6 SSGB Amity Bsi Msa P

Module-6
Measurements Systems Analysis
Copyright 2012 BSI. All rights

reserved.
Measurements Systems Analysis - Agenda
1.
2.
3.
4.
Is our data accurate?
Repeatability & Reproducibility
Accuracy & Precision
DMAIC
Measurements System Variation
Bias, Linearity, Stability, Repeatability, Reproducibility, Calibration, Gauge R&R

Variable Gauge R&R
Parts, Operators, Variation
Is the gauge good?
Workshop
Attribute Gauge R&R
Workshop
Appendix
Analysis of Variance (ANOVA)
Copyright2012BSI.Allrights
reserved.
Is our Data Accurate?

Gauge R & R is a means of assessing the repeatability and reproducibility of our
measurement systems.
Gauge R & R studies are carried out in order to discover how much of the process variation
is due to the measurement device and measurement methods.
Dimension
Copyright
2012BSI.Allrights
reserved.
DMAIC Improvement Process
Define

Select Project
Define Project
Objective
Form the Team
Measure
Define Measures (ys)
Analyse
Improve
Identify Potential xs
C1 C2 C3
Evaluate Measurement
System
Effect

Map the Process

Identify Customer
Requirements

Determine Process
Stability
Determine Process
Capability
LSL
15

Identify Priorities
Update Project File
Phase Review
20
USL
25
30
35
Set Targets for

Measures
Phase Review
Copyright
2012BSI.Allrights
reserved.
. .. .
.
. .. .. . x
y=f(x1,x2,..)
C4 C5 C6
Characterise xs
Analyse xs
Run
1
2
3
4
5
6
7
8
1
1
1
1
2
2
2
2
1
1
2
2
1
1
2
2
1
1
2
2
2
2
1
1
1
2
1
2
1
2
1
2
1
2
1
2
2
1
2
1
1
2
2
1
1
2
2
1
1
2
2
1
2
1
1
2
Select Critical xs
Optimise xs

Phase Review
Control Critical xs
10.2
Upper Control Limit
10.0
9.8
Lower Control Limit
9.6
1
10
15
20
Monitor ys
y
Set Tolerances for xs
Verify Improvement
xx
x
x x x
x x
x
x
x
Control
LSL
15
20
Validate Control
Plan
Close Project
USL
25
30
35
Phase Review
Phase Review
4
Measurement Accuracy & Precision
Measurement is accurate but not

precise
Measurement is precise but not

accurate
Measurement is accurate and

precise
Copyright
2012BSI.Allrights
reserved.
Measurement System Variation
Bias
Accuracy
Measurement
System
Variation
Linearity
Stability
Repeatability
Precision
Reproducibility
Copyright
2012BSI.Allrights
reserved.
Bias
Bias
True
Value
Observed
Average
Bias is the difference between the observed average of the measurements and the true value.
Copyright
2012BSI.Allrights
reserved.
Measured Value
Linearity
Non-Linearity
Gauge is measuring lower than true
value at high end
Reference Value
Linearity is the difference in bias values over the expected operating range of the
measurement gauge.
Copyright
2012BSI.Allrights
reserved.
Stability
Stability
Time
1
Time
2
Stability is the variation (differences) in the average over extended periods of time using
the same gauge and appraiser to repeatedly measure the same part
Copyright
2012BSI.Allrights
reserved.
Repeatability
Repeatability
Repeatability is the variation between successive measurements of the same part, same
characteristic, by the same person using the same gauge.
Copyright
2012BSI.Allrights
reserved.
10
Reproducibility
Reproducibility
Operator
1
Operator
2
Reproducibility is the difference in the average of the measurements made by different

people using the same instrument when measuring the identical characteristic on the
same pieces.
Copyright
2012BSI.Allrights
reserved.
11
Measurement System Variation
Bias
Accuracy
Linearity
Calibration
Stability
Repeatability
Gauge R&R
Precision
Reproducibility
C
opyright2012BSI.A
lrights
reserved.
12
Calibration
The Bias of a gauge can be assessed by repeat measurements of a known reference unit
This can be extended across the operating range of the gauge in a Gauge Linearity Study
The Stability of the gauge can be assessed by control charting a reference unit
Should not routinely recalibrate, instead if reference unit tests outside the control limits, then
re-calibrate
If measurement device requires frequent recalibration, attempt to improve stability
reserved.
13
Gauge R & R
Gauge R & R is a means of assessing the repeatability and reproducibility of our measurement
systems.
Gauge R & R studies are carried out in order to discover how much of the process variation is
due to the measurement device and measurement methods.
?
Dimension
Copyright
2012BSI.Allrights
reserved.
14
Variable Gauge R&R

