SPC
&
PROCESS CAPABILITY
By C.S Sarangi
PROCESS
MAN
METHOD
MACHINE
MATERIAL
EQUIPMENT
ENVIRONMENT
4M & 2E - ALL WORK TOGETHER TO
PRODUCE OUTPUT.
TRADITIONAL APPROACH FOR
CONTROL
PRODUCTION MANUFACTURES - Q.C
INSPECTS FINAL PRODUCT
DETECTION STRATEGY
WASTES MAN POWER USED TO
MANUFACTURE SUB STANDARD
PRODUCT
WASTES MANPOWER TO INSPECT GOOD
LOT
WASTES RESOURCES USED
STATISTICAL
PROCESS CONTROL
TO MONITOR A PROCESS TO PREDICT &
PREVENT NON CONFORMANCE
APPLIES TO PRESET & ONGOING
PROCESS
ECONOMICALLY SOUND DECISSION -
AVOIDS OVER & UNDER CONTROL
FEEDBACK SYSTEM - PREVENTION
STRATEGY- UNUSABLE PARTS NOT
PRODUCED
VARIATION
NO TWO ITEMS IDENTICAL
MUST BE WITHIN REQUIREMENT
COMMON CAUSES (CHANCE
CAUSES) & SPECIAL CAUSES
(ASSIGNABLE CAUSES)
COMMON CAUSES
MANY SMALL SOURCES
STABLE
OUTPUT PREDICTABLE
PERMANENT UNLESS ACTION TAKEN
PROCESS UNDER STATISTICAL
CONTROL
MANAGERIAL ACTION NEEDED
EXAMPLE-COMMON CAUSES
POOR DESIGN OF PRODUCT
SMALL INACCURACIES IN MEASUREMENT
SMALL VARIATIONS IN INCOMING MATERIAL
PROCEDURES NOT WELL SUITED TO THE REQUIREMENT
SETTINGS OF MACHINE CHRONICALLY INACCURATE
INADEQUATE ILLUMINATION
VIBRATION, HUMIDITY
OPERATOR FATIGUE
NORMAL WEAR & TEAR OF TOOLS WITH PRESENT
CONDITION ETC.
SPECIAL CAUSES
ONE OR FEW MAJOR SOURCES
MAY BE IRREGULAR
OUTPUT UNPREDICTABLE
MAY REAPPEAR IF ACTION NOT
TAKEN
MOST CORRECTED AT SHOP LEVEL.
EXAMPLE -SPECIAL CAUSES
• FIXTURE NOT CLEANED PRIOR TO USE
• TOOL NOT SHARPENED WHEN DULL
• MACHINE NOT LUBRICATED AS PER SCHEDULE
• OPERATOR NOT FOLLOWING PROCEDURES
• PASSING KNOWN FAULTY PARTS
• OPERATOR TIRED FROM THE NIGHT BEFORE
• NOT SETTING MACHINE CORRECTLY
• BEING OVER CRITICAL IN REJECTING PARTS
• NOT READING MEASUREMENT TOOLS CORRECTLY
ETC.
CONTROL CHARTS
SPC TOOL TO ANALYSE , MAINTAIN & IMPROVE
PROCESS
CONTROLLED & UNCONTROLLED VARIATION - A.
SCHEWHART
VARIABLES (X bar, R & )
ATTRIBUTES (P , NP , C)
STANDARD ASSUMPTIONS
FOR VARIABLE CONTROL CHARTS
DATA ARE NORMALLY DISTRIBUTED
DATA ARE INDEPENDANT
NORMAL DISTRIBUTION
68-95-99.7 Rule
68% of
the data
1σ
95% of the data
2σ
99.7% of the data
3σ
Normal Distribution Curve – Bell Shaped – Max data points at centre
X bar-R CONTROL CHARTS
R CHART FOR MEASUREMENT OF DISPERSION /
SPREAD
X̅ CHART FOR CENTRAL TENDENCY
BOTH CHARTS USED SIMULTANEOUSLY.
Why ?
# River Example
#Avg. of 4 & 6 is 5. Also Avg. of 2 & 8 is also 5
BENEFITS OF CONTROL CHARTS
EASILY USED BY OPERATOR
CONSISTENT PERFORMANCE
PROVIDE COMMON LANGUAGE FOR
PERFORMANCE DISCUSSION
DISTINGUISH COMMON & SPECIAL CAUSES
HIGHER QUALITY, LOWER COST, SCOPE FOR
IMPROVEMENT
BEFORE USE OF Xbar-R CONTROL CHARTS
PROCESS MUST BE UNDERSTOOD
CHARACTERISTIC DETERMINED
EXTERNAL CAUSE REDUCED
CURRENT PROCESS SETTING NOTED
EVENTS NOTED WHEN OCCUR
RELIABLE MEASUREMENT SYSTEM.
