Training Program

On
Measurement System Analysis
&
Statistical Process Control
as per AIAG guidelines

Presented by Octagon Consultants ,

3, Kanchan Apartment , 19Girija Society , MIT
College Road, Pune – 411 029.
020 –25439282 , 9325084622
[email protected]
1
 Section I
Introduction to Process variations
and Basic Statistics.

2
3
 Process Input Process Output

Design Specifications Finish Production

Raw materials

Variations in a process (manufacturing and or measuring) is inherent.

If no special cause exist the variations will be due to Natural Causes.

Actual Process Variation

(machine variation)
- Production
Observed Process Variation Gauge Variation
(effect of measurement)

4
 VARIATION

No two products or characteristics are exactly alike.

Process contains many sources of variability.
The difference among products may be large, or they may be
immeasurably small, but they are always present.

5
 CAUSES OF VARIATION
Chance (or Common ) causes
Many sources of little variation within a process that has a stable an repeatable
distribution over time
The process output is predictable
Resolution of these causes usually requires action on the system
Management decision (80%) based on local feed-back (20%)

Special (or Assignable) causes

Single source causing high variation that makes the process distribution change
The process output is unpredictable
Resolution of theses causes usually requires local action especially during the early
process improvement efforts

6
 SPC & MSA
Controlling the process variations by using statistics is called
Statistical Process Control (SPC).

Analyzing the Measurement System Variations by using statistical

techniques (MSA) for determining the reliability of measurement results .

STATISTICS : Deals with designing for collection of data, data analysis,

interpretation and initiating action based on the analysis .

PROCESS : Converting an input into an output by using Man,

Machine, Material, Method and Environment.

CONTROL : Ensuring to make a variable to be with in the stated

limits.

7
Product Inspection Vrs SPC – Approach.

Product Inspection .
Using statistics for Product Inspection of final
product to detect Waste –
Detection - to Tolerate Waste

SPC – Statistical Process Control.

Using statistics for controlling process variations
to prevent waste.

Prevention - to Avoids Waste

8
 Total Process Variations Vs
Measurement Variations.
S. P.C. M. S. A.
( Statistical Process Control) ( measurement System Analysis)
Capability study of Capability study of
total process measurement process

σto ta lp r os sc e= σ 2
m es +σ
2
m fg σ mes = σ bias 2 + σ R & R 2

Process capability Index Cp Measurement Process capability

Index Cg
Tolerance 0.3Tolerance
Cp = Cg =
6σ totalproce ss 6σ mes
or
6σ R & R
%GRR = x100
Tolerance

9
 Total Process Variations Measurement Process Variations

To evaluate σ Total process data to be To evaluateσ mes data to be

collected by measuring n parts in collected by measuring same part
the process . n times .

To have total process under To have measurement process

control under control

Acceptance limit Acceptance limit

Cp index should be grater than 1.0 6 σ mes should be less than 30% of
Target will be to achieve C p > 1.33, tolerance most acceptable should
1.67 , 2.0 be less than 10%

It means 6 σ total process should be

less than tolerance and most
acceptable be less than
75% of tolerance for Cp > 1.33.

10
Process under development
From the initial condition of out of control to
a controlled condition and marching
towards reduced random variations.

11
 Frequency of Variation – Normal Curve

The Curve
which shows,
Frequency of variation
of a number of
observations for the
various values around the
Arithmetic Mean.

12
 Frequency of occurrence against the actual variation
– graph will be a normal curve
F
r
e
q
u
e
n
c
y

-3 -2 -1 0 1 2 3
68.26 % Variation
95.46 %

99.73 %

Area under the normal curve - in terms of standard deviation

13
 Important Characteristics – Normal curve
For n-number of observations

Average Value

Standard Deviation
For n>20

For n<20

For observations taken from g- no. of

subgroups & m- subgroup size

Estimated Standard R
σˆ =
Deviation d 2*
d2* corresponds to g & m
14
Example - Forming Normal Distribution Curve

Following are the 10 observations of a Internal Dia. Measurement of

Nominal Size 30.00 mm
30.006 , 30.005 , 30.005 , 30.003 , 30.005 , 30.007 , 30.004 , 30.006 , 30.005 , 30.004 mm

15
Example-Forming Normal Distribution Curve
(Contd.)
Where n = 10

Average Value.

= 30.005 mm

Standard Deviation

= 1.2 µm

Estimated Standard Deviation

R
σˆ = * = 1.25 µm
d2
d2* = 3.17905
corresponds to g =1 & m= 10

16
Types of Measurement Variation

17
 VARIATION
Individual measured values may all be different, as a group
they tend to form a pattern that can be described as
distribution. This distribution can be characterized by :
Location
(typical value)

Spread (Span of values from

smallest to largest)

Shape
(the pattern of
variation - whether
it is symmetrical,
skewed, etc)
18
 Section - II
Measurement System Analysis

19
 Measurement Needs in Industries

Measurements in Industry can be also classified as

• Laboratory measurements : Calibration / Testing.
• Shop Floor Measurements : Majority times it is Mass Gauging.

→ Both of above measurements requires reliability for the Zero Defect Production.

20
 Measurement Needs in Industries

Measurements are either carried out for

Product Conformity Test

or
For Monitoring the Manufacturing Process.

Reliability of measurement results is must for

zero defect Production.

21
Basic Elements of measuring system

22
Measurement Systems Variables

These are some of the variables in a measurement

system. What others can you think of?

23
 Bias and Repeatability

(Bias)

24
Relationship between Bias and Repeatability

25
 What is meant by Reliability in Measurements ?
Measurement is the Process and is subjected to variation.
To have Knowledge & Control on Measurement Process Variation is the
Measurement Reliability.
Knowledge of possible Measurement unreliability
(uncertainty) helps us in taking Proper Decision on Product Acceptance for the its
zero defect. Control on Measurement Reliability is essential for Controlling total
manufacturing Process variation to achieve zero defect Production.

Effect of Measurement on Quality

Measurement leads to knowledge, knowledge leads to understanding and
understanding lead to improvement (decision). Therefore, any improvement in an
activity, such as in a manufacturing process, depends on its ability to measure.

• Effect of Unreliable measurement.

- Effect of wrong decision could be disastrous, if the measurement results are

not reliable.

- Level of confidence on the decision is directly related to quality of

measurement result.
26
 Types of Measurement Variation

Location Variation Width Variation

Accuracy Closeness to the true Precision Closeness of repeated
value measurements

Bias : Closeness to the reference Repeatability:conditions constant

value. Reproducibility : changing either of the
conditions
Uniformity: change in repeatability over
Linearity: Change in bias over a
a measuring range.
measuring range.

Consistency: Change in
Stability : Change in bias over a time.
repeatability over a time.

27
Measurement system eats your Tolerance

USL LSL
Design Tolerance

+U -U

Tolerance Available

28
 Reliable Measurement Uncertainty Reliable Measurement Uncertainty Reliable
Measurement Measurem Measurement
Supplier Customer ent Customer Supplier
risk in risk in risk in risk in
Part is truly accepting accepting Part is accepting accepting Part is truly
rejected rejection part truly part rejection rejected
due to lower acceptable due to Higher
size size

-U +U -U +U
T
LSL USL
Effect of measurement uncertainty on product reliability.
29
Available tolerance for the manufacturing is equal to Specified
tolerance minus measurement Uncertainty.
ISO - 14253.

TS (USL) - U = Upper Manufacturing Tolerance (UMT)

TS (LSL) + U = Lower Manufacturing Tolerance (LMT)

Available Tolerance (workable tolerance) = Design Tolerance – 2U

30
 Measurement Capability

Short term
Capability of measurement system is an estimate of combined
variation of measurement errors based on short term assessment.
This includes :
• Uncorrected Bias or Linearity
• Repeatability and Repeatability.

Capability is expressed as :

σ Capability
2
= σ Bias
2
/ Linerarity
+σ
2
GRR

31
 Measurement Performance

Long term
The net effect of significant and determinable sources of variation over
time. Performance includes error components this includes:
• Capability. ( Short term errors)
• Stability and consistency (Repeatability over time).

Performance is expressed as :

σ Performanc
2
e
= σ 2
Capability
+σ
2
Stability
+σ
2
Consistency

32
 Measurement Uncertainty Vs MSA

Measurement uncertainty is focused at reducing all the factors that

impact measurement to a set of confidence intervals, to show that,
measurement could lie somewhere between those values.

MSA is concerned with understanding the impact of variables on the

measurement result and trying to reduce its variation so that correct
decision can be made on qualification of parts produced in a
manufacturing process.

33
 Measurement Issues

Various issues regarding Measurement are highlighted by

AIAG IN MSA Manual. The important once among them
are :
Discrimination: Discrimination of a measurement system
is its capability to detect part to part variation. The
measurement system should be selected such that it
divides the tolerance into ten parts or more.

34
 Measurement Issues (Contd.)

Distinct Categories: The number of distinct categories

gives us an indication as to how many categories the total
variation can be classified into. If the number of distinct
categories is 1, the measurement system is not able to
discriminate within the process variation. The number of
distinct data categories should be at lest 5.

35
 Measurement Issues (Contd.)

Sensitivity: The measurement system needs to be

sensitive enough to detect the changes in product and
process variation.
Stability: The measurement system variation should be
under random causes alone, and not governed by any
external (assignable) causes.

