Documente Academic
Documente Profesional
Documente Cultură
Basic Metrology
&
Reliability of Measurements
PQSPC-Pune.
Process Input Process Output
Raw materials
PQSPC-Pune.
Total Process Capability Vs
Measurement Capability
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
PQSPC-Pune.
Why total Process Capability Study? Why Measurement Capability Study?
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%
PQSPC-Pune.
Basic Elements of measuring system
PQSPC-Pune.
Measurement Needs in Industries
PQSPC-Pune.
Measurement Needs in Industries (Contd.)
→ Both of above measurements requires reliability for the Zero Defect Production.
PQSPC-Pune.
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.
Actual value
It is the value obtained with an error with-in the permissible limit of tolerances .
Approximate value
It is the value obtained with an error exceeding the permissible error of measurement and requiring
refinement.
PQSPC-Pune.
Basic Metrology - Terms & Definitions (Contd.)
Accuracy
The closeness of agreement to the true value of measurand.
Repeatability or Precision
The closeness of agreement between results of successive measurements of the same value of a
quantity carried out under identical conditions at short intervals of times.
Reproducibility
The closeness of agreement between corrected results of measurements of the same value of a
quantity when the measurements are made under different conditions (such as different methods,
different instruments, different observers, etc.)
Error of measurement
It is the difference between the true value of the size being measured and the value found by
measurement. Or
Discrepancy between the measured and true values of the quantity.
Bias
The closeness of agreement to the reference value.
Uncertainty of measurement
It is the range of values with in which the true value is estimated to lie . PQSPC-Pune.
Bias and Repeatability
(Bias)
PQSPC-Pune.
Relationship between Bias and Repeatability
PQSPC-Pune.
Basic Metrology - Terms & Definitions (Contd.)
Variance
A measure of dispersion which is the sum of the squared deviations of observation from their average
divided by the number of observations.
Alternatively, if the same quantity is measured by any other observer using an unbiased instrument, a
large number of times, then 95% times values of the measurand would lie within
-U +U
95% X 95%
PQSPC-Pune.
Basic Metrology - Terms & Definitions (Contd.)
Calibration
Calibration is set of operations which under specified conditions, establish the relationship between
values indicated by measuring system and the corresponding known values of a measurand.
• It is essentially comparison with higher standard which is traceable to national / international
standard
• Higher standard means
– Higher accuracy (1:10 preferably, but not less than 1:3)
– Better resolution (1:10 or 1:5)
PQSPC-Pune.
Measurement Systems Variables
PQSPC-Pune.
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.
PQSPC-Pune.
Important Characteristics – Normal curve
For n-number of observations
Average Value
Standard Deviation
For n>20
For n<20
Estimated Standard R
σˆ =
Deviation d 2*
d2* corresponds to g & m
PQSPC-Pune.
Example - Forming Normal Distribution Curve
PQSPC-Pune.
Example-Forming Normal Distribution Curve
(Contd.)
Where n = 10
Average Value.
= 30.005 mm
Standard Deviation
= 1.2 µm
R
σˆ = * = 1.25 µm
d2
d2* = 3.17905
corresponds to g =1 & m= 10
PQSPC-Pune.
Types of Measurement Variation
PQSPC-Pune.
Types of Measurement Variation
PQSPC-Pune.
Basic Types of Errors and their characteristics.
PQSPC-Pune.
Basic Types of Errors and their characteristics.
(Contd.)
PQSPC-Pune.
Basic Types of Errors and their characteristics.
(Contd.)
Characteristics of random errors:
• These occur randomly and the specific cases of such errors cannot be determined.
• These are caused by friction and play in the instrument’s linkages, estimation of
reading by judging fractional part of a scale division, by errors in positioning the
measured object, etc.
• These are variable in magnitude and sign.
• Random errors of larger absolute value are rarer than those of smaller values.
• The probability of occurrence is equal for positive and negative errors of the same
absolute value since random errors follow normal frequency distribution.
• The arithmetic mean of random errors in a given series of measurements
approaches zero as the number of measurement increases. The most reliable value
of the size being sought in a series of measurements in the arithmetic mean of the
results obtained.
• The main characteristic of random errors, which is used to determine the maximum
measuring error, is the standard deviation. The maximum error for a given method of
measurement is determined as three times the standard deviation.
• The maximum error determines the spread of possible random error values.
PQSPC-Pune.
Sources of Errors
I – Instrument
1 – Instrumental error
2 – Uncertainty of calibration
W – Work piece 3 – Instrumental Res.