reserved.
Variable Gauge R&R
Requirements:
A minimum of two operators (recommend 3 or 4)
At least 10 parts which should be chosen to represent the full range of manufacturing variation
(it may be acceptable to use fewer parts in some special cases)
Each part should be measured two or three times in a random order
Operators should not be aware of the previous result when measuring the same part
Part 10
Part 1
Part 5
Part 4
Part 3
Part 2
reserved.
16
Variable Gauge R&R
There are two methods available:

1. Analysis of Variance (ANOVA)
2. X-Bar and R
The ANOVA method is:

the recommended approach
takes into account any interactive effect between operator and part
Copyright
2012BSI.Allrights
reserved.
17
Variable Gauge R&R
Part-to-Part
Variation
Operator
Overall
Variation
Reproducibility
Operator
by part
Interaction
Measurement
System
Variation
Repeatability
We want the Part-to-Part component to be large!
reserved.
18
Variable Gauge R&R - Example

Part
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
Copyright
2012BSI.Allrights
reserved.
Operator 1
0.65
0.60
1.00
1.00
0.85
0.80
0.85
0.95
0.55
0.45
1.00
1.00
0.95
0.95
0.85
0.80
1.00
1.00
0.60
0.70
Operator 2
0.55
0.55
1.05
0.95
0.80
0.75
0.80
0.75
0.40
0.40
1.00
1.05
0.95
0.90
0.75
0.70
1.00
0.95
0.55
0.50
Operator 3
0.50
0.55
1.05
1.00
0.80
0.80
0.80
0.80
0.45
0.50
1.00
1.05
0.95
0.95
0.80
0.80
1.05
1.05
0.85
0.80
19
Variable Gauge R&R - Minitab
Open Worksheet: Gauge R&R
Each operator measures

each part twice
The Part numbers being

measured
Operators performing
measurements
Individual measurements
In Minitab the data is entered in

single columns
Copyright
2012BSI.Allrights
reserved.
20
Stat>Quality Tools>Gage Study>Gage R& R (Crossed)

Enter Part, Operator, Measurement
Check ANOVA Method
C
opyright2012BSI.A
lrights
reserved.
21

Two-Way ANOVA Table With Interaction
Source
Part
Operator
Part * Operator
Repeatability
Total
DF
9
2
18
30
59
SS
2.05871
0.04800
0.10367
0.03875
2.24913
MS
0.228745
0.024000
0.005759
0.001292
F
39.7178
4.1672
4.4588
Gage R&R
P
0.000
0.033
0.000
p < 0.05 so all terms significant
Source
Total Gage R&R
Repeatability
Reproducibility
Operator
Operator*Part
Part-To-Part
Total Variation
VarComp
0.0044375
0.0012917
0.0031458
0.0009120
0.0022338
0.0371644
0.0416019
reserved.
%Contribution
(of VarComp)
10.67
3.10
7.56
2.19
5.37
89.33
100.00
Estimates of Variances for each

source
22
Variance Component Estimates
Part-to-Part
Variation
0.0371644
Overall
Variation
0.0416019
Operator
0.0009120
Reproducibility
0.0031458
Measurement
System
Variation
0.0044375
Operator
by part
Interaction
0.0022338
Repeatability
0.0012917
Variances are additive!
reserved.
23
Variable Gage R&R Standard Deviations
This is the gauge

standard deviation,
R&R = 0.066615
Remember that
standard deviations
are not additive!
Source
Study Var
%Study Var
StdDev (SD)
(6 * SD)
(%SV)
0.066615
0.39969
32.66
Repeatability
0.035940
0.21564
17.62
Reproducibility
0.056088
0.33653
27.50
Operator
0.030200
0.18120
14.81
Operator*Part
0.047263
0.28358
23.17
Part-To-Part
0.192781
1.15668
94.52
Total Variation
0.203965
1.22379
100.00
Total Gage R&R
C
opyright2012BSI.A
lrights
reserved.
24
Interpreting the Results
We would like the total measurement system variation (Gauge

