form a structure like a thin rope since it is impos-
Tennis string diameters vary from about 1.20 mm to sible to join the ribbons end-to-end. The rope is dried,
about 1.40 mm and are specified by a gauge number. mechanically polished to form a smooth string and
The nominal diameters are approximately as follows: then coated to make it more water and abrasion resist-
18G =#20 mm, 17G = 1.25 mm, 16G = 1.30 mm, 15L ant and hence more durable. Natural gut has the prop-
= 1.35 mm and 15G = 1.40 mm. The actual gauges vary erty that it does not get significantly stiffer at tensions
from one manufacturer to the next and they vary along above the normal strung tension, whereas synthetics
the string (especially on textured or rough surfaces) do. As a result, natural gut is generally softer. However,
and they will vary slightly from one batch to the next. it can play as a stiff string if strung at high tension,
Some strings are elliptical in cross section so the diam- since it holds its tension better than most synthetics.
eter is different depending on which diameter is meas-
ured. If you are concerned about actual string gauge, Hardcourt
then you will need to measure it with a micrometer.
Thin gauge strings usually wear out faster. They are A court with a hard surface. It may be solid concrete
not necessarily softer than thicker gauges. Thin strings or asphalt or have an upper layer of rubber and acrylic
stretch further when they are installed, but they tend to mixed with sand to make the surface slightly rough.
stretch less than thick strings when they are used to hit
a ball since they get stiffer quicker at high operating
tensions. Thin strings tend to hold tension better than
 Glossary 397

Head-heavy why a racquet is made of layers of graphite arranged

at different angles in different parts of the racquet,
If the balance point of a racquet is midway along its
depending on the expected directions of forces in that
length, the racquet is neutrally balanced. If the balance
area.
point is closer to the tip of the racquet, the racquet is
head heavy. If the balance point is closer to the butt
Kevlar
end of the handle, the racquet is head light or handle
heavy. A very stiff and strong synthetic material developed by
Dupont used in some racquet frames and to make ten-
Headsize nis strings.
Modern racquets have a much larger head than old
wood racquets. Head sizes are classified as midsize (< Kinetic energy
95 in.* / 613 cm?), midplus (95-104 in.? / 613-671 cm2.), The kinetic energy E (in Joules) of an object of mass
oversize (105-117 in.? / 677-755 cm2), or superoversize M (kg) is the energy due to its motion and is given by
Catz in2 7-755 cm’). E = Mv?’/2 where v is the speed of its center of mass
(in m/s). A tennis ball of mass 0.057 kg served at a
Horizontal velocity speed of 50 m/s has kinetic energy E = 71.25 Joules.
An 80 kg person jogging at 5 m/s has kinetic energy E
When a ball flies through the air, it has a speed in both
= 1000 Joules. If the object is rotating about a fixed
the horizontal and vertical directions. When _ it
axis, then E = Im?/2 where I is the moment of inertia
bounces, the friction force changes the horizontal
velocity. The relationship between the horizontal and about that axis, and @ is the angular velocity due to
vertical speeds after the bounce, as well as the spin, is
rotation (measured in radians/sec). One revolution/sec
what characterizes the speed and angle of the bounce. is equal to 6.28 radians/sec.

Impulse Layup
When a racquet hits a ball, we say that the ball was By-hand process of placing different layers of unidi-
acted upon by an impulsive (or sudden) force. The rectional graphite at different angles to each other in a
racquet. A racquet is composed of 7-12 layers, with
impulse is defined as the force F multiplied by the
various layers lying between 0 and 90 degrees to each
duration of the force t, and is equal to the change in
other. This is important to the strength and stiffness of
the momentum of the ball. That is,
the racquet because these depend on the direction of
Impulse = Ft = mv, - mv, = change in momentum. the graphite fibers at any given location.

Inertia Lockout

This is the property of racquets and balls to resist any See constant pull stringing machine.
change in their motion. An object’s mass is a measure
of this inertia. Magnus force

When a ball is spinning, the flow of air around the ball

Inelastic
is altered in such a way that the air pressure is reduced
An object that does not return to its original form when at the bottom of a ball with topspin or at the top of a
it is stretched by a force and then released. ball with backspin. As a result, a downwards force is
exerted by the air on a ball with topspin and an
Inelastic collision upwards force is exerted on a ball with backspin. This
force is called the Magnus force. A ball hit with topspin
See elastic collision. is forced downwards onto the court with a higher ver-
tical speed than a ball without spin, so 1t bounces high-
Isotropic er (assuming that both balls are hit with the same
The properties of the material are the same in all direc- speed).
tions. Metals, like titanium, are isotropic. Graphite is
Mass
not. Because graphite is not isotropic, its stiffness and
strength depend on the direction of the long axis of The mass of an object is usually measured in kg or gm
the material compared to the stressing force. That is or oz or lb. It represents the resistance of the object to
398 Appendix A

acceleration. A heavy object is more difficult to accel- from the tip and the other about 15 cm from the butt
erate than a light object and will require a larger force end of the handle. If the ball hits at the node, the rac-
to accelerate it to the same speed in the same time. If quet will not vibrate. If it hits anywhere else, the rac-
a light and a heavy object are subject to the same force quet will vibrate everywhere but the nodes.
acting for the same length of time, the light object will
accelerate to a higher speed. The relation between Open Face
these quantities is given by F = Ma where F = force, M
= mass and a = acceleration. Describes racquet head orientation at time of ball con-
tact. An open face has the top of the string plane tilt-
Mer ed backward and upward when the ball contact is
made.
The smallest unit that is repeated over and over (thou-
sands of times) to make a polymer, which is also called Open stance
a long chain molecule. Sometimes called a “monomer.”
Players these days often stand facing the net when
Midsize returning the ball, rather than side-on. Open stance
means facing the net, in which case the racquet is
Generally speaking, racquets with a headsize < 95 in.* swung around in front of the body rather than in a
/ 613 cm?. Midplus racquets have a headsize between direction towards the net.
95-104 in.? / 613-671 cm?.
Overgrip
Modulus
A thin grip that is wrapped on top of an existing grip.
Describes the relationship of stress to strain on an Cheaper than replacing the whole grip with a replace-
object when a pull of a given force is applied along its ment grip. You can choose between soft, tacky, sticky,
length: Modulus = stress/strain. Modulus is a measure smooth, perforated, embossed, ribbed, firm or cush-
of a material’s stiffness. ioned and various combinations of these properties. A
smooth, firm grip allows you to feel everything that's
Moment of Inertia happening to the handle, a bit like a manual rather
than an automatic.
The moment of inertia of an object represents its resist-
ance to change in rotation and is commonly known as
Oversize
the swingweight when referring to a racquet. If the rac-
quet has mass M (kg) and length L (cm), then the Generally speaking, racquets with a headsize between
swingweight or moment of inertia about an axis 105-117 1n.27/677-755-cm-.
through the butt end of the handle is approximately
ML?/3 (kg-cm?). For example, if M = 0.3 kg and L = 70 Pace
cm, then the swingweight is approximately 490
kg-cm*. The swingweight about an axis 4 inches from The International Tennis Federation (ITF) definition to
describe the speed of a court is called Pace. The for-
the butt end is smaller than this by about 180 kg-cm?.
The value of the swingweight will be larger if the rac- mula is slightly complicated, but if the ball slides
throughout the bounce, then Pace = 1000. - COF)
quet is head heavy, and smaller if the racquet is head-
light. where COF is the coefficient of sliding friction. For a
grass court, COF is about 0.6 and hence Pace = 100 x
Momentum 0.4 = 40. For a clay court COF is about 0.8 and hence
Pace will be about 100 x 0.2 = 20.
The momentum of an object is its mass times its veloc-
ity. When two objects have a head-on collision, the Peak force
total momentum of the two objects after the collision
The maximum force obtained during a collision. When
is the same as the total momentum before the collision
even if some of the kinetic energy is lost. the ball hits the strings, for example, the force starts at
0, climbs to a peak just as the strings achieve maxi-
Node
mum deflection as the ball comes to a rest, and then
goes back to 0 as the ball leaves the strings.
A point on a vibrating object that does not move. In a
racquet this occurs in two locations, one about 15 cm
 Glossary 399

Peak tension ing and give almost all that energy back to the ball.

Upon impact, the strings stretch and therefore increase

Power
in tension. At maximum deflection, they will achieve a
peak tension, whereupon the tension will decrease as There is no such thing as a 100 Watt or a 0.2
the ball is ejected. The string will not return to its Horsepower racquet or string. Racquets and _ strings
before-collision tension. It will be slightly less after don't generate any power by themselves. All the
each impact. power comes from the player, but the racquet and
strings allow some of that power to be transferred to
Player Rating the ball with an efficiency that depends on the partic-
ular racquet and strings. Power is one of those terms
In the USA, players are rated on a scale from 1 to 7. A
used in tennis, like feel or playability, that can't easily
7.0 player is one of the top 100 players in the world.
be defined, but everyone has a good idea what it
A 1 or 2 is a beginner.
means. Essentially it means that if you can hit a ball
faster with one racquet or string than another, then the
Period
racquet or string is more powerful or more efficient in
The time it takes a vibrating object to complete one transferring energy from the arm to the ball. However,
back-and-forth cycle. Typically about 0.007 seconds a precise definition of racquet power is complicated by
for a vibrating racquet. the fact that it depends on the impact point and it
depends on how the racquet is swung. A ball hit from
Piezoelectric the tip won't travel at the same speed as a ball struck
in the middle of the strings since the tip and the mid-
A material that is piezoelectric generates an electric dle travel at different speeds. The difference will
charge when a force deforms it. The charge produced depend on how much wrist action the player uses.
is proportional to the force. The opposite is also true.
A piezo material is deformed when it is subjected to an Pressureless ball
electric charge, and the deformation is proportional to
the amount of charge. A ball that is sold in a cardboard box instead of a pres-
surised container since there is no excess air inside the
Polymer ball. The rubber wall is slightly thicker and stiffer than
in a pressurized ball. It bounces the same as a pres-
Long chain molecules made from the combinations of surized ball and lasts a lot longer, but it sounds slight-
thousands of repetitions of a basic molecular unit ly different, since it tends to vibrate a bit more. The
called a mer. “Poly” means many and “mer” means ball never goes flat, but the cloth cover will wear.
unit. Most string materials are polymers as are the
resins in racquets. Pressurized ball