36
 Guidelines for Conducting Measurement System
Analysis Studies
When Should MSA Studies be performed?
• Before being placed in a production environment. The gage should not be used until the study
results are acceptable.
• A repeat study should be done sooner as initiated by the following situations cases:
a) The gage is found to be out of calibration during regular gage maintenance.
b) A change has been done on the gage such as a component replaced, which might affect the
performance of the gage
c) A major repair has been performed on the gage.
d) The gage will be used as part of a process or equipment characterization, a process
capability study, implementation of SPC in a process area or running an experiment (i.e., a
DOE).
e) The gage measurement system does not have adequate discrimination.
f) The measurement system is statistically unstable over time.
g) The measurement variation has significantly increased over time.
h) Providing a comparison of one measuring device against another.
i) The gage is being used over a new range of the characteristic not covered by
previous MSA studies.
j) There is significantly reduced total variation in the process or product.

37
 Guidelines for Conducting Measurement System
Analysis Studies
Preparation for an MSA Gage Capability Study
• Determine if repeated measurements can be made of the same characteristic on the
same part (and at the same location within the part if applicable).
• The gage must be calibrated, repaired or adjusted prior to the MSA study, not during
the study.
• Determine the sampling plan of the gage capability study.
• Specifically, if multiple operators, parts and trials are required in the gage capability
design, some factors to be considered are:
Whenever possible, the operators chosen should be selected from those who normally
operate the gage. If these individuals are not available, then personnel should be
properly trained in the correct usage of the gage, to be as close to the actual daily
usage of the gage as possible.

38
 Guidelines for Conducting Measurement System
Analysis Studies
Preparation for an MSA Gage Capability Study - Contd.
• The sample parts must be selected from the process, which represent its entire
operating range. This is sometimes done by taking one sample per day for several
days. This is necessary because the parts will be treated in the analysis as if they
represent the full range of product variation that exists in production. In case of
simulation, the simulation must be representative of the parts/process variation.
• The sample parts must be labeled but the measurements must be made in complete
random order. To avoid any possible bias, the operator who is measuring a part should
not know which numbered part is being measured.
• Criticality: critical dimensions require more parts and/or trials to have a higher degree
of confidence.
• Part configuration: Bulky or heavy parts may dictate fewer samples and more trials.

39
Guidelines for Conducting Measurement System
Analysis Studies
Flow Diagram for MSA studies

Define study to be made Bias / Linearity & R&R

Collect Data in Random manner

Analysis the Data

Interpret result to decide System Capability

Conduct Periodically, the Stability study to

monitor System Performance.

40
 Reference manual-3rd Edition (MSA)

Variable Measurement System Study.

Stability Study
Bias study Independent Sample Method
Control Chart Method
Linearity Study
Repeatability & Reproducibility Study Range Method
Range & Average Method
Analysis of Variances ( ANOVA) Method

Attribute Measurement System Study

Risk Analysis Method Cross –Tab method
Signal Detection Method
Analytic method Gauge Performance Curve.

41
 Reference manual-3rd Edition (Contd.)
(for Destructive Non-Destructive, Measurement)
Complex or Non – Replicable Measurement System Study

Stability Studies Single Part , Single measurement per cycle

n ≥ 3 parts, Single Measurement Per Cycle per Part
Large Sample from Stable Process
Split Specimens ( General), Single Specimen Per Cycle
Test Stands.

Variability Studies Standard GRR Studies

Multiple Readings with P ≥ 2 Instruments
Split Specimens ( m = 2)
Split Specimens ( General)
Time Series Analysis
Linear Analysis
Time versus Characteristic ( Property ) Degradation

42
 Bias Study

Bias
Bias is the difference between the observed average of
Measurements and the Reference Value.
If the Bias is relatively large look for following causes,
• Error in the master
• worn components
• Instrument made to the wrong dimension.
• Instrument measuring to a wrong characteristic.
• Instrument not calibrated properly.
• Instrument used improperly by appraiser

43
 Bias Study (Contd…)

Independent Sample Method

Bias test is conducted by repeatedly measuring a component with
known reference value. Number of readings should be grater than 10.
The average of observed readings is calculated for n readings as :
n
∑ xi
X = i =1
n
Bias = Observed Average - Ref. Value

σ repeatability = Max (xi

)− min(xi )
*
d2

44
 Bias Study (Contd.)

Independent Sample Method

σr
The uncertainty for Bias is given by σb =
n
Bias is acceptable or rejected on the basis of statistic as follows,

= bias
t
σb
Bias is acceptable at α level if zero falls within 1-α confidence
bounds around the bias value as follows,

d σ
2 b   d σ  
Bias − * t( ν ,1− ) 
α ≤ zero ≤ Bias + *2 b
t(ν ,1−α2 ) 
d2  2  d2  

45
 Bias Study (Contd.)

Independent Sample Method

Example
Single Part was chosen within the operating range of the
measurement system .
The part was measured in standards Room - to establish a Ref Value of a
sample part - 6.00 mm.
The part was then measured 15 times by the lead operator.

46
 Independent Sample Method (Contd.)
Ref. Value Bias n (m) = 15, df = 10.8,
6.00 - - d2* =3.55333 d2= 3.47193
5.8 -0.2 n
1
5.7 -0.3 x = ∑ xi = 6.0067
5.9 -0.1 ni =1
T 5.9 -0.1
Bias = x − Re f = 6.0067 − 6.00 = 0.0067
R 6.0 0.0
A 6.1 0.1
( )− min (x )
σ repeatability = max xi
I 6.0 0.0 *
i

d
L 6.1 0.1 2

−
S 6.4 0.4 σ = 6.4 5.6 = 0.22514
r 3.55333
6.3 0.3
σr
6.0 0.0 σb = = 0.22514 = 0.05813
6.1 0.1 n 15
6.2 0.2 bias = 0.0067 =
5.6 -0.4 tstatistic = 0.1153
σ b 0.05813
6.0 0.0

47
 Independent Sample Method (Contd.)

Significant ‘t’ Value (2 tailed) corresponds to 95% confidence interval

( α =0.05) and for df =10.8
t = 2.206
95%Confidence interval of Bias

Lower d 2σ b   3.472 × 0.05813

Bias − t  = 0.0067 − × 2.206
d 2*  ( )
α

3.55333
ν ,1−
2

LowerLimit = −0.1185

Upper d 2σ b   3.472 × 0.05813

Bias + t
 (ν ,1− 2 ) 
α
= 0. 0067 + × 2.206
d 2*   3.55333

UpperLimit = 0.1319

lowerbound (−0.1185) 〈 zero 〈Upperbound (0.1319)

Bias is insignificant i.e no chances for bias correction.
Capability will be based on R&R study.
48
Variable Measurement system Study:
Bias Study -Control Chart Method

49
 Variable Measurement system Study:
Bias Study -Control Chart Method (Contd.)
R
Compute σ repeatability from control chart parameter, σ repeatability =
d 2*
σr
The uncertainty for Bias is given by σb =
g
Bias is acceptable or rejected on the basis of t statistic as follows ,

= bias
t
σb
Bias is acceptable at α level if zero falls within 1-α confidence bounds
around the bias value as follows,

d σ
2 b   d σ
2 b  
Bias − * t( ν ,1− ) 
α ≤ zero ≤ Bias + t
*  (ν ,1−α ) 
d2  2 
d2  2 
50
 Linearity Study

Linearity is the difference in the Bias Values through a

expected operating range of the Gauge.

51
 Linearity Study (Contd.)

The linearity is determined by slope of Regression line

Lower the slope grater the linearity and vice versa.

52
 Linearity Study (Contd.)

Non -Linearity in Measurement System is caused by followings:

• Instrument not calibrated properly at both lower and upper end of

the operating range.

• Error in maximum or minimum master.

• Worn out Instrument.

• Internal instrument design characteristics.

53
 Variable Measurement system Study:
Linearity Study

Linearity test is performed as follows:

1. Select more than 5 parts ( g) for evaluation.
2. Measure each part under layout conditions and establish its
reference value.
3. Each part is measured more than 10 times (m) by instrument
under test.
4. Calculate Bias at every level and also average bias at all levels.

54
 Variable Measurement system Study:
Linearity Study (Contd.)

Best fit line is fitted as

For ‘reference value’ against ‘Bias Value’.
Regression parameters ‘slope’ and ‘intercepts’ are calculated as
 1 
∑ xy −  ∑ x∑ y 
 gm 
a= = Slope b = y − ax
1
∑x −
2
(∑ x)2
gm
55
 Variable Measurement system Study:
Linearity Study (Contd.)
Acceptance Criteria :
For a given intercept, α level confidence bands are computed as :
∑ y 2
− b∑ yi − a ∑ xi yi
s = i

gm − 2

2  2 
Lower bound :
 
 1
b + axo − t(gm−2,1−α )
x −x 
+ o
( )
 s


2  gm

2
∑ xi − x   ( )



( ) 
2
 1 x − x
2

 
Upper bound : b + axo + t( ) gm +
o
s

gm − 2 ,1−
α
2
 ( )
∑ xi − x 
2



If Bias = 0, the line should lie entirely within the confidence bands:

56
 Linearity Study

Example
Gauge /Instrument under test - Plunger type Dial Gauge .
Range 10 mm / Rev. 0.1mm

Parts selected - 5 Nos , ( was measured in standards Room) Ref

value of parts selected - 2.0 , 4.0 , 6.0 , 8.0, 10.0
Following are the observations of each part measured 12 times by
one appraiser.