R&R) to be as small as possible.
Calculate the percentage of the process tolerance taken up by the
measurement system variation, represented by 6 x the gauge
standard deviation. This is known as %Precision/Tolerance or
%P/T.
The Process Tolerance is equivalent to the difference between the
upper and lower specification limits (USL LSL).
6 R & R
%P/T = 100%
Process Tolerance
Copyright
2012BSI.Allrights
reserved.
25
Is the Gauge Good?
% P/T
(6
R&R/Process Tolerance)
Acceptability
0 - 10%
Very Good (Six Sigma Gauge)
10 - 30%
May be Acceptable
>30%
Probably Not Acceptable
The interpretation will also depend on the current level of process variation
Note that these guidelines are as recommended in Measurement Systems Analysis Third Edition published in March 2002 as part of
QS-9000 and developed in conjunction with AIAG.
reserved.
26
Is the Gauge Good?
% R&R
If the %P/T is greater than 10%, then a secondary calculation can be used to decide
whether the gauge can be used during the DMAIC activity.
Comparing R&R to the current process variation indicates whether the measurement
device is currently causing a problem. This is known as %R&R.
We need an independent estimate of the process (total) variation (the value from the Gauge
R&R is based on only a few samples)
We would like the measurement standard deviation to be less than the total standard
deviation
%R & R =
reserved.
R &R
100% < 50%
Process(total)
27
Interpreting the Results
1. Comparing the gauge variation to the process tolerance:
6 0.0666
%P/T =
100% =
100% = 28.5%
1.4
Tolerance
R &R
This is greater than 10% so the gauge will not be good enough for six sigma. As the process
improves the gauge will become a problem. To improve this gauge we should start by
addressing the reproducibility.
2. Comparing the gauge variation to the process variation:
0.0666
%R & R =
100% =
100% = 37%
0.18
Process(total)
R &R
This is less than 50% so the gauge is not the limiting factor at the moment. We can use this
gauge for process improvement.
Copyright
2012BSI.Allrights
reserved.
28

Open Worksheet: GaugeR&R
Stat>Quality Tools>Gage Study>Gage R& R (Crossed)
Enter Part, Operators, Measurement
Check ANOVA Method
Select Options:
Study Variation: 6
Process Tolerance: 1.4
Historical standard deviation: 0.18
Two-Way ANOVA Table With Interaction
Source
Part
Operator
Part * Operator
Repeatability
Total
Copyright
2012BSI.Allrights
reserved.
DF
9
2
18
30
59
SS
2.05871
0.04800
0.10367
0.03875
2.24913
MS
0.228745
0.024000
0.005759
0.001292
F
39.7178
4.1672
4.4588
P
0.000
0.033
0.000
29
Components of Variation
Gage R&R
Source
Total Gage R&R
Repeatability
Reproducibility
Operator
Operator*Part
Part-To-Part
Total Variation
Source
Total Gage R&R
Repeatability
Reproducibility
Operator
Operator*Part
Part-To-Part
Total Variation
VarComp
0.0044375
0.0012917
0.0031458
0.0009120
0.0022338
0.0371644
0.0416019
%Contribution
(of VarComp)
10.67
3.10
7.56
2.19
5.37
89.33
100.00
StdDev (SD)
0.066615
0.035940
0.056088
0.030200
0.047263
0.192781
0.203965
Study Var
(6 * SD)
0.39969
0.21564
0.33653
0.18120
0.28358
1.15668
1.22379
%Study Var
(%SV)
32.66
17.62
27.50
14.81
23.17
94.52
100.00
%Tolerance
(SV/Toler)
28.55
15.40
24.04
12.94
20.26
82.62
87.41
%Process
(SV/Proc)
37.01
19.97
31.16
16.78
26.26
107.10
113.31
Number of Distinct Categories = 4
reserved.
30
Gage R&R (ANOVA) for Measurement

