Sensors and

transducers

2.1 Sensors and The term sensor is used for an element which produces a signal
transducers relating to the quantity being measured Thus in the case of; say,
an electrical resistance temperature element, the quantity being
measured is temperature and the sensor transforms an input of
temperature into a change in resistance. The term transducer is
often used in place of the term sensor. Transducers are defined as
elements that when subject to some physical change experience a
related change. Thus sensors are transducers. However, a
measurement system may use transducers, in addition to the
sensor, in other parts of the system to convert signals in one form
to another form.
This chapter is about transducers and in particular those used as
sensors. The terminology that is used to specify the performance
characteristics of transducers is defined and examples of
transducers commonly used in engineering are discussed.

2.2 Performance The following terms are used to define the performance of
transducers, and often measurement systems as a whole.
terminology
1 Range and span The range of a transducer defines the limits
between which the input can vary. The span is the maximum
value of the input minus the minimum value. Thus, for
example, a load cell for the measurement of forces might
have a range of 0 to 50 kN and a span of 50 kN.

2 Error Error is the difference between the result of the

measurement and the true value of the quantity being
measured.

Error = measured value - true value

Thus if a measurement system gives a temperature reading of

25°C when the actual temperature is 24"C,then the error is
+l°C.If the actual temperature had been 26OC then the error
18 Mechatronics Sensors and transducers 19

would have been -1°C. A sensor might give a resistance straight line relationship against which the error is specified.
change of 10.2 1(2 when the true change should have been One method is to draw the straight line joining the output
10.5 Q. The error is -0.3 Q. values at the end points of the range; another is to find the
straight line by using the method of least squares to
Accuracy Accuracy is the extent to which the value indicated determine the best fit line when all data values are considered
by a measurement system might be wrong. It is thus the equally likely to be in error, another is to find the straight
summation of all the possible errors that are likely to occur, line by using the method of least squares to determine the
as well as the accuracy to which the transducer has been best fit line which passes through the zero point. Figure 2.2
calibrated. A temperature-measuring instrument might, for illustrates these three methods and how they can affect the
example, be specified as having an accuracy of *2"C. This 0 100 non-linearity error quoted. The error is generally quoted as a
would mean that the reading given by the instrument can be lnput % percentage of the fidl range output. For example, a
expected to lie within + or -2°C of the true value. Accuracy is (a) transducer for the measurement of pressure might be quoted
often expressed as a percentage of the full range output or as having a non-linearity error of *OS% of the 111range.
full-scale deflection. The percentage of full-scale deflection
term results from when the outputs of measuring systems 7 Repeatability/reprodducibility The terms repeatability and
were displayed almost exclusively on a circular or linear reproducibility of a transducer are used to describe its ability
scale. A sensor might, for example, be specified as having an to give the same output for repeated applications of the same
accuracy of *5% of full range output. Thus if the range of the input value. The error resulting from the same output not
sensor was, say, 0 to 200°C, then the reading given can be being given with repeated applications is usually expressed as
expected to be within + or -10°C of the true reading. a percentage of the fill range output.

Sensitivity The sensitivity is the relationship indicating how Repeatability =

max. - min. values given
much output you get per unit input, i.e. ouputhput. For full range x 100
example, a resistance thermometer may have a sensitivity of
0.5 W°C. This term is also frequently used to indicate the A transducer for the measurement of angular velocity
sensitivity to inputs other than that being measured, i.e. typically might be quoted as having a repeatability of iO.Ol%
environmental changes. Thus there can be the sensitivity of of the 111 range at a particular angular velocity.
the transducer to temperature changes in the environment or
perhaps fluctuations in the mains voltage supply. A 8 Stability The stability of a transducer is its ability to give the
transducer for the measurement of pressure might be quoted same output when used to measure a constant input over a
as having a temperature sensitivity of M.1% of the reading period of time. The term dr19 is often used to describe the
per OC change in temperature. 0 100 change in output that occurs over time. The drift may be
lnput % expressed as a percentage of the full range output. The term
Hysteresis error Transducers can give different outputs from (c) zero drift is used for the changes that occur in output when
the same value of quantity being measured according to Fig. 2.2 Non-linearity error using: there is zero input.
whether that value has been reached by a continuously (a) endrange values,
increasing change or a continuously decreasing change. This (b) best straight line for all values, 9 Dead band/time The dead band or dead space of a transducer
effect is called hysteresis. Figure 2.1 shows such an output (c) best straight line through zero point is the range of input values for which there is no output. For
with the hysteresis error as the maximum difference in output example, bearing friction in a flow meter using a rotor might
for increasing and decreasing values. mean that there is no output until the input has reached a
particular velocity threshold. The dead time is the length of
Non-linearity error For many transducers a linear relation- time from the application of an input until the output begins
ship between the input and output is assumed over the to respond and change.
working range, i.e. a graph of output plotted against input is
assumed to give a straight line. Few transducers, however, 10 Resolution When the input varies continuously over the
have a truly linear relationship and thus errors occur as a range, the output signals for some sensors may change in
result of the assumption of linearity. The error is defined as small steps. A wire-wound potentiometer is an example of
0 Value of measured the maximum difference from the straight line. Various such a sensor, the output going up in steps as the
quantity methods are used for the numerical expression of the potentiometer slider moves from one wire turn to the next.
Fig. 2.1 Hysteresis non-linearity error. The differences occur in determining the The resolution is the smallest change in the input value that
 Sensors and transducers 21
20 Mechatronics
suddenly changed from 0 to a constant value, or a ramp input
will produce an observable change in the output. For a when the input is changed at a steady rate, or a sinusoidal input
wire-wound potentiometer the resolution might be specified
of a specified frequency. Thus we might find the folIowing terms
as, say, 0.5" or perhaps a percentage of the full-scale
(see Chapter 8 for a more detailed discussion of dynamic
deflection. For a sensor giving a digital output the smallest
systems):
change in output signal is 1 bit. Thus for a sensor giving a
data word of N bits, i.e. a total of 2Nbits, the resolution is
Response time This is the time which elapses after a