57
 Linearity Study (Contd.)

Ref. Value 2.00 4.00 6.00 8.00 10.00

1 2.7 5.1 5.8 7.6 9.1
2 2.5 3.9 5.7 7.7 9.3
3 2.4 4.2 5.9 7.8 9.5
4 2.5 5.0 5.9 7.7 9.3
5 2.7 3.8 6.0 7.8 9.4
6 2.3 3.9 6.1 7.8 9.5
7 2.5 3.9 6.0 7.8 9.5
8 2.5 3.9 6.1 7.7 9.5
9 2.4 3.9 6.4 7.8 9.6
10 2.4 4.0 6.3 7.5 9.2
11 2.6 4.1 6.0 7.6 9.3
12 2.4 3.8 6.1 7.7 9.4

Part Avg. 2.4920 4.1250 6.0250 7.7080 9.3830

Bias 0.4917 0.1250 0.0250 -0.2917 -0.6167
Range 0.4000 1.3000 0.7000 0.3000 0.5000

58
 Linearity Study (Contd.)

a = -0.13167
b = 0.73667
Xo 2.0 4.0 6.0 8.0 10.0
Σ yi 2 3.07 2.39 0.43 1.13 4.8
Σ yi 5.90 1.5 0.3 -3.5 -7.4
Σ xi yi 11.8 6.0 1.8 -28.0 -74.0
s 0.069 0.1891 0.08769 0.019291 0.0935670
Upper 0.562485 0.3391 -0.03069 -0.30350 -0.45914
Lower 0.3841 0.08076 -0.0760 -0.32988 -0.700927
Avg. Bias 0.491667 0.125 0.025 -0.29167 -0.61667

At Xo = 6.0 & Xo = 8 avg. bias values are going out

of upper & lower band. This is why the system has a
linearity problem.

59
 Linearity Study (Contd.)

Linearity plot –
plotted in GAGEpack a Gauge management software – by PQ Systems USA.

60
 Repeatability & Reproducibility Study

Elements of observed Variations

Observed variation comprises of a combination of the following :

V observed = V instrument + V appraiser + V Part to part

61
 Repeatability & Reproducibility Study (Contd.)

Repeatability is the Variation in measurements obtained with one

measuring Instrument when used several times by one appraiser while
measuring the identical characteristics on the Same Part .

62
 Repeatability & Reproducibility Study (Contd.)

Reproducibility is the variation in average of the measurements made by

different appraisers using the same measuring instrument When
measuring the identical characteristics on the same Part.

63
Repeatability & Reproducibility Study (Contd.)

64
 Range method – for R & R test

• select two apprisers .

• Select min 5 sample Parts , most preferable 10 parts , which will
represent the entire process variation.
• Measure each part one time by a operator A and similarly one time
by operator B, in a random order.
• Tabulate the results and calculate the contribution of sources of
variation.
Note - Before starting the test , the exact location of measurement
should be marked on the sample parts , taken for the test. This
is to avoid within part variation. Both the measurements has
to be taken at same the location.

65
Range method – for R & R test (Contd.)
Part Appraiser A Appraiser B Range
1 0.85 0.80 0.05
2 0.75 0.70 0.05
3 1.00 0.95 0.05
4 0.45 0.55 0.10
5 0.50 0.60 0.10

∑R
R= i = 0.35 = 0.07
5 5
R 0.07
GRR = = = 0.0588
d *2 1.19

d*2 = 1.19 Corresponds to

m=2 and g=5 number of parts.

66
 Range and Average method – for R & R test

• Set up a team of 2 or 3 appraisers for the test .

• Select 7-10 sample Parts , most preferable 10 parts , which will
represent the entire process variation.
• Measure each part a number of times ( typically2 to 3 ) in a random
order.
• Tabulate the results and calculate the contribution of sources of
variation.
Note - Before starting the test , the exact location of measurement
should be marked on the sample parts, taken for the test. This
is to avoid within part variation. All the measurements has to
be taken at same location for all trials and by all operators .

67
 Repeatability & Reproducibility Study
Range and Average Method

Parts Average
Appriser / Trail 1 2 3 4 5 6 7 8 9 10
A

Average Xa
Range Ra
B

Average Xb
Range Rb
C

Average Xc
Range Rc
Part Average Rp

68
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Evaluation is done by multiple appraisers measuring more than 5 parts 2 to 3
times each.

Trails 2 3
Repeatability
EV = R × K 1 K1 0.8862 0.5908

2(EV )2 Appraisers 2 3
Reproducibility AV = ( X diff × K 2 ) − K2 0.7071 0.5231
n×r
Where , n - parts , r - trials

R&R GRR = AV 2 + EV 2
69
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Calculate Study Variation.

Part Variation PV = R p × K 3
Parts 2 3 4 5 6 7 8 9 10
K3 0.7071 0.5231 0.4467 0.4030 0.3742 0.3534 0.3375 0.3249 0.3146

Total Study Variation TV = GRR 2 + PV 2

E.V., A.V., R&R can be expressed as % of

*Study Variation *Process Variation * Design Tolerance.

70
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)

 EV   6 × EV 
% Equipment Variation % EV = 100 ×   or % EV = 100 ×  
 TV   Tolerance 

 AV   6 × AV 
% Appraiser Variation % AV = 100 ×   or % AV = 100 ×  
 TV   Tolerance 

 GRR   6 × GRR 
% GRR %GRR = 100 ×   or %GRR = 100 ×  
 TV   Tolerance 

% Part Variation  PV   6 × PV 
% PV = 100 ×   or % PV = 100 ×  
 TV   Tolerance 

Calculate Number of distinct data categories (ndc) has to be minimum 5

 PV 
ndc = 1.41 ×  
 GRR 
71
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)

Measurement System must have Discrimination ability to

distinguish between part to to part variation ( process variation )

ndc must be minimum 5.

Guidelines for Acceptance of system based on %GRR.

Under 10% - measurement system acceptable.

10% to 30% - may be acceptable based on gauge cost.

Over 30% - measurement system not acceptable.

72
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Gauge R & R Study will provide information concerning the cause of
measurement system variation.

e.g. if repeatability is large compared to reproducibility the reason may be –

•The instrument needs maintenance.
•The gauge may need to be redesigned to be more rigid.
•The clamping or location for gauging need to be improved.

•There is excessive with-in part variation.

e.g. if reproducibility is large compared to repeatability the reason may be –

•The appraiser needs to be better trained in how to use and use the gauge instrument .
•Calibrations on the gauge Dial are not clear.
•There is excessive with-in part variation.
•A fixture of some sort may be needed to help the appraiser use the gauge more
consistently.

73
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Parts Average
Appriser / Trail 1 2 3 4 5 6 7 8 9 10
A 0.29 -0.56 1.34 0.47 -0.8 0.02 0.59 -0.31 2.26 -1.36 0.194
0.41 -0.68 1.17 0.5 -0.92 -0.11 0.75 -0.2 1.99 -1.25 0.166
0.64 -0.58 1.27 0.64 -0.84 -0.21 0.66 -0.17 2.01 -1.31 0.211
Average 0.447 -0.607 1.260 0.537 -0.853 -0.100 0.667 -0.227 2.087 -1.307 Xa = 0.1903
Range 0.35 0.12 0.17 0.17 0.12 0.23 0.16 0.14 0.27 0.11 Ra = 0.184
B 0.08 -0.47 1.19 0.01 -0.56 -0.2 0.47 -0.63 1.8 -1.68 0.001
0.25 -1.22 0.94 1.03 -1.2 0.22 0.55 0.08 2.12 -1.62 0.115
0.07 -0.68 1.34 0.2 -1.28 0.06 0.83 -0.34 2.19 -1.5 0.089
Average 0.133 -0.790 1.157 0.413 -1.013 0.027 0.617 -0.297 2.037 -1.600 X b = 0.068
Range 0.18 0.75 0.4 1.02 0.72 0.42 0.36 0.71 0.39 0.18 R b = 0.513
C 0.04 -1.38 0.88 0.14 -1.46 -0.29 0.02 -0.46 1.77 -1.49 -0.223
-0.11 -1.13 1.09 0.2 -1.07 -0.67 0.01 -0.56 1.45 -1.77 -0.256
-0.15 -0.96 0.67 0.11 -1.45 -0.49 0.21 -0.49 1.87 -2.16 -0.284
Average -0.073 -1.157 0.880 0.150 -1.327 -0.483 0.080 -0.503 1.697 -1.807 X c = -0.2543
Range 0.19 0.42 0.42 0.09 0.39 0.38 0.2 0.1 0.42 0.67 R c = 0.328

Part Average 0.17 -0.85 1.10 0.37 -1.06 -0.19 0.45 -0.34 1.94 -1.57 R p = 3.511

74
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
R +R +R
R= a b C = 0.184 + 0.513 + 0.328 = 0.3417
3 3
X diff = max X − min X = 0.1903 − ( −0.2543) = 0.4446

Repeatability EV = 0.3417 * 0.5908 = 0.20188

2
(0.20188)
Reproducibility AV = (0.446 × 0.5231)2 − = 0.22963
10 × 3

R&R GRR = 0.20188 2 + 0.22963 2 = 0.30575

75
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
 EV   0 . 20188 
% Equipment Variation % EV = 100 ×   = % EV = 100 ×   = 17 . 62 %
 TV   1 . 14610 

 AV   0 . 22963 
% Appraiser Variation % AV = 100 ×   = % AV = 100 ×   = 20 . 04 %
 TV   1 . 147610 

 GRR   0 . 30575 
% GRR %GRR = 100 ×   = % GRR = 100 ×   = 26 . 68 %
 TV   1 . 14610 

% Part Variation  PV   1 . 10456 

% PV = 100 ×   = % PV = 100 ×   = 96 . 38 %
 TV   1 . 14610 

 PV   1.10456  ndc grater than 5 indicates measurement

ndc = 1.41×   = 1. 41×   = 5.094
 GRR   0.30575  system has discrimination ability .