Reported by :
Tolerance:
M isc:
G age name:
D ate of study :
120
% Contribution
% Study Var
% Process
100
% Tolerance
Percent
80
60
40
20
0
Gage R&R

reserved.
Repeat
Reprod
Part-to-Part
31
Part to Part Measurements

Reported by :
Tolerance:
M isc:
Gage name:
Date of study :
Measurement by Part
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
1
10
Part

reserved.
32
Operator by Part Interaction

Reported by :
Tolerance:
M isc:
Gage name:
Date of study :
Operator * Part Interaction

1.1
Operator
1.0
1
2
3
A verage
0.9
0.8
0.7
0.6
0.5
0.4
1

reserved.
6
Part
10
33
Measurements by Operator

Reported by :
Tolerance:
M isc:
Gage name:
Date of study :
Measurement by Operator
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
1

reserved.
2
Operator
34
Xbar and R Chart by Operator

Reported by :
Tolerance:
M isc:
Gage name:
Date of study :
Xbar Chart by Operator
Sample M
Mean
1.0
_
_
UC L=0.8796
X=0.8075
LCL=0.7354
0.8
0.6
0.4
R Chart by Operator
Sample Range
1
0.12
3
UC L=0.1252
0.08
0.04
_
R=0.0383
0.00
LCL=0

reserved.
35
Rounding Errors
Rounding is another component of measurement variation which needs to be minimised

It can be shown that to avoid rounding error getting in the way of achieving six sigma quality, it
is necessary to have a minimum of 14 discrete values between the upper and lower specification
For one-side specifications, there need to be at least 7 discrete values between the process
average and the specification limit
reserved.
UG3637
Rounding Errors - Interpolating
If possible interpolate between graduation marks

For example, thermometers are frequently marked to the nearest degree but can be read to the
nearest 0.2 degrees, even if the last digit is not entirely accurate
Interpolating frequently reduces and never increases the measurement variation
reserved.
37
Improving the Measurement System
Gauge incapable:
Repeatability (Gauge)
Take multiple measurements and use average (short term fix)
Mistake proofing (e.g. provision of tooling to hold part during measurement)
May need maintenance
Reproducibility (Operators)
Use 1 operator (short term fix during improvement only)
Have several operators measure the part and take the average (short term fix)
Ensure consistency (training, SOPs, WIS, )
Mistake proofing (e.g. provision of tooling to hold part during measurement)
Calibrations on the gauge dial may not be clear
Reproducibility Operator x Part Interaction

Identify cause of interaction and then as Operator
reserved.
38
Destructive Gauge R&R
Destructive gauge testing means that it is impossible to carry out repeat tests!
To complete an assessment of a destructive gauge it is therefore necessary to assume

homogeneity within batches.
If there is much more difference in parts between batches than within batches, then a standard
variable Gauge R & R may be sufficient.
reserved.
39
Workshop Variable Gauge R&R
Using the provided measuring device and products carry out a Gauge R&R
Use three operators and measure each part twice
Ensure that the order of measuring is randomised
Analyse the data using Minitab

What could you do, if anything to improve the Measurement System?
Prepare a short report detailing your findings
Copyright
2012BSI.Allrights
reserved.
40
Attribute Gauge R&R

reserved.
Attribute Gauge R&R
A Gauge R&R study can also be carried out on attribute data

Using attribute data, we would have a problem with the measurement system if:
Operators disagree with each others evaluation of a piece
The same operator gains different results from a repeat evaluation of the same piece
reserved.
42
Attribute Measurement System
An attribute measurement system compares each part to a standard and either accepts or rejects
the part.
The screen effectiveness is the ability of the attribute measurement system to properly
discriminate good from bad.
Screen effectiveness of 100% is desirable.
reserved.
43
Conducting Attribute Gauge R&R
1. Select a minimum of 30 parts from the process. These parts should represent the full
spectrum of process variation (good parts, defective parts, borderline parts).
2. An expert inspector performs an evaluation of each part, classifying it as Good or Not
Good.
3. Independently and in a random order, each of 2 or 3 operators should assess the parts as
Good or Not Good.
4. Calculate effectiveness scores.
reserved.
44
Attribute Gauge R&R