-
generally expressed as 1/2N.
constant input, a step input, is applied to the transducer up to
11 Output impedance When a sensor giving an electrical output the point at which the transducer gives an output
is interfaced with an electronic circuit it is necessary to know corresponding to some specified percentage, e.g. 95% of the
value of the input (Fig. 2.3). For example, if a mercury-
the output impedance since this impedance is being
in-glass thermometer is put into a hot liquid there can be
connected in either series or parallel with that circuit. The 0 Time 95% Time quite an appreciable time lapse, perhaps as much as 100 s or
inclusion of the sensor can thus sigruficantly modify the constant response more, before the thermometer indicates 95% of the actual
behaviour of the system to which it is connected. See Section time temperature of the liquid.
4.1.1 for a discussion of loading. Fig. 2.3 Response to a step input
Time constant This is the 63.2% response time. A
To illustrate the above, consider the sigruficance of the terms in
thermocouple in air might have a time constant of perhaps
the following specification of a strain gauge pressure transducer:
40 to 100 s. The time constant is a measure of the inertia of
Ranges: 70 to 1000 kPa, 2000 to 70 000 kPa the sensor and so how fast it will react to changes in its
Supply voltage: 10 V d.c. or a.c. r.m.s. input; the bigger the time constant the slower will be its
Full range output: 40 mV reaction to a changing input signal. See Section 10.2.3 for a
Non-linearity and hysteresis: a . 5 % full range output mathematical discussion of the time constant in terms of the
behaviour of a system when subject to a step input.
Temperature range: -54°C to +120°C when operating
Thermal zero shift: 0.030% full range outputPC
Rise time This is the time taken for the output to rise to some
specified percentage of the steady-state output. Often the rise
The range indicates that the transducer can be used to measure
time refers to the time taken for the output to rise from 10%
pressures between 70 and 1000 kPa or 2000 and 70 000 kPa It
of the steady-state value to 90 or 95% of the steady-state
requires a supply of 10 V dc. or ax. r.m.s. for its operation and
value.
will give an output of 40 mV when the pressure on the lower
range is 1000 kPa and on the upper range 70 000 kPa.
Settling time This is the time taken for the output to settle to
Non-linearity and hysteresis will lead to errors of 34.5% of 1000,
within some percentage, e.g. 2%, of the steady-state value.
i.e. &5 kPa on the lower range and @.5% of 70 000, i.e.
*350 Wa on the upper range. The transducer can be used
between the temperatures of -54 and +120°C. When the To illustrate the above, consider the following data which
indicates how an instrument reading changed with time, being
temperature changes by 1°C the output of the transducer for zero
obtained from a thermometer plunged into a liquid at time t = 0.
input will change by 0.030% of 1000 = 0.3 kPa on the lower
The 95% response time is required.
range and 0.030% of 70 000 = 21 P a on the upper range.
Time (s) 0 30 60 90 120 150 180
2.2.1 Static and dynamic characteristics
Temp.("C) 20 28 34 39 43 46 49
The static characteristics are the values given when steady-state
conditions occur, i.e. the values given when the transducer has Time (s) 210 240 270 300 330 360
settled down after having received some input. The terminology Temp.("C) 51 53 54 55 55 55
defhed above refers to such a state. The dynamic characteristics
refer to the behaviour between the time that the input value Figure 2.4 shows the graph of how the temperature indicated by
changes and the time that the value given by the transducer settles the thermometer varies with time. The steady-state value is 55°C
down to the steady-state value. Dynanuc characteristics are stated Time (s)
and so, since 95% of 55 is 52.25"C, the 95% response time is
in terms of the response of the transducer to inputs in particular Fig. 2.4 Thermometer in liquid about 228 s.
forms. For example, this might be a step input when the input is
22 Mechatronics Sensors and transducers 23
The following sections give examples of transducers grouped monitored by a sensor. The movement of the shaft may be used to
according to what they are being used to measure. The cause changes in electrical voltage, resistance, capacitance, or
measurements considered are those frequently encountered in mutual inductance. For angular displacement methods involving
mechanical engineering, namely: displacement, proximity, mechanical connection the rotation of a shaft might directly drive,
velocity, force, pressure, fluid flow, liquid level, temperature, and through gears, the rotation of the transducer element. Non-
light intensity. For a more comprehensivecoverage of transducers contacting sensors involve the presence in the vicinity of
the reader is referred to more specialist texts, e.g. Instrumentation the measured object causing a change in the air pressure in the
Reference Book edited by B.E. Noltingk (Buttenvorth 1988, sensor, or perhaps a change in inductance or capacitance. The
1995), Measurement and Instrumentation Systems by W . Bolton following are examples of commonly used displacement sensors.
(Newnes 1996) and the concise Newnes Instrumentation and
Measurement Pocket Book by W . Bolton (Newnes 1991, 1996, 2.3.1 Potentiometer sensor
2000). Examples of the types of specifications that might be
expected fiom electrical transducers are given in Transducer A potentiometer consists of a resistance element with a sliding
Handbook by H.B. Boyle (Newnes 1992). contact which can be moved over the length of the element. Such
elements can be used for linear or rotary displacements, the
displacement being converted into a potential difference. The
Displacement sensors are concemed with the measurement of the rotary potentiometer consists of a circular wire-wound track or a
2.3 Displacement, film of conductive plastic over which a rotatable sliding contact
amount by which some object has been moved; position sensors
position and are concemed with the determination of the position of some can be rotated (Fig. 2.5). The track may be a single turn or
proximity object with reference to some reference point. Proximity sensors helical. With a constant input voltage V,, between terminals 1 and
are a form of position sensor and are wed to determine when an 3, the output voltage V, between terminals 2 and 3 is a fraction of
object has moved to within some particular critical distance of the the input voltage, the fraction depending on the ratio of the
sensor. They are essentially devices which give on-off outputs. resistance RB between terminals 2 and 3 compared with the total
In selecting a displacement, position or proximity sensor, resistance R I between
~ terminals 1 and 3, i.e. V& = RdR13 If
consideration has to be given to: A rotary potentiometer the track has a constant resistance per unit length, i.e. per unit
1 angle, then the output is proportional to the angle through which
The size of the displacement are we talking of fractions of a the slider has rotated. Hence an angular displacement can be
millimetre, many millimetres or perhaps metres? For a converted into a potential difference.
proximity sensor, how close should the object be before it is With a wire-wound track the slider in moving from one turn to
detected? the other will change the voltage output in steps, each step being
Whether the displacement is linear or angular; linear a movement of one turn.If the potentiometer has N turns then the
displacement sensors might be used to monitor the thickness
3
I - - - L 3
resolution, as a percentage, is 1001N. Thus the resolution of a
or other dimensions of sheet materials, the separation of wire track is limited by the diameter of the wire used and
The circuit when connectedto a load
rollers, the position or presence of a part, the size of a part, typically ranges from about 1.5 mm for a coarsely wound track to
etc. while angular displacement methods might be used to 0.5 mm for a finely wound one. Errors due to non-linearity of the
monitor the angular displacement of shafts. track tend to range from less than 0.1% to about 1%. The track
The resolution required. resistance tends to range from about 20 C2 to 200 kR. Conductive
The accuracy required. plastic has ideally infinite resolution, errors due to non-linearity
What material the measured object is made of; some sensors of track of the order of 0.05% and resistance values from about
d l only work with ferromagnetic materials, some with only 500 to SO kn. The conductive plastic has a higher temperature
metals, some with only insulators. coefficientof resistance than the wire and so temperature changes
The cost. RLin parallel have a greater effect on accuracy.
with xR, An important effect to be considered with a potentiometer is the
Displacement and position sensors can be grouped into two The circuit as a potential divider effect of a load RL connected across the output. The potential
basic types: contact sensors in which the measured object comes difference across the load VLis only directly proportional to V, if
into mechanical contact with the sensor or non-contacting where Fig. 2.5 Rotary potentiometer the load resistance is infinite. For finite loads, however, the effect
there is no physical contact between the measured object and the of the load is to transform what was a linear relationship between
sensor. For those linear displacement methods involving contact, output voltage and angle into a non-linear relationship. The
there is usually a sensing shaft which is in direct contact with the resistance RLis in parallel with the fraction x of the potentiometer
object being monitored. The displacement of this shaft is then
24 Mechatronics Sensors and transducers 25

resistance R,. This combined resistance is R$RJ(RL + d p ) . The semiconductor strain gauges have gauge factors of about +I00 or
total resistance across the source voltage is thus: more for p-type silicon and -100 or more for n-type silicon. The
gauge factor is normally supplied by the manufacturer of the
strain gauges from a calibration made of a sample of strain
gauges taken fiom a batch. The calibration involves subjecting
The circuit is a potential divider circuit and thus the voltage the sample gauges to known strains and measuring their changes
across the load is the fraction the resistance across the load is of in resistance. A problem with all strain gauges is that their
the total resistance across which the applied voltage is connected: resistance not only changes with strain but also with temperature.
Ways of eliminating the temperature effect have to be used and
are discussed in Chapter 3. Semiconductor strain gauges have a
much greater sensitivity to temperature than metal strain gauges.
To illustrate the above, consider an electrical resistance strain
gauge with a resistance of 100 IR and a gauge factor of 2.0. What
is the change in resistance of the gauge when it is subject to a

62
strain of 0.0011 The hctional change in resistance is equal to the
If the load is of infinite resistance then we have VL = xK. Thus
the error introduced by the load having a finite resistance is: gauge factor multiplied by the strain, thus:
Strain gauges
re xvs Change in resistance = 2.0 x 0.001 x 100 = 0.2 !2
error = xK - VL= xVS- (RpIR~)x(l
- X )+ 1
One form of displacement sensor has strain gauges attached to
flexible elements in the form of cantilevers, rings or U-shapes
Fig. 2.7 Straingauged elements (Fig. 2.7). When the flexible element is bent or deformed as a
result of forces being applied by a contact point being displaced,
To illustrate the above, consider the non-linearity error with a then the electrical resistance strain gauges mounted on the
\ Connection potentiometer of resistance 500 S2, when at a displacement of half element are strained and so give a resistance change which can be
(a) leads its maximum slider travel, which results fiom there being a load monitored. The change in resistance is thus a measure of the
of resistance 10 kR. The supply voltage is 4 V. Using the displacement or deformation of the flexible element. Such
equation derived above: arrangements are typically used for linear displacements of the
order of 1 rnm to 30 rnm and have a non-linearity error of about *
1% of full range.