Conclusion :
Measurement System is marginally Acceptable as % GRR is less
than 30% .
76
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Gauge R & R Study Graphical Analysis – Average Control Charts

The Average of multiple readings

by each appraiser on each part.

UCL X = X + A2 R
= 0 . 001444 + 1 . 023 × 0 . 3417
= 0 . 351

LCL X = X − A2 R
= 0 . 001444 − 1 . 023 × 0 . 3417
= − 0 . 3481

•Control limits give a visual indication of measurement variation in comparison to

the part to part variation.
•Narrower the control limits I.e. more averages out side the limits the better the
measurement System
77
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Gauge R & R Study Graphical Analysis – Range Control Charts

The Ranges of multiple readings

by each appraiser on each part.

UCL R = D4 R
= 2 . 574 × 0 . 3417
= 0 . 8795

LCL R = D3R
= 0 . 0 × 0 . 3417
= 0

•Range chart reflects the ability of the measurement system to re-produce the
`same measurement, it represents the consistency of the measurement process.
•Point out side the control limits requires to investigate special cause I.e. to assess
the ability of the respective operator to test samples consistently.
78
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Gauge R & R Study Graphical Analysis – Appraiser Bias Control Charts

The Average of multiple readings

by each appraiser on all parts.

79
 Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Gauge R & R Study Graphical Analysis – Appraiser Inconsistency Control Charts

The Average of multiple ranges

by each appraiser.

80
 Stability Study

Stability is the total variation in the measurements obtained with a

measurement system on the same master or part when measuring a
single characteristic over an extended time period.

(For Monitoring Instrument over a period of Time)

81
 Stability Study (Contd.)

Conducting Stability Test

• Measure a Sample Part minimum 10 Times.

• Plot a X-MR chart or X-bar R chart and plot control limits to

determine the control limits of natural variation .

• Preserve the sample part carefully.

• On a periodic basis measure the same part at least three times, and
check that whether all readings are with in the control limits.

82
Stability Study (Contd.)

Conduct Bench mark test

Apply Control Limits

Monitor Periodically With respect

to the Set Control Limits

83
 Stability Study (Contd.)

Example
Gauge/ Instrument under test - Vernier Caliper L.C. 0.1mm.
A Bench Mark test is performed by measuring sample part 20 times on
17/02/2000 at different time , and control limits are calculated .

Bench Mark test :

Date 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00
Time 0.00 1.00 2.00 3.00 4.00 5.00
Reading 20.1 20.1 19.9 20.0 20.1 20.2

Date 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00
Time 6.00 7.00 8.00 9.00 10.00 11.00 12.00
Reading 20.0 20.1 20.0 19.9 20.2 20.3 20.1

Date 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/ 02 /00 17/02/00
Time 13.00 14.00 15.00 16.00 17.00 18.00 19.00
Reading 19.9 20.0 20.0 20.2 20.1 20.2 20.1

84
 Stability Study (Contd.)
R 0 .4
Calculate : X Bar = 20.08 mm σ= = = 0.11
d 2 3.80537
X Chart MR Chart
UCL = X + 3σ UCL R = MR × D4
= 20.08 + 3(0.11) = 20.41 = 0.1263 × 3.267 = 0.413

LCL = X − 3σ LCL R = MR × D 3
= 20.08 − 3(0.11) = 19.75 = 0 .1263 × 0 = 0

85
Stability Study (Contd.)

86
 Stability Study (Contd.)

On 19/10/2000 the same part is measured 8 times the results

were found as follows ,
20.2 , 20.1 , 20.2 , 20.3 , 20.1 , 19.9 , 20.0 mm

All measurements are within U.C.L. & L.C.L.means system is

stable from period 17/02/2000 up to 19/10/2000.

87
 Attribute Measurement System Study:
Risk analysis Method
Performing Study :
Each appraiser measures 25-50 parts upto 2 to 3 times each:
Tabulation :
Part A1 A2 A3 B1 B2 B3 C1 C2 C3 Ref. Code
1 1 1 1 1 1 1 1 1 1 1 +
2 0 0 0 0 0 0 0 0 0 0 -
3 1 1 0 1 1 0 1 0 0 1 x
Legend :
1 : GOOD PARTS
0 : NOT GOOD PARTS
+ : Appraisers agree on GOOD PART.
– : Appraisers agree on NOT GOOD PART.
X : Appraisers do not agree .

88
 Attribute Measurement System Study :
Study Data Set
Part A-1 A-2 A-3 B-1 B-2 B-3 C-1 C-2 C-3 Reference Reference value Code Part A-1 A-2 A-3 B-1 B-2 B-3 C-1 C-2 C-3 Reference Reference value Code
1 1 1 1 1 1 1 1 1 1 1 0.476901 + 26 0 1 0 0 0 0 0 0 1 0 0.547204 X
2 1 1 1 1 1 1 1 1 1 1 0.509015 + 27 1 1 1 1 1 1 1 1 1 1 0.502436 +
3 0 0 0 0 0 0 0 0 0 0 0.576459 - 28 1 1 1 1 1 1 1 1 1 1 0.521642 +
4 0 0 0 0 0 0 0 0 0 0 0.566152 - 29 1 1 1 1 1 1 1 1 1 1 0.523754 +
5 0 0 0 0 0 0 0 0 0 0 0.57036 - 30 0 0 0 0 0 1 0 0 0 0 0.561457 X
6 1 1 0 1 1 0 1 0 0 1 0.544951 X 31 1 1 1 1 1 1 1 1 1 1 0.503091 +
7 1 1 1 1 1 1 1 0 1 1 0.465454 X 32 1 1 1 1 1 1 1 1 1 1 0.50585 +
8 1 1 1 1 1 1 1 1 1 1 0.502295 + 33 1 1 1 1 1 1 1 1 1 1 0.487613 +
9 0 0 0 0 0 0 0 0 0 0 0.437817 - 34 0 0 1 0 0 1 0 1 1 0 0.449696 X
10 1 1 1 1 1 1 1 1 1 1 0.515573 + 35 1 1 1 1 1 1 1 1 1 1 0.498698 +
11 1 1 1 1 1 1 1 1 1 1 0.488905 + 36 1 1 0 1 1 1 1 0 1 1 0.543077 X
12 0 0 0 0 0 0 0 1 0 0 0.559918 X 37 0 0 0 0 0 0 0 0 0 0 0.409238 -
13 1 1 1 1 1 1 1 1 1 1 0.542704 + 38 1 1 1 1 1 1 1 1 1 1 0.488184 +
14 1 1 0 1 1 1 1 0 0 1 0.454518 X 39 0 0 0 0 0 0 0 0 0 0 0.427687 -
15 1 1 1 1 1 1 1 1 1 1 0.517377 + 40 1 1 1 1 1 1 1 1 1 1 0.501132 +
16 1 1 1 1 1 1 1 1 1 1 0.531939 + 41 1 1 1 1 1 1 1 1 1 1 0.513779 +
17 1 1 1 1 1 1 1 1 1 1 0.519694 + 42 0 0 0 0 0 0 0 0 0 0 0.566575 -
18 1 1 1 1 1 1 1 1 1 1 0.484167 + 43 1 0 1 1 1 1 1 1 0 1 0.46241 X
19 1 1 1 1 1 1 1 1 1 1 0.520496 + 44 1 1 1 1 1 1 1 1 1 1 0.470832 +
20 1 1 1 1 1 1 1 1 1 1 0.477236 + 45 0 0 0 0 0 0 0 0 0 0 0.412453 -
21 1 1 0 1 0 1 0 1 0 1 0.45231 X 46 1 1 1 1 1 1 1 1 1 1 0.493441 +
22 0 0 1 0 1 0 1 1 0 0 0.545604 X 47 1 1 1 1 1 1 1 1 1 1 0.486379 +
23 1 1 1 1 1 1 1 1 1 1 0.529065 + 48 0 0 0 0 0 0 0 0 0 0 0.587893 -
24 1 1 1 1 1 1 1 1 1 1 0.514192 + 49 1 1 1 1 1 1 1 1 1 1 0.483803 +
25 0 0 0 0 0 0 0 0 0 0 0.599581 - 50 0 0 0 0 0 0 0 0 0 0 0.446697 -

89
 Attribute Measurement System Study :
Study Data Set (Contd.)

Analysis :
Kappa is a measure of inter-rater agreement
Calculate Kappa measure to determine the agreement of each
appraiser to the other.
B
A * B Crosstabulation Total
.00 1.00
.00 Count 44 6 50
A Expected Count 15.7 34.3 50.0
1.00 Count 3 97 100
Expected Count 31.3 68.7 100.0
Total Count 47 103 150
Expected Count 47.0 103.0 150.0

90
Attribute Measurement System Study :
Study Data Set (Contd.)
47 x 50
Expectedco untA 00 B 00 = . = 15 .7
150
Po = Sum of observed proportions in the diagonal cells
Pe = Sum of expected proportions in the diagonal cells.

Po − Pe
kappa =
1 − Pe
kappaA* B =
(44 + 97) − (15.7 + 68.7) = 0.86
150− (15.7 + 68.7)
kappa〉 0.75 _____ good − aggreement. − BetweenA& B

91
 Attribute Measurement System Study :
Study Data Set (Contd.)

Also calculate Kappa measure to determine the agreement of each

appraiser to the reference decision.
REF
A * REF Crosstabulation Total
.00 1.00
.00 Count 45 5 50
A Expected Count 16.0 34.0 50.0
1.00 Count 3 97 100
Expected Count 32.0 68.0 100.0
Total Count 48 102 150
Expected Count 48.0 102.0 150.0

Po = Sum of observed proportions in the diagonal cells

Pe = Sum of expected proportions in the diagonal cells.