STAGES - INITIAL SET UP
DATA COLLECTION BY SETTING
FREQUENCY & SUBGROUP SIZE
FINDING OUT UCL & LCL
HOMOGENIZATION
FIXING UCL & LCL (USING STANDARD
RANGE & MEAN) FOR FUTURE
CONTROL
STAGES - FUTURE USE
CHART WITH FIXED UCL & LCL USED
FOR ON-GOING CONTROL
UNDERSTANDING OF OCCURANCE
OF SPECIAL CAUSES
REMOVE SPECIAL CAUSES
STAGES - IMPROVEMENT
PROCESS STABLE FOR LONG
PERIOD
COMMON CAUSES IDENTIFIED
COMMON CAUSES MINIMISED
CONTROL LIMITS REVISED
MANAGERIAL ACTION NEEDED FOR
IMPROVEMENT
INITIAL SET-UP &
HOMOGENIZATION OF R CHART
RANGE CHART TO BE SET-UP FIRST
MINIMUM 100 SAMPLES STUDIED
R FOR EACH SUBGROUP CALCULATED
(R̅ = XMAX - XMIN ).
AVERAGE RANGE CALCULATED
(R̅ = R / K , K IS NO. OF SUBGROUPS)
UCLR= D4R̅ , LCLR = D3R̅ . D4 & D3 ARE
CONSTANTS FROM TABLE RELATED TO SUB
GROUP SIZE
HOMOGENIZATION OF RANGE CHART
DRAW CHART WITH R̅ ,UCL & LCL
COMPARE SUBGROUP RANGES
OUT OF CONTROL POINTS-PRESENCE OF SPECIAL
CAUSE - ELIMINATED
REST SUB GROUPS PLOTTED
REPEATED TILL ALL POINTS WITHIN CONTROL
LIMITS
RANGE FOR HOMOGENIZED SUBGROUPS CALLED
STANDARD RANGE
POINTS AFFECTED BY SPECIAL CAUSE
ELIMINATED -ONLY COMMON CAUSES REMAIN
HOMOGENIZATION OF X-bar CHART
HOMOGENISED SUB GROUPS FROM RANGE
CHART TO BE STUDIED
SUBGROUP AVERAGES CALCULATED
(X̅=X1+X2+ --Xn / n )
OVERALL AVERAGE X CALCULATED
(X̿ = (X̅) / K , K - NO. OF SUBGROUPS)
UCL = X̿ + A2 R̅ , LCL = X̿ - A2 R̅
HOMOGENIZATION OF X-bar CHART
X̅ FOR ALL SUBGROUP PLOTTED
OUT OF CONTROL POINTS ELIMINATED TILL ALL
POINTS REMAIN WITHIN CONTROL LIMITS
NOT MORE THAN 25% IN X̅ & R CHART.
MEAN VALUE CALLED STANDARD MEAN
IF STANDARD MEAN IS FAR AWAY FROM SPEC.
MEAN , THEN SPEC. MEAN & STANDARD RANGE TO
BE USED FOR DRAWING CONTROL LIMITS -
PROCESS TO BE CENTRED - FOR ON GOING
PROCESS CONTROL.
Control Chart Sheet
Sub group 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Sam ple 1 0.555 0.555 0.565 0.555 0.555 0.555 0.57 0.565 0.57 0.555 0.58 0.575 0.56 0.57 0.57 0.575 0.555 0.555 0.555 0.56 0.565 0.565 0.565 0.56 0.565
Sam ple 2 0.55 0.56 0.55 0.555 0.565 0.56 0.56 0.56 0.575 0.56 0.575 0.565 0.565 0.575 0.57 0.57 0.55 0.57 0.565 0.565 0.57 0.56 0.57 0.57 0.57
Sam ple 3 0.56 0.55 0.555 0.555 0.565 0.565 0.55 0.57 0.575 0.575 0.56 0.57 0.575 0.575 0.565 0.57 0.56 0.55 0.555 0.575 0.595 0.56 0.56 0.575 0.565
Sam ple 4 0.55 0.56 0.555 0.56 0.575 0.565 0.565 0.57 0.56 0.565 0.56 0.57 0.565 0.56 0.56 0.565 0.56 0.57 0.56 0.57 0.56 0.565 0.56 0.575 0.56
Sam ple 5 0.55 0.56 0.57 0.57 0.56 0.56 0.57 0.57 0.56 0.565 0.555 0.565 0.565 0.575 0.56 0.565 0.56 0.565 0.555 0.555 0.59 0.565 0.565 0.555 0.555