P – Person
1 – Form error. 4 – Repeatability. 1 – Repeatability
2 – Surface finish. (Skill of operator for positioning and
measuring pressure etc.)
Instrumental Resolution
(When instrument is under calibration)
S – Standard
SWIPE E – Environmental Factors
1 – Dev. From nominal Dim. Temperature.
The result of any measurement does not indicate true value, but only an approximation
or estimate of the true value of the specific quantity i.e., measurand. Therefore, the
result is complete only when accompanied by a quantitative statement of its uncertainty.
Uncertainty is defined as how close the measurement result is likely to be to the true
value with reference to the value assigned to the national prototype standard for that
parameter with certain probability or confidence level.
• This is expressed in quantitative aspect of measurement quality.
• This means traceability is automatically established when measurement Uncertainty
is quoted.
• Uncertainty of measurement determines the measurement capability of a
measurement standard or a measuring equipment.
PQSPC-Pune.
Measurement Uncertainty
PQSPC-Pune.
Uncertainty in measurement will occur due to
combined effect of
PQSPC-Pune.
II. Why Uncertainty of Measurement?
The Standard S
The Work-Piece W
The Environment E
The factors effecting these elements cause errors at the end of measurement
process and hence the measurement result never gives the true value of a
quantity (measurand) under measurement.
Both systematic and random errors affect the observed results and contribute to this
uncertainty. These contribution have been referred to as systematic and random
component of uncertainty receptively.
PQSPC-Pune.
Measurement system eats your Tolerance
USL LSL
Design Tolerance
+U -U
Tolerance Available
PQSPC-Pune.
Available tolerance for the manufacturing is equal to Specified
tolerance minus measurement Uncertainty.
ISO - 14253.
PQSPC-Pune.
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
/ Linearity
+σ
2
GRR
PQSPC-Pune.
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 Performance
2
e
= σ 2
Capability
+σ
2
Stability
+σ
2
Consistency
y
PQSPC-Pune.
CLAUSE 7.6 : ISO 9001 : 2000 Control of
Monitoring & Measuring Equipments
PQSPC-Pune.
CLAUSE 7.6 : ISO 9001 : 2000 Control of
Monitoring & Measuring Equipments
Where necessary to ensure valid results, measuring equipment shall
PQSPC-Pune.
Clause 7.6 : ISO/TS 16949 : 2002 - Control of
Monitoring and Measuring Devices
7.6 7.6.1 Measurement system analysis
Statistical studies shall be conducted to analyse the variation present in the results of
each type of measuring and test equipment system. This requirement shall apply to
measurement systems referenced in the control plan. The analytical methods and
acceptance criteria used shall conform to those in customer reference manuals on
measurement systems analysis. Other analytical methods and acceptance criteria may
be used if approved by the customer.
PQSPC-Pune.
Techniques for Evaluation and Control of
Measurement Reliability.
PQSPC-Pune.
Measurement Uncertainty Vs MSA
PQSPC-Pune.
Section II
Measurement System Analysis
PQSPC-Pune.
Measurement Issues
PQSPC-Pune.
Measurement Issues (Contd.)
PQSPC-Pune.
Measurement Issues (Contd.)
PQSPC-Pune.
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.
PQSPC-Pune.
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.
PQSPC-Pune.
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.
PQSPC-Pune.
Guidelines for Conducting Measurement System
Analysis Studies
Flow Diagram for MSA studies
PQSPC-Pune.
Reference manual-3rd Edition (Contd.)
(for Destructive Non-Destructive, Measurement)
Complex or Non – Replicable Measurement System Study
PQSPC-Pune.
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
PQSPC-Pune.
Bias Study (Contd…)
PQSPC-Pune.
Bias Study (Contd.)
= 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
PQSPC-Pune.
Bias Study (Contd.)
PQSPC-Pune.
Independent Sample Method (Contd.)
Ref. Value Bias n (m) = 15, df = 10.8,
6.0 - - 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
PQSPC-Pune.
Independent Sample Method (Contd.)
3.55333
ν ,1−
2
PQSPC-Pune.
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
PQSPC-Pune.
Linearity Study
PQSPC-Pune.
Linearity Study (Contd.)
PQSPC-Pune.
Linearity Study (Contd.)
PQSPC-Pune.
Variable Measurement system Study:
Linearity Study
PQSPC-Pune.
Variable Measurement system Study:
Linearity Study (Contd.)
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:
PQSPC-Pune.