Po − Pe
kappa =
1 − Pe
kappa〉 0.75 _____ good − aggreement .

92
 Risk Analysis Method
Signal Detection Method
Parts with known values ( Ref. values) are evaluated by different operators.
The readings are divided in to three zones.
1. Zone where all appraisers are unanimous on acceptance.
2. Zone where all appraisers are unanimous on rejection.
3. Zone of difference between appraisers.
d = Average width of the zone of non-unanimity.
2× d
%GRR = × 100
Tolerance
Zone of
L. S. L. U. S. L.
Non- unanimity.

_ + _

93
 Risk Analysis Method (Contd.)
di=difference between the size of last part accepted by all appraisers
and the size of first part rejected by all.
Ref. Value
0.599581
Code
-
Ref. Value
0.503091
Code
+
d = average (di)
0.587893
0.576459
-
-
0.502436
0.502295
+
+
is an estimation
0.570360 - 0.501132 + of region II area.
0.566575 - 0.498698 +
0.566152 - 0.493441 +
0.561457 X 0.488905 +
0.559918 X 0.488184 + d LSL
= 0.470832 − 0.446697
0.547204 X 0.487613 +
0.545604 X 0.486379 + d LSL
= 0.024132
0.544951 X 0.484167 +
0.543077 X 0.483803 +
0.542704
0.531939
+
+
0.477263
0.476901
+
+ d USL
= 0.566152 − 0.542704
0.529065
0.523754
+
+
0.470832
0.465454
+
X d USL
= 0.023448
0.521642 + 0.462410 X
0.520496 + 0.454518 X
0.519694 + 0.452310 X
0.024132 + 0.023448
0.517377
0.515573
+
+
0.449696
0.446697
X
-
d=
0.514192 + 0.437817 -
2
0.513779 + 0.427687 - d = 0.0237915
0.509015 + 0.412453 -
0.505850 + 0.409238 -

94
 Risk Analysis Method (Contd.)

2 × 0 . 02379
% GRR = × 100 = 25 %
0 . 599581 − 0 . 409238

The goodness of estimation depends on the sample

size and how closely sample represents the process.

95
Analytic Method – Gauge Performance Curve

Ga u g e P e r f o r m a n c e C u r v e
Gauge
performance
R e f e r e n c e Va l u e o f curve is used to
me a s ure d P a rt determine the
Bias and
repeatability of
1 20 the attribute
1 00 LSL USL measurement
80
system.
60 a + 0 .5 a
Pa* = if 〈 0.5
40 m m
20
a − 0 .5 a
0 Pa* = if 〉 0.5
m m
-20
a
Pa* = 0.5if = 0.5
m

96
Analytic Method – Gauge Performance Curve (Contd.)

Bias = lowerLimit − X T (atPa' = 0.5)

X T (atPa' = 0.995) − X T (atPa' = 0.005)

Re peatability =
1.08

31.3 × Bias If t > t ( 0.025 ,19 ) then

t= Bias is significantly different
Re peatabiity from zero. Gauge is not
suitable for use.

97
Analytic Method – Gauge Performance Curve (Contd.)

Minimum 8 parts should be selected at as nearly equidistant intervals

as practical. The max. & min value should represent the process
range.
8 parts must go through the gage, m=20 times, and the number of
accepts (a)

Criteria
1. Smallest part must have value a = 0
2. Largest part must have value a = 20,
3. min 6 other parts should have 1 ≤ a ≤19

98
Analytic Method – Gauge Performance Curve (Contd.)

XT a Pa' Bias = lowerLimit − X T (atPa' = 0.5)

-0.016 0 0.025
-0.015 1 0.075 Bias = −0.010 − ( −0.0123) = 0.0023
-0.014 3 0.175
-0.013 5 0.275
-0.012 8 0.425 − .0084 − (−0.0163)
-0.011 16 0.775 Re peatability =
-0.0105 18 0.875
1.08
-0.01 20 0.975
-0.008 20 1.0000 0.0079 =
Re peatability = 0.0073
1.08
31.3 × Bias
t=
Re peatabiity
as t > t ( 0.025 ,19 ) = 2.093
Bias is significantly different
31.3 × 0.0023 from zero. Gauge is not
t= = 9.86
0.0073 suitable for use.

99
Analytic Method – Gauge Performance Curve (Contd.)
Attribute Gauge Performance Curve
Plotted on Normal Probability Paper.

100
 Section - III
Statistical Process Control

101
 Goal of Process Control System.

Is to make economically sound decisions about actions affecting the

process. This means balancing the consequences of

Taking action when action is not necessary ( over control or “tempering” )

Verses
Failing to take action when action is necessary ( under control)

Important function of process control system is to provide a

statistical signal when special causes of variation are present , and
to avoid giving false signals when they are not present.

102
 Process Control And Process Capability.
Process Control : A process is said to be operating under statistical
control when the only sources of variations are from common
causes.
Process Capability : The ability of process to meet customer
expectations. Determined by the variation that comes from
common cause.

Basis of continual Improvement :

The process must first be brought in to statistical control by
detecting and acting upon special causes of variation. Then its
performance is predictable, and its capability to meet customer
expectations can be assessed.

103
 Process Improvement Cycle.

Take Decision Act Plan Which process to improve

Analysis Data Check Do Collection of Data

104
Process Quality Improvement.

105
 Guidelines for implementation of Statistical Process Control
Established Process
New Process

Take a Sample of at least

50 consecutive parts Complete an on-
going control chart
Find and remove
special cause Complete variable Find and remove
control chart special cause

Is the process No
No Is the process Stable?
Stable?
Improve Yes
Yes Improve
Process Process
Complete a capability
Complete a capability
study using Cp and Cpk
study using Pp and Ppk

No Yes Is the process No

Is the process
capable? capable?

Yes

106
 Steps for use SPC Tools

SPC tools are used for determining

Process Stability and Process Capability.

• Plot Control Chart and analysis to find out process stability .

• Analyze Histogram to access systems current performance, as well as

it helps as an indicator after an improvement has been tested.

• Perform Process Capability Analysis to find out ability of process to

meet the specifications.

107
 Control Charts

It is the basic tool of SPC , used to find out whether system is under
control or not ?
It consists Center line with straight line, Upper and lower control
limits with dashed line.
Phases are
– Collection (Gather data and plot on a chart)
– Control ( Calculate trial control limits from process data
and Identify special causes of variation and act upon them)
– Analysis and improvement (Quantify common cause of
variation; take action to reduce it)

These three phases are repeated for continual

process improvement

108
 Benefits of Control charts

• Help to understand the process variation

• Inform when action should be taken or not
• Show evidence that the process is in control
• Are part of the improvement process
• Provided a common language for communicating information
• Indicate the extent of common causes of variation (but the causes themselves
need more detailed analysis to isolate)
• Detect special causes of variation (and allow to take appropriate action)
Type of Control Charts
Variable charts Attribute charts
X bar- R , X bar –s , p , np - Defectives.
X Median - R , X - MR c , u - Defects.

109
 Selection of control chart.
Which Control
Chart ?

Is it Yes
Variable data?

Is it homogeneous
data in nature or not
No
conducive to
(Attribute data) No
subgroup sampling

Is the interest in in No Is the interest in Can sub group avg. be

non-conformities nonconforming units conveniently computed
Yes

Yes
Is the sub group
Is the sample size Is the sample size size 9 or more
Yes No
constant constant No
Yes Yes
Yes
No No

Use c or Use Use Use np or Use charts for Use Use Use
u chart u chart p chart p chart individuals X-MR Median chart X-bar -S chart X-bar -R chart

Attribute charts Variable charts

 Control charts for Variables

Used when:
"Process measurements are available"
Variable charts explain process data in two terms of :
- Location (average)
- Spread (Range or STD deviation)
Note - The shape can not be identified with control charts.
Kind of Charts
X - R charts (average and range)
X -S chart (average and STD deviation)
X -R chart (median and range)
X - MR chart (individuals and moving range)

Advantages Generally small subgroup size (4 to 5, max 10)

111
 Control charts for Attributes

Used when:
"Output results only indicates two possibilities as Pass or Fail”
Attribute charts explain process data in only one terms of : Proportion or Number
Kind of Charts
p - chart (Proportion of NC units) - Sample of equal size are NOT required
np - chart (number of NC units) - Samples of equal size
c - chart (number of NC) - sample of equal size
u - chart (number of NC / unit) - sample of equal size are NOT required
Advantages
- Easy to measure
- Inspection records are mostly already available

112
 Control charts for variables
Preparation for use of control charts

- Establish an environment suitable for action

- Define the process
- Determine characteristics to be managed
Considerations :
- The customer’s needs
- Current and potential problem areas
- Correlation between characteristics
- Define the measurement system
- Minimize unnecessary variation

113
 Guidelines for Data collection for plotting Control Charts.
Process Data is collected in number of subgroups, each subgroup is of defined sample size,
and these sample groups are selected at a defined period. ( frequency)

Sub groups

Frequency

114
 Control charts for Variables

Reason for subgroups instead of individual readings

- The average of a subgroup is probably closer to the process average than an

individual reading

- The variation (σ X ) of subgroup averages is smaller than the variation of the

individual readings (σX ).
σX
σX =
n

The subgroup averages follow a normal distribution more and more with an
increasing subgroup size, starting from n=5.

