Section I

Basic Metrology
&
Reliability of Measurements

PQSPC-Pune.
 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)

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

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

PQSPC-Pune.
Why total Process Capability Study? Why Measurement Capability Study?

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.

PQSPC-Pune.
Basic Elements of measuring system

PQSPC-Pune.
 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.

PQSPC-Pune.
 Measurement Needs in Industries (Contd.)

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.

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.

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. PQSPC-Pune.
Basic Metrology - Terms & Definitions
Nominal / Basic value
It is the basic value on which tolerances are given.

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.

Exact value / Reference value

It is the value obtained with the highest metrological accuracy attainable in practice.
True Value
It is the value obtained by the measurements when measurement errors will become equal to zero.

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.

Confidence Level (Coverage Factor)

It is a measure of degree of difficulty with which a result is expressed. Precisely, it is the band of
range of uncertainty within which the value of measurand lies e.g. If X is the measured value of the
measurand with uncertainty U at a confidence level of 95%. It means that the probability for true value
lying between X + U and X - U is 95%.

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

These are some of the variables in a measurement

system. What others can you think of?

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

For observations taken from g- no. of

subgroups & m- subgroup size

Estimated Standard R
σˆ =
Deviation d 2*
d2* corresponds to g & m
PQSPC-Pune.
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

PQSPC-Pune.
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

PQSPC-Pune.
Types of Measurement Variation

PQSPC-Pune.
 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

Linearity: Change in bias over a Uniformity: change in repeatability over

measuring range. a measuring range.

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

over a time.

PQSPC-Pune.
 Basic Types of Errors and their characteristics.

1. Controllable (Static Errors) – One which remains constant during a

number of identical measurements or which varies in a predictable manner
when conditions change.

2. Random Error – One which varies in an unpredictable manner during a

number a of identical measurements.

PQSPC-Pune.
Basic Types of Errors and their characteristics.
(Contd.)

Characteristics of controllable errors:

• These are controllable in both their magnitude and sense.
• These can be determined and reduced, if attempts are made to
analyse them.
• These errors either have a constant value or a value changing
according to a definite law.
• These can be due to thermal effects during measurements,
mechanical measurement errors due to misalignment of gauge the
component, calibration errors of reference standards, errors due to
stylus pressure etc.

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.

2 – Uncertainty of calibration. Humidity.

3 – Wear in standard. Cleanliness.

4 – Form error. Vibration.

5 – Surface Finish. P – Procedure
1 – Method of measurement

2- Software used for processing of

Measurement Data.
PQSPC-Pune.
 I. What is Uncertainty of measurement?

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

Uncertainty is a range assigned to a measurement result that

describes, within a defined level of confidence , the range
expected to contain the true measurement result. It is reported
as a bilateral quantity.
Uncertainty (U) is expressed as :
True Value = Measured Value ± U
U is the expanded uncertainty of measurand and measurement
result.

PQSPC-Pune.
Uncertainty in measurement will occur due to

combined effect of

Un-corrected Controllable/ Static errors

&
Random errors.

PQSPC-Pune.
 II. Why Uncertainty of Measurement?

• Uncertainty of Measurement determines the measurement capability of

measuring instrument
• 6 Basic metrological elements effected by various factors

The Standard S

The Work-Piece W

The Measuring Instrument I

The Person / The Procedure P

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.

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

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

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

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

The organization shall determine the monitoring and measurement to be

undertaken and the monitoring and measuring devices needed to provide
evidence of conformity of product to determined requirements.
The organization shall establish processes to ensure that monitoring and
measurement can be carried out and are carried out in a manner that is
consistent with the monitoring and measurement requirement.

PQSPC-Pune.
 CLAUSE 7.6 : ISO 9001 : 2000 Control of
Monitoring & Measuring Equipments
Where necessary to ensure valid results, measuring equipment shall

a) be calibrated or verified at specified intervals, or prior to use, against measurement

standards traceable to international or national measurement standards; where no
such standards exist, the basis used for calibration or verification shall be recorded;

b) be adjusted or-adjusted as necessary;

c) be identified to enable the calibration status to be determined;

d) be safeguarded from adjustments that would invalidate the measurement result;

e) be protected from damage and deterioration during handling, maintenance and

storage.

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.

ISO - GUM: 1995.

ISO Guidelines for Uncertainty of Measurements.

Guidelines published by ISO (International Organization for Standardization) for
evaluation of measurement Uncertainties.
Equivalent National document NABL -141.

AIAG - Reference Manual - Measurement System Analysis.

Guidelines published by Automotive Industries Action Group for assessment of
measurement system Capability and Performance.

PQSPC-Pune.
 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.

PQSPC-Pune.
 Section II
Measurement System Analysis

PQSPC-Pune.
 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.

PQSPC-Pune.
 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.

PQSPC-Pune.
 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.

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

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.
PQSPC-Pune.
 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.

PQSPC-Pune.
 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

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…)

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

PQSPC-Pune.
 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  
PQSPC-Pune.
 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.

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.)

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

Lower Limit= −0.1185

Upper d 2σ b   3.472 × 0.05813

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

Upper Limit= 0.1319

Lower bound(−0.1185) 〈 zero 〈Upper bound(0.1319)

Gauging system is acceptable with respect to bias test
PQSPC-Pune.
Variable Measurement system Study:
Bias Study -Control Chart Method

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

Linearity is the difference in the Bias Values through a

expected operating range of the Gauge.

PQSPC-Pune.
 Linearity Study (Contd.)

The linearity is determined by slope of Regression line

Lower the slope grater the linearity and vice versa.

PQSPC-Pune.
 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.

PQSPC-Pune.
 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.

PQSPC-Pune.
 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
PQSPC-Pune.
 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:
PQSPC-Pune.
 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.

PQSPC-Pune.
 Linearity Study (Contd.)

Ref. Value 2.0 4.0 6.0 8.0 10.0

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

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.)

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

PQSPC-Pune.
 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

PQSPC-Pune.
 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 .

PQSPC-Pune.
 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.

PQSPC-Pune.
 Repeatability & Reproducibility Study (Contd.)

Part to Part variation : In the indication of manufacturing process

variation .

PQSPC-Pune.
 Range method – for R & R test

• Select two appraisers .

• 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.

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

d*2 = 1.19 Corresponds to

m=2 and g=5 number of parts.
PQSPC-Pune.
 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 .

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

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.

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 

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

 PV 
ndc = 1.41 ×  
 GRR  PQSPC-Pune.
 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.

PQSPC-Pune.
 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.
PQSPC-Pune.
 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

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

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

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 

% 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% . PQSPC-Pune.
 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 PQSPC-Pune.
 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. PQSPC-Pune.
 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.

PQSPC-Pune.
 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.

PQSPC-Pune.
 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)

PQSPC-Pune.
 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.

PQSPC-Pune.
Stability Study (Contd.)

Conduct Bench mark test

Apply Control Limits

Monitor Periodically With respect

to the Set Control Limits

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 .

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

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.)

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

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.)

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 .
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

The goodness of estimation depends on the sample

size and how closely sample represents the process.

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.)

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.

PQSPC-Pune.
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

PQSPC-Pune.
Analytic Method – Gauge Performance Curve (Contd.)

XT a P d' 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.
PQSPC-Pune.
Analytic Method – Gauge Performance Curve (Contd.)
Attribute Gauge Performance Curve
Plotted on Normal Probability Paper.

PQSPC-Pune.
 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

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

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

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.