Linearity Study
Example
Gauge /Instrument under test - Plunger type Dial Gauge .
Range 10 mm / Rev. 0.1mm
PQSPC-Pune.
Linearity Study (Contd.)
PQSPC-Pune.
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
PQSPC-Pune.
Linearity Study (Contd.)
PQSPC-Pune.
Repeatability & Reproducibility Study (Contd.)
PQSPC-Pune.
Repeatability & Reproducibility Study (Contd.)
PQSPC-Pune.
Repeatability & Reproducibility Study (Contd.)
PQSPC-Pune.
Range method – for R & R test
PQSPC-Pune.
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
PQSPC-Pune.
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
PQSPC-Pune.
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
PQSPC-Pune.
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
PQSPC-Pune.
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
PV
ndc = 1.41 ×
GRR PQSPC-Pune.
Repeatability & Reproducibility Study
Range and Average Method (Contd.)
PQSPC-Pune.
Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Gauge R & R Study will provide information concerning the cause of
measurement system variation.
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
PQSPC-Pune.
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
2
(0.20188)
Reproducibility AV = (0.446 × 0.5231)2 − = 0.22963
10 × 3
PQSPC-Pune.
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
Conclusion :
Measurement System is marginally Acceptable as % GRR is less
than 30% . PQSPC-Pune.
Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Gauge R & R Study Graphical Analysis – Average Control Charts
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
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. PQSPC-Pune.
Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Gauge R & R Study Graphical Analysis – Appraiser Bias Control Charts
PQSPC-Pune.
Repeatability & Reproducibility Study
Range and Average Method (Contd.)
Gauge R & R Study Graphical Analysis – Appraiser Inconsistency Control Charts
PQSPC-Pune.
Stability Study
PQSPC-Pune.
Stability Study (Contd.)
• On a periodic basis measure the same part at least three times, and
check that whether all readings are with in the control limits.
PQSPC-Pune.
Stability Study (Contd.)
PQSPC-Pune.
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 .
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
PQSPC-Pune.
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
PQSPC-Pune.
Stability Study (Contd.)
PQSPC-Pune.
Stability Study (Contd.)
Conclusion:
All measurements are within U.C.L. & L.C.L.means system is
stable from period 17/02/2000 up to 19/10/2000.
PQSPC-Pune.
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 .
PQSPC-Pune.
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 -
PQSPC-Pune.
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
PQSPC-Pune.
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
PQSPC-Pune.
Attribute Measurement System Study :
Study Data Set (Contd.)
Po − Pe
kappa =
1 − Pe
kappa〉 0.75 _____ good − aggreement .
PQSPC-Pune.
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.
_ + _
PQSPC-Pune.
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 Code Ref. Value Code
0.599581 - 0.503091 + d = average (di)
0.587893 - 0.502436 +
0.576459 - 0.502295 + is an estimation
0.570360 - 0.501132 +
0.566575 - 0.498698 + of region II area.
0.566152 - 0.493441 +
0.561457 X 0.488905 +
0.559918
0.547204
X
X
0.488184
0.487613
+
+
d LSL
= 0.470832 − 0.446697
0.545604
0.544951
X
X
0.486379
0.484167
+
+
d LSL
= 0.024132
0.543077 X 0.483803 +
0.542704 + 0.477263 +
0.531939
0.529065
+
+
0.476901
0.470832
+
+
d USL
= 0.566152 − 0.542704
0.523754
0.521642
+
+
0.465454
0.462410
X
X
d USL
= 0.023448
0.520496 + 0.454518 X
0.519694 + 0.452310 X
0.517377 + 0.449696 X 0.024132 + 0.023448
0.515573 + 0.446697 - d=
0.514192
0.513779
+
+
0.437817
0.427687
-
-
2
0.509015 + 0.412453 - d = 0.0237915
0.505850 + 0.409238 -
PQSPC-Pune.
Risk Analysis Method (Contd.)
2 × 0 . 02379
% GRR = × 100 = 25 %
0 . 599581 − 0 . 409238
PQSPC-Pune.
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
PQSPC-Pune.
Analytic Method – Gauge Performance Curve (Contd.)
PQSPC-Pune.
Analytic Method – Gauge Performance Curve (Contd.)
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
PQSPC-Pune.
Analytic Method – Gauge Performance Curve (Contd.)
PQSPC-Pune.
Percentage points of the two tailed
t distribution at α = 0.05 i.e. at 95%C.L.
PQSPC-Pune.
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
PQSPC-Pune.
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.