115
 Guidelines for sampling - Process Data collection.
Section of Subgroup Size , Frequency and Number of Subgroups.
Control charts for variables.
Subgroup Size : The intention should be the pieces within each
subgroup would be produced under very similar production
conditions over a very short interval .
If the variation within subgroup represents the piece to piece
variability over a very short period of time, then any unusual
variation between subgroups would reflect changes in the process
that should be investigated for appropriate action.
Sub Group Frequency : The goal is to detect changes in the
process over time. Sub groups should be collected often enough,
and at appropriate times, that they can reflect the potential
opportunities for the change.
Number of Subgroups: The criteria from process standpoint ,
enough subgroups should be gathered to assure that the major
sources of variation have had an opportunity to appear.

116
 Guidelines for sampling - Process Data collection.
Section of Subgroup Size , Frequency and Number of Subgroups.
Control charts for variables.
Subgroup size n- generally should not be less than 4 more
preferably it should be 4 or 5 or 6

No. of subgroups k - generally selected more than 20 , the more is

preferable.

Relationship between
n – Subgroup size & K – no. of subgroups
Smaller batch size, more number of batches ( Higher frequency )
– Easier to identify assignable cause
To plot initial control limits minimum 100 samples are advisable
For 100 samples
n=4 , n=5 ,
K=25 K=20

117
 Average and Range charts ( X and R)
1. Gather Data
- Subgroup size
Initially 4 to 5 consecutively produced pieces from a single process stream.
Must remain constant for all subgroups
All pieces would be produced under very similar conditions
- Subgroup frequency
Initially short intervals
Can be lowered once the process has demonstrated stability
Changes in the process over time must be detectable
- Number of subgroups
Initially minimum 25 subgroups providing minimum 100 readings
Major sources of variation must have an opportunity to appear
If stable process. A good estimation must be about the process location
and spread

contd ---

118
 Average and Range charts ( X and R)
- Setup Control charts and Record raw data
Individual reading :X
Sample size :n

- Calculate the Average and Range of each subgroup

X 1 + X 2 + X 3 + − − − − + Xn
Sample average : X=
n

Range : R = X max − X min

- Select scales for the control charts
X -chart - 2 times difference between highest and lowest X
R-chart - 2 times largest R Number of subgroups :k

- Plot the averages and ranges on the control charts

119
 Average and Range charts ( X and R)

2. Calculate control limits

- Calculate average range and process average
Average Range: Average Process:

R1 + R2 + R 3 + − − − − − + Rk X1 + X 2 + X 3 + − − − − − + X k
R= X=
k k
- Calculate the control limits
Range Chart : UCLR = D 4 * R Average Chart : UCL , LCL ( X )
LCLR = D3 * R
= X ± A2 R

- Draw lines for the Averages and control limits on the charts
- Estimated process standard deviation = R
σˆ =
d2

(Note : There is no lower control limit for ranges for sample size below 7)

120
 Average and Range charts ( X and R)

- Extend control limits for Ongoing Control

R
Estimated process standard deviation = σˆ =
d2

Calculate control limits for new subgroup size (if changed)

Rnew = σˆ × d 2
R – bar Chart X-bar Chart

UCLR = D 4 × Rnew UCL X & LCL X = X ± A2 × Rnew

LCLR = D3 × Rnew

121
 The X and S chart for average and standard deviation

2
Sample standard deviation : ∑ ( Xi − X )
S=
n −1

Where , Xi = Sub group’s individual Value , X-bar = average , n = sample size.

Control limits
Deviation Average
UCLs = B 4 × S
UCL X & LCL X = X ± A3 × S
LCLs = B3 × S
(Note : There is no lower control limit for ranges for sample size below 6)

S
Estimated process standard deviation = σˆ =
C4

123
124
 ~ and R chart for Median and Range
The X
Sample size : Odd sizes are most convenient
~
Median : X
~ ~ ~ ~
Median average : ~ X 1 + X 2 + X 3 + − − − − + Xn
X=
k
(Point all readings in the chart and circle the median points and connect these median points)
Control limits
Deviation : Average :

UCLR = D 4 × R ~
UCLX~ & LCLX~ = X ± A2 × R
LCLR = D3 × R
(Note : There is no lower control limit for ranges for sample size below 7)

R
Estimated process standard deviation = σˆ =
d2

125
 The X - MR chart for Individuals and Moving Range

Used in case of Subgroup variation approximately equal to zero. (homogenous

batch measurement) and single destructive tests
X - MR is less sensitive as X bar - R chart
Take care of :
- not symmetrical distributions
- piece to piece repeatability
- number of subgroup >100 before X-bar can have substantial variability
Moving Range : MR = difference between each successive pair of readings
(1-2, 2-3, 3-4, …) or in rare case larger moving groups(1-2-3, 2-3-4, 3-4-5,…)

Process average : X =
X 1 + X 2 + X 3 + − − − − + Xk
k
Range average : MR1 + MR 2 + MR3 + − − − − + MRK − 1
MR=
k −1
(k-1 for successive pair reading, k-2 for 3 successive readings, etc)

126
 The X - MR chart for Individuals and Moving Range

Control limits
Range : Average :

UCLMR = D 4 × M R UCLX & LCLX = X ± E 2 × M R

LCLMR = D3 × M R
=

(Note : There is no lower control limit for ranges for sample size below 7)

R
Estimated process standard deviation =
σˆ =
d2

127
 Control charts for Attributes
Gather Data
- Sub-Group Size :
np and p charts
Large enough to disclose sufficient non conformities per subgroup
Small enough to avoid a too long period of process operation
Sub group sizes are generally 50 to 200 or more

c and u charts
The subgroup size is the inspected area

- Subgroup frequency :
Consensus between fast feedback and a large subgroup size

- Number of subgroups :
Minimum 25 to give a good test for stability and if stable, a reliable estimate
of process performance

128
 The p chart for proportion nonconforming
Subgroup sample size: n is not necessary constant (variation below +/- 25%)
Number of NC found : np

np
Proportion NC found p=
n
n1 p1 + n2 p 2 + n3 p3 + − − − − + nkpk
Process average p=
n1 + n2 + n3 + − − − + nk
Scale for the control chart : 1.5 to 2 times p max
Control limits :
p(1 − p)
UCLp, LCLp = p ± 3
n
(note : n is the average sample size) If LCL < 0, then LCL = 0

(Where, n= particular subgroup sample size)

Process capability = (1 − p) ×100%

129
 The np chart for number nonconforming

Subgroup sample size: n = constant

Number of NC found : np = defective

np1 + np 2 + np3 + − − − − + npk

Process average np =
k

k = number of subgroups
Control limits :
np
UCLnp, LCLnp = n p ± 3 n p(1 − )
n
= n p ± 3 n p − (1 − p)

Process capability = (1- np )*100 %

131
 The c chart for number nonconforming

Sample size has to be constant

NC per sample :c
Number of samples : k
c1 + c 2 + c3 + − − − − +ck
Process average : c=
k
Control limits

UCLc, LCLc = c ± 3 c

Process capability = (1- c )*100 %

133
 The u chart for number nonconforming

Sample size expressed on a per unit base (eg. Per sq.meter) : n

c
NC per sample u =:
n

c1 + c 2 + c3 + − − − − +ck
Process average : u=
n1 + n 2 + n3 + − − − − + nk

u
Control limits : UCLu, LCLu = u ± 3
n

Note : n is the average sample size 0.75 max ≤ n ≥ 1.25n min

Process capability = (1- u )*100 %

135
Interpret for Process control - to identify any instability and to take appropriate action.
Start with a
completed
control chart

Point
outside the Yes
control Limit?

No
7 or more
points in a row above Yes
or below the
center line?

No

7 or more
points in a row going Yes
in one direction, up or
down?

No

Nonrandom Yes
patterns present

No

Declare the Declare the

system stable or system unstable or out
in control . of control .
136
 Interpret for Process control - to identify any instability and to take appropriate action.

Analyze the data plots for non-control

a. Points beyond the control limits

b. 7 points in a row on one side of the average

137
 Interpret for Process control - to identify any instability and to take appropriate action.

c. 7 points in a row that are consistently

increasing or consistently decreasing

Normality test
About half of the points
should fall above the average
and half below

2/3 of the plotted points should

lie within the middle third of the
region between control limits

138
Interpret for Process control - to identify any instability and to take appropriate action.
Nonrandom Patterns
e. Too far from the
d. Too close to average
the average

f. Cycles

• Find and Address special causes

• Recalculate control limits

(Exclude all subgroups which are in out of limits and recalculate the
control limits of both range and average.)

-Repeat the recalculation of control limits up to all points comes within control
limits 139
Control chart - Interpretation

Central location and variability –

Variable control charts gives two
types of information about each set of
of data or subgroup: the central
location and the variability.

140
 Process Capability and Process Performance

- Indices of process variation only, relative to specifications :Cp and Pp

- Indices of process variation and centering combined, relative to
specifications : CPU, CPL, Cpk and Ppk
- Ratios of process variation only, relative to specifications : CR and PR

Inherent process variation - That portion of process variation due to common causes
only. This variation can be estimated from control charts by R
σˆ =
d2
Total process variation - This is the variation due to both common and special
causes. This variation may be estimated by s, the sample standard deviation, using
individual readings obtained from either a detailed control
chart or a process study. ∑ ( Xi − X )
2

S=
n −1

Process capability - The 6σ range of a process’s inherent variation, for statistically

stable processes only.
Process performance - The 6 σ range of a process’s total variation.