Metal As a percentage of the full range reading, this is 0.625%. 2.3.3 Capacitive element
foil

-
The capacitance C of a parallel plate capacitor is given by
2.3.2 Strain-gauged element
The electrical resistance strain gauge (Fig. 2.6) is a metal wire,
(a) I 1 Plate moves
andchanges d
.

metal foil strip, or a strip of semiconductor material which is

wafer-like and can be stuck onto surfaces like a postage stamp.
A overlap
area where G is the relative permittivity of the dielectric between the
When subject to strain, its resistance R changes, the fractional plates, ~a a constant called the permittivity of free space, A the
change in resistance ARIR being proportional to the strain E, i.e. area of overlap between the two plates and d the plate separation.
@) and changes A Capacitive sensors for the monitoring of linear displacements
& might thus take the forms shown in Figure 2.8. In (a) one of the
plater is moved by the displacement so that the plate separation
where G, the constant of proportionality, is termed the gauge changes; in @) the displacement causes the area of overlap to
factor. Since strain is the ratio (change in lengthloriginal length) (Deielccrti moves change; in (c) the displacement causes the dielectric between the
then the resistance change of a strain gauge is a measurement of plates to change.
(4 the change in length of the element to which the strain gauge is For the displacement changing the plate separation (Fig. 2.8(a)),
attached. The gauge factor of metal wire or foil strain gauges with Fig. 2.8 Forms of capacitive if the separation d is increased by a displacement x then the
Fig. 2.6 Strain gauges: (a) met
wire, (b) metal foil, (c) semiconductor the metals generally used is about 2.0. Silicon p- and n-type sensing element capacitame becomes:
i
 Sensors and transducers 27
26 Mechatronics

Hence the change in capacitance AC as a fraction of the initial

capacitance is given by: difference between

There is thus a non-linear relationship between the change in

Sc, capacitance AC and the displacement x. This non-linearity can be Displacement moves
Capacitor overcome by using what is termed a push-pull displacement Fig. 2.11 LVDT from central position
sensor (Fig. 2.9). This has three plates with the upper pair
forming one capacitor and the lower pair another capacitor. The
displacement moves the central plate between the two other However, when the core is displaced from the central position
plates. The result of, for example, the centraI plate moving down- there is a greater amount of magnetic core in one coil than the
Fig. 2.9 Push-pull sensor
wards is to increase the plate separation of the upper capacitor other, e.g. more in secondary coil 2 than coil 1. The result is that
dnd decrease the separation of the lower capacitor. We thus have: a greater e.m.f. is induced in one coil than the other. There is
then a net output from the two coils. Since a greater displacement
means even more core in one coil than the other, the output, the
Core more I Core more difference between the two e.m.f.s increases the greater the
in 2 than 1 I in 1 than 2 displacement being monitored (Fig. 2.12).
The e.m.f. induced in a secondary coil by a changing current i in
central the primary coil is given by:
When Cl is in one arm of an a.c. bridge and C2 in the other, then
the resulting out-of-balance voltage is proportional to x. Such a
sensor is typically used for monitoring displacements from a few
millimetres to hundreds of rnillimetres. Non-linearity and Displacement
where M is the mutual inductance, its value depending on the
*
hysteresis are about 0.01% of fill range.
number of turns on the coils and the ferromagnetic core. Thus, for
One form of capacitive proximity sensor consists of a single
a sinusoidal input current of i = 1 sin of to the primary coil, the
capacitor plate probe with the other plate being formed by the
-1w- e.m.f.s induced in the two secondary coils 1 and 2 can be
object, which has to be metallic and earthed (Fig. 2.10). As the
represented by:
object approaches so the 'plate separation' of the capacitor
Fig. 2.10 Capacitive proximity changes, becoming sigdicant and detectable when the object is Fig. 2.12 LVDT output
sensor close to the probe.
where the values of kl,kz and 4 depend on the degree of coupling
2.3.4 Differential transformers
between the primary and secondary coils for a particular core
The linear variable differential transformer, generally referred to position. 4 is the phase difference between the primary
by the abbreviation LVDT, consists of three coils symmetrically alternating voltage and the secondary alternating voltages.
spaced along an insulated tube (Fig. 2.11). The central coil is the Because the two outputs are in series, their difference is the
primary coil and the other two are identical secondary coils which output:
are connected in series in such a way that their outputs oppose
each other. A magnetic core is moved through the central tube as output voltage = vl - vz = (kl - k2)sin(ot- 4)
a result of the displacement being monitored.
When there is an alternating voltage input to the primary coil, When the core is equally in both coils, kl equals k2 and so the
alternating e.m.f.s are induced in the secondaxy coils. With the output voltage is zero. When the core is more in 1 than in 2 we
magnetic core central, the amount of magnetic material in each of have kl > k2 and:
the secondary coils is the same. Thus the e.m.f.s induced in each
coil are the same. Since they are so connected that their outputs output voltage = (kl- k2) sin(ot - 4)
oppose each other, the net result is zero output.
 Sensors and transducers 29
28 Mechatronics
magnetic field. This distorts the magnetic field responsible for
Phase sensitive their production. As a result, the impedance of the coil changes
LVDT demodulator Low pass filter and so the amplitude of the alternating current. At some preset
Ievel, this change can be used to trigger a switch. Figure 2.15

as. d.c. -
n Displacement
shows the basic form of such a sensor, it is used for the detection
of non-magnetic but conductive materials. They have the
advantages of being relatively inexpensive, small in size, with
. -
Reference ~ehsor high reliability and can have high sensitivity to small
Fig. 2.13 LVDT d.~.output coil coil
Conducting displacements.
object
When the core is more in 2 than in 1 we have kl < k2. A
2.3.6 Inductive proximity switch
Fig. 2.15 Eddy current sensor
consequence of kl being less than k2 is that there is a phase This consists of a coil wound round a core. When the end of the
change of 180° in the output when the core moves from more in 1 coil is close to a metal object its inductance changes. This change
to more in 2. Thus: can be monitored by its effect on a resonant circuit and the
change used to trigger a switch. It can only be used for the
output voltage detection of metal objects and is best with ferrous metals.
= -(kl - kt) sin(wt - 4) = (k2 - k 1) sin[ot + (TI
- $)I
2.3.7 Optical encoders
Figure 2.12 shows how the size and phase of the output changes
with the displacement of the core. An encoder is a device that provides a digital output as a result of
With this form of output, the same amplitude output voltage is a linear or angular displacement Position encoders can be
produced for two different displacements. To give an output grouped into two categories: incremental encoders that detect
voltage which is unique to each value of displacement we need to changes in rotation from some datum position and absolute
distinguish between where the amplitudes are the same but there encoders which give the actual angular position.
is a phase difference of 180". A phase sensitive demodulator, with Figure 2.16(a) shows the basic form of an incremental encoder
a low pass filter, is used to convert the output into a dc. voltage for the measurement of angular displacement. A beam of light
which gives a unique value for each displacement (Fig. 2.13). passes through slots in a disc and is detected by a suitable light
Such circuits are available as integrated circuits. sensor. When the disc is rotated, a pulsed output is produced by
Typically, LVDTs have operating ranges from about i 2 mm to the sensor with the number of pulses being proportional to the
*400 mm with non-linearity errors of about i0.25%. LVDTs are angle through which the disc rotates. Thus the angular position of
very widely used as primary transducers for monitoring the disc, and hence the shaft rotating it, can be determined by the
displacements. The free end of the core may be spring loaded for number of pulses produced since some datum position. In practice
contact with the surface being monitored, or threaded for three concentric tracks with three sensors are used (Fig. 2.16(b)).
mechanical connection. They are also used as secondary The inner track has just one hole and is used to locate the 'home'
transducers in the measurement of force, weight and pressure; position of the disc. The other two tracks have a series of equally
these variables are transformed into displacements which can spaced holes that go completely round the disc but with the holes
" then be monitored by LVDTs. in the middle track offset from the holes in the outer track by
Secondary 1 Secondary 2 A rotary variable differential transformer (RVDT) can be used one-half the width of a hole. This offset enabIes the direction of
for the measurement of rotation (Fig. 2.14); it operates on the rotation to be determined. In a clockwise direction the pulses in
Fig. 2.14 RVDT
same principle as the LVDT. -The core is a cardioid-shaped piece the outer track lead those in the inner, in the anticlockwise
of magnetic material and rotation causes more of it to pass into direction they lag. The resolution is determined by the number of
one secondary coil than the other. The range of operation is slots on the disc. With 60 slots occurring with 1 revolution then,
typically MOO with a linearity error of about M.5%of the range. since 1 revolution is a rotation of 360°, the resolution is 360160 =
6".
2.3.5 Eddy current proximity sensors Figure 2.17 shows the basic form of an absolute encoder for the
measurement of angular displacement. This gives an output in the
If a coil is supplied with an alternating current, an alternating Fig. 2.16 Incrementalencoder: form of a binary number of several digits, each such number
magnetic field is produced. If there is a metal object in close (a) the basic principle, (b) concentric representing a particular angular position. The rotating disc has
proximity to this alternating magnetic field, then eddy currents tracks three concentric circles of slots and three sensors to detect the
are induced in it. The eddy currents themselves produce a
30 Mechatronics Sensors and transducers 31
light pulses. The slots are arranged in such a way that the
sequential output from the sensors is a number in the binary code.
Typical encoders tend to have up to 10 or 12 tracks. The number Object blocking escaping air increases
Air dragged out of port and so drop in system
of bits in the binary number will be equal to the number of tracks.
Thus with 10 tracks there will be 10 bits and so the number of
positions that can be detected is 21°, i.e. 1024, a resolution of

