Statistical Process Control

P i r z a d a
Common Causes of Variation

Common cause variability is a source of variation caused by

unknown factors (intrinsic) that result in a steady but
random distribution of output around the average of the data.
Common cause variation determines how well the process can
perform.

Common cause variation is also called random variation, noise,

non-controllable variation, within-group variation, or inherent
variation.
Special Causes of Variation

Special cause variation is caused by known factors that result in

a non-random distribution of output. Also referred to as
“assignable” variation.

Example: A shift in output caused by a specific factor such as

environmental conditions or process input parameters. It can
be accounted for directly and potentially removed and is a
measure of process control.
Special Causes of Variation:

Types
Special Causes: One Point Outside
Control Limit

Special Cause _
One Point Outside Control Limit (- in x or R-charts)

Reasons
_
• Error inx or R of sample
• Sudden power surge
• A broken tool
• Measurement error incomplete/ omitted process
Special Causes: Sudden Shift
Sudden Shift in the Process
_ Average
- in x
or R-charts
Special Cause
 Process average suddenly shifted

Reasons

• New operator
• New machine setting
• New setup
Special Causes: Cycles

Special Cause
Short repeated patterns

Reasons:
• Operator rotation
• Fatigue
• Different tool
• Time effects
Special Causes: Trends

Special Cause
Gradual effect of a cause

Reasons
• Dirt/ chip buildup
• Tool wear
• Aging of equipment
Special Causes: Mixture
Hugging the Control Limits- Mixture

Special Cause
Combination of two pattern

Reasons:

Different lot of materials or machines

Can split in two separate patterns

Special Causes: Hugging the Center Line

Special Cause
Points clustering the centerline

Reasons
 Control limits too wide
 Systematic error in sample
 Miscalculation of control limits
Special Causes: Instability

 Unnatural and erratic fluctuations

 Usually on both sides and outside limits

Reasons
• Difficult to identify
• Over adjustment
Statistical Process Control

A methodology for monitoring a process to identify special

causes of variation and signaling the need to take corrective
action when appropriate.

If the variation in the process is due to common cause alone, the

process is said to be in statistical control.
Capability and Control
The Control

A phenomenon will be said to be controlled when, we can

predict, at least within limits, how the phenomenon may be
expected to vary in the future, i.e. we can state, at least
approximately, the probability that the observed phenomenon
will fall within the given limits.

Control is not the complete absence of variation.

 Variation is predictable. Process mean and variance are
constant over time.
 Not necessarily a sign of good management,
Shewhart
Control and Capability

Two different concepts

 The issue of control should
be addressed first

 Then attack the common In control Out of control

causes to improve its
capability Capable
Ideal

Not
capable
Statistical Process Control

Features of SPC

 Relies on control charts

 An evidence of quality capability
 Ineffective for approaching six sigma
 Good for early stages of quality enhancement
Statistical Process Control

Quality control measurements (2 types of indicators)

Variables
Continuous data, (e.g. length or weight).
Measure of degree of conformance.
Attribute
A characteristic either present or ascent, (e.g. within
tolerance or not). Usually easier data collection
Measure of conformance.
SPC Methodology - I

Preparation 1. Identify variables to be measured

2. Decide basis, size & frequency of sampling
3. Set-up control chart
Data Collection 1. Record the data
2. Calculate, averages, ranges, proportions etc
3. Plot charts
Control Limits 1. Draw central line (process average)
2. Compute upper & lower control limits
SPC Methodology - II

Analysis and 1. Choose measurable attributes

interpretation 2. Determine basis, size and frequency of
sampling
3. Set-up control chart

Using as problem- 1. Continue data collection & plotting

solving tool 2. Identify out of control situations

Process capability 1. Determine Process capability using

evaluation. control chart data.
Theory of SPC

Assumptions for the Criteria

1.Distribution of sample means is normal

- (Central limit theorem)
2. The UCLA & LCL are + 3σ around mean. ( ~ 99.1% values)
3. The normal distribution being symmetric , (mean is median)
dispersion around mean is also symmetric.
4. As ~ 68% values are within + σ, most values are close to
center line.
Statistical Process Control
Control Charts

Constructing Charts

1. 25-30 samples of 3-10 measurements are collected

2. Is samples fall out of CL, special causes exist. The process is
non-representative of SPC. Eliminate bias, recalculation needed.
3. R-charts are analyzed first. Why?