Minitab Data Layout:
Open Worksheet: Attribute Gage R&R
Column containing
parts being
assessed
Text column
containing operator
performing
measurements
reserved.
Text column
containing expert
assessment (can use
words or numbers but
must be consistent)
Text column
containing results of
measurements (can
use words or numbers
but must be
consistent)
45
Attribute Gauge R&R

Stat>Quality Tools>Attribute Agreement Analysis
Enter Results in Attribute Column, Part in
Samples, Appraiser in Appraisers and Expert in
Known standard/attribute
Click on Results button and select Percentages

reserved.
46
Attribute Gauge R&R - Results
Attribute Agreement Analysis

Within Appraiser
Assessment Agreement
Appraiser # Inspected # Matched Percent (%)
95.0% CI
A
30
28
93.3 ( 77.9, 99.2)
B
30
30
100.0 ( 90.5, 100.0)
C
30
30
100.0 ( 90.5, 100.0)
# Matched: Appraiser agrees with him/herself across trials.
Each Appraiser vs Standard
Appraiser A was not

consistent on two out of
thirty parts inspected
Appraiser A disagreed
with expert on two parts,
Appraiser B and C
disagreed with expert on
one part
Appraiser # Inspected # Matched Percent (%)

95.0% CI
A
30
28
93.3 ( 77.9, 99.2)
B
30
29
96.7 ( 82.8, 99.9)
C
30
29
96.7 ( 82.8, 99.9)
# Matched: Appraiser's assessment across trials agrees with standard.

reserved.
47

Assessment Disagreement
# Not Good/
# Good/
Appraiser
Good
Percent (%)
Not Good Percent (%) # Mixed Percent (%)
A
0
0.0
0
0.0
2
6.7
B
1
6.7
0
0.0
0
0.0
C
1
6.7
0
0.0
Appraiser B0 assessed 0.0
one
# Not Good/Good: Assessments across trials = Not Good / standard = Good.
part as Not
Good
when the
# Good/Not Good: Assessments across trials = Good / standard
= Not
Good.
# Mixed: Assessments across trials are not identical.
standard (expert) assessed it
as Good
Between Appraisers
Appraiser A,B and C agreed
# Inspected # Matched Percent (%)
95.0% CI
on 26 out of 30 parts
30
26
86.7 ( 69.3, 96.2)
inspected
# Matched: All appraisers' assessments agree with each other.
All Appraisers vs Standard
Appraiser A,B and C all
agreed with the standard on
# Inspected # Matched Percent (%)
95.0% CI
26 out of 30 parts inspected
30
26
86.7 ( 69.3, 96.2)
# Matched: All appraisers' assessments agree with standard.

reserved.
48
Date of study :
Reported by :
Name of product:
Misc:
Within Appraisers
Appraiser vs Standard
95.0% C I
P ercent
100
95
Percent
95
Percent
95.0% C I
P ercent
100
90
90
85
85
80
80
B
Appraiser

reserved.
B
Appraiser
49
The target effectiveness is always 100%
Possible Corrective Actions include:

Operator Training
Clarification of Standards
Simplification of Standards
Conversion to Variable Data
reserved.
50
Workshop Attribute Gauge R&R
From your team select two expert inspectors.

The experts should select 20 sweets, roughly half good (pass) and half bad (fail).
Some sweets should be borderline.
Carry out a Gauge R&R
Use two operators and measure each part twice (if more time available use three operators)
Ensure that the order of measuring is randomised
Analyse the data

What could you do, if anything, to improve the Measurement System?
Prepare a short report detailing your findings.
Copyright
2012BSI.Allrights
reserved.
51
Measurement Systems Analysis - Summary
Measurement errors can account for a large proportion of the variation in our measures (ys)
We must evaluate our measurement systems before assessing process stability or process
capability
Errors in measurement systems can come from a variety of sources
Action should be taken to improve the capability of our measurement systems if they are found
to be inadequate
reserved.
52
DMAIC Improvement Process
Define

Select Project
Define Project
Objective
Form the Team
Measure
Define Measures (ys)
Analyse
Improve
Identify Potential xs
C1 C2 C3
Evaluate Measurement
System
Effect