LEDs Sensors 1 ri . Escaping

..............B;..<

1'
%<.+:.>. air

I I I I
it, Low-pressure Low-pressure ; Escaping
air inlet air inlet air

Fig. 2.19 Pneumatic proximity sensor

Such sensors are used for the measurement of displacements of

Fig. 2.17 A 3-bit absolute encoder fractions of millimetres in ranges which typically are about 3 to
12 mm.
The normal form of binary code is generally not used because 2.3.9 Proximity switches
changing from one binary number to the next can result in more
( Normal binary Gray code than one bit changing and if, through some misalignment, one of There are a number of forms of switch which can be activated by
the bits changes fractionally before the others then an the presence of an object in order to give a proximity sensor with
intermediate binary number is momentarily indicated and so can an output which iseither on or off.
lead to false counting. To overcome this the Gray code is The rnicvoswitch is a small electrical switch which requires
generally used. With this code only one bit changes in moving physical contact and a small operating force to close the contacts.
from one number to the next. Figure 2.18 shows the tracks with For example, in the case of determining the presence of an item
normal binary code and the Gray code. on a conveyor belf this might be actuated by the weight of the
Optical encoders, e.g. HEDS-5000 from Hewlett Packard, are item on the belt depressing the belt and hence a spring-loaded
supplied for mounting on shafts and contain an LED light source platform under it, with the movement of this platform then
and a code wheel. Interface integrated circuits are also available closing the switch. Figure 2.20 shows examples of ways such
to decode the encoder and give a binary output suitable for a switches can be actuated.
microprocessor. For an absolute encoder with 7 tracks on its code
Fig. 2.18 Binary and Gray codes disc, each track will give one of bits in the binary number and
thus we have 27positions specified, i.e. 128.
Button to operate
2.3.8 Pneumatic sensors I switch

Pneumatic sensors involve the use of compressed air,

displacement or the proximity of an object being transformed into
a change in air pressure. Figure 2.19 shows the basic form of
such a sensor. Low-pressure air is allowed to escape through a
port in the front of the sensor. This escaping air, in the absence of
(a) Switch contacts
any close-by object, escapes and in doing so also reduces the
pressure in the nearby sensor output port. However, if there is a Fig. 2.20 (a) Lever-operated,
close-by object, the air cannot so readily escape and the result is (b) roller-operated, (c) cam-operated
that the pressure increases in the sensor output port. The output switches
pressure from the sensor thus depends on the proximity of objects.
32 Mechatronics Sensors and transducers 33

Figure 2.21 shows the basic form of a reed switch. It consists of 10 mV per mT (1 mV per gauss) when there is a supply voltage
Magnet two magnetic switch contacts sealed in a glass tube. When a of 5 V. The threshold Hall effect sensor Allegro UGN3l32U gives
'3' magnet is brought close to the switch, the magnetic reeds are an output which switches from virtually zero to about 145 mV
attracted to each other and close the switch contacts. It is a when the magnetic f l w density is about 3 mT (30 gauss). Hall
noncontact proximity switch. Such a switch is very widely used effect sensors have the advantages of being able to operate as
for checking the closure of doors. It is also used with such devices switches which can operate up to 100 lcHz repetition rate, cost
springy strips Electricalcontacts as tachometers which involve the rotation of a toothed wheel past less than electromechanical switches and do not suffer from the
the reed switch. If one of the teeth has a magnet attached to if problems associated with such switches of contact bounce
Fig. 2.21 Reed switch
then every time it passes the switch it will momentarily close the occurring and hence a sequence of contacts rather than a singIe
contacts and hence pioduce a currenthroltage pulse in the clear contact. The Hall effect sensor is immune to environmental
Photodetector associated electrical circuit contaminants and can be used under severe service conditions.
Photosensitive devices can be used to detect the presence of an
opaque object by it breaking a beam of light, or infrared radiation,
falling on such a device or by detecting the light reflected back by
Object breaks the object (Fig. 2.22).
the beam
(4 2.3.10 Hall effect sensors
When a beam of charged particles passes through a magnetic
field, forces act on the particles and the beam is deflected from its 0 Flux density
Fig. 2.24 Hall effect sensors:
+ght line path. A current flowing in a conductor is like a beam
of moving charges and thus can be deflected by a magnetic field. (a) linear, (b) threshold
This effect was discovered by E.R Hall in 1879 and is called the
0"' the light
Hall effect. Consider electrons moving in a conductive plate with Such sensors can be used as position, displacement and
a magnetic field applied at right angles to the plane of the plate proximity sensors if the object being sensed is fitted with a small
Fig. 2.22 Using photoelectric (Fig. 2.23). As a consequence of the magnetic field, the moving Ground Output permanent magnet. As an illustration, such a sensor can be used
sensors to detect objects by (a) the electrons are deflected to one side of the plate and thus that side , , * PUS ............ Hall to determine the. level of fuel in an automobiIe fuel tank. A
object breaking the beam. (b) it becomes negatively charged while the opposite side becomes .......
....... sensor magnet is attached to a float and as the level of fuel changes so
reflecting light positively charged since the electrons are directed away from it. the float distance from the Hall sensor changes (Fig. 2.25). The
This charge separation produces an electric field in the material. result is a Hall voltage output which is a measure of the distance
The charge separation continues until the forces on the charged Float of the float from the sensor and hence the level of fuel in the tank.
Magnetic i particles from the electric field just balance the forces produced Another application of Hall effect sensors is in brushless d.c.
by the magnetic field. The result is a transverse potential motors. With such motors it is necessary to determine when the
difference V given by: permanent magnet rotor is correctly aligned with the windings on
the stator so that the current through the windings can be
switched on at the right instant to maintain the rotor rotation.
Fig. 2.25 Fluid level detector Hall effect sensors are used to detect when the alignment is right.

' Potential difference

produced by deflection
where B is the magnetic flux density at right angles to the plate, I
the current through if t the plate thickness and KH a c 0 1 . t
2.4 Velocity and motion The following are examples of sensors that can be used to monitor
called the Hall coeflcient. Thus if a constant current source is
of electrons linear and angular velocities and detect motion. The application
used with a particular sensor, the Hall voltage is a measure of the
4 magnetic flux density. of motion detectors includes security systems used to detect
Hall effect sensors are generally supplied as an integrated circuit intruders and interactive toys and appliances, e.g. the cash
~ e g a q ~ :
charged 2;
charged with the necessary signal processing circuitry. There are two basic machine screen which becomes active when you get near to it.
.&j forms of such sensor, linear where the output varies in a
reasonably linear manner with the magnetic flux density (Fig. 2.4.1 Incremental encoder
2.24(a)) and threshold where the output shows a sharp drop at a
The incremental encoder described in Section 2.3.7 can be used
Current particular magnetic flux density (Fig. 2.24@)). The linear output
for a measurement of angular velocity, the number of pulses
Hall effect sensor 634SS2 gives an output which is fairly linear
produced per second being determined.
Fig. 2.23 Hall effect over the range 4 0 to +40 mT (-400 to + 400 gauss) at about
34 Mechatronics Sensors and transducers 35