Follow example discussed on page 695

Control Charts

Control Chart for Variable Data

 Variable data are those measured along a continuous scale.

 Most commonly represented by x-bar charts and R- charts
 x-bar charts monitor the centering of the process
 R-charts measure the monitoring of the variation
 Standard deviations are better measure of variability.
SPC

Criteria for Statistical Control Process

1. No points are outside control limits.

2. The number of points above and below the center line as
about the same.
3. The points seem to fall randomly above and below the center
line.
4. Most points, but not all, are near the center line, and only a
few are close to the control limits.
SPC
Process Monitoring & Control

Used to monitor performance and

identify special causes.
Employees running the process use these Control charts
indicate when to
charts.
take action and
They should be involved and trained when to leave a
Revisions can be made to improve process alone.
process capability.
X-bar Charts &
R- Charts
 Statistical Process Control

Estimating UCL and LCL

The constants
_
A2, D3 and D4
xi : mean depend on the
 sample size.
x : overall mean
Values are
R : range given in
 Appendix B,
R
: average range Page A-3
 Statistical Process Control

Constructing X and R-charts


_
_ UCLR  D 4 R
  xi _
x LCLR  D 3 R
k  _
_
Ri UCL_  x  A2 R
R x

k 
LCL_  x  A2 R
_

x
Silicon Wafer Example No. Data Avg. Range
1 41 70 22 44 48
Page 698 Evans 2 78 53 68 66 25
3 84 34 48 55 50
4 60 36 25 40 35
5 46 47 29 41 18
6 64 16 56 45 48
7 43 53 64 53 21
8 37 43 30 37 13
9 50 29 57 45 28
10 57 83 32 57 51
11 24 42 39 35 18
12 78 48 39 55 39
13 51 57 50 53 7
14 41 29 35 35 12
15 56 64 36 52 28
16 46 41 16 34 30
17 99 86 98 94 13
18 71 54 39 55 32
19 41 2 53 32 51
20 41 39 36 39 5
21 22 40 46 36 24
22 62 70 46 59 24
23 64 52 57 58 12
24 44 38 60 47 22
25 41 63 62 55 22

Avg. 49 27
Silicon Wafer Example
Page 698 Evans
 R
R  27
k

 x
X  48.8
k

From Appendix B of A2 R  1.023.(27)  27.6
the Text book, the
 
values of coefficients
for n=3 (sample size)
UCLx  X  A2 R
are:  
  LCLx  X  A2 R
A2 = 1.023 
D3 = 0 UCLR  D4 R  2.574.(27)  69.5
D4 = 2.574 
LCLR  D3 R  0.(27)  0
 Variable Control R-Chart

80

70

60

50

40

30

20

10

0
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
 Variable Control Chart X-bar

100

90

80

70

60

50

40

30

20

10
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
 Variable Control Chart X-bar & R
100

90

80

70

60

50

40

30

20

10
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

60
50
40
30
20
10
0
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
 Variable Control Chart R (after removing #17)

60

50

40

30

20

10

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
 Variable Control Chart X-bar (after removing #17)
80

70

60

50

40

30

20

10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
 Variable Control Chart X-bar & R ( after removing #17)
80

70

60

50

40

30

20

10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
60

50

40

30

20

10

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Special Control Charts

x-bar and s-charts

Sample St. Dev. (s) as alternative to R

Construction:
 Compute s for individual samples
 Find s-bar (Mean of all s values)
 Find UCLS and LCLS using B3 and B4
 For x-bar chart use A3