Potential energy A ball that is pressurized with air to increase its stiff-
ness so that it bounces to the correct height. The air
When a ball or any other object of mass M is lifted off gradually leaks out over time or after a few sets.
the ground to a height h, the work done to lift it is
given by the force Mg times the distance h. This is RA
stored as potential energy Mgh. When the ball is
released, the stored potential energy is converted to See DA units.
kinetic energy and the potential energy drops to zero
at ground level. The ball hits the ground with a speed Radar gun
v given by Mv’/2 = Mgh. The kinetic energy of the ball The speed of a ball can be measured remotely using a
is converted to elastic potential energy in the ball as radar gun. The gun emits a beam of microwaves which
the ball squashes. The ball comes to rest momentarily bounces off both the racquet and the ball and it returns
and loses all its kinetic energy in doing so. If the ball to a microwave detector in the gun. The frequency of
has a stiffness k and is squashed by an amount x, the the microwave beam is shifted by any moving object
stored elastic energy is kx?/2 = Mv?/2. Some of this by an amount that depends on the speed of the object.
energy is lost in heating up the ball, but the rest is If the detector is set to detect the maximum frequency
converted back to kinetic energy as the ball bounces shift, it will record the speed of the ball rather than the
off the ground. Similarly, when the strings impact a speed of the racquet. The gun actually records v cos(®)
ball the strings store elastic potential energy by stretch-
400 Appendix A

where v is the speed of the ball and 0 is the angle depending on the racquet and where the impact was.
between the path of the ball and the path of the
microwave beam. If the ball is headed say 5 degrees Short side
away from the gun, then the recorded speed is 0.996v.
A racquet stringer starts with a length of string about
If 8 = 10 degrees, then the gun records 0.985v.
40 feet (12 m) long. About 1/3 of the string (the short
side or the short end) is fed through one of the grom-
Rebound Ace
met holes so that the other 2/3 (the long side) remains
This is the court surface used for the Australian Open. outside the racquet. The stringer can string the short
It consists of a layer of rubber about 6 mm thick glued end first without having to pull the whole string
to a concrete or asphalt base with a 1 mm thick and through each hole.
slightly rough acrylic upper surface. The speed of the
court can be adjusted slightly by repainting the upper Spaghetti strings
surface with fine sand mixed in with the paint.
These originated in Germany around 1975, but they
were quickly banned by the International Tennis
Recoilweight
Federation because they allowed a player to put exces-
Not an accepted physics term, this is the tennis play- sive spin on the ball. The original version was strung
er’s version of moment of inertia about the center of with only about 5 cross strings and used spaghetti tub-
mass. It describes the racquet’s resistance to rotational ing to cover the main strings to allow them to slide
acceleration about the center of mass as the axis -of more easily over the cross strings. The main strings
rotation. This is usually a little less than half the swing- were then tied together to allow the mains to move
weight. Whenever the ball hits any place besides the freely as a single unit across the cross strings. As a
center of mass, it tends to rotate the racquet back- result, the mains stretched elastically in a direction par-
wards. The higher the recoilweight, the less the rota- allel to the string plane (as well as in the direction per-
tional acceleration for a given impact velocity. pendicular to the string plane like other strings) and
imparted extra spin as they returned to their original
Resilience position.

If an object is stretched or squashed, its length or

Speed
diameter will change. If the force on the object is then
reduced to zero, the object will return to its original How fast an object like a ball travels from one location
shape if it is perfectly resiliant. Materials such as plas- to another. It does not take into account the direction
ticene are not resilient since they do not return to their of motion, just the distance traveled per unit of time:
original shape but remain plastically deformed. A ten- speed = distance/time.
nis string under tension for some time does not return
to its original length if the tension is reduced to zero. Speed of a court
However, a tennis string is perfectly resiliant if it
remains under tension in a racquet since it always The speed of a court refers to the change in the hori-
returns to the same length after stretching due to an zontal speed of the ball when it bounces. There is a
impact with a ball. The main strings often move side- relatively large change on a slow court like clay and a
ways and get stuck at a slightly longer length, but they relatively small change on a fast court such as grass.
are still perfectly resiliant in a direction perpendicular The coefficient of sliding friction (between the ball and
to the string plane. Some materials take time to recov- the surface) is about 0.8 for clay and about 0.6 for
er to their original shape and are therefore not very grass. However, the speed difference between clay and
resiliant. They make good vibration dampeners. grass is not as large as one might expect. A ball served
on clay takes only about 0.02 seconds longer to cross
Shock the baseline than on grass, assuming it is served at
around 100 mph or more.
When a ball impacts a racquet, the racquet translates
backwards, it rotates, and it bends creating a bending A bigger time difference will be noticed if the ball is
wave. All three of these motions arrive at the hand at hit at a slower speed, partly because the ball is slow-
the same time about 2 milliseconds after impact. er, but also because the ball will be incident on the
Together, they jolt the hand and this is what a player court at a higher angle. The ball slows down by a larg-
feels as shock. The hand stops the rotation and trans- er fraction when it is incident at a larger angle.
lation, but the vibration continues for about 25-50 ms,
 Glossary 401

Standing wave light). Most racquets vibrate at around 120 to 150 Hz.
Stiff racquets are like tight guitar strings. The higher
A standing wave is one that does not propagate like a
the tension or the stiffer the racquet the higher the
normal wave, but “stands still”. A standing wave is
vibration frequency.
generated when a propagating wave reflects back-
wards off the ends of the medium supporting the wave
Strain
and meets the incoming wave. The two waves inter-
fere constructively at some points and destructively at Strain is related to stress and is given by the extension
points in between, with the result that the medium or compression of an object divided by the length of
supporting the wave (racquet, guitar string, tennis the object. If a 10 inch length of tennis string stretches
string, etc.) vibrates back and forth rapidly. Points that by 1 inch, then the strain = 1/10 = 0.1. If the stress
vibrate the most (tip and butt end of the handle) are increases, then the strain will increase. The ratio of the
described as antinodes, and points that don't vibrate at stress to the strain is a constant for any given material,
all are called nodes. independent of the diameter or the length of the mate-
rial.
Stiffness
Stress
If a tennis string is anchored at one end and pulled by
a force F at the other end, the string will stretch by an Stress is an engineering term describing the tension or
amount x. The stiffness k of the string is defined by the compression or shear force acting per unit area on an
relation F = kx (or k = F/x) and is measured in l|b/in. object. For a tennis string, the stress is given by the
or N/m. If k is very large, then the string is very stiff. string tension divided by the cross-sectional area of the
If k is small, then the string is softer or more elastic. string, and is measured in pounds per square inch or
Since a long or a small diameter string will stretch fur- Newton per square metre.
ther than a short or fat string, k depends on the length
of the string and its diameter as well as the type of String dampener
string. A stretched string also has a certain stiffness
when it is pushed sideways. If a sideways force F at
A small object located on the strings close to the frame
whose function is to dampen vibrations of the strings
the middle of the string causes the string to move side-
so that the impact sounds like a thud instead of a ping.
ways by a distance y, then the transverse stiffness is
Andre Agassi uses a rubber band. About half the pro-
given by k, = F/y. In this case, k, depends mainly on
fessionals use them, and the other half don't.
the string tension and the string length, but at large
values of F or y it also depends on the string stiffness Stringbed
for a longitudinal stretch. The transverse stiffness of
the whole string plane depends on all these factors, as The stringbed is simply the bed of strings used to hit
well as the total number of strings and on the diame- the ball.
ter of the object used to push on the strings.
String saver
The longitudinal stiffness of a string may remain con-
The main strings move across the cross strings every
stant or it may change as x gets bigger. If it changes,
time a ball is hit, eventually wearing agroove in the
then it is useful to define the dynamic stiffness by k =
main strings. The main strings usually break first, at
DF/Dx, where DF is the increase in the force (starting
one of the grooves. String savers are small circular
at a certain specified non-zero value), and Dx is the
plastic inserts that lie between the mains and the cross-
increase in the length of the string.
es to reduce wear. A drop of superglue would proba-
bly do the same job by preventing sideways move-
The stiffness of a ball or a racquet can be defined in
ment.
the same way. If a ball or racquet squashes or bends
by a distance x when subject to a force F, then the stiff-
Sweetspot
ness k = F/x. However, racquet stiffness and stringbed
stiffness is more commonly measured in DA or RA The impact point on the strings that feels best when
units. A simple way to judge the stiffness of a racquet you hit a ball. Points nearby feel almost as good,
is to tap the tip with your finger and listen to the han- which enables manufacturers to claim that their sweet
dle vibrate (holding the frame lightly about 6 in. from spot is bigger than a single point. The sweetspot is
the butt end with the handle close to your ear). If it located near the middle of the strings in a modern rac-
vibrates at around 200 Hz, it is very stiff (and/or very quet.
402 Appendix A

Swingweight the court with topspin, then it is usually incident on

the strings of a racquet with backspin. The ball does-
When a racquet is swung it may feel light or heavy. n't change its spin direction just before it hits the
Part of the heaviness depends on its weight, but one strings, but the point on the ball that contacts the
racquet can feel heavier than another even if both rac- strings does so in the same manner as a ball hit with
quets have the same weight. A racquet with a heavy backspin contacts the court surface. If a ball is thrown
head and a light handle will feel heavier when it is on the floor and lands under a table, then the ball
swung than one with a light head and a heavy handle. bounces off the floor with topspin, but it approaches
The heaviness depends on how the weight is distrib- the underneath side of the table with backspin.
uted along the racquet and it is measured in terms of
the swingweight. The technical name for swingweight Torque
is the moment of inertia. The actual value of the swing-
weight depends on the location of the axis of rotation. Defined as the force times the distance of the lever arm
Measurements of swingweight are normally done by to the axis of rotation. A torque tends to rotationally
the swinging the racquet about an axis in the handle accelerate an object. For instance, when a_ ball
located 4 inches from the butt end. bounces, the friction force of the court surface exerts a
torque equal to the frictional force times the radius of
Synthetic gut the ball. This torque causes the ball to change its rate
of spin.
A synthetic string, usually nylon, which first appeared
in the 1960's as a more durable replacement for natu- Twistweight
ral gut.
More correctly known as the polar moment of inertia,
Tension twistweight describes the racquet’s resistance to rota-
tional motion around the longitudinal (length-wise)
A tennis string is relatively soft and floppy when it is axis. It is the same as swingweight, but around a dif-
not under tension. If the string is tied down at one end
ferent axis. The twistweights for most racquets are
and pulled at the other end, then the string will stretch
between 13 and 24 kge-cm?. Swingweights, on the other
and it will be under tension. The tension force acts in
a direction along the string. A string under tension will
hand, typically lie between 250 and 350 kg-cm?, indi-
resist any attempt to push it sideways. cating that it is much more difficult to rotationally
accelerate a racquet around an axis 4 inches from the
Throat butt than it is to do the same about the long axis.