2.4.2 Tachogenerator
The tachogenerator is used to measure angular velocity. One
form, the variable reluctance tachogenerator, consists of a
toothed wheel of ferromagnetic material which is attached to the
Pick-up coil rotating shaft (Fig. 2.26). A pick-up coil is wound on a permanent
magnet. As the wheel rotates, so the teeth move past the coil and
1 \\\\ 1 the air gap between the coil and the ferromagnetic material Fig. 2.28 Polarising a pyroelectric
changes. We have a magnetic circuit with an air gap which
material
- - output
periodically changes. Thus the flux linked by a pick-up coil
changes. The resulting cyclic change in the flux linked produces
Toothed wheel an alternating e.m.f. in the coil. 2.4.3 Pyroelectric sensors
If the wheel contains n teeth and rotates with an angular velocity
Fig. 2.26 Variable reluctance o , then the flux change with time for the coil can be considered Pyroelectric materials, e.g. lithium tantalate, are crystalline
tachogenerator to be of the form: materials which generate charge in response to heat flow. When
Amount of Curie temperature such a material is heated to a temperature just below the Curie
@ = @o + @a cos not polarisation temperature, this being about 610°C for lithium tantalate, in an
or charge electric field and the material cooled while remaining in the field,
at crystal
where 00is the mean value of the flux and ma the amplitude of I electric dipoles within the material line up and it becomes
surfaces
the flux variation. The induced e.m.f. e in the N turns of the polarised (Fig. 2.28, (a) leading to (b)). When the field is then
pick-up coil is thus: removed the material retains its polarisation; the effect is rather
like magnetising a piece of iron by exposing it to a magnetic field.
e = -N=dm = -N$@~ + a a cosnot) = NQanosinnot Temperature
When the pyroelectric material is exposed to infrared radiation,
its temperature rises and this reduces the amount of polarisation
Fig. 2.29 Effect of temperature in the material, the dipoles being shaken up more and losing their
and so we can write: on amount of polarisation alignment (Fig. 2.29).
A pyroelectric sensor consists of a polarised pyroelectric crystal
e = Em sin o t
with thin metal film electrodes on opposite faces. Because the
No incident I
infrared .Charges
crystaI is polarised with charged surfaces, ions are drawn fi-om
where the maximum value of the induced e.m.f. Em, is N @ a o
the surrounding air and electrons from any measurement circuit
and so is a measure of the angular velocity. balanced
+++++ connected to the sensor to balance the surface charge (Fig.
Instead of using the maximum value of the e.m.f. as a measure
2.30(a)). If infrared radiation is incident on the crystal and
of the angular velocity, a pulse-shaping signal conditioner can be
changes its temperature, the polarisation in the crystal is reduced
used to transform the output into a sequence of pulses which can Infrared
and consequently that is a reduction in the charge at the surfaces
be counted by a counter, the number counted in a particular time radiation I Excess
of the crystal. There is then an excess of charge on the metal
interval being a measure of the angular velocity. charge on
electrodes electrodes over that needed to balance the charge on the crystal
Another form of tachogenerator is essentially an ax. generator.
surfaces (Fig. 2.30@)). This charge leaks away through the
It consists of a coil, termed the rotor, which rotates with the
measurement circuit until the charge on the crystal once again is
rotating shaft. This coil rotates in the magnetic field produced by
balanced by the charge on the electrodes. The pyroelectric sensor
a stationary permanent magnet or electromagnet (Fig. 2.27) and
Fig. 2.30 Pyroelectric sensor thus behaves as a charge generator which generates charge when
so an alternating e.m.f. is induced in it. The amplitude or
there is a change in its temperature as a result of the incidence of
frequency of this alternating e.m.f. can be used as a measure of
infrared radiation. For the linear part of the graph in Figure 2.29,
the angular velocity of the rotor. The output may be rectified to
Rotating when there is a temperature change the change in charge Aq is
coil
give a d.c. voltage with a size which is proportional to the angular
proportional to the change in temperature At:
velocity. Non-linearity for such sensors is typically of the order of
Fig. 2.27 A.C. generator a . 1 5 % of the full range and the sensors typically used for
form of tachogenerator rotations up to about 10 000 revlminute.
where k, is a sensitivity constant for the crystal. Figure 2.31
shows the equivalent circuit of a pyroelectric sensor, it effectively
Fig. 2.31 Equivalent circuit being a capacitor charged by the excess charge with a resistance R
36 Mechatronics Sensors and transducers 37
to represent either internal leakage resistance or that combined range and repeatability error rt0.02% of full range. Strain gauge
with the input resistance of an exqernal circuit. load cells based on the bending of a strain-gauged metal element
To detect the motion of a human or other heat source, the tend to be used for smaller forces, e.g. with ranges varying from 0
Infra-
sensing element has to distinguish between general background to 5 N up to 0 to 50 kN. Errors are typically a non-linearity error
red
3 heat radiation and that given by a moving heat source. A single of about 3~0.03%of full range, hysteresis error *0.02% of full
pyroelectric sensor would not be capable of this and so a dual range and repeatability error 2=0.02%of full range.
element is used (Fig. 2.32). One form has the sensing element
with a single front electrode but two, separated, back electrodes.
The result is two sensors which can be connected so that when Many of the devices used to monitor fluid pressure in industrial
both receive the same heat signal their outputs cancel. When a 2.6 Fluid pressure
processes involve the monitoring of the elastic deformation of
Fig. 2.32 Dual pyroelectric heat source moves so that the heat radiation moves from one of
, Pressure
diaphragms, capsules, bellows and tubes. The types of pressure
sensor the sensing elements to the other then the resulting current measurements that can be required are: absolute pressure where
through the resistor alternates from being first in one direction the pressure is measured relative to zero-pressure, i.e. a vacuum,
and then reversed to the other direction. Typically a moving differential pressure where a pressure difference is measured and
human gives an alternating current of the order of 10-l2 A. The gauge pressure where the pressure is measured relative to the
resistance R has thus to be very high to give a sigruficant voltage. barometric pressure.
For example, 50 GG with such a current gives 50 mV. For this , Pressure For a diaphragm (Fig. 2.34(a) and @)), when there is a
reason a JFET transistor is included in the circuit as a voltage difference in pressure between the two sides then the centre of the
follower in order to bring the output impedance down to a few diaphragm becomes displaced. Cormgations in the diaphragm
m. result in a greater sensitivity. This movement can be monitored by
A focusing device is needed to direct the infrared radiation onto some form of displacement sensor, e.g. a strain gauge, as
the sensor. While parabolic mirrors can be used a more illustrated in Figure 2.35. A specially designed strain gauge is
Fig. 2.34 Diaphragms: (a) flat,
commonly used method is a Fresnel plastic lens. Such a lens also often used, consisting of four strain gauges with two measuring
(b) cormgated
protects the front surface of the sensor and is the form commonly the strain in a circumferential direction while two measure strain
used for sensors to trigger intruder alarms or witch on a light in a radial direction. The four strain gauges are then connected to
when someone approaches. form the arms of a Wheatstone bridge (see Chapter 3). While
Diaphragm
Strain gauge strain gauges can be stuck on a diaphragm, an alternative is to
create a silicon diaphragm with the strain gauges as specially
2.5 Force A spring balance is an example of a force sensor in which a force, doped areas of the diaphragm.
a weight, is applied to the scale pan and causes a displacement, Another form of silicon diaphragm pressure sensor is used for
i.e. the spring stretches. The displacement is then a measure of the Motorola MPX pressure sensors. The strain gauge element is
the force. Forces are commonly measured by the measurement of integrated, together with a resistive network, in a single silicon
displacements, the following method illustrating this. Pressure diaphragm chip. When a current is passed through the sorain
Fig. 2.35 Diaphragm pressure gauge gauge element and pressure applied at right angles to it, a voltage
2.5.1 'Straingauge load cell is produced in a transverse direction (Fig. 2.36). This, together
, Force A very commonly used form of force-measuring transducer is with signal conditioning and temperature compensation circuitry,
is packaged as the MPX sensor. The output voltage is directly
I based on the use of electrical resistance strain gauges to monitor
the strain produced in some member when stretched, compressed proportional to the pressure. Such sensors are available for use for
or bent by the application of the force. The arrangement is Transversevoltage taps the measurement of absolute pressure (the MX numbering system
Strain generally referred to as a load cell. Figure 2.33 shows an example ends with A, AP, AS or ASX)), differential pressure (the MX
of such a cell. This is a cylindrical tube to which strain gauges numbering system ends with D or DP) and gauge pressure (the
have been attached. When forces are applied to the cylinder to MX numbering system ends with GP, GVP, GS, GVS, GSV or
compress it, then the strain gauges give a resistance change GVSX). For example, the MPX2100 series has a pressure range
which is a measure of the strain and hence the applied forces. of 100 kPa and with a supply voltage of 16 V, d.c., gives in the
Since temperature also produces a resistance change, the signal absolute pressure and differential pressure forms a voltage output
I / Active element over the full range of 40 mV. The response time, 10 to 90%, for a
conditioning circuit used has to be able to eliminate the effects
due to temperature (see Section 3.5.1). Typically such load cells Fig- 2.36 Pressure sensor element step change from 0 to 100 kPa is about 1.0 ms and the output
are used for forces up to about 10 MN, the non-linearity error impedance is of the order of 1.4 to 3.0 kn. The absolute pressure
Strain gauge load cell being about M.03% of full range, hysteresis error *0.02% of full sensors are used for such applications as altimeters and
Fig. 2.33
barometers, the differential pressure sensors for air flow
38 Mechatronics Sensors and transducers 39

measurements and the gauge pressure sensors for engine pressure 2.6.1 Piezoelectric sensors
and tyre pressure.
Capsules (Fig. 2.37(a)) can be considered to be just two corrug- Piezoelectric materials when stretched or compressed generate
ated diaphragms combined and give even greater sensitivity. A
I Force electric charges with one face of the material becoming positively
stack of capsules is just a bellows (Fig. 2.37@)) and even more charged and the opposite face negatively charged (Fig. 2.40). As a
sensitive. Figure 2.38 shows how a bellows can be combined with Surfaces result a voltage is produced. Piezoelectric materials are ionic
a LVDT to give a pressure sensor with an electrical output. become crystals wluch when stretched or compressed results in the charge
Diaphragms, capsules and bellows are made from such materials charged distribution in the crystal changing so that there is a net
as stainless steel, phosphor bronze, and nickel, with rubber and displacement of charge with one face of the material becoming
nylon also being used for some diaphragms. Pressures in the I positively charged and the other negatively charged. The net
charge q on a surface is proportional to the amount x by which
range of about lo3 to lo8Pa can be monitored with such sensors.
Fig. 2.40 Piezoelectricity the charges have been displaced, and since the displacement is
Iron rod proportional to the applied force F:

Secondary
Fig. 2.37 (a) Capsule, coils coil where k is a constant and S a constant termed the charge
(b) bellows \ sensitivity. The charge sensitivity depends on the material
concerned and the orientation of its crystals. Quartz has a charge
sensitivity of 2.2 p C N when the crystal is cut in one particular
direction and the forces applied in a specific direction; barium
Bellows
titanate has a much higher charge sensitivity of the order of
130 pC/N and lead zirconate-titanate about 265 pC/N.
Metal electrodes are deposited on opposite faces of the

Fig. 2.38 LVDT vfih bellows

't'
Pressure
piezoelectric crystal (Fig. 2.41). The capacitance C of the
piezoelectric material between the plates is:

A different form of deformation is obtained using a tube with an Fig. 2.41 Piezoelectriccapacitor
elliptical cross-section (Fig. 2.39(a)). Increasing the pressure in where G is the relative permittivity of the material, A is area and t
such a tube causes it to tend to a more circular cross-section. its thickness. Since the charge q = Cv,where v is the potential
When such a tube is in the form of a C-shaped tube (Fig. 2.39(b)), difference produced across a capacitor, then:
this being generally known as a Bourdon tube, the C opens up to
some extent when the pressure in the tube increases. A helical
form of such a tube (Fig. 2.39(c)) gives a greater sensitivity. The
tubes are made from such materials as stainless steel and
phosphor bronze and are used for pressures in the range lo3 to The force F is applied over an area A and so the applied pressure
lo8P a p is FIA and if we write Sv= (Sl~o~r), this being termed the
voltage sensitivityfactor:

(a) Tube cross-section The voltage is proportional to the applied pressure. The voltage
sensitivityfor quartz is about 0.055 Vlm Pa. For barium titanate it
is about 0.01 1 Vlm Pa
Piezoelectric sensors are used for the measurement of pressure,
Fig. 2.39 Tube pressure force and acceleration. The applications have, however, to be
sensors such that the charge produced by the pressure does not have much
time to leak off and thus tends to be used mainly for transient
rather than steady pressures.
40 Mechatronics Sensors and transducers 41

The equivalent electrical circuit for a piezoelectric sensor is a constriction, vz the velocity, P2 the pressure and AZ the
charge generator in parallel with capacitance C, and in parallel cross-sectional area at the constriction, p the fluid density, then
with the resistance R, arising from leakage through the dielectric Bernoulli's equation gives
(Fig. 2.42(a)). When the sensor is connected via a cable, of
capacitance C,, to an amplifier of input capacitance CA and
resistance RA we have effectively the circuit shown in Figure
2.42(b) and a total circuit capacitance of C, + C, + CA in parallel
with resistance of R&I(RA + Rs).When the sensor is subject to Since the mass of liquid passing per second through the tube prior
pressure it becomes charged, but because of the resistance the Fig. 2.44 Fluid flow through a to the constriction must equal that passing through the tube at the
capacitor will discharge with time. The time taken for the constriction constriction, we have Alvlp = Azvzp. But the quantity Q of liquid
discharge will depend on the time constant of the circuit. passing through the tube per second is A ~ v l A2v2.
= Hence

Pressure difference

to display Thus the quantity of fluid flowing through the pipe per second is
s A proportional to d@ressure difference). Measurements of the

-f--+-
/n pressure difference can thus be used to give a measure of the rate
Sensor Cable Amplifier
of flow. There are many devices based on this principle, and the
(4 ("1 Orifice following example of the orifice plate is probably one of the most
plate common.
Fig. 2.42 (a) Sensor equivalent circuit,
(b) sensor connected to charge amplifier Fig. 2.45 Orifice plate 2.7.1 Orifice plate
The orifice plate (Fig. 2.45) is simply a disc, with a central hole,
which is placed in the tube through which the fluid is flowing.
2.6.2 Tactile sensor The pressure difference is measured between a point equal to the
A tactile sensor is a particular form of pressure sensor. Such a diameter of the tube upstream and a point equal to half the diam-
sensor is used on the 'fingertips' of robotic 'hands' to determine eter downstream. The orifice plate is simple, cheap, with no
when a 'hand7 has come into contact with an object. They are also Turbine moving parts, and is widely used. It, however, does not work well

-73-
used for 'touch display' screens where a physical contact has to be with slurries. The accuracy is typically about *IS% of full range,
sensed. One form of tactile sensor uses piezoelectric it is non-linear, and does produce quite an appreciable pressure
polyvinylidene fluoride (PVDF) film. Two layers of the film are loss in the system to which it is connected.
d p used and are separated by a soft film which transmits vibrations
(Fig. 2.43). The lower PVDF film has an alternating voltage 2.7.2 Turbine meter
applied to it and this results in mechanical oscillations of the film
(the piezoelectric effect described above in reverse). The + The turbine flowmeter (Fig. 2.46) consists of a multi-bladed rotor
that is supported centrally in the pipe along which the flow
film intermediate film transmits these vibrations to the upper PVDF
occurs. The fluid flow results in rotation of the rotor, the angular
film.As a consequence of the piezoelectric effect, these vibrations velocity being approximately proportional to the flow rate. The
Fig. 2.43 PVDF tactile sensor
cause an alternating voltage to be produced across the upper film. mte of revolution of the rotor can be determined using a magnetic
When pressure is applied to the upper PVDF film its vibrations Magnetic pick-up coil pick-up. The pulses are counted and so the number of revolutions
are affected and the output alternating voltage is changed. of the rotor can be determined. The meter is expensive with an
Fig. 2.46 Turbine flowmeter
accuracy of typically about M.3%.

2.7 Liquid flow The traditional methods of measuring the flow rate of liquids
involves devices based on the measurement of the pressure drop 2.8 Liquid level The level of liquid in a vessel can be measured directly by
occurring when the fluid flows through a constriction (Fig. 2.44). monitoring the position of the liquid surface or indirectly by
For a horizontal tube, where vl is the fluid velocity, PI the measuring some variabIe related to the height. Direct methods
pressure and A , the cross-sectional area of the tube prior to the can involve floats; indirect methods include the monitoring of the
42 Mechatronics Sensors and transducers 43

- weight of the vessel by, perhaps, load cells. The weight of the 2.9.2 Resistance temperature detectors (RTDs)
liquid is Ahpg, where A is the cross-sectional area of the vessel, h The resistance of most metals increases, over a limited
the height of liquid, p its density and g the acceleration due to temperature range, in a reasonably linear way with temperature
gravity. Thus changes in the height of liquid give weight changes. (Fig. 2.50). For such a linear relationship:
More commonly, indirect methods involve the measurement of
the pressure at some point in the liquid, the pressure due to a R, = Ro(1 + at)
column of liquid of height h being hpg, where p is the liquid
density. where R, is the resistance at a temperature t OC, ROthe resistance
at O°C and a a constant for the metal termed the temperature
2.8.1 Floats coefficient of resistance. Resistance temperature detectors (RTDs)
A direct method of monitoring the level of liquid in a vessel is by are simple resistive elements in the form of coils of wire of such
monitoring the movement of a float. Figure 2.47 illustrates this metals as platinum, nickel or nickel-copper alloys; platinum is
with a simple float system. The displacement of the float causes a the most widely used. Thin film pIatinum elements are often
lever arm to rotate and so move a slider across a potentiometer. 0 400 800 made by depositing the metal on a suitable substrate, wire-wound
The result is an output of a voltage related to the height of liquid. Temperature OC elements involving a platinum wire held by a high temperature
Other forms of this involve the lever causing the core in a LVDT Fig. 2-50 Variation of resistance
glass adhesive inside a ceramic tube. Such detectors are highly
to become displaced, or stretch or compress a strain-gauged stable and give reproducible responses over long periods of time.
with temperature for metals
element. They tend to have response times of the order of 0.5 to 5 s or
Atmosphere more.
2.8.2 Differential pressure
Thermistor 2.9.3 Thermistors
Figure 2.48 shows two forms of level measurement based on the