Values of A3, B3 and B4 from Table

x-bar s-Chart: Mechanical Part Example
1 2 3 4 5 6 7 8 9 10 Avg. St. Dv.
1 8 6 9 7 9 2 7 9 7 6.5 2.838
9 4 0 3 0 0 3 4 8 3 3.4 3.134
0 8 0 0 3 1 2 0 2 3 1.9 2.470
1 1 0 2 1 1 2 0 0 1 0.9 0.738
-3 -1 0 -4 0 1 0 -2 0 -2 -1.1 1.595
-6 2 0 0 2 -1 2 0 -3 0 -0.4 2.503
-3 -1 0 -2 -1 -1 -3 0 -2 -2 -1.5 1.080
0 -2 -3 -1 -2 1 -3 0 -3 -2 -1.5 1.434
2 0 -1 -1 -3 0 1 -3 -1 0 -0.6 1.578
0 0 -2 -1 -1 0 -1 -2 -2 0 -0.9 0.876
-3 -2 2 -1 1 -2 -2 -1 1 1 -0.6 1.713
-12 2 0 -4 -1 4 2 -3 -4 0 -1.6 4.526
-6 -3 0 0 -8 -4 -6 -1 -1 -2 -3.1 2.807
-3 -5 5 0 -5 1 5 -4 -1 -5 -1.2 3.910
-1 -1 -1 -2 -1 0 -2 -1 0 -1 -1 0.667
-1 -2 -2 0 -4 0 -2 -4 -1 0 -1.6 1.506
-2 2 -1 0 -1 -1 2 -1 1 -2 -0.3 1.494
0 4 0 0 0 3 0 0 1 0 0.8 1.476
0 3 -3 3 3 1 0 1 2 -2 0.8 2.098
1 2 1 1 -3 2 1 -2 3 0 0.6 1.838
1 2 2 1 2 2 1 1 1 2 1.5 0.527
-1 0 2 -1 2 2 -1 0 0 -1 0.2 1.317
0 0 -1 0 1 0 0 0 -1 0 -0.1 0.568
1 0 0 1 1 2 0 0 -1 0 0.4 0.843
2 2 1 2 -1 2 2 1 -1 2 1.2 1.229
x-bar s-chart
Mechanical Part Example

No. of samples: 25 Sample size: 10 Grand average: 0.18

Average St. Dev. 1.7905259

A3: 0.98
B3: 0.28
B4: 1.72
d2: 3.08

For X-bar chart

Center*: 0.18
LCL: -1.64
UCL: 1.854
x-bar s-chart
Mechanical Part Example

No. of samples: 25 Sample size:10 Grand Average: 0.18

Average St. Dev: 1.7905259

A3: 0.98
B3: 0.28
B4: 1.72
d2: 3.08
For S chart
Center*: 1.791
LCL: 1.791.(0.28) =3.073
UCL: 1.791.(1.22) =0.501
x-bar Chart
Mechanical part example

0
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

-2

-4
s-Chart
Mechanical part example

The LCL as calculated in

5.000
the previous slide should
be ~ 0.5. LCL drawn in
4.000
this graph is wrong.
3.000

2.000

1.000

0.000
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

-1.000

-2.000
Attribute Control Charts

Fraction nonconforming (p) charts)

Difficult to measure attributes, good-bad, pass fail,

Construction n: sample size

y: no. non-conforming
 Take 25-30 samples.
p : fraction nonconforming
 Large enough sample
to have several defect =y/n
 UCL, LCL p-bar + 3s p-bar = average p
s (p-bar) =
Sqrt [p-bar (1-p-bar)/ n]
Page 722
p-Chart
1 3 100 0.03 0.01467 0 0.022 0.066
2 1 100 0.01 0.01467 0 0.022 0.066
3 0 100 0 0.01467 0 0.022 0.066
4 0 100 0 0.01467 0 0.022 0.066
5 2 100 0.02 0.01467 0 0.022 0.066
6 5 100 0.05 0.01467 0 0.022 0.066
7 3 100 0.03 0.01467 0 0.022 0.066
8 6 100 0.06 0.01467 0 0.022 0.066
9 1 100 0.01 0.01467 0 0.022 0.066
10 4 100 0.04 0.01467 0 0.022 0.066
11 0 100 0 0.01467 0 0.022 0.066
12 2 100 0.02 0.01467 0 0.022 0.066
13 1 100 0.01 0.01467 0 0.022 0.066
14 3 100 0.03 0.01467 0 0.022 0.066
15 4 100 0.04 0.01467 0 0.022 0.066
16 1 100 0.01 0.01467 0 0.022 0.066
17 1 100 0.01 0.01467 0 0.022 0.066
18 2 100 0.02 0.01467 0 0.022 0.066
19 5 100 0.05 0.01467 0 0.022 0.066
20 2 100 0.02 0.01467 0 0.022 0.066
21 3 100 0.03 0.01467 0 0.022 0.066
22 4 100 0.04 0.01467 0 0.022 0.066
23 1 100 0.01 0.01467 0 0.022 0.066
24 0 100 0 0.01467 0 0.022 0.066
25 1 100 0.01 0.01467 0 0.022 0.066
p-chart Formulae