The middle section of a racquet near the top end of the Vector quantity
handle.
Quantity that has both a magnitude and direction.
Titanium Examples are force, velocity, acceleration, torque.
Vector quantities are depicted by arrows. The arrow
Titanium is used mostly as titanium oxide to make indicates the direction and the length indicates the
white paint. As a metal or an alloy it is very strong and magnitude.
light and withstands high temperatures better than
most other metals so it is suitable in aircraft engines. Velocity
It's use in tennis racquets, balls or strings is less direct-
ly associated with performance, but it seems to work The speed and direction of an object’s motion. If a ball
well as a selling point. and racquet approach each other with the same speed
but from opposite directions, then one has a positive
Topspin velocity and the other a negative velocity, even though
their speeds are the same.
A ball can spin about a horizontal axis at right angles
to its direction of motion in one of two directions. If Vertical speed
the ball has topspin, then the top of the ball rotates in
the same direction as the direction of motion. The bot- See horizontal speed.
tom of the ball rotates in the opposite direction to the
direction of motion. A ball rolling forwards on the Vibration
ground has topspin. A ball that bounces off the court A racquet will vibrate back and forth about 140 times
almost always bounces with topspin. If it bounces off per second when it strikes a ball. The vibration fre-
 Glossary 403

quency is said to be 140 Hz. Very stiff racquets vibrate and 11.5 inches (29.21cm.) in overall width. It says
faster, up to about 200 Hz. Old wood racquets vibrate nothing about the thickness of the racquet head in a
at about 90 Hz. The strings themselves vibrate at about direction perpendicular to the string plane. The thick-
500 Hz. ness is typically about 20 mm, but anything over about
22 mm is regarded as a widebody racquet. An increase
Wave in frame thickness adds to the stiffness of the head.

A wave is a disturbance that propagates from one

Work
point to another. Sound waves travel through the air at
a speed of 340 m/s. Sound travels through water at In general useage, the term “work” implies that some
1450 m/s and through solids at about 6000 m/s. Light effort is required. In physics, work has a_ specific
and radio waves are electromagnetic waves that travel meaning. When a force F is applied to move an object
through air or a vacuum at 300,000,000 m/s. Waves at through a distance d in the direction of the applied
the beach travel on the surface of the water at a speed force, then the work done is defined to be W = Fd. A
of a few m/s. An impact of a ball and a racquet sends number of different things can happen to the object as
a transverse bending wave along the racquet at a a result of this work, depending on the circumstances.
speed of around 200 m/s. The racquet bends locally If an object of mass M is accelerated in a horizontal
around the impact point, displacing adjacent bits of the direction from rest to a speed v, then the kinetic ener-
racquet from each other in a direction transverse to the gy of the object will be equal to the work done, ice.,
direction of propagation of the wave. The adjacent bits W = Fd = }Mv*. If the object is lifted through a height
in turn pull on neighboring bits, but it takes time for h by a force Mg without any increase in its speed, then
each bit to react to a bending force since each bit takes the potential energy of the object increases by an
time to accelerate up to maximum speed. It is this amount W = Fd = Mgh. The work done by a player in
delayed reaction time that causes the bending wave to swinging a racquet results in an increase in the speed
travel along the racquet at relatively low speed (com- and kinetic energy of the racquet.
pared with sound or light). The stiffer and lighter the
racquet the faster is the bending wave and the faster
the racquet vibrates. If the racquet vibrates at, say, 140
Hz, the associated bending wave travels up and down
the racquet 140 times each second.

Weight

When you weigh yourself on a set of bathroom scales

you do so to measure your weight in Ib or kg. At least
that's what most people do. Scientists and engineers
have stolen this common word to mean something
else, in order to distinguish mass from weight. In sci-
entific language, weight is the force Mg (measured in
Newton) acting on a mass M (measured in kg) due to
the pull of the earth. A mass of 1 kg weighs 9.8
Newton. If bathroom scales gave a reading in Newton,
then scientistists and engineers would agree that they
are used to measure weight. But bathroom scales give
a reading in lb or kg, so people actually mass them-
selves rather than weigh themselves. In the USA, most
people get around the problem by measuring mass in
lb and weight or force in Ib, but that is not scientifi-
cally correct. This book is scientifically incorrect when-
ever it quotes string tension in kg or Ib since tension
is 2 force, Hola mass.

Widebody

The rules of tennis specify that the strung surface shall

not exceed 15.5 inches (39.37cm.) in overall length,
 Units
Conversions

To convert from USCS to Sl, multiply the USCS units by the conversion factor. To covert from SI to
USCS, divided the SI units by the conversion factor.

U.S. Customary System (USCS) Conversion factor International System (Sl)

Acceleration
foot per second squared ft/s® 0.305 meter per second. squared m/s

Energy
foot-pound ft-lb 1.356 joule

Force
pound Ib 4.448 newton

Force per Length

pound per inch Ib/in. WAS WI newton per meter N/m

Length
inch in. 2.540 centimeters cm
foot ft 0.305 meter
mile mi 1.609 kilometer
406 Appendix B

U.S. Customary System (USCS) Conversion factor International System (Sl)

Mass
pound Ib ie kilogram kg

Torque
pound-foot lb-ft 1.356 newton meter Nem
pound-inch Ib-in. 0.113 newton meter Nem

Stress
pound per square inch psi 6894.760 pascal Pa

Velocity
foot per second ft/s 0.305 meter per second m/s
inch per second in./s 0.025 meter per second m/s
mile per hour mph 0.447 meter per second m/s
mile per hour mph 1.609 kilometer per hour km/hr

Other Useful Data

Acceleration due to Earth’s gravity = 9.8 m/s* = 32.13 ft/s”

Angles
1 rad = 57.30°
360° = 2x rad
180° = x rad

Area

1 square inch = 6.452 square cm

1 SC) Chin S ©. NO Se} ne
90 sq in. = 580.6 sq cm
100 sq in. = 645.2 sq cm
110°sqiins = 70957 sq cm

Energy
1 Joule = 0.7376 ft-lb
1 ft-lb = 1.856 Joules

Force

1 Newton = 0.2248 Ib
1 Ib = 4.448 Newton (lb here is a unit of force)

In the USA one pound can be a unit of mass or force in everyday use.

Length

1 inch = 25.4 mm
12 inch = 1 foot = 304.8 mm
il Clan) = 0) trntan
1 metre = 100 cm =1000 mm = 39.370 inches = 3.281 feet
1 km = 1000 m = 0.6214 mile
1 mile = 5280 feet = 1.609 km
Length of typical racquet = 27 inch = 685.8 mm
3 feet = 0.9144 m = net height in middle
 Units and Conversions 407

3.5 feet = 1.067 m = net height at posts

4.5 feet =1.372 m = distance between tramlines
21 feet = 6.401 m = distance from net to serve line
PGES OOn) (na) = distance of net posts from centre of net.
PT Wal = (yi) (an) = width of singles court
36 feet = 10.973 m width of doubles court
39 feet = 11.887 m distance of baseline to net
78 feet = 23.774 m length of court

Mass
i [<e) = OOO) CG = 2.205 lo =, 2H Oy
Deal OrOZe— Or 5SOnkG
IeOZe—223755) Oi

Speed

1 m/s = 3.6 km/hr = 3.281 ft/sec = 2.237 mph

100 mph = 160.9 km/hr = 44.70 m/s
140 mph = 225.3 km/hr Il 62.58 m/s = fastest serve speed up to 1996.
100 km/hr = 62.15 mph = 27.778 m/s
50 m/s = 111.85 mph

Spin
1 revolution/sec = 60 rpm = 6.28 radian/sec
1000 rpm = 16.67 rev/sec = 104.7 radian/sec
3000 rpm = 50 rev/sec = 314.16 radian/sec

String Gauge

1e8G = 1-20) mm
WAG) = WS Tainan
16G = 1,00 mm
Wok = Wseiotnolan
15G= 1-40 mm

String Tension

50 Ib = 22.676 kg
60 Ib = 27.211 kg
70 Ib = 31.746 kg
20 kg = 44.100 Ib
25 kg = 55.125 Ib
30 kg = 66.150 Ib
Time
i) (oaks = COLO) S
 ——w ar a «wet £4 |
ae » “7S, of cer ts =e =e

jemia® « 7 2984@ ze °D
<7 i “So eee
7 S Ss - _ > soot TF

: : ae) « eae
: a ee ee
a _ > an ae 3 ae)

re etl
r , on) eee -

» 2, ea 30 « ae
an «2 2 «se8 7
,
— - y

ees
wa

eal a ea “Ah GA,

* =w Zi Aquaa
ra ‘12 62D « dyn OF
: ; we P _ —a we
— me E> sai s ge an
—s
x ——— rity.
Sey,
‘ ® , mim

= > ims > 12.41

«= am
ee a i oa48 : , ¢ wet Se
The following are the equipment 2002 rules for balls (rule 3) and racquets (rule 4), as specified by the
International Tennis Federation.