-- -- -----
+
measurement of differential pressure. In Figure 2.48(a), the
differ- ential pressure cell determines the pressure difference
between the liquid at the base of the vessel and atmospheric flq Thermistor
Thermistors are small pieces of material made from mixtures of
metal oxides, such as those of cluomium, cobalt, iron, manganese
and nickel. These oxides are semiconductors. The material is

0
I--I pressure, the vessel being open to atmospheric pressure. With a formed into various forms of element, such as beads, discs and
(b) Differential pressure cell closed or open vessel the system illustrated in (b) can be used. rods (Fig. 2.51). The resistance of conventional metal-oxide
The differential pressure cell monitors the difference in pressure thermistors decreases in a very non-linear manner with an
Fig. 2.48 Using a differential Rod
between the base of the vessel and the air or gas above the surface Di= increase in temperature, as illustrated in Figure 2.52. Such
pressure sensor
of the liquid. thermistors having negative temperature coefficients (NTC).
Positive temperature coefficient (PTC) thermistors are, however,
available. The change in resistance per degree change in
2.9 Temperature Changes that are commonly used to moGtor temperature are the temperature is considerably larger than that which occurs with
expansion or contraction of solids, liquids or gases, the change in ~ead metals. The resistance-temperature relationship for a thermistor
electrical resistance of conductors and semiconductors and can be described by an equation of the form
Fig. 2.51 Thermistors
thermoelectric e.m.f.s. The following are some of the methods
D
Electrical that are commonly used with temperature control systems. I

2.9.1 Bimetallic strips where R, is the resistance at temperature t, with K and B being
constants. Thermistors have many advantages when compared
This device consists of two different metal strips bonded together. with other temperature sensors. They are rugged and can be very
The metals have different coefficients of expansion and when the small, so enabling temperatures to be monitored at virtually a
temperature changes the composite strip bends into a curved strip, point. Because of their small size they respond very rapidly to
with the higher coefficient metal on the outside of the curve. This changes in temperature. They give very large changes in
deformation may be used as a temperature-controlled switch, as in resistance per degree change in temperature. Their main
the simple thermostat which was commonly used with domestic disadvantage is their non-linearity.
heating systems (Fig. 2.49). The small magnet enables the sensor
to exhibit hysteresis, meaning that the switch contacts close at a
Small magnet/ '' different temperature from that at which they open.
Fig. 2.49 Bimetallic thermostat
 Sensors and transducers 45
44 Mechatronics
temperature. A common method is to use two transistors with
different collector currents and determine the difference in the
baseemitter voltages between them, this difference being directly
proportional to the temperature on the Kelvin scale. Such
transistors can be combined with other circuit components on a
single chip to give a temperature sensor with its associated signal
LM35 conditioning, e.g. LM35 (Fig. 2.53). This sensor can be used in
Output the range 4 0 ° C to 110°C and gives an output of 10 mVPC.

2.9.5 Thermocouples
If two different metals are joined together, a potential difference
Fig. 2.53 LM35 occurs across the junction. The potential difference depends on
Fig. 2.52 Variation of resistance the metals used and the temperature of the junction. A
with temperature for a typical thermocouple is a complete circuit involving two such junctions
thermistor
Temperature O C (Fig. 2.54). If both junctions are at the same temperature there is
no net e.m.f. If, however, there is a difference in temperature
between the two junctions, there is an e.m.f. The value of this
e.m.f. E depends on the two metals concerned and the
Metal A e.m.f. Metal A
2.9.4 Thermodiodes and transistors temperatures t of both junctions. Usually one junction is held at
A junction semiconductor diode is vr;idely used as a temperature
sensor. When the temperature of doped semiconductors changes,
the mobility of their charge carriers changes and this affects the
rate at which electrons and holes can diffuse across a p-n
junction. Thus when a p-n junction has a potential difference V
across it, the current I through the junction is a function of the
temperature, being given by:
0
Hot junction

Fig. 2.54
Metal B

Referencejunction

A thermocouple
0°C and then, to a reasonable ex%ent,the following relationship
holds:

E=at+bP

where a and b are constants for the metals concerned. Commonly

used thermocouples are shorm in Table 2.1, with the temperature
ranges over which they are generally used and typical
sensitivities. These commonly used thermocouples are given
reference letters. For example, the iron-constantan thermo-
couple is called a type J thermocouple. Figure 2.55 shows how the
where T is the temperature on the Kelvin scale, e the charge on
e.m.f. varies with temperature for a number of commonly used
an electron, and k and I. are constants. By taking logarithms we
pairs of metals.
can write the equation in terms of the voltage as:
Table 2.1 Thermocouples

Ref. Materials Range OC pV/OC

Thus, for a constant current, we have V proportional to the
Platinum 30% rhodium/platinum
temperature on the Kelvin scale and so a measurement of the
6% rhodium
potential difference across a diode at constant current can be used
as a measure of the temperature. Such a sensor is compact like a ChromeVconstantan
thermistor but has the great advantage of giving a response which Ironkonstantan
is a linear function of temperature. Diodes for use as temperature ChromeValumel
sensors, together with the necessary signal conditioning, are Nirosil/nisil
supplied as integrated circuits, e.g. LM3911, and give a very Platindplatinum 13% rhodium
small compact sensor. The output voltage from LM3911 is Platindplatinum 10% rhodium
proportional to the temperature at the rate of 10 mV/"C.
In a similar manner to the thermodiode, for a thermo-transistor Copper/constantan
the voltage across the junction between the base and the emitter
depends on the temperature and can be used as a measure of
46 Mechatronics Sensors and transducers 47

Metal
Copper

Copper

Output of
Resistance compensation p.d.
element
\ Compensation circuit
Fig. 2.57 Cold junction i o c k at constant
compensation temperature

Fig. 2.55 Thermoelectric W 200 400 600 800 1000 1200 Thermocouples are generally mounted in a sheath to give them
e.m.f.-temperature graphs Temperature OC mechanical and chemical protection. The type of sheath used
depends on the temperatures at which the thermocouple is to be
used. In some cases the sheath is packed with a mineral which is
A thermocouple circuit can have other metals in the circuit and a good conductor of heat and a good electrical insulator. The
they will have no effect on the thermoelectric e.m.f. provided all response time of an unsheathed thermocouple is very fast. With a
their junctions are at the same temperature. A thermocouple can sheath this may be increased to as much as a few seconds if a
be used with the reference junction at a temperature other than large sheath is used. In some instances a group of thermocouples
0°C. The standard tables, however, assume a 0°C junction and are connected in series so that there are perhaps ten or more hot
hence a correction has to be applied before the tables can be used. junctions sensing the temperature. The e.m.f. produced by each is
The correction is applied using what is known as the law of added together. Such an arrangement is known as a thermopile.
intermediate temperatures, namely To illustrate the above, consider a type E thermocouple which is
The above is equivalent to: to be used for the measurement of temperature with a cold
junction at 20°C. What will be the thermoelectric e.m.f. at 200°C?
The following is data from standard tables.
The e.m.f. E,,o at temperature t when the cold junction is at O°C
equals the e.m.f. E;. I at the intermediate temperature I plus the Temp. PC) 0 20 200
e.m.f. EI,O at temperature I when the cold junction is at 0°C (Fig.
e.m.f. (mV) 0 1.192 13.419
Fig. 2.56 Law of intermediate 2.56). To maintain one junction of a thermocouple at O°C, i.e.
temperatures have it immersed in a mixhue of ice and water, is often not
convenient. A compensation circuit can, however, be used to Using the law of intermediate temperatures
provide an e.m.f. which varies with the temperature of the cold
junction in such a way that when it is added to the thermocouple
e.m.f. it generates a combined e.m.f. which is the same as would
have been generated if the cold junction had been at 0°C (Fig. Note that this is not the e.m.f. given by the tables for a temp-
2.57). The compensating e.m.f can be provided by the voltage erature of 180°C with a cold junction at O°C, namely 11.949 mV.
drop across a resistance thermometer element.
The base-metal thermocouples, E, J, K and T, are relatively
cheap but deteriorate with age. They have accuracies which are 1 2.10 Light sensors Photodiodes are semiconductorjunction diodes (see Section 7.3.1
typically about *l to 3%. Noble-metal thermocouples, e.g. R, are
more expensive but are more stable with longer life. They have
1 for discussion of diodes) which are connected into a circuit in
reverse bias, so giving a very high resistance, so that when light
accuracies of the order of *l% or better. i falls on the junction the diode resistance drops and the current in
the circuit rises appreciably. For example, the current in the
absence of light with a reverse bias of 3 V might be 25 pA and
48 Mechatronics Sensors and transducers 49