X - BAR 0.553 0.557 0.559 0.559 0.564 0.561 0.563 0.567 0.568 0.564 0.566 0.569 0.566 0.571 0.565 0.569 0.557 0.562 0.558 0.565 0.576 0.563 0.564 0.567 0.563
RANGE 0.01 0.01 0.02 0.015 0.02 0.01 0.02 0.01 0.015 0.02 0.025 0.01 0.015 0.015 0.01 0.01 0.01 0.02 0.01 0.02 0.035 0.005 0.01 0.02 0.015
AV. X-BAR= 0.564 AV.RANGE= 0.015 UCL-R= 0.032 LCL-R= 0 UCL-X= 0.573 LCL-X= 0.555 SIGMA= 0.00653 Cp= 2.55
UCL-R LCL - R R R Bar
RANGE CHART
0.040
0.032
0.030
0.020 0.015
0.010
0.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 21 2 2 2 25
UCL-x bar LCL - x bar Av. X bar X BAR
X-BAR CHART
0.580
0.573
0.570
0.56
0.560
0.555
0.550
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
ON-GOING
PROCESS CONTROL
UCL & LCL FOR HOMOGENIZED
SUBGROUPS TO BE SET AS LIMITS
FOR REGULAR PLOTTING
SUBGROUP & FREQUENCY IN LINE
WITH INITIAL STUDY
OUT OF CONTROL SITUATIONS
STUDIED & ACTIONS TAKEN
CHART DIPLAYED IN SHOPFLOOR
OUT OF CONTROL
SITUATIONS
OUT OF CONTROL LIMITS IN BOTH CHART
RUN OF SEVEN POINTS ABOVE OR BELOW
CENTRAL LINE
RUN OF SEVEN POINTS UP OR DOWN
POINTS VERY NEAR TO CENTRAL LINE
POINTS VERY NEAR TO CONTROL LIMITS
POINT OUT OF
CONTROL LIMIT
AVERAGE CHART RANGE CHART
LIMITS LIMITS
MISCALCULATED MISCALCULATED
PROCESS SHIFTED- PIECE TO PIECE
EITHER AT THAT VARIABILITY
INCREASED
POINT OR PART OF A
MEASUREMENT
TREND SYSTEM HAS
MEASUREMENT CHANGED.
SYSTEM HAS
CHANGED.
SEVEN POINTS ABOVE OR
BELOW CENTRAL LINE
AVERAGE CHART RANGE CHART
PROCESS AVERAGE HAS ALL ABOVE CENTRAL LINE
CHANGED - SETTING NOT IN
LOOSE FIXTURING - GREATER
THE CENTRE
SPREAD
MEASUREMENT SYSTEM
SHIFT IN PROCESS ELEMENT
HAS CHANGED.
CHANGE IN MEASUREMENT
SYSTEM
ALL BELOW
GOOD SIGN
CHANGE IN MEASUREMENT
SYSTEM
SEVEN POINTS IN A ROW
UP OR DOWN
AVERAGE CHART RANGE CHART
DETORIARATION OF IMPROVEMENT OR
M/C DETERIORATION OF
TIRED OPERATOR OPERATOR SKILL
TOOL WEAR TIRED OPERATOR
CHANGE IN
INCOMING MAT.
QUALITY.
POINTS VERY NEAR TO
CENTRAL LINE
AVERAGE CHART RANGE CHART
MISCALCULATION OF
CONTROL LIMITS SAMPLING METHOD
IMPROVEMENT DONE IN STRATIFIED
PROCESS SINCE LIMITS IMPROVEMENT DONE
WERE CALCULATED
IN PROCESS SINCE
INSPECTOR MAY NOT BE
MAKING CHECKS. LIMITS WERE
CALCULATED
INSPECTOR MAY NOT
BE MAKING CHECKS.
POINTS VERY NEAR TO
CONTROL LIMITS
AVERAGE CHART RANGE CHART
OVER CONTROL MIXTURE OF MAT. OF
LARGE DISTINCTLY
DIFFERENT QUALITY.
SYSTEMATIC
DIFFERENCES IN
MAT. QUALITY
LARGE
SYSTEMATIC
DIFFERENCES IN
TEST METHOD AND
EQUIPMENT.
PROCESS CAPABILITY
MEASURE OF INHERENT VARIATION
VARIATION CORRESPONDING TO BEST
PERFORMANCE
MINIMUM 25 SUBGROUPS
SHORT TERM STUDY - MINIMUM
EXTERNAL INFLUENCE - M/C CAP.
LONG TERM STUDY - ALL SOURCES OF
VARIATION EXPERIENCED.