3. THE BALL
Balls that are approved for play under the Rules of Tennis must comply with the following specifications:

a. The ball shall have a uniform outer surface consisting of a fabric cover and shall be white or yellow in
colour. If there are any seams they shall be stitchless.

b. The ball shall conform to the requirements specified in Appendix I (Regulations for making tests speci-
fied in Rule 3) Section iv for size and be more than 1.975 ounces (506.0 grams) and less than 2.095 ounces
(59.4 grams) in weight.

c. More than one type of ball is specified. Each ball shall have a bound of more than 53 inches (134.62cm)
and less than 58 inches (147.32cm) when dropped 100 inches (254.00cm) upon a flat, rigid surface e.g. con-
crete. Ball Type 1 (fast speed) shall have a forward deformation of more than .195 inches (.495 cm) and less
than .235 inches (.597 cm) and return deformation of more than .295 inches (.749 cm) and less than .380
inches (.965 cm) at 18 lb. (8.165kg) load. Ball Types 2 (medium speed) and 3 (slow speed) shall have a for-
ward deformation of more than .220 inches (.559cm) and less than .290 inches (.737cm) and return defor-
mation of more than .315 inches (.800cm) and less than .425 inches (1.080cm) at 18 lb. (8.165kg) load. The
two deformation figures shall be the averages of three individual readings along three axes of the ball and
no two individual readings shall differ by more than .030 inches (.076cm) in each case.
410 Appendix C

d. For play above 4,000 feet (1219m) in altitude above sea level, two addi-
tional types of ball may be used.

i. The first type is identical to Ball Type 2 (medium speed) as defined

above except that the ball shall have a bound of more than 48 inches
(121.92cm) and less than 53 inches (134.62cm) and shall have an inter-
nal pressure that is greater than the external pressure. This type of ten-
nis ball is commonly known as a pressurised ball.

ii. The second type is identical to Ball Type 2 (medium speed) as

defined above except that the ball shall have a bound of more than 53
inches (134.62cm) and less than 58 inches (147.32cm) and shall have
an internal pressure that is approximately equal to the external pres-
sure and have been acclimatised for 60 days or more at the altitude of
the specific tournament. This type of tennis ball is commonly known
as a zero-pressure or non-pressurised ball.

iii. The third type of ball which is recommended for use for play on
any court surface type above 4,000 feet (1219 m) in altitude is the Ball
Type 3 (slow speed), as defined above.

e. All tests for bound, size and deformation shall be made in accordance
with the regulations in Appendix I.

f. The International Tennis Federation shall rule on the question of

whether any ball or prototype complies with the above specifications or is
otherwise approved, for play. Such ruling may be taken on its own initia-
tive, or upon application by any party with a bona fide interest therein,
including any player, equipment manufacturer or National Association or
members thereof. Such rulings and applications shall be made in accor-
dance with the applicable Review and Hearing Procedures of the
International Tennis Federation (see Appendix IID.

Note 1: Any ball to be used in a tournament which is played under the Rules
of Tennis, must be named on the official ITF list of approved balls issued by the
International Tennis Federation.

Case 1: Which ball type should be used on which court surface?

Decision. 3 different types of ball are approved for play under the Rules of
Tennis, however:

a. Ball Type 1 (fast speed) is intended for play on slow pace court surfaces
(see Appendix I).

b. Ball Type 2 (medium speed) is intended for play on medium/medium-

fast pace court surfaces (see Appendix 1).

c. Ball Type 3 (slow speed) is intended for play on fast pace court surfaces
(see Appendix D.
 Equipment Rules 411

4. THE RACKET

Rackets failing to comply with the following specifications are not approved
for play under the Rules of Tennis:

a. ihe hitting surface of the racket shall be flat and consist of a pattern of
crossed strings connected to a frame and alternately interlaced or bonded
where they cross; and the stringing pattern shall be generally uniform, and in
particular not less dense in the centre than in any other area. The racket shall
be designed and strung such that the playing characteristics are identical on
both faces.

The strings shall be free of attached objects and protrusions other than those
utilised solely and specifically to limit or prevent wear and tear orvibration, and
which are reasonable in size and placement for such purposes.

b. The frame of the racket shall not exceed 29 inches (73.66cm) in overall
length, including the handle. The frame of the racket shall not exceed 121/2
inches (31.75cm) in overall width. The hitting surface shall not exceed 151/2
inches (39.37cm) in overall length, and 111/2 inches (29.21cm) in overall
width.

c. The frame, including the handle, shall be free of attached objects and
devices other than those utilised solely and specifically to limit or prevent wear
and tear or vibration, or to distribute weight. Any objects and devices must be
reasonable in size and placement for such purposes.

d. The frame, including the handle, and the strings, shall be free of any device
which makes it possible to change materially the shape of the racket, or to
change the weight distribution in the direction of the longitudinal axis of the
racket which would alter the swing moment of inertia, or to deliberately
change any physical property which may affect the performance of the racket
during the playing of a point. No energy source that in any way changes or
affects the playing characteristics of a racket may be built into or attached to a
racket.

The International Tennis Federation shall rule on the question of whether any
racket or prototype complies with the above specifications or is otherwise
approved, or not approved, for play. Such ruling may be undertaken on its
own initiative, or upon application by any party with a bona fide interest there-
in, including any player, equipment manufacturer or National Association or
members thereof. Such rulings and applications shall be made in accordance
with the applicable Review and Hearing Procedures of the International Tennis
Federation (see Appendix IID).

Case 1. Can there be more than one set of strings on the hitting surface of rack-
et?

Decision. No. The rule clearly mentions a pattern, and not patterns, of crossed
strings.
412 Appendix C

Case 2. Is the stringing pattern of a racket considered to be generally uniform

and flat if the strings are on more than one plane?

Decision. No.

Case 3. Can vibration dampening devices be placed on the strings of a racket

and if so, where can they be placed?

Decision. Yes; but such devices may be placed only outside the pattern of the
crossed strings.

Case 4. In the course of play, a player accidentally breaks the strings of his
racket. Can he continue to play with the racket in this condition?

Decision. Yes.

Case 5. Can a battery that affects playing characteristics be incorporated into a

racket?

Decision. No. A battery is prohibited because it is an energy source, as are solar

cells, and other similar devices.
 pendix D
The ———

The simplest model of a bouncing ball is the one described by Howard Brody in his 1984 paper “That's how
the ball bounces.” In that model, Howard assumed that the ball and the surface were both rigid so that there
was no deformation of the ball or the surface. In that case, the horizontal friction force acting backwards on the
ball is given by

= -mdv,/dt

the vertical or normal reaction force is given by

N = mdv Jat

and the rotation speed is given by

FR = Idw/dt

where x and y are the horizontal and vertical coordinates of the center of mass of the ball, is the angular
velocity and I = amR* is the moment of inertia of the ball. For a thin spherical shell, @ = 2/3, but for a tennis
ball with a 6 mm thick wall, a = 0.55. Suppose that the ball spends a time t in contact with the surface before
it bounces. During that time the average values of F and N are respectively F, and N,. We can integrate the above
414 Appendix D

equations over the whole contact period, giving

Pa -m(V,» - vo (Ds)

Nie m(v,, = vy) (DP)

F Rt = I(w, - w,) (D.3)

where subscripts 1 and 2 denote conditions before and after the bounce
respectively.

Provided that the ball slides throughout the bounce, then F = uN at all times
and F, = uN, where uw is the coefficient of sliding friction, so

(Vxq ~ Vig) = MW Wyo - Vy) = ARG, - @)

The ratio e = - v,, wy is called the coefficient of restitution. It is a positive

number with a value of about 0.75 (given that v,, is negative if the ball is inci-
dent downwards onto the surface). Hence v y2 Wi
= -( + eV, sO

V2 cs Vi MW dd v e) V4

and

Ona = NaCl e)v,,/(aR)

It is easier to think about these results if we change the sign of Vy,1 SO that vy,
is a positive number instead of a negative number, then e = Vy2 /Vy, and

Vee Vo lee De Dee) tan») (D.4)

where 6, is the angle of incidence, tan(0,) = va NEN

and

Oo es Ce e)v,,/(aR) (D.5)

While the ball is sliding, the horizontal speed decreases and the rotation speed
increases. The ball will start to roll at the end of the bounce period if Vena RG.
In that case (v,, now being a positive number)

Vp WO + iv = lo, 4+ w+ eWv,,/(aR)] = Rw, + w (1 + ev,,/a

SO. Vit ROp= Ws Od + Ov, 4/4

 The Mathematics of Ball Bounce 415

or

b= a(v,, - Ro JAG + 2) + av, (D.6)

This gives the value of the coefficient of friction that will cause the ball to start
rolling just at the end of the bounce. If the coefficient of friction is larger than
this, then the ball will start rolling before the end of the bounce period. The
friction force will suddenly drop to zero when the ball starts to roll and there
will be no further change in the horizontal speed or rotation speed of the ball.

Suppose that the ball starts to roll at a time t. Suppose also that the horizontal
speed at that time is v, . Since the friction force drops to zero at this time, the
horizontal speed when the ball bounces is also v, , so we can let v,, = v, where
V,, is the horizontal speed of the ball after it bounces. If we replace t in equa-
tions D.1 - D.3 by t, then we will arrive at exactly the same equations as D.4
and D.5, except that u is given by Equation D.6. If u is bigger than the value
given by D.6, then the ball will start rolling at an earlier time, but there is no
change in v,,, since the condition for the ball to roll is given by v,, = Rw, which
leads directly to Equation D.6.