when illuminated by 25 000 lumens/m2 the current rises to 375 system after signal conditioning, and thus there has to be a
pA. The resistance of the device with no light is 3/(25 x lo4) = suitable marriage between sensor and signal conditioner.
120 kf2 and with light is 3/(375 x lo4) = 8 162. A photodiode can To illustrate the above, consider the selection of a sensor for the
thus be used as a variable resistance device controlled by the light measurement of the level of a corrosive acid in a vessel. The level
incident on it. Photodiodes have a very fast response to light. can vary from 0 to 2 m in a circular vessel which has a diameter
The phototransistors (see Section 7.3.3 for a discussion of of 1 m. The empty vessel has a weight of 100 kg. The minimum
transistors) have a light-sensitive collector-base p n junction. variation in level to be detected is 10 cm. The acid has a density
When there is no incident light there is a very small of 1050 kg/m3. The output from the sensor is to be electrical.
collector-toemitter current. When light is incident, a base current Because of the corrosive nature of the acid an indirect method of
is produced that is directly proportional to the light intensity. This determining the level seems appropriate. Thus it is possible to use
leads to the production of a collector current which is then a a load cell, or load cells, to monitor the weight of the vessel. Such
b measure of the light intensity. Phototransistors are often available cells would give an electrical output. The weight of the liquid
as integrated packages witl~the phototransistor connected in a changes from 0 when empty to, when full, 1050 x 2 x n(12/4) x
Fig. 2.58 -Photo Darlington Darlington arrangement with a conventional transistor (Fig. 9.8 = 16.2 kN. Adding this to the weight of the empty vessel gives
2.58). Because this arrangement gives a higher current gain, the a weight that varies from about 1 to 17 W. The resolution
device gives a much greater collector current for a given light required is for a change of level of 10 cm, i.e. a change in weight
intensity. of 0.10 x 1050 x n(12/4) x 9.8 = 0.8 W. If three load cells are
A photoresistor has a resistance which depends on the intensity used to support the tank then each will require a range of about 0
of the light falling on it, decreasing linearly as the intensity to 6 kN with a resolution of 0.27 kN.Manufacturers' catalogues
increases. The cadmium sulphide photoresistor is most responsive can then be consulted to see if such load cells can be obtained.
to light having wavelengths shorter than about 515 nm and the
cadmium selinide photoresistor for wavelengths less than about
700 nm. 2.12 Inputting data by Mechanical switches consist of one or more pairs of contacts
An array of light sensors is often required in a small space in switches which can be mechanically closed or opened and in doing so
order to determine the variations of light intensity across that make or break electrical circuits. Thus 0 or 1 signals can be
space, e.g. in the automatic camera to determine the exposure that transmitted by the act of opening or closing a switch.
will be most appropriate to take account of the varying light Mechanical switches are specified in terms of their number of
intensities across the image. To this end array devices are poles and throws. Poles are the number of separate circuits that
available with large numbers of photodiodes in an array. can be completed by the same switclung action and throws are the
number of individual contacts for each pole. Figure 2.59(a) shows
a single pole-single throw (SPST) switch, Figure 2.59(b) a single
2.11 Selection of sensors In selecting a sensor for a particular application there are a pole-double throw (DPDT) switch and Figure 2.59(c) a double
number of factors that need to be considered: pole-double throw @PDT) switch.

1 The nature of the measurement required, e.g. the variable to 2.12.1 Debouncing
be measured, its nominal value, the range of values, the
accuracy required, the required speed of measurement, the A problem that occurs with mechanical switches is switch
reliability required, the environmental conditions under bounce. When a mechanical switch is switched to close the
Fig. 2.59 Switches: (a) SPST, contacts, we have one contact being moved towards the other. It
which the measurement is to be made.
(b) SPDT, (c) DPDT hits the other and, because the contacting elements are elastic,
2 The nature of the output required from the sensor, this
determining the signal conditioning requirements in order to bounces. It may bounce a number of times (Fig. 2.60) before
Connection Final
give suitable output signals from the measurement. finally settling to its closed state after, typically, some 20 rns.
closed
3 Then possible sensors can be identified, taking into account made+ Bounces contack Each of the contacts during tlus bouncing time can register as a
such factors as their range, accuracy, linearity, speed of separate contact Thus, to a microprocessor, it might appear that
response, reliability, maintainability, life, power supply perhaps two or more separate witch actions have occurred.
requirements, ruggedness, availability, cost. Similarly, when a mechanical switch is opened, bouncing can
Initially, occur. To overcome this problem either hardware or software can
The selection of sensors cannot be taken in isolation from a no connection be used.
consideration of the form of output that is required from the Fig. 2.60 Switch bounce on closing With software, the microprocessor is programmed to detect if
a switch the switch is closed and then wait, say, 20 ms. After checking that
50 Mechatronics Sensors and transducers 51
bouncing has ceased and the switch is in the same closed position, bottom one to make the connection and then opens when the
the next part of the program can take place. The hardware pressure is released.
solution to the bounce problem is based on the use of a flip-flop. While each switch in such arrays could be connected to
Figure 2.61 shows a circuit for debouncing a SPDT switch which individually give signals when closed, a more economical method
is based on the use of a SR flip-flop (see Section 14.7). As shown, is to connect them in an array in that an individual output is not
we have S at 0 and R at 1 with an output of 0.When the switch is needed for each key but each key gives a unique row-column
moved to its lower position, initially S becomes 1 and R becomes combination. Figure 2.66 shows the connections for a 16-way
0. This gives an output of 1. Bouncing in changing S from 1to 0 keypad.
to 1 to 0, etc. gives no change in the output. Such a flip-flop can
be derived from two NOR or two ? N gates. A SPDT switch
can be debounced by the use of a D flip-flop (see Section 14.7).
Fig. 2.61 Debouncing a SPDT Figure 2.62 shows the circuit. The output from such a flip-flop
switch only changes when the clock signal changes. Thus by choosing a
clock period which is greater than the time for which the bounces
last, say 20 ms,the bounce signals will be ignored.

-1Fr Clock

signal input
An alternative method of debouncing using hardware is to use a
Schrnitt trigger. This device has the 'hysteresis' characteristic
shown in Figure 2.63. When the input voltage is beyond an upper
switching threshold and giving a low output, then the input
voltage needs to fall below the lower threshold before the output
Fig. 2.62 Debouncing a SPDT can switch to high. Conversely, when the input voltage is below
switch the lower switching threshold and giving a high, then the input
needs to rise above the upper threshold before the output can
switch to low. Such a device can be used to sharpen slowly
High changing signals; when the signal passes the switching threshold
it becomes a sharply defined edge between two well-defined logic
- levels. The circuit shown in Figure 2.64 can be used for
-0'
a
a debouncing, note the circuit symbol for a Schrnitt trigger. With
the switch open, the capacitor becomes charged and the voltage
applied to the Schrnitt trigger becomes high and so it gives a low
output. When the switch is closed, the capacitor discharges very
Low rapidly and so the first bounce discharges the capacitor; the
4
Lower
4
Upper
Input Schrnitt trigger thus switches to give a high output. Successive
threshol threshol switch bounces do not have time to recharge the capacitor to the
required threshold value and so further bounces do not switch the
Fig. 2.63 Schrnitt trigger Schmitt trigger.

2.1 2.2 Keypads

A keypad is an array of switches, perhaps the keyboard of a
computer or the touch input membrane pad for some device such
as a microwave oven. A contact type key of the form generally
used with a keyboard is shown in Figure 2.65(a), depressing the
key plunger forces the contacts together with the spring returning
the key to tile off position when the key is released. A typical
I A membrane switch (Fig. 2.65(b)) is built up from two wafer-thin
Switch -
- plastic films on which conductive layers have been printed These
layers are separated by a spacer layer. When the switch area of
Fig. 2.64 Debouncing the membrane is pressed, the top contact layer closes with the