_
Central line= p
_ p1  p 2  p  ....  p k
p
k

_
UCL  p  3.s _  
p p(1  p )
_
s_ 
LCL  p  3.s _
p n
p
 Plot of p-chart
0.07

0.06

0.05

0.04

0.03

0.02

0.01

0
0 5 10 15 20 25
c-chart Formulae

 c1  c2  c3  ...ck
Central line (CL): c 
k


UCL  c  3.s 
C 


s  c
C
LCL  c  3.s 
C
Attribute Control Charts

Average no of defects (c) chart

Defect: non-conforming characteristic

 c-chart: Total number of defects/ unit: s c
 Sample size constant
 Follows Poisson distribution

 u-chart: Average number of defects/ unit

 Sample size variable.
Sample Defects
No c CL UCL LCL
1 2 1.8 5.82 0
2 3 1.8 5.82 0
3 0 1.8 5.82 0
4 1 1.8 5.82 0
5 3 1.8 5.82 0
7
6 5 1.8 5.82 0 c-chart
7 3 1.8 5.82 0 6
8 1 1.8 5.82 0
5
9 2 1.8 5.82 0
10 2 1.8 5.82 0 4
11 0 1.8 5.82 0
12 1 1.8 5.82 0 3

13 0 1.8 5.82 0 2
14 2 1.8 5.82 0
15 4 1.8 5.82 0 1

16 1 1.8 5.82 0
0
17 2 1.8 5.82 0 0 5 10 15 20 25

18 0 1.8 5.82 0
19 3 1.8 5.82 0
20 2 1.8 5.82 0
21 1 1.8 5.82 0
22 4 1.8 5.82 0
23 0 1.8 5.82 0
24 0 1.8 5.82 0
25 3 1.8 5.82 0
Average 1.8
2 Special Control Charts
Moving Range Chart
for x-Chart n = 1
 R- chart   x-chart
S# Value Range Cl LCL UCL   S# Value CL UCL LCL
1 3.75   1 3.75 3.498 4.344 2.562
2 3.8 0.05 0.35 0 1.15   2 3.8 3.498 4.344 2.562
3 3.7 0.1 0.35 0 1.15   3 3.7 3.498 4.344 2.562
4 3.2 0.5 0.35 0 1.15   4 3.2 3.498 4.344 2.562
5 3.5 0.3 0.35 0 1.15   5 3.5 3.498 4.344 2.562
6 3.05 0.45 0.35 0 1.15   6 3.05 3.498 4.344 2.562
7 3.5 0.45 0.35 0 1.15   7 3.5 3.498 4.344 2.562
8 3.25 0.25 0.35 0 1.15   8 3.25 3.498 4.344 2.562
9 3.6 0.35 0.35 0 1.15   9 3.6 3.498 4.344 2.562
10 3.1 0.5 0.35 0 1.15   10 3.1 3.498 4.344 2.562
11 4 0.9 0.35 0 1.15   11 4 3.498 4.344 2.562
12 4 0 0.35 0 1.15   12 4 3.498 4.344 2.562
13 3.5 0.5 0.35 0 1.15   13 3.5 3.498 4.344 2.562
14 3 0.5 0.35 0 1.15   14 3 3.498 4.344 2.562
15 3.8 0.8 0.35 0 1.15   15 3.8 3.498 4.344 2.562
16 3.4 0.4 0.35 0 1.15   16 3.4 3.498 4.344 2.562
17 3.6 0.2 0.35 0 1.15   17 3.6 3.498 4.344 2.562
18 3.1 0.5 0.35 0 1.15   18 3.1 3.498 4.344 2.562
19 3.55 0.45 0.35 0 1.15   19 3.55 3.498 4.344 2.562
20 3.65 0.1 0.35 0 1.15   20 3.65 3.498 4.344 2.562
21 3.45 0.2 0.35 0 1.15   21 3.45 3.498 4.344 2.562
22 3.3 0.15 0.35 0 1.15   22 3.3 3.498 4.344 2.562
23 3.75 0.45 0.35 0 1.15   23 3.75 3.498 4.344 2.562
24 3.5 0.25 0.35 0 1.15   24 3.5 3.498 4.344 2.562
25 3.4 0.1 0.35 0 1.15   25 3.4 3.498 4.344 2.562
x-chart: Moving Range Chart