Map the Process

Identify Customer
Requirements

Determine Process
Stability
Determine Process
Capability
LSL
15

Identify Priorities
Update Project File
Phase Review
20
USL
25
30
35
Set Targets for

Measures
Phase Review
Copyright
2012BSI.Allrights
reserved.
. .. .
.
. .. .. . x
y=f(x1,x2,..)
C4 C5 C6
Characterise xs
Analyse xs
Run
1
2
3
4
5
6
7
8
1
1
1
1
2
2
2
2
1
1
2
2
1
1
2
2
1
1
2
2
2
2
1
1
1
2
1
2
1
2
1
2
1
2
1
2
2
1
2
1
1
2
2
1
1
2
2
1
1
2
2
1
2
1
1
2
Select Critical xs
Optimise xs

Phase Review
Control Critical xs
10.2
Upper Control Limit
10.0
9.8
Lower Control Limit
9.6
1
10
15
20
Monitor ys
y
Set Tolerances for xs
Verify Improvement
xx
x
x x x
x x
x
x
x
Control
LSL
15
20
Validate Control
Plan
Close Project
USL
25
30
35
Phase Review
Phase Review
53
Appendix - ANOVA

reserved.
ANOVA Table - Construction
Construction of an Analysis of Variance (ANOVA) table requires the following:
1. Identification of the Sources (Components) of Variation
2. Calculation of the Sum of Squares due to each Source of Variation
3. Assignment of the appropriate Degrees of Freedom
4. Calculation of the Mean Squares
5. Calculation of the F-Ratio
reserved.
55
1. Components of Variation
Analysis of Variance (ANOVA) allows the decomposition of the

variability in the Gauge R&R study.
The components of variation in the Gauge R&R study are:
2Part
= Variation due to the different parts
= Variation due to different operators
2Operator
2Operator x Part = Variation due to the interaction between operator
and part
2Repeatability = Variation due to gauge repeatability
= 2Part + 2Operator + 2Operator x Part + 2Repeatability
2Total
Copyright
2012BSI.Allrights
reserved.
56
2. Calculation of the Sum of Squares
The total sum of squares is calculated as follows:
(
y)
= (y y ) = y
n
SSTotal
Strictly speaking the sum of squares column is the sum of squares

around the mean, known as the corrected sum of squares. We always
use the corrected sum of squares when estimating variation.
y =0.65 + 0.60 + 1.00 + 1.00 + ............. + 0.80 = 48.45

y = 0.65 + 0.60 + 1.00 + 1.00 + ........... + 0.80 = 41.3725
(
y)
(
48.45)
=y
= 41.3725
= 41.3725 39.1234 = 2.2491
SS
2
Total
Copyright
2012BSI.Allrights
reserved.
60
57
Calculation of the Sum of Squares
The sum of squares due to parts is calculated as follows:

SS Part
2
2
2
2
(
P1 ) + (P2 ) + (P3 ) + ......... + (P10 ) ( y )
=
SS Part
2
2
2
2
2
(
3.40) + (6.05) + (4.80) + ......... + (4.00) (48.45)
=
np
60
SS Part = 41.1821 39.1234 = 2.0587
Where:
P1, P2, P3..P10 are the Sums for each Part
i.e the Sum of the 6 measurements made on each part.
np is the number of individual measurements of each part.
Copyright
2012BSI.Allrights
reserved.
58
The sum of squares due to operators is calculated as follows:
SS Operator
2
2
2
(
O1 ) + (O2 ) + (O3 ) ( y )
=
SS Operator
2
2
2
2
(
16.55 ) + (15.35 ) + (16.55 ) (48.45 )
=
no
20
60
SS Operator = 39.1714 39.1234 = 0.0480
Where:
O1, O2, O3 are the Sums for each Operator
i.e the sum of the 20 measurements made by each operator.
no is the number of measurements made by each operator.
Copyright
2012BSI.Allrights
reserved.
59
The sum of squares due to the interaction between operators and