ASSUMPTION
PROCESS UNDER STATISTICAL
CONTROL
INDIVIDUAL MEASUREMENTS
CONFORM NORMAL DISTRIBUTION
SPECIFICATIONS DEFINED
MEASUREMENT VARIATION
RELATIVELY SMALL
CALCULATION OF P.C
STANDARD DEVIATION , = R̅ / d2 , d2 RELATED TO
SUBGROUP SIZE
P.C = 6
P.C INCEX Cp = USL - LSL /6
PERFORMANCE INDEX
Cpk =MIN. (USL - X̿ )/3 &(X̿ - LSL)/ 3
INTERPRETATION OF
Cp & Cpk
• Cp =Cpk =1 - PROCESS IS CENTRED &
0.27% DEFECTIVE PARTS PRODUCED
• Cp Cpk - PROCESS NOT CENTRED -
SETTING TO BE ADJUSTED
• Cp , Cpk 1.33 - IDEAL CONDITION
• Cpk CAN NEVER BE MORE THAN Cp.
• Cpk MUST BE CALCULATED TO KNOW
ACTUAL PERFORMANCE.
Formulae Recall
SPC Formulae
1. Upper Control limit -X̅ Chart , UCL x = X̿ + A2R̅
2. Lower Control limit -X̅ Chart , LCL x = X̿ - A2R̅
3. Upper Control limit-R Chart, UCL R = D4R̅
4. Lower Control Limit-R Chart, LCL R = D3R̅
5. Standard Deviation, σ = R̅ / d2
6. Process Capability = 6*σ
7. Cp = Tolerance / 6*σ
8. Cpk = Min. of ( USL-X̿)/3*σ , ( X̿ - LSL)/3*σ
Statistical Table – Constants for Control Chart
No. of samples in
A2 D3 D4 d2
sub Group
2 1.88 0 3.268 1.128
3 1.023 0 2.574 1.693
4 0.729 0 2.282 2.059
5 0.577 0 2.114 2.326
6 0.483 0 2.004 2.534
7 0.419 0.076 1.924 2.707
8 0.373 0.136 1.864 2.847
9 0.337 0.184 1.816 2.970
10 0.308 0.223 1.777 3.078
11 0.285 0.256 1.744 3.173
12 0.266 0.284 1.717 3.258
13 0.249 0.308 1.692 3.336
14 0.235 0.329 1.671 3.407
15 0.223 0.348 1.652 3.472
Process Capability : Practice Paper
Parameter - I.D Spec:450-550 mm M/C: CNC 2 OPT - XX
Sub group 1 2 3 4 5 6 7 8 9 10
1
500 490 515 500 515 505 525 500 520 505
2
500 515 510 520 510 525 525 515 500 525
3
515 510 510 535 520 510 510 510 515 525
4
490 515 525 515 505 525 525 515 515 510
5
515 510 515 515 495 480 525 500 475 525
X – BAR (X̅)
RANGE (R)
X̿ : R̅ : UCL X-Bar: LCL X-Bar: UCL R:
LCL R:
Range Chart
X̅ Chart
Standard Deviation(σ) : Process Capability : Cp : Cpk
Process Capability : Practice Paper Answer
Parameter - I.D Spec:450-550 mm M/C: CNC 2 OPT - XX
Sub group 1 2 3 4 5 6 7 8 9 10
1
500 490 515 500 515 505 525 500 520 505
2
500 515 510 520 510 525 525 515 500 525
3
515 510 510 535 520 510 510 510 515 525
4
490 515 525 515 505 525 525 515 515 510
5
515 510 515 515 495 480 525 500 475 525
X - BAR 504 508 515 517 509 509 522 508 505 518
RANGE 25 25 15 35 25 45 15 15 45 20
X̿ :511.50 R̅: 26.50 UCL X-Bar: 526.87 LCL X-Bar: 496.13 UCL R: 56.05 LCL R: 0
UCL-R LCL - R R-BAR
60.000 RANGE CHART
40.000
20.000
0.000
1 2 3 4 5 6 7 8 9 10
UCL-x bar LCL - x bar Av. X bar X BAR
530.000 X-BAR CHART
520.000
510.000
500.000
490.000
480.000
1 2 3 4 5 6 7 8 9 10
Standard Deviation(σ) : 11.392 Process Capability : 68.352 Cp : 1.46 Cpk
Cp, SIGMA LEVEL & PPM REJECTIONS
Cp Sigma PPM Rejn – Ideal PPM Rejn – Practical Yield
Level (Cp=Cpk) (1.5σ shift of centre)
0.33 1σ 317400 690000 31%
0.67 2σ 45600 308537 69%
1 3σ 2700 66807 93.3
1.33 4σ 64 6210 99.3
1.67 5σ 1 233 99.98
2 6σ 0.002 3.4 99.9997
TO IMPROVE
PROCESS CAPABILITY
TO IMPROVE PROCESS CAPABILITY (AND HENCE
PERFORMANCE OF PROCESS) , CONCENTRATE ON
REDUCING COMMON CAUSES
SOME TIMES REQUIRES MANAGEMENT ACTION ON
THE SYSTEM