Hence v,,, will be given by v,, = Rw,, exactly as if the ball started rolling at the
end of the bounce period. The relevant values of v,, and w, can be found
using Equation D.6, which can be expressed as

1 Cs Wy = Oye RO, 1 en)

Hence

We = 1 ine) vi ie = ORC ve)

and

@, = 0, + (V,, - ROP/AG + ORI

For example, if w, = 0 and o = 0.55, then v,, /v,, = 0.645 and @, = v,,/(..55R)
Sun) c= Vcr 1,99. = Ven &

The above model works fine if the ball slides throughout the bounce and it is
roughly correct when the ball bites. However, experimental results show that
the ball slides over a slightly larger range of incident angles than predicted
above since the normal reaction force acts ahead of the center of the ball when
the ball is squashed flat on the bottom. In addition, v,, /v,, drops below 0.645
when G0) and when the ball bites.
 a

1e@

oe by) art a, oe
. P
A j ; { iu it‘ ail oe sip =e
iv i lw tremeT? . ait)

a
ge i i . | toes

| 7 et! warn
a = i
7 hit" 9 SW © 00 0D Uy dd
ee ip dus alrrat of) @ ge®

| <= | 74) Coe

7 ; i) by omy, 0.0) eee
ee
| > tert oli rif Fh Th we Ti’ Ge
‘
a |: nit, vq seal
:
4 vr a
_ < reel ciel Cepia wall
a aueet nu ges

lsd digits al
—
— itp

B4t °
he ®

; ee aj iui UH

ee i
we we team ved ats
a eee ia ees
a - ~ -_ ™ ann © Z one Gable oo
4 5 oer win An I\e / Nth etal all wrovtia * 4 -,
~ “ OG mae 4 ~~ aba reicote uh .-
OS 20> aa! “SY oO KD i iit

es
aad ; ar.
ne = — Po _

=o gups * ir av

=
wa _
<> _ ;
 10’ or more

or more

Los > aeons a eee

<—______ 78’ (23.774 m)

All measurements are made to the outside of the lines

 The following books and journal articles are suggested for readers interested in further information on the
physics and technology of tennis. They do not represent all that has been written on the subject but each con-
tains its own list of references to previous work (and each of those in turn will contain their own list of refer-
ences). Articles on tennis can be traced back to Sir Isaac Newton and Lord Raleigh. Taken together, the reader
will discover that there is a vast amount of information now available and a lot that is still not properly under-
stood. Articles labelled with a ¢ are especially recommended.

_ Aerodynamics _ oN

Cooke, A.J. (2000) An overview of tennis ball aerodynamics, Sports Engineering, 3, 123-129.

De Mestre, N. (1991) The Mathematics of Projectiles in Sport, Cambridge University Press, New York.

Mehta, R.D. (1985) Aerodynamics of sports balls, Annual Review of Fluid Mechanics, 17, 151-189.

¢ Mehta, R.D. and Pallis, J.M. (2001) The aerodynamics of a tennis ball, Sports Engineering, 4, 177-189.

Stepanek, A. (1988) The aerodynamics of tennis balls-the topspin lob, American Journal of Physics, 50, 138-142.

Zayas, J.M. (1986) Experimental determination of the coefficient of drag of a tennis ball, American Journal of
Physics, 54, 622-625.
420 Appendix F

_Ball Bounce _
Brody, H. (1984) That’s how the ball bounces, The Physics Teacher, 22, 494-
497,

Brody, H. (1990) The tennis-ball bounce test, The Physics Teacher, 28, 407-409.

Cross, R.C. (1999) Dynamic properties of tennis balls, Sports Engineering, 2, 23-
Sek

Cross, R.C. (1999) The bounce of a ball, American Journal of Physics, 67, 222-
22am

Cross, R.C. (2000) The coefficient of restitution for collisions of happy balls,
unhappy balls and tennis balls, American Journal of Physics, 68, 1025-1031.

e Cross, R.C. (2002) Grip-slip behaviour of a bouncing ball, American Journal

of Physics, 70, in press.

Cross, R.C. (2002) Measurements of the horizontal coefficient of restitution for

a superball and a tennis ball, American Journal of Physics, 70, 482-489.

Garwin, R.L. (1969) Kinematics of an ultraelastic rough ball, American Journal

of Physics, 37, 88-92.

Groppel, J.L., Dillman, C.J. and Lardner, TJ. (1983) Derivation and validation
of equations of motion to predict ball spin upon impact in tennis, Journal of
Sports sciences, 1, 111-120.

Groppel, J.L., Shin, I-S., Spotts, J. and Hill, B. (1987) Effects of different string
tension patterns and racket motion on tennis racket-ball impact, International
Journal of Sport Biomechanics, 3, 142-158.

Maw, N., Barber, J.R. and Fawcett, J.N. (1976) The oblique impact of elastic
spheres, Wear, 38, 101-114.

Maw, N., Barber, J.R. and Fawcett, JIN. (1981) The role of elastic tangential
compliance in oblique impact, Journal of Lubrication Technology, 103, 74-80.

Rose, P., Coe, A. and Haake, S.J. (2000) The variation of static and dynamic ten-
nis ball properties with temperature, in Tennis Science and Technology (Eds
S.J. Haake and A. Coe), Blackwell Science. pp.169-174.

Bahamonde, R.E. (2000) Changes in angular momentum during the tennis

serve, Journal of Sports Sciences, 18, 579-592.

Elliott, B. and Christmass, M. (1995) A comparison of the high and low back-
spin backhand drives in tennis using different grips, Journal of Sports Sciences,
13, 141-151.

Elliott, B.C, Marshall, R.N. and Noffal, GJ. (1995) Contributions of Upper Limb
Segment Rotations During the Power Serve in Tennis, Journal of Applied
Biomechanics, 11, 433-442.
 References 421

Elliott, B.C, Marshall, R.N. and Noffal, GJ. (1996) The role of upper limb seg-
ment rotations in the development of racket-head speed in the squash fore-
hand, Journal of Sports Sciences, 14, 159-165.

e Tennis Science and Technology, (Eds S.J. Haake and A. Coe), Blackwell
Science. Published as the proceedings of the 1st International Congress on
Tennis Science and Technology, London, August 2000.

Arthur, C. (1992) Anyone for slower tennis?, New Scientist, 2 May, 24-28.

Barger, V. and Olsson, M. (1973) The tennis racket theorem, in Classical

Mechanics, McGraw-Hill, New York. pp 226-232

Bowers, C. (2002) The Book of Tennis, JWM Publishing Ltd, London (see
www .thebookoftennis.com).

e Brody, H. (1990) Tennis Science for Tennis Players, University of

Pennsylvania Press, Philadelphia.

e Brody, H. and Braden, V. (1992) The Science and Myths of Tennis (video).

Brody, H. and Cross, R. (2000) Proposals to slow the serve in tennis, in Tennis
Science and Technology, (Eds S.J. Haake and A. Coe), Blackwell Science, pp.
261-268.

Haake, SJ., Chadwick, S.G., Dignall, RJ., Goodwill, S. and Rode, P. (2000)
Engineering tennis - slowing the game down, Sports Engineering, 3, 131-143.

Brody, H. (1992) Radar Guns and Tuning Forks, The Physics Teacher, 30, 102.

Knudson, D. and White, S.C. (1989) Forces on the hand in the tennis forehand
drive: Application of force sensing resistors, International Journal of Sport
Biomechanics, 5, 324-331., 9, 143-148.

Brannigan, M. and Adalai, S. (1981) Mathematical modelling and simulation of

a tennis racket, Medicine and Science in Sports and Exercise, 13, 44-53.

¢ Brody, H. (1979) Physics of the tennis racket, American Journal of Physics,

47, 482-487.

Brody, H. (1985) The moment of inertia of a tennis racket, The Physics Teacher,
23, 213-216.

Brody, H. (1987) Models of Tennis Racket Impacts, International Journal of

Sports Biomechanics, 3, 293-290.

Brody, H. (1989) Vibration damping of tennis rackets, International Journal of

Sport Biomechanics, 5, 451-450.
422 Appendix F

Brody, H. (1995) How would a physicist design a tennis racket?, Physics Today,
48, March, 26-31.

Brody, H. (1996) The modern tennis racket, Proceedings of the Ist

International Conference on the Engineering of Sport, Sheffield, UK. Edited by
S. Haake. Published by A. Balkema, Rotterdam pp.79-82.

Brody, H. (1997) The physics of tennis. III. The ball-racket interaction.,

American Journal of Physics, 65, 981-987.

Brody, H. (2000) An Overview of Racket Technology, in Tennis Science and

Technology (Eds S.J. Haake and A. Coe), Blackwell Science pp. 43-48,

Brody, H. (2000) Player sensitivity to the moments of inertia of a tennis rack-

et., Sports Engineering, 3, 145-148.

Brody, H. and Smith, S. (2001) Revolutionary Rackets, in World-Class Tennis

Technique Roetert and Groppel editors, Human Kinetics, Champaign IL.

Casolo, F. and Ruggieri, G. (1991) Dynamic analysis of the ball-racket impact

in the game of tennis, Meccanica, 26, 67-73.

Cross, R.C. (1997) The dead spot of a tennis racket, American Journal of
Physics, 65, 754-764.

¢ Cross, R.C. (1999) Impact of a ball with a bat or racket, American Journal of
Physics, 67, 692-702.

Cross, R.C. (2000) Effects of friction between the ball and strings in tennis,
Sports Engineering, 3, 85-97.

Cross, R.C. (2000) Flexible beam analysis of the effects of string tension and
frame stiffness on racket performance, Sports Engineering, 3, 111-122.

Cross, R.C. (2001) Customising a tennis racket by adding weights, Sports

Engineering, 4, 1-14.

Elliot, B.C. 1981) Tennis racquet selection: a factor in early skill development,
The Australian Journal of Sport Sciences, Sept., 23-25.