parts is calculated as follows:
SSOperatorPart
2
2
2
(
O1 P1 ) + (O1 P2 ) + ..........(O3 P10 ) ( y )
=
SSOperatorPart
2
2
2
2
(
1.25) + (2.00 ) + ........(1.65) (48.45)
=
0.0480 2.0587
nO P
2
SSOperator SS Part
60
SSOperatorPart = 41.3338 39.1234 0.0480 2.0587 = 0.1037
Where:
O1P1, O1P2,.O3P10 are the Sums for each Operator & Part Combination
i.e the sum of the 2 measurements made by each operator on each part.
nOxP is the number of measurements made by each operator on each part.
Copyright
2012BSI.Allrights
reserved.
60
The sum of squares due to repeatability is obtained by

subtraction:
SS Repeatability = SSTotal SS Part SSOperator SSOperator Part
SS Repeatability = 2.2491 2.0587 0.0480 0.1037 = 0.0387
Copyright
2012BSI.Allrights
reserved.
61
Source of Variation
Sum of Squares
Between Parts
Between Operators
Operator x Part
Repeatability
2.0587
0.0480
0.1037
0.0387
Total
2.2491
Copyright
2012BSI.Allrights
reserved.
62
3. Degrees of Freedom
Degrees of Freedom is a statistical concept relating to the

number of paired comparisons required to distinguish
between items.
For example, we need to find the tallest person out of 3
people. 2 comparisons would be required:
Person 1 v Person 2
Tallest v Person 3
We would then know who the tallest person is.
Copyright
2012BSI.Allrights
reserved.
63
Rules for Degrees of Freedom
The following rules apply to Degrees of Freedom:

DF for a Factor (Main Effect) = (Number of Levels) 1
DF for interactions = Product of the DF of the Factors involved
DF for Repeatability = (Product of Factor Levels) x (Repeats 1)
Total DF = (Number of Individual Results) - 1
Copyright
2012BSI.Allrights
reserved.
64
Degrees of Freedom
Source of Variation
Sum of Squares
Degrees
of
Freedom
Between Parts
Between Operators
Operator x Part
Repeatability
2.0587
0.0480
0.1037
0.0387
9
2
18
30
Total
2.2491
59
Copyright
2012BSI.Allrights
reserved.
65
4. Calculation of the Mean Squares
The Mean Square is calculated as follows:

Mean Square = (Sum of Squares) / (Degrees of Freedom)
Source of Variation
Sum of Squares
DF
Mean Square
Between Parts
Between Operators
Operator x Part
Repeatability
2.0587
0.0480
0.1037
0.0387
9
2
18
30
0.2287
0.0240
0.0058
0.0013
Total
2.2491
59
Copyright
2012BSI.Allrights
reserved.
66
5. Calculation of the F-Ratio
Source of Variation
Sum of Squares
DF
Mean Square
F-Ratio
Between Parts
Between Operators
Operator x Part
Repeatability
2.0587
0.0480
0.1037
0.0387
9
2
18
30
0.2287
0.0240
0.0058
0.0013
39.43
4.14
4.46
Total
2.2491
59
The F-Ratio is used to test the significance of each source of variation.

F-Ratio for Parts = (MSParts) / (MSOperator x Part)
F-Ratio for Operators = (MSOperators) / (MSOperator x Part)
F-Ratio for Operator x Part = (MSOperators x Parts) / (MSRepeatability)
Copyright
2012BSI.Allrights
reserved.
67
Estimating Components of Variation
The Mean Square column is expected to contain the following

components of variation. This expected mean square is only applicable
to this current study, where we have 3 operators, 10 parts and 2 repeat
measurements. For other studies, the number of the components will
change. (Fortunately, Minitab can do this for us!)
Source
Mean Square
Expected Mean Square
Parts
0.2287
62Part + 22OperatorPart + 2Repeatability
Operators
0.0240
202Operator + 22OperatorPart + 2Repeatability
Operator x Part 0.0058

Repeatability
0.0013
Copyright
2012BSI.Allrights
reserved.
22OperatorPart + 2Repeatability
2
Repeatability
68
Mean Square
Source
Parts
0.2287
62Part + 2 2OperatorPart + 2Repeatability
Operators
0.0240
20 2Operator + 2 2OperatorPart + 2Repeatability
Operator x Part 0.0058