141
 Process Capability and Process Performance

Cp: This is the capability index which is defined as the tolerance width divided by the
process capability, irrespective of process centering. Typically, this is expressed as
where,
USL − LSL σˆ =
R
Cp =
6σ d2

Pp: This is the performance index which is defined as the tolerance width divided by the
process performance, irrespective of process centering.

USL − LSL where, ∑ ( Xi − X )

2
Pp = σ=
6σ n −1

142
Process Capability and Process Performance

143
 Process Capability and Process Performance

Cpk = This is the capability index which accounts for process centering and is defined
as the minimum of CPU or CPL

X − LSL USL − X R
CPL = CPU = σˆ =
3σ 3σ where, d2

Ppk= This is the performance index which accounts for process centering and is defend as
the minimum

USL − X X − LSL ∑ ( Xi − X )
2

Ppk = minimum of OR where, σ=

3σ 3σ n −1
Ratios

CR: This is the capability ratio and is simply reciprocal of Cp

PR: This is the performance ratio and is simply the reciprocal of Pp

144
Process Capability and Process Performance

145
 Percentage points of the two tailed
t distribution at α = 0.05 i.e. at 95%C.L.

d.f. α=0.05 d.f. α=0.05 d.f. α=0.05

1 12.70615 11 2.200986 21 2.079614
2 4.302656 12 2.178813 22 2.073875
3 3.182449 13 2.160368 23 2.068655
4 2.776451 14 2.144789 24 2.063898
5 2.570578 15 2.131451 25 2.059537
6 2.446914 16 2.119905 26 2.055531
7 2.364623 17 2.109819 27 2.051829
8 2.306006 18 2.100924 28 2.048409
9 2.262159 19 2.093025 29 2.045231
10 2.228139 20 2.085962 30 2.042270

146
 Values associated with the Distribution of the Average Range
Subgroup Size (m)
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
1 1.0 2.0 2.9 3.8 4.7 5.5 6.3 7.0 7.7 8.3 9.0 9.6 10.2 10.8 11.3 11.9 12.4 12.9
1.41421 1.91155 2.23887 2.48124 2.67253 2.82981 2.96288 3.07794 3.17905 3.36909 3.35016 3.42378 3.49116 3.55333 3.61071 3.66422 3.71424 3.76118
2 1.9 3.8 5.7 7.5 9.2 10.8 12.3 13.8 15.1 16.5 17.8 19.0 20.2 21.3 22.4 23.5 24.5 25.5
1.27931 1.80538 2.15069 2.40484 2.60438 2.76779 2.90562 3.02446 3.12869 3.22134 3.30463 3.37017 3.44922 3.51287 3.57156 3.62625 3.67734 3.72524
3 2.8 5.7 8.4 11.1 13.6 16.0 18.3 20.5 22.6 24.6 26.5 28.4 30.1 31.9 33.5 35.1 36.7 38.2
1.23105 1.76858 2.12049 2.37883 2.58127 2.74681 2.88628 3.00643 3.11173 3.20526 3.28931 3.36550 3.43512 3.49927 3.55842 3.61351 3.66495 3.71319
4 3.7 7.5 11.2 14.7 18.1 21.3 24.4 27.3 30.1 32.7 35.3 37.7 40.1 42.4 44.6 46.7 48.8 50.8
1.20621 1.74989 2.10522 2.36571 2.56964 2.73626 2.87656 2.99737 3.10321 3.19720 3.28163 3.35815 3.42805 3.49246 3.55183 3.60712 3.65875 3.70715
5 4.6 9.3 13.9 18.4 22.6 26.6 30.4 34.0 37.5 40.8 44.0 47.1 50.1 52.9 55.7 58.4 61.0 63.5
1.19105 1.73857 2.09601 2.35781 2.56263 2.72991 2.87071 2.99192 3.09808 3.19235 3.27701 3.35372 3.42381 3.48836 3.54787 3.60328 3.65502 3.70352
6 5.5 11.1 16.7 22.0 27.0 31.8 36.4 40.8 45.0 49.0 52.8 56.5 60.1 63.5 66.8 70.0 73.1 76.1
1.18083 1.73099 2.08985 2.35253 2.55795 2.72567 2.86680 2.98829 3.09467 3.18911 3.27392 3.32077 3.42097 3.48563 3.54522 3.60072 3.65253 3.70109
7 6.4 12.9 19.4 25.6 31.5 37.7 42.5 47.6 52.4 57.1 61.6 65.9 70.0 74.0 77.9 81.6 85.3 88.8
1.17348 1.72555 2.08543 2.34875 2.55460 2.72263 2.86401 2.98568 3.09222 3.18679 3.27172 3.34866 3.41894 3.48368 3.54333 3.59888 3.65075 3.69936
Numbers of Subgroups (g)

8 7.2 14.8 22.1 29.2 36.0 42.4 48.5 54.3 59.9 65.2 70.3 75.2 80.0 84.6 89.0 93.3 974.0 101.4
1.16794 1.72147 2.08212 2.34591 2.55208 2.72036 2.86192 2.98373 3.09039 3.18506 3.27006 3.34708 3.41742 3.48221 3.54192 3.59751 3.64941 3.69806
9 8.1 16.6 24.9 32.9 40.4 47.7 54.5 61.1 67.3 73.3 79.1 84.6 90.0 95.1 100.1 104.9 109.5 114.1
1.16361 1.71828 2.07953 2.34370 2.55013 2.71858 2.86028 2.98221 3.08896 3.18370 3.26878 3.34585 3.41624 3.48107 3.51081 3.59644 3.64838 3.69705
10 9.0 18.4 22.6 36.5 44.9 52.9 60.6 67.7 74.8 81.5 87.9 94.0 99.9 105.6 111.2 116.5 121.7 126.7
1.16014 1.71573 2.07746 2.34192 2.54856 2.71717 2.85898 2.98100 3.08781 3.18262 3.26775 3.34486 3.41529 3.48016 3.53993 3.59559 3.64755 3.69625
11 9.9 20.2 30.4 40.1 49.4 58.2 66.6 74.6 82.2 89.6 96.6 103.4 109.9 116.2 122.3 128.1 133.8 139.4
1.15729 1.71363 2.07577 2.34048 2.54728 2.71600 2.85791 2.98000 3.08688 3.18174 3.26690 3.34406 3.41452 3.47941 3.53921 3.59489 3.64687 3.69558
12 10.7 22.0 33.1 43.7 53.8 63.5 72.6 81.3 89.7 97.7 105.4 112.7 119.9 126.7 133.3 139.8 146.0 152.0
1.15490 1.71189 2.07436 2.33927 2.54621 2.71504 2.85702 2.97917 3.08610 3.18100 3.26620 3.34339 3.41387 3.47879 3.53861 3.59430 3.64630 3.69503
13 11.6 23.8 35.8 47.3 58.3 68.7 78.6 88.1 97.1 105.8 114.1 122.1 129.8 137.3 144.4 151.4 158.1 164.7
1.15289 1.71041 2.07316 2.33824 2.54530 2.71422 2.85627 2.97847 3.08544 3.18037 3.26561 3.34282 3.41333 3.47826 3.53810 3.59381 3.64582 3.69457
14 12.5 25.7 38.6 51.0 62.8 74.0 84.7 94.9 104.6 113.9 122.9 131.5 139.8 147.8 155.5 163.0 170.3 177.3
1.15115 1.70914 2.07213 2.33737 2.54452 2.71351 2.85562 2.97787 3.08487 3.17984 3.26510 3.34233 3.41286 3.47781 3.53766 3.59339 3.64541 3.69417
15 13.4 27.5 41.3 54.6 67.2 79.3 90.7 101.6 112.1 122.1 131.7 140.9 149.8 158.3 166.6 174.6 182.4 190.0
1.14965 1.70804 2.07125 2.33661 2.54385 2.71290 2.85506 2.97735 3.08438 3.17938 3.26465 3.34191 3.41245 3.47742 3.53728 3.59302 3.64505 3.69382
16 14.3 29.3 44.1 58.2 71.7 84.5 96.7 108.4 119.5 130.2 140.4 150.2 159.7 168.9 177.7 186.3 194.6 202.6
1.14833 1.70708 2.07047 2.33594 2.54326 2.71237 2.85457 2.97689 3.08395 3.17897 3.26427 3.34540 3.41210 3.47707 3.53695 3.59270 3.64474 3.69351
17 15.1 31.1 46.8 61.8 76.2 89.8 102.8 115.1 127.0 138.3 149.2 159.6 169.7 179.4 188.8 197.9 206.7 215.2
1.14717 1.70623 2.06978 2.33535 2.54274 2.71190 2.85413 2.97649 3.08358 3.17891 3.26393 3.34121 3.41178 3.47677 3.53666 3.59242 3.64447 3.69325
18 16.0 32.9 49.5 65.5 80.6 95.1 108.8 121.9 134.4 146.4 157.9 169.0 179.7 190.0 199.9 209.5 218.8 227.9
1.14613 1.70547 2.06917 2.33483 2.54228 2.71148 2.85375 2.97613 3.08324 3.17829 3.26362 3.34092 3.41150 3.47650 3.53640 3.59216 3.64422 3.69301
19 16.9 34.7 52.3 69.1 85.1 100.3 114.8 128.7 141.9 154.5 166.7 178.4 189.6 200.5 211.0 221.1 231.0 240.5
1.14520 1.70480 2.06862 2.33436 2.54187 2.71111 2.85341 2.97581 3.08294 3.17801 3.26335 3.34066 3.41125 3.47626 3.53617 3.59194 3.64400 3.69280
20 17.8 36.5 55.0 72.7 89.6 105.6 120.9 135.4 149.3 162.7 175.5 187.8 199.6 211.0 222.1 232.8 243.1 253.2
1.14437 1.70419 2.06813 2.33391 2.54149 2.71077 2.85310 2.97552 3.08267 3.17775 3.26311 3.34042 3.41103 3.47605 3.53596 3.59174 3.64380 3.92600
d2 1.12838 1.69257 2.05875 2.32593 2.53441 2.70436 2.8472 2.97003 3.07751 3.17287 3.25846 3.33598 3.40676 3.47193 3.53198 3.58788 3.64006 3.68896
cd 0.876 1.815 2.738 3.623 4.4658 5.2673 6.0305 6.7582 7.4539 8.1207 8.7602 9.3751 9.9679 10.5396 11.0913 11.6259 12.1440 12.6468

Table entries: 1 st line of each degrees of freedom ( v ) and the 2 nd line of each cell is d :d is the infinity value of d : Additional values of v can be built up
from the constant difference cd.