Elliott, B.C. (1981) Tennis: the influence of grip tightness on reaction impulse
and rebound velocity, Medicine and Science in Sports and Exercise, 14, 348-
aya

Elliott, B.C., Blanksby, B.A. and Ellis, R. (1980) Vibration and rebound veloci-
ty characteristics of conventional and oversized tennis rackets, Research
Quarterly for Exercise and Sport, 51, 608-615.

Grabiner, M.D., Groppel, J.L and Campbell, K.R. (1983) Resultant tennis ball
velocity as a function of off-center impact and grip firmness, Medicine and
Science in Sports and Exercise, 15, 542-544.

Groppel, J.L., Shin, I-S., Thomas, J. and Welk, GJ. (1987) The effects of string
type and tension on impact in midsized and oversized tennis racquets,
International Journal of Sport Biomechanics, 3, 40-46.
 References 423

Hatze, H. (1993) The relationship between the coefficient of restitution and

energy losses in tennis rackets, Journal of Applied Biomechanics, 9, 124-142.

Kotze, J., Mitchell, S.R. and Rothberg, S.J. (2000) The role of the racket in high-
speed tennis serves, Sports Engineering, 3, 67-84.

Kuebler, S., (2000) Book of Tennis Rackets, Kuebler GmbH, Germany (see
www .racketantiques.com)

Leigh, D.C. and Lu, W.-Y. (1992) Dynamics of the interactions between ball.
strings and racket in tennis, International Journal of Sport Biomechanics, 8,
181-206.

Mitchell, S.R., Jones, R. and King, M. (2000) Head speed vs racket inertia in the
tennis serve, Sports Engineering, 3, 99-110.

Widing, M.A. and Moeinzadeh, M.H. (1990) Finite element modeling of a ten-
nis racket with variable string patterns and tensions, International Journal of
Sports Biomechanics, 6, 78-91.

Adair, R.K. (1995) The physics of baseball, Physics Today, May, 26-31.

Adair, R.K. (2002) The physics of baseball, HarperCollins, New York, 3rd edi-
tion.

Armenti, A. (1992) The Physics of Sports, AIP Press, American Institute of

Physics, New York.

Bahill, T. and Karnavas, W. (1991) The ideal baseball bat, New Scientist, 6 April.

Brancazio, P. (1984) Sport Science, Simon and Schuster, New York.

Brody, H. (1990) Models of baseball bats, American Journal of Physics, 58, 756-
758.

Daish, C.B. (1972) The Physics of Ball games, The English Universities Press
Ltd, London.

Jorgensen, T.P. (1999) The Physics of Golf, Springer-Verlag, New York.

Kagan, D. (1990) The effects of coefficient of restitution variations on long fly

balls, American Journal of Physics, 58, 151-154.

Kirkpatrick, P. (1963) Batting the ball, American Journal of Physics, 31, 606-613.

Nathan, A.M. (2000) Dynamics of the baseball-bat collision, American Journal

of Physics, 68, 979-990.

Noble, L. and Walker, H. (1994) Baseball bat inertial and vibrational character-
istics and discomfort following ball-bat impacts, J. Applied Biomechanics, 10,
132-144.
424 Appendix F

Van Zandt, L.L. (1992) The dynamical theory of the baseball bat, American
Journal of Physics, 60, 172-181.

Watts. G. and Bahill, A.T. (1990) Keep your eye on the ball, the science and
folklore of baseball, W.H. Freeman & Co, New York.

Carter, W.H. and Crews, S.L. (1974) An analysis of the game of tennis, The
American Statistician, 28, 130-134.

Fischer, G. (1980) Exercise in probability and statistics, or the probability of

winning at tennis, American Journal of Physics, 48, 14-19.

Magnus, J.R. and Klaassen J.G.M. (1999) The effect of new balls in tennis:
four years at Wimbledon, The Statistician, 48, 239-240.

Magnus, J.R. and Klaassen,J.G.M. (1999) On the advantage of serving first

in a tennis set: four years at Wimbledon, The Statistician, 48, 247-250.

Pollard, G.H. (1983) An analysis of classical and tie-breaker tennis, Australian

Journal of Statistics, 25, 496-505.

Bower, R. and Cross, R.C. (2003) Sensitivity of tennis players to changes in

string tension in a tennis racquet, Journal of Science and Medicine in Sport, in
press.

Bower, R. and Sinclair, P. (1999) Tennis racquet stiffness and string tension
effects on rebound velocity and angle for an oblique impact, Journal of Human
Movement Studies, 37, 271-286.

Brody, H. and Knudson, D. (2000) A model of tennis stroke accuracy relative

to string tension, International Sports Journal, 4, 38-45.

Cross, R and R. Bower (2001) Measurements of string tension in a tennis rac-

quet, Sports Engineering, 4, 165-175.

Cross, R.C. (2000) Tension loss along a string, American Journal of Physics, 68,
NSPS.

Cross, R.C. (2001) Why bows get stiffer and racquets get softer when the strings
are added, American Journal of Physics, 69, 907-910.

Cross, R.C., Lindsey, C. and Andruczyk, D. (2000) Laboratory testing of tennis

strings, Sports Engineering, 3, 219-230.

Knudson, D. (1991) Effect of string type and tension on ball vertical angle of
pees in static tennis impacts, Journal of Human Movement Studies, 20, 39-
Te .

Knudson, D. (1993) Effect of string tension and impact location on ball

rebound accuracy in static tennis impacts, Journal of Applied Biomechanics, 9
143-148. a
 References 425

Knudson, D. (1997) Effect of string tension on rebound accuracy in tennis

impacts, International Sports Journal, 1, 108-112.

Putnam, C.A. and Baker, J.A. (1984) Spin imparted to a tennis ball during
impact with conventionally and diagonally strung rackets, Research Quarterly
for Exercise and Sport, 55, 261-266.

Stroede, C.L., Noble, L. and Walker, H.S. (1999) The effect of tennis racket
string vibration dampers on racket handle vibrations and discomfort following
impacts, Journal of Sports Sciences, 17, 379-385.

Thornhill, P.M., Baker, J.S. and Cooper, S-M. (1993) Interaction between tennis
string type and tension and the subsequent effect upon ball rebound velocity,
Journal of Human Movement Studies, 24, 157-167.

Blievernicht, J.G. (1968) Accuracy of the tennis forehand drive; cinemato-

graphic analysis, Research Quarterly, 39, 776-779.

Brody, H. (1998) Improving Your Serve, in The Engineering of Sport (S. Haake
editor) Sheffield England. pp.311-316.

Brody, H. (2000) Unforced Errors and Error Reduction in Tennis, in Tennis

Science and Technology (Eds S.J. Haake and A. Coe), Blackwell Science. pp.
305-312.

Adair, R.K. (2001) Comment on the sweet spot of a baseball bat, American
Journal of Physics, 69, 229-232.

Brody, H. (1981) Physics of the tennis racket II: The "sweet spot", American
Journal of Physics, 49, 816-819.

Brody, H. (1986) The sweet spot of a baseball bat, American Journal of Physics,
54, 640-643.

Cross, R.C. (1998) The sweet spot of a baseball bat, American Journal of
Physics, 66, 772-779.

¢ Cross, R.C. (1998) The sweet spots of a tennis racquet, Sports Engineering,
1, 63-78.

Cross, R.C. (2001) Response to “Comment on the sweet spot of a baseball bat",
American Journal of Physics, 69, 231-232.

Hatze, H. (1994) Impact probability distribution, sweet spot, and the concept
of an effective power region in tennis rackets., Journal of Applied
Biomechanics, 10, 43-50.

Brody, H. (1992) Measuring the Softness of an Athletic Surface, The Physics

Teacher, 30, 28-31.
426 Appendix F

Cross, R.C. (2001) The dual function of sand on a clay tennis court, The Physics
Teacher, 39, 330-331.

¢ Cross, R.C. (2003) Measurements of the horizontal and vertical speeds of ten-
nis courts, Sports Engineering, 5, in press.
 427

A Ball speed
Accelerometer 87 and swing speed 116
ACOR. See Apparent coefficient of restitution ball speed formula using COR 119-120
Air resistance how to use formulas 120
ball 331 ball speed formula using ACOR 125-126, 165
in freefall 150 effect of COR 129
Air resistance of string 239-240 effect of racquet mass 116, 128-130

Amplitude. See vibration energy and ball speed 116

Angular momentum. See momentum free fall speed formula 118
Angular speed. See rotation maximum possible serve speed 129-130
Anisotropic 15, 18 measuring groundstroke speed 172-173, 388
Apparent coefficient of restitution 6 serve speed
ACOR bounce height formula 122 and ACOR 292
and energy return 121-123 increasing ball bounce 293-294
bounce height scenarios 122 lowering tension 294, 296-297

definition 121 measuring serve speed 169-172

formula 121 radar gun 169-170
frames of reference 165-166 stop watch 170-172

measuring 122-123 VCR/camcorder 170

power point 58 serve speed formula using ACOR 124-125, 202-203

variation with impact location 122, 293 tension 153-154, 294,296-297

Approach shot 218-219 Balls

Atomic bonds 253-255, 268 air resistance 331

ball stiffness 298
B big balls 331-332
Balance point bounce
center of mass 9 fast courts and slow courts 341-342
choking up 27 horizontal velocity 340-341

customizing 28 vertical velocity 339-340

customizing table 29 force at impact 151

defined 24 high altitude 330-331

determining 25 larger balls 188
feel 9 pressureless 330
formula 25 pressurized 330
length 24 stiffness 150-151
spreading fingers and 27 tests for bounce
torque 26-28 compression test 333-334
for spinning balls 336-337
428 Index

slow vs fast compression 334-336 linear momentum 52

Type 2 ball 188 locating by tapping 55

Type 3 ball 188 racquet motion 75-76

vibration 336 racquet recoil 52
Baseball swing 5 racquet rotation 52
Basketball 362 sweetspot 52-55
Bending wave. See vibration Coefficient of friction (COF) 340, 357