R-chart x-chart
_
_
CL  x
CL  R
_ _
_
UCL  x  3. * R/ d 2
UCL  D 4 * R
_ _
_
LCL  x  3. * R/ d 2
UCL  D3 * R

^ _
  R/ d 2
Moving Range Chart
1.4

1.2

0.8

0.6

0.4

0.2

0
0 5 10 15 20 25 30
x-chart: Moving Range Chart

4.25

3.75

3.25

2.75

2.25
1 6 11 16 21
Variable Size p-chart
Sample size # Defectives y/n      
n y p CL UCPL LCL
137 18 0.131387 0.094519 0.167869 0.021168
158 20 0.126582 0.094519 0.162821 0.026217
92 14 0.152174 0.094519 0.184028 0.005009
122 6 0.04918 0.094519 0.172248 0.01679
86 11 0.127907 0.094519 0.187098 0.001939
187 22 0.117647 0.094519 0.157302 0.031736
156 6 0.038462 0.094519 0.163257 0.02578
117 9 0.076923 0.094519 0.173891 0.015146
110 14 0.127273 0.094519 0.176378 0.01266
142 12 0.084507 0.094519 0.166566 0.022471
140 8 0.057143 0.094519 0.167079 0.021958
179 13 0.072626 0.094519 0.158689 0.030348
196 5 0.02551 0.094519 0.155843 0.033194
163 15 0.092025 0.094519 0.161765 0.027272
140 25 0.178571 0.094519 0.167079 0.021958
135 12 0.088889 0.094519 0.168411 0.020627
186 16 0.086022 0.094519 0.15747 0.031567
193 12 0.062176 0.094519 0.156318 0.032719
181 15 0.082873 0.094519 0.158334 0.030704
160 18 0.1125 0.094519 0.162393 0.026645
Average: 0.094519
 Variable Size p-Chart
0.2

0.18

0.16

0.14

0.12

0.1

0.08

0.06

0.04

0.02

0
0 2 4 6 8 10 12 14 16 18 20
Quality Rating System

Rating the defects

A: Very serious Demerits: 100

B: Serious Demerits: 50
C: Moderately serious Demerits: 10
D: Not serious Demerits: 1
SPC
Control Chart Applications in Service Organizations
Organization Quality Measure
Hospital Lab test accuracy, Insurance claim accuracy
Bank Check-processing accuracy
Insurance company Claim-processing response time, Billing accuracy
Post office Sorting accuracy, Time of delivery, % mail delivered in
time
Ambulance Response time
Police Department Incidence of crime, No. of traffic citations
Hotel Proportion of rooms cleaned, No. of complaints
Transportation Proportion of freight cars correctly routed
Auto service % job completed as promised, % parts out of stock
 Table of Coefficients

n x-charts s-charts R-charts

A A2 A3 c4 B3 B4 B5 B6 d2 d3 D1 D2 D3 D4

2 2.121 1.881 2.659 0.7979 0 3.267 0 2.606 1.128 0.853 0 3.686 0 3.267

3 1.732 1.023 1.954 0.8862 0 2.568 0 2.276 1.693 0.888 0 4.358 0 2.574

4 1.500 0.729 1.628 0.9213 0 2.266 0 2.088 2.059 0.880 0 4.698 0 2.282

5 1.342 0.577 1.427 0.9400 0 2.089 0 1.964 2.326 0.864 0 4.918 0 2.114

10 0.949 0.308 0.975 0.9727 0.284 1.716 0.276 1.669 3.078 0.797 0.687 5.469 0.223 1.777