Note : The notation used in this table follows that of Acheson Duncan Control and Industrial Statistics . 5th edition. McGraw - Hill. 1986
v ( R d ) / is distributed approximately as a % distribution with v of freedom where R is the average range of g subgroups of size m.
 TABLE OF CONSTANTS

n 2 3 4 5 6 7 8 9 10
D4 3.27 2.57 2.28 2.11 2 1.92 1.86 1.82 1.78
D3 - - - - - 0.08 0.14 0.18 0.22
A2 1.88 1.02 0.73 0.58 0.48 0.42 0.37 0.34 0.31
A2m 1.88 1.19 0.8 0.69 0.55 0.51 0.43 0.41 0.36
d2 1.13 1.69 2.06 2.33 2.53 2.7 2.85 2.97 3.08

B4 3.27 2.57 2.27 2.09 1.97 1.88 1.82 1.76 1.72

B3 - - - - 0.03 0.12 0.19 0.24 0.28
A3 2.66 1.95 1.63 1.43 1.29 1.18 1.1 1.03 0.98
c4 0.798 0.886 0.921 0.94 0.952 0.959 0.965 0.969 0.973

E2 2.66 1.77 1.46 1.29 1.18 1.11 1.05 1.01 0.98

Where n = subgroup size

148
z x.xo x.x1 x.x2 x.x3 x.x4 x.x5 x.x6 x.x7 x.x8 x.x9 z x.xo x.x1 x.x2 x.x3 x.x4 x.x5 x.x6 x.x7 x.x8 x.x9

4.0 0.00003

3.9 0.00005 0.00005 0.00004 0.00004 0.00004 0.00004 0.00004 0.00004 0.00003 0.00003 1.9 0.0287 0.0281 0.0274 0.0268 0.0262 0.0256 0.025 0.0244 0.0239 0.0233
3.8 0.00007 0.00007 0.00007 0.00006 0.00006 0.00006 0.00006 0.00005 0.00005 0.00005 1.8 0.0359 0.0351 0.0344 0.0336 0.0329 0.0322 0.0314 0.0307 0.0301 0.0294
3.7 0.00011 0.0001 0.0001 0.0001 0.00009 0.00009 0.00008 0.00008 0.00008 0.00008 1.7 0.0446 0.0436 0.0427 0.0418 0.0409 0.0401 0.0392 0.0384 0.0375 0.0367
3.6 0.00016 0.00015 0.00015 0.00014 0.00014 0.00013 0.00013 0.00012 0.00012 0.00011 1.6 0.0548 0.0537 0.0526 0.0516 0.0505 0.0495 0.0485 0.0475 0.0465 0.0455
3.5 0.00023 0.00022 0.00022 0.00021 0.0002 0.00019 0.00019 0.00018 0.00017 0.00017 1.5 0.0668 0.0655 0.0643 0.063 0.0618 0.0606 0.0594 0.0582 0.0571 0.0559

3.4 0.00034 0.00032 0.00031 0.0003 0.00029 0.00028 0.00027 0.00026 0.00025 0.00024 1.4 0.0808 0.0793 0.0778 0.0764 0.0749 0.0735 0.0721 0.0708 0.0694 0.0681
3.3 0.00048 0.00047 0.00045 0.00043 0.00042 0.0004 0.00039 0.00038 0.00036 0.00035 1.3 0.0968 0.0951 0.0934 0.0918 0.0901 0.0885 0.0869 0.0853 0.0838 0.0823
3.2 0.00069 0.00066 0.00064 0.00062 0.0006 0.00058 0.00056 0.00054 0.00052 0.0005 1.2 0.1151 0.1131 0.1112 0.1093 0.1075 0.1056 0.1038 0.102 0.1003 0.0985
3.1 0.00097 0.00094 0.0009 0.00087 0.00084 0.00082 0.00079 0.00076 0.00074 0.00071 1.1 0.1357 0.1335 0.1314 0.1292 0.1271 0.1251 0.123 0.121 0.119 0.117
3.0 0.00135 0.00131 0.00126 0.00122 0.00118 0.00114 0.00111 0.00107 0.00104 0.001 1.0 0.1587 0.1562 0.1539 0.1515 0.1492 0.1469 0.1446 0.1423 0.1401 0.1379

2.9 0.0019 0.0018 0.0018 0.0017 0.0016 0.0016 0.0015 0.0015 0.0014 0.0014 0.9 0.1841 0.1814 0.1788 0.1762 0.1736 0.1711 0.1685 0.166 0.1635 0.1611
2.8 0.0026 0.0025 0.0024 0.0023 0.0023 0.0022 0.0021 0.0021 0.002 0.0019 0.8 0.2119 0.209 0.2061 0.2033 0.2005 0.1977 0.1949 0.1922 0.1894 0.1867
2.7 0.0035 0.0034 0.0033 0.0032 0.0031 0.003 0.0029 0.0028 0.0027 0.0026 0.7 0.242 0.2389 0.2358 0.2327 0.2297 0.2266 0.2236 0.2206 0.2177 0.2148
2.6 0.0047 0.0045 0.0044 0.0043 0.0041 0.004 0.0039 0.0038 0.0037 0.0036 0.6 0.2743 0.2709 0.2676 0.2643 0.2611 0.2578 0.2546 0.2514 0.2483 0.2451
2.5 0.0062 0.006 0.0059 0.0057 0.0055 0.0054 0.0052 0.0051 0.0049 0.0048 0.5 0.3085 0.305 0.3015 0.2981 0.2846 0.2912 0.2877 0.2843 0.281 0.2776

2.4 0.0082 0.008 0.0078 0.0075 0.0073 0.0071 0.0069 0.0068 0.0066 0.0064 0.4 0.3446 0.3409 0.3372 0.3336 0.33 0.3264 0.3228 0.3192 0.3156 0.3121
2.3 0.0107 0.0104 0.0102 0.0099 0.0096 0.0094 0.0091 0.0089 0.0087 0.0084 0.3 0.3821 0.3783 0.3745 0.3707 0.3669 0.3632 0.3594 0.3557 0.352 0.3483
2.2 0.0139 0.0136 0.0132 0.0129 0.0125 0.0122 0.0119 0.0116 0.0113 0.011 0.2 0.4207 0.4168 0.4129 0.409 0.4052 0.4013 0.3974 0.3936 0.3897 0.3859
2.1 0.0179 0.0174 0.017 0.0166 0.0162 0.0158 0.0154 0.015 0.0146 0.0143 0.1 0.4602 0.4562 0.4522 0.4483 0.4443 0.4404 0.4364 0.4325 0.4286 0.4247
2.0 0.0228 0.0222 0.0217 0.0212 0.0207 0.0202 0.0197 0.0192 0.0188 0.0183 0.0 0.5 0.496 0.492 0.488 0.484 0.4801 0.4761 0.4721 0.4681 0.4641
 SPECIAL CAUSE CRITERIA

1 1 Point more than 3 std.deviation from centerline

2 7 Points in a row on same side of center line
3 6 points in a row, alternating up and down

4 14 points in a row, alternating up and down

5 2 out of 3 points > 2 std. Deviations from center line
(same side)
6 4 out of 5 points > 1 std deviation from centerline

7 15 points in a row within 1 std. Deviation of centerline

(either side)
8 8 points in a row >1 std deviation from centerline
(either side)
 AVERAGE RUN LENGTH – Out of control

SHIFT IN TARGET
Sigma xbar ARL
0 370.4
0.1 352.9
0.2 308.4
0.3 253.1
0.5 155.2
1.0 43.9
1.5 15
2.0 6.3
3.0 2.0
4.0 1.2
6-Sigma concept and Process Capabilities

6 Sigma calculation of PPM

Sigma Area under the curve

-7.5 3.219646771413E-14 Area between -3 sigrma to 3 sigma PPM

-6 9.901218733788E-10
-5 2.87104999663E-07 0.997300066 2699.9344
-4 3.1686E-05
-3 0.001349967
-2 0.022750062 Area between -6 sigrma to 6 sigma
-1 0.15865526
0 0.5 9.999999980198E-01 0.0020
1 0.84134474
2 0.977249938
3 0.998650033 Area between -7.5 sigma to 4.5 sigma
4 0.999968314 ( considering the centre shift of 1.5 sigma )
4.5 0.999996599
5 0.999999713 9.999965991969E-01 3.4008
153