Bounce Coefficient of restitution. Also see energy

and COR 349-350 bounce effect 349-350
and impact location 157 bounce height formula 118
biting 359-366 calculating from bounce height 118
basketball 362 defined 117-119
braking motion 364-365 formulas 117, 118
experiment to measure 359-360 Collisions
friction force 360-364 elastic collision 114, 116, 118
high friction surface 361-362 inelastic collision 118, 121
low friction surface 360-361 involving rotating racquet 138-140
superball 362 Composite 16
biting vs rolling 371 Compression springs 147
fast courts and slow courts 341-342 Conjugate point. See center of percussion
horizontal velocity 340-341 Conservation of energy. See energy
metal ball on strings 156-157 Conservation of momentum. See momentum
off racquet 369-380 Control
bounce for a serve 377-379 asymmetric string deformation 247
calculating bounce 376-377 dwell time 246
speed, spin and angle 373-376 tension 246
rolling 347-348 Copolymers 257
rules of spin 370 COF. See coefficient of friction
sliding 346-347 COP. See center of percussion
factors determining 346 COR. See coefficient of restitution
sliding vs biting 371 Court speed 350-357
tests for balls 333-336 measuring coefficient of friction 357
vertical and horizontal speeds 339-340, 349-350 measuring different surfaces 350-353
factors determining 350 measuring spin 354
Breaking energy 262 measuring vertical speed 354-356
pace 350
C perception of 366
Carbon. See graphite Sestee machine 350
Cantilever beam 78 Courts
Carbonization 15 clay court 356
Center of mass 9 Deco Turf 352
Center of percussion Rebound Ace 352, 356
angular momentum 52 Covalent bonds 254
conjugate point 53 Customization
feel 7 balance point 176-177
force on hand 54 change in balance formula 176
formulas 54 locating 176
 Index 429

matching racquets 176-177 formula 114

lead tape 175 difference from momentum 115-117
length 34 elastic energy 130-133
power 178 amount stored derivation 131-132
shifting power point 178 amount stored formula 132
weight locations 178 energy sharing ball and strings 152-153
stability 178-179 energy in strings formula 295
string tension 179 kinetic energy 112-114
weight locations 178-179 loss 6
sweetspots 179-180 loss strings and ball 160
COP shifting 179-180 potential energy 149-150
node shifting 179 racquet energy 298-299
swingweight 177-178 sharing between ball and strings 295
axis of rotation 177-178 sharing between racquet and ball 297
change in swingweight formula 177 stored vs lost 294-295
tables 29, 32 tension and energy storage 296
work-energy theorem 112-113
D Epoxy. See resin
Dampening. See vibration Extension spring 147
Deadspot 59
effective mass 59 F

Diving board mode 78 Fiber count 269

Drag force 383-384 First moment 26
drag coefficient 384 Force
drag crisis 384 first law of motion 100-101
formula 384 formula 23
Draw ratio 270 frictional force 340
Drop shot mass and inertia 23-24
hitting drop shots 221-223 normal force 340
spin 222 Force gauge 87
trajectories 222-223 Frame distortion 314-315
returning drop shots 223-224 Frames of reference 165
racquet speed 223 changing 123
trajectories 223 Free beam 78-79
Dwell time. See impact duration Free fall 150, 382-383
Friction
E and biting balls 360-364
Effective mass. See mass coefficient of friction 340, 357
Elastic energy. See energy factors determining 345-346
Elastomer 69 in stringing 315
Energy 6 rolling friction 344
and racquet mass 116 sliding friction 344
coefficient of restitution 117-119 static friction 344
bounce height formula 118 Fundamental mode. See vibration
calculating from bounce height 118
formulas 117, 118
conservation of energy 114
430 Index

Impulse. See momentum

G
Inertia 100
Gauge
and elasticity 245 Inertial mass 100

and strength 245 Inertial motion 100

defined 244 Isotropic 15, 18

spin 370
stiffness 321-326 L
Laws of motion. See Newton’s laws
Graphite
carbon fiber 15-16 Layup of racquet 18

composite 16 Lift coefficient 385

fiber/matrix ratio 16 Linear momentum. See momentum

lay-up of racquet 18
making graphite 15-17 M
prepreg 16-17 Magnus force 384-385

raw materials 15 formula 385

Gravity lift coefficient 385

force in freefall 150 Mass

Grip Size 5 effective mass

Grommets and impact point 140-142

and string length 248-249 deadspot 59

purpose 248 estimating ACOR from 142-143

size 248 formula 140
Gut formula for impacts along center line 144-145
manufacturing 267 formula for impacts anywhere 145-146
properties 266-267 in formulas 99
inertial mass 100
H mass vs weight 35
Headsize Modes of oscillation. See vibration
polar moment 10 Modulus
sweetspot 10 and strength 17
Huygens, Christian 113 categories 14-15
Hysteresis curve 334 defined 14
formula 14
| how to make 15-17
Impact duration of strings 244
and bounce 157 Moment 26
and control 241-242 Moment of inertia. Also see swingweight
and serve speed 203 calculating for common objects 136-137
and string deformation 241 formula 136
and swing speed 241 resistance to rotation 136-137
shock 92-93 Momentum
Impact location angular momentum
and court type 190 conservation of 138
and stroke type 190-191 definition 137
determining 190 formula 137
in transverse direction 191 definition 103
spin 191 impulse/change in momentum theorem 103-108
 Index 431

formulas 104, 111 in modern racquets 185-186

law of conservation of momentum 108-109 mass 160
formula 108 maximum possible 163
linear momentum 52 power point 161
racquet speed 160-161
N stiffness 166-167
Newton's first law 100-101 string gauge 288-289
Newton's second law 101 string length 287-288
formula 101 string tension 283-287
Newton's third law 101-103 bow and arrow analogy 286-287
formulas 102 steel balls and bounce height 283-285
Node tennis balls and bounce height 285-287
amplitude 55 swing 161
feel 7 Power point 80
fundamental mode 55 and impact location 59
locating the node 57 and impact point 190
node line 57 and the node 80
sweetspot 55-58 and racquet rotation 135
vibration modes 55 apparent coefficient of restitution 58
Notching 245-246 sweetspot 8, 58-59
Prepreg 16-17
O curing 16
Oscilloscope 64, 87 thin prepreg 17
unidirectional graphite prepreg 16
P Prestretching 306
Pace of court 350 Proportional stringing 303
Patterns, string 246 Pull angle 308-309, 315
Parallel axis theorem 38, 45 Pythagorean theorem 154
Period 74, 77
Pickup weight 26 R
Piezoelectric disk 87 Racquet speed
Piezoelectric 20-21 final racquet speed formula 120
Player's racquets 4 moving racquet vs racquet at rest 165-166
Poisson's ratio 245 swinging machines 166
Polar moment. See twistweight Radar gun 169-170
Polyacrylonitrile (PAN) 15 Radian 134
Polymers 252-253 Recoilweight
Potential energy formula 50
bow and arrow 149 measuring 50
elastic potential energy 149 stability 38
in springs 150 Recreational racquets 4
KE converted to PE 150 Resin
PE converted to KE 150 epoxy 15, 16
stored in a mass 149-150 fiber reinforced matrix 15
Power in string construction 268-269
factors 160-161 strength of 16
from strings 242-243 thermoplastic resins 15
432 Index

thermosetting resins 15 recoil of handle 91-92

Restorative force 62 relation to vibration 68-69
Rotation stiffness 81-82
and effective mass 140-142 string tension 82-83
and maximum power point 135 Sizing 15
and moment of inertia 136 Slice shots 217-218
and swingweight 136 Spaghetti strings 348-349
angular speed 134 Spin
angular vs actual speed 134-135 and gauge 370
radian 134 and tension 370
rebound angle 5 approach shot 218-219
bounce 214
S bounce angles 215-216
Serve drop shots 222
effect of impact duration 203 durability of string and spin 245
maximum speed 129-130 impact location 191
measuring serve speed 169-172 off racquet 373-376
motion 201-202 margin of error 215
arm rotation 203-204 returning spin shots 216-217
receiving serve 209 returning with same spin 214
location and time 209-211 reversing spin 214
serve speed formula 6, 124-125, 202-203 rules of 370
See also ball speed slicing the slice 217-218
serve speed strategy 204-205 spin after bounce 216
first and second serve speed 206-207 spin into the strings 371-373
second serve speed 205 stroke speed 214
vertical angle of acceptance topspin vs backspin 213-215
defined 194 watching opponent's spin 215
effect of higher ball toss on 197-198 Springs
effect of impact height on 195-196 ball as spring 150-151
effect of location on court on 198-199 ball stiffness 150-151
effect of longer racquet on 199-200 force at impact 151
effect of serve speed on 195 compression springs 147
effect of swingweight on 199 diameter and stiffness 149
effect of topspin on 196 energy stored formula 150
Serve speed formula 6. See ball speed extension/tension springs 147
Sestee machine 350 force formula 150
Shock 64 in parallel 149, 326
and feel 90-91 in series 148
absorption 69 length and stiffness 148
elastomer 69 spring constant 147
force on hand 95-96 stiffness 147
impact location 81 stiffness and force 148
isolation 69 Stiffness formula 148
maximum shock 89 string as spring 151-152
measurements 83-86 potential energy 152
minimum shock 89 tension and ball speed 153-154
 Index 433