0.0013
Repeatability
2
Repeatability
2
Operator Part
+ 2Repeatability
2Repeatability
= 0.0013
2
2
+
= 0.0058
2Operator
Part
Repeatability
2
2Operator
Part = 0.0058 0.0013 = 0.0045
2
Operator
Part = 0.00225
Copyright
2012BSI.Allrights
reserved.
69

Mean Square
Source
Parts
0.2287
Operators
0.0240
Operator x Part
0.0058
Repeatability
0.0013
2
Operator Part
+ 2Repeatability
2Repeatability
2
2
2
+ 2Operator
+
= 0.0240
20Operator
Part
Repeatability
20
2
Operator
2
Operator
2
2
= 0.0240 2Operator
Repeatability
Part
(0.0240) 2(0.00225) 0.0013

=
= 0.00091
20
reserved.
70

Mean Square
Source
Parts
0.2287
Operators
0.0240
Operator x Part
0.0058
Repeatability
0.0013
2
Operator Part
+ 2Repeatability
2Repeatability
2
2
2
6Part + 2Operator Part + Repeatability = 0.2287
62Part = 0.2287 22Operator Part 2Repeatability
2
Part
0.2287 2(0.00225) 0.0013

=
= 0.03715
6
Copyright
2012BSI.Allrights
reserved.
71
2Part = 0. 03715
2
= 0. 00091
Operator
2
Operator
Part = 0. 00225
2
= 0. 00130
Repeatabil
ity
2
2
2
+ Operator
+
2Total = 2Part + Operator
Part
Repeatability
2Total = 0. 03715 + 0. 00091 + 0. 00225 + 0. 00130 = 0. 04161

We have established estimates of each of the components of variation!
Copyright
2012BSI.Allrights
reserved.
72
Variance Component Estimates
Part-to-Part
Variation
0.03715
Overall
Variation
0.04161
Operator
0.00091
Reproducibility
0.00316
Measurement
System
Variation
0.00446
Operator
by part
Interaction
0.00225
Repeatability
0.00130
Variances are additive!

Copyright
2012BSI.Allrights
reserved.
73

6 SSGB Amity Bsi Msa P

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

6 SSGB Amity Bsi Msa P

Uploaded by

Copyright:

Available Formats

Module-6

Measurements Systems Analysis

Copyright 2012 BSI. All rights

Measurements Systems Analysis - Agenda

Is our data accurate?

Repeatability & Reproducibility

Accuracy & Precision

Measurements System Variation

Bias, Linearity, Stability, Repeatability, Reproducibility, Calibration, Gauge R&R

Parts, Operators, Variation

Is the gauge good?

Analysis of Variance (ANOVA)

Is our Data Accurate?

DMAIC Improvement Process

Define Measures (ys)

Map the Process

Set Targets for

Upper Control Limit

Lower Control Limit

Measurement Accuracy & Precision

Measurement is accurate but not

Measurement is precise but not

Measurement is accurate and

Measurement System Variation

Reproducibility is the difference in the average of the measurements made by different

Measurement System Variation

If measurement device requires frequent recalibration, attempt to improve stability

Variable Gauge R&R

Copyright 2012 BSI. All rights

Variable Gauge R&R

Variable Gauge R&R

There are two methods available:

The ANOVA method is:

Variable Gauge R&R

We want the Part-to-Part component to be large!

Variable Gauge R&R - Example

Variable Gauge R&R - Minitab

Open Worksheet: Gauge R&R

Each operator measures

The Part numbers being

In Minitab the data is entered in

Variable Gauge R&R - Minitab

Stat>Quality Tools>Gage Study>Gage R& R (Crossed)

Variable Gauge R&R - Minitab

p < 0.05 so all terms significant

Estimates of Variances for each

Variance Component Estimates

Variable Gage R&R Standard Deviations

This is the gauge

Total Gage R&R

Interpreting the Results

We would like the total measurement system variation (Gauge

Is the Gauge Good?

Very Good (Six Sigma Gauge)

Probably Not Acceptable

Is the Gauge Good?

100% < 50%

Interpreting the Results

1. Comparing the gauge variation to the process tolerance:

2. Comparing the gauge variation to the process variation:

Variable Gauge R&R - Minitab

Number of Distinct Categories = 4