tension and stiffness 152 copolymers 257

Stability 38-42 functional groups 255
and weight 39 impact behavior 257-258
comfort 40 kevlar 256-257
control 40 material formulas 256
effects of 39 nylon 256
in modern racquets 185 polymer blends or alloys 257
polarization index 43 polymers 252-253
power 40 backbone 253
recoilweight 38 crystallinity 253, 256
stability 38 defined 252
stability table 43 polyethylene 252
sweetspot 41 polymerization 252
swingweight 38 responses to tension 253
twistweight 38 viscoelasticity 257
weight polarization 38-39 construction
Standing waves 76 coatings 269-270
Static moment 26 durability 269
Stevens machine 334 lubrication 269
Stiffness. core 265-266
and fiber direction 17 draw ratio 270
and vibration 66 fiber count 268
by weight 37, 38, 262-263 resin 268-269
deflection vs stiffness 275, 278 bonding 268
duration vs stiffness 276, 278 effect on performance 269
dynamic stiffness 272-274 types 268
elongation vs stiffness 273 void volume 268-269
force vs stiffness 275 twist angles 267-268
peak force vs stiffness 275 durability 267
peak tension vs stiffness 275 elasticity 267-268
sideways in string 277 energy return 268
springs 148, 152 wraps 266-267
tension increase vs stiffness 275 durability 245-246
vibration 93-94 and spin 245
cross-section 10 impact location 245
vibration frequency 11 notching 245-246
widebody 11 materials
Strain 14 breaking energy 262
Strength by weight 262-263 families 259-260
Stress 14 stiffness 260
String stiffness by weight 262-263
chemistry strength 260
bonding 253-255 strength by weight 262-263
between molecules 254-255 tension/stretch curves 260-262
covalent bonds 254 patterns 246
secondary bonds 255 properties 271-282
within a molecule 254 deflection - linear vs perpendicular 277
434 Index

deflection vs stiffness 275 and stokes 30-31

deflection vs stiffness at different tensions 278 customization table 32
duration vs stiffness at different tensions 278 customizing 31-32
dynamic stiffness 272-274 explained 28-30
defined 272 feel 9-10
fast vs slow stretches 273 formula 30
material dependence 273-274 mass distribution 30
measuring 272-273 measuring 46-47
elongation vs stiffness 273 pendulum formula 46
force vs stiffness at different tensions 278 stability 38
impact force and duration 274-276
+
impact tension loss 277
measuring 271-272 Tension

new strings vs old strings 279, 292 determining stringbed tension 316-317
peak force vs stiffness 275 determining tension in strings 317
peak tension vs stiffness 275 effect on ball speed 153-154
power 279-282 explained 302-303
and age of string 279 increase at impact 160
effect of dwell time on 280 measuring
psychological perceptions of 280 calibration 314
trajectory and power 280-282 diagnostic machines 311-312
static tension loss 277 ERT700 313-314
tension increase vs stiffness 275 Stringmeter 312-313, 319-320
tension rise vs stiffness at different tensions 278 player's perceptions of 315-316
Stop watch 170-172 spin 370
String plane tension loss
deformation 240 and tension head location 309
and tension 240 bond slipping 314
measuring 240 creep
String vibration 94-95 how it happens 303-305
Strokes what it is 303
classic 184 during stringing 305-306
influence of courts 186 effect of prestretching 306
modern 184 effect of pull angle 308-309
western grip 184 effect of temperature 309
Superball 362 effect of time 307-308
Sweetspot frame distortion 314-315
center of percussion (COP) 7, 52-55 friction 315
customizing 190 impact tension loss 277
feel 88-89 pull angle 315
in modern racquets 184-185 static tension loss 277
node 7, 55-58 tension vs elongation 323-324
power point (COR) 8 shock 82-83
size 89-90 Tennis elbow 40, 67-68
Swing speed 116, 214, 241 Thermoplastic 15
Swingweight Thermoset 15
and customizing length 34 Thin prepreg. See graphite
 Index 435

Tie-breaks 187 string tension 234-235

Time lateral errors 229-231
first serve importance in doubles 227 timing errors 231
going to the net 226 acceptance window and hitting height 232
hitting deep 226 acceptance window and shot selection 233
hitting early 226 acceptance window and spin 232-233
hitting hard 225-226 acceptance window and velocity 231-232
hitting location 226-227 Unidirectional graphite. See graphite
lobs and spin 227
Titanium 18-19 V
and golf 19 VCR 170
mechanical properties 19 Vertical angle of acceptance. See serve
Topspin and the serve
topspin and window of acceptance 196 defined 194
topspin vs backspin 213-215 effect of higher ball toss on 197-198
Torque 26, 41 effect of impact height on 195-196
Tows 15 effect of location on court on 198-199
Trajectories effect of longer racquet on 199-200
and power 280-282 effect of serve speed on 195
drag force 383-384 effect of swingweight on 199
drop shot 222-223 effect of topspin on 196
free fall 382 and timing errors
horizontal with freefall 382-383 acceptance window and hitting height 232
Magnus force 384-385 acceptance window and shot selection 233
serve trajectories 385-386 acceptance window and spin 232-233
trajectory equations 387-388 acceptance window and velocity 231-232
Transverse wave 77 Vibration 73
Twist angles in string 267-268 adding weights 66
Twistweight 39-40 amplitude 62, 64, 65
head width 40 bending wave 68, 79
maneuverability 40 center of percussion 64, 75-76
measuring 48-49 cycle 74
pendulum formula 48 dampening 56, 57, 64, 67
polar moment 39, 235-236 hand as dampener 67
stability 38 stiffness effects 67
tennis elbow 40 vibration dampeners 67
torsion pendulum 39, 49 description 62
experiments 83
U feel 4
Unforced errors frame bending 73-74
ball angle 230 frequency 57, 64, 65-66, 76-77
ball angle and head speed 230 frequency and amplitude 86-88
depth errors 231-234 fundamental mode 76
equipment and error reduction 234-236 hand-held vs clamped racquet 62
balance and sweetspot 236 high frequency demonstration 66
headsize and polar moment 235-236 injury 67-68
spin 235 maximum vibration 89
436 Index

measurements 83-86, 87 Weight

minimum vibration 89 and force 24
modes of a beam 77-79 center of mass 26
cantilever beam 78 pickup weight 26
diving board mode 78 first moment 26
free beam 78-79 static moment 26
modes of oscillation 62, 63 power 5-6
node 64, 65, 74 stiffness-to-weight ratio 37, 38
period 74, 77 strength-to-weight ratio 37, 38
restorative force 62 technological advances 37-38
shapes 74-75 torque 26
standing waves 76 weight vs mass 35
stiffness 66, 93-94 Wood racquets 183
string dampening 57 and speed of the game 186
string vibration 94-95, 240-241, 313-314 Work
strung vs unstrung 66 and energy 112
transverse wave 77 formula 112, 149
vibration vs impact location 56 work-energy formula 113
wave speed 77 work-energy theorem derivation 112-113
wave speed formula 77 Wraps 266-267
Viscoelasticity 257
Void volume 268-269 Y
Young's modulus 77
W
Wave speed 77
 437

> € >
Authors

aunt
cease
ecuenscennenenceteteate
ntaamnmnmuaemnerntcetctscc0/97
eecaesectantetammmmmmneetrecesenctoceeteecctitie SSB
INN
BRASS
DIUM
SEUSS
eile
i

Howard Brody, Crawford Lindsey, Rod Cross

Dr. Howard Brody is an emeritus professor of physics at the University of Pennsylvania, where he was inter-
im varsity tennis coach for part of the 1991 season. He played varsity tennis and earned his bachelor’s degree
at MIT and his master’s and doctoral degrees at Cal Tech. For many years his primary field of research was ele-
mentary particle physics and high energy nuclear physics. Professor Brody is a member of the International
Tennis Federation Technical Commission, the USTA Sports Science Committee, science advisor to the
Professional Tennis Registry, technical advisor to the United States Racquet Stringers Association, on the Editorial
Board of the Journal of Sports Engineering, and on the technical advisory panel of Tennis Magazine. His book
Tennis Science for Tennis Players was published by the University of Pennsylvania Press in 1987. He and Vic
Braden are featured in a video “The Science and Myths of Tennis.” He has received the USPTR Plagenhoef award
for sports science in 1996 and the International Tennis Hall of Fame Educational Merit Award for the year 2000.

Rod Cross is Associate Professor in Physics at the University of Sydney, Australia. Up until 1995 Rod's main
research interest was the physics of Alfven waves in plasmas at temperatures above one million degrees. He
switched to the physics of sport in 1996 when government research funding became scarce. His paper in the
American Journal of Physics on the sweetspots of a baseball bat was voted as one of the top 75 papers in physics
in the USA in 1998. Rod's main research interest now is the physics of tennis, which he has played competitively
since 1950. He has published about 25 papers on this subject in physics and engineering journals and has writ-
ten extensively for USRSA's Racquet Tech on the physical properties of tennis strings and tennis racquets.

Crawford Lindsey is the publisher and managing editor of Racquet Tech magazine and The Stringer's Digest
(publications of the USRSA). He is also the web developer/master of www.racquettech.com, for which he has
developed extensive software programs dedicated to racquet selection, customization, optimization and string-
ing. His bachelor's degree is from Princeton University and he has done further studies in publishing, education
and business. He was a product manager for squash and racquetball racquets as well as tennis and ski acces-
sories at Head Sports. His book, The Book of Squash, was published in 1987, and he has authored two other
books on unrelated subjects. He has written extensively on the science and technology of tennis in Racquet Tech
magazine
 he life's purpose of a tennis racquet is to change the speed and direction of a tennis ball. This
momentous event takes place in approximately five-one-thousandths of a second. During that
time, the ball and the racquet launch into a frenzy of technology-enhanced physics:

The strings stretch, deflect, vibrate, store and return energy.

The ball squashes, slides, bites, reverses direction, ejects,
and flies away like vibrating jello.
The racquet decelerates, rotates, bends, twists and vibrates.
Bending waves race up and down the racquet, slamming the handle into the hand.
Pain, frustration, anger, joy, or elation explode forth from the collision.

Players paying upwards of $300 per racquet and $50 for restringings can spend hundreds of dollars
every year trying to influence any of these events in their favor. This book explains the what, how and
why of those fateful collisions on each side of the net, as well as the flight and the bounce of the ball
in-between.

y & Rod Cross

USA $29.95
ISBN 0-9722759-0-8

- RACQUET TECH PUBLISHING oo : | "uf

. AN Imprint.of the USRS. 2 pac ee