Lesson 05 Measurement System Analysis

(Part 2: Variable Data)

E339 Quality and Reliability Engineering

E339 Quality and Reliability Engineering
– Topic Tree L01: Quality Control and Reliability Overview
L02-03: Geometric Dimensioning and Part 1: Concepts 2
Tolerancing Part 2: Applications
Legend
Part 1: Concepts
Face-to-Face
eLearning
L01-L14 L04-07: Measurement System Analysis
(Variables Gage Study,
Part 2: Variable Data 5
Part 3: Attribute Data
Unscheduled
#
Asynchronous
Introduction to Quality Control, Attribute Agreement Analysis)
Part 4: Applications
e-Learning
(UNS01, 02…) Measurement System Analysis, Part 1: Concepts
students should L08-11: Statistical Process Control Part 2: Variable Data 9
plan to cover Statistical Process Control (Variable, Attribute Data Analysis) Part 3: Attribute Data
ahead of L0#
Part 4: Applications
and Capability
Part 1: Concepts
L12-13: Process Capability Analysis
E339 Quality Part 2: Applications 13
and Reliability L14: LLP1 Consolidation

Engineering Part 1: Sampling Plans 15

Part 2: Sampling Risks 16
L15-18: Acceptance Sampling Plans
Part 3: Rectifying Inspection
Part 4: Comparing Plans
L15-L26 Part 1: Reliability Statement
Part 2: Reliability Measurement
Acceptance Sampling, L19-22: Reliability Concepts
Part 3: Reliability Distributions
Reliability Concepts and Part 4: Failure Modes 22
System Reliability Part 1: Reliability Block Diagram
L23-25: System Reliability Part 2: Fault Tree Analysis
Part 3: Dynamic Tests 25
L26: LLP2 Consolidation 2
Gage Repeatability and Reproducibility (GR&R)
What (dimensions) to measure?

(b) Vertical Prong Height (a) Vertical Prong Width (e) Bottom Virtual
(8 ribs positions) (4 pairs of ribs) Circle Diameter

(d) Bottom
Inner Ribs
Width (4 pairs)

(c) Prong Width

(Front, Left Prongs)

3
Gage Repeatability and Reproducibility (GR&R)
What (dimensions) to measure?

4
Gage Repeatability and Reproducibility (GR&R)
What (equipment) measures what (dimensions) well?

Romer arm
Dial Gauge
(iv)
(v)

(iii)

(ii)
(i) Vernier
Micrometer Calliper Coordinate Measurement Machine 5
Gage Repeatability and Reproducibility (GR&R)
What (equipment) measures what (dimensions) well?

Measurement System / Equipment

Note 1: A standard
micrometer can provide
precise measurement
but for very small
objects that fit between
its anvil and spindle
(typically 25mm).

Vernier Calliper Micrometer Dial Gauge Romer Arm Probe CMM

(a) Vertical Prong Width
(4 pairs of ribs)
X (‘outer jaw’) X (a) X X
(b) Vertical Prong Height
X (‘tail’) (b) X X X
Dimensions

(at 8 ribs’ positions)

(c) Prong Width
(Front and Left prongs)
X (‘outer jaw’) (c) X X X
(d) Bottom Inner Ribs Width
(4 pairs)
X (‘inner jaw’) (d) X X
(e) Bottom Virtual Circle Diameter
(4 circles)
X (e) X
6
Gage Repeatability and Reproducibility (GR&R)
What (analysis) it takes to assess a measurement system well?

Measurements

Operators Parts Repeat1 Repeat2 Repeat3

1 1
1 2
1 3
1 4
1 5
1
1
6
7
GRR Table (Excel) to Minitab
1 8
2 1
2 2
2 3
2 4
2 5
2 6
2 7
2 8 7
Bias & Linearity Studies
What (analysis) it takes to assess a measurement system well?
PART REFERENCE MEASUREMENT
1 e.g. 25.0
1 25.0
1 25.0
1 25.0
1 25.0
1 25.0
1 25.0
1 25.0
1 25.0
1 25.0
2 20.5
Measurement System | Sine Block Dimensions (mm) 2 20.5
2 20.5
Romer Arm: 50.0 75.0 100.0
2 20.5
Micrometer: 14.5 20.5 25.0 2 20.5
Vernier Calliper: 10.0 15.0 20.0 2 20.5
2 20.5
Vernier Calliper: 9.0 14.0 19.0
2 20.5
Dial Gauge: 8.0 13.0 18.0 2 20.5
2 20.5
3 14.5
3 14.5
3 14.5
Objective of study is to verify whether 3 14.5
measuring equipment gives the same 3 14.5
3 14.5
readings as master standards. If it is 3 14.5
different, measurement equipment may 3 14.5
3 14.5
have an issue. 3 14.5 8
Bias & Linearity Studies – Vernier Caliper
General guidelines for good
measurement systems:
% Linearity < ___, % Bias < ___

Enter process variation

9
Gage Repeatability and Reproducibility (GR&R)

• MINITAB > Stat > Quality Tools > Gage Study > Gage R&R Study (Crossed)
• Method and Analysis default is ANOVA
• Study variation = 6 (standard deviations)
• Upper spec – Lower Spec = refer to tolerance attached to each dimension
• Process Tolerances = 1
• Alpha = 0.05

10
Gage Repeatability and Reproducibility (GR&R)
– Vernier Caliper (Interaction term)
Interpreting the Output

• MINITAB actually does an

analysis of variance
• (ANOVA) to calculate variance
components, which are then
used to estimate the percent
variation due to the measuring
system. The percent variation
appears in columns at the end
of the Gage R&R output.

• When the P-value for the interaction

term is > 0.05, a second ANOVA
table is generated that omits the
interaction term from the model.

• In this case, P-value = 0.65 (> 0.05 -

Second ANOVA table > Interaction between Vernier
Caliper and Operator is not
significant), therefore second
ANOVA without interaction is
established and this is the correct
ANOVA table to be used.

11
Gage Repeatability and Reproducibility (GR&R)
– Vernier Caliper (%Contribution)
Interpreting the Output
M, Sum of repeatability & Variance Component
reproducibility • Variance Component (VarComp) is actually the base

Variance
values for calculating % Gage R&R using the ANOVA
method.
Components
Equipment error • Estimates for a Gage R&R study are obtained by
calculating the variance component for each term
(Operator, Part, Oper*Part) and for equipment error
Operator variation (Repeatability). Repeatability, Operator, and
Reproducibility components are summed to obtain a total
variability due to the measuring system (i.e., Total Gage
Source VarComp %Contribution
R&R).
(of VarComp)
Total Gage R&R 0.0008060 7.88
Repeatability 0.0007649 7.48 Why Use Variance Components?
Reproducibility 0.0000411 0.40 • Variance components are easy to understand and to use
OPERATORS_Ve 0.0000411 0.40 when we are just concerned with estimating variation
Part-To-Part 0.0094165 92.12 contributed by the measurement system (i.e., Total Gage
Total Variation 0.0102226 100.00 R&R) to the variation we see within our study (i.e., Total
Variation).

% Contribution
• While an ANOVA table compares each term to error, we
really want to compare each component (equipment error,
Operator, Reproducibility, part) variation to Total Variation,
P, Part-To-Part i.e., %Contribution.
• (In terms of % Contribution (7.88% which is < 10%),
T, Sum of total gage R&R and part-to-part this Measurement System is Marginally Acceptable
depending on application, the cost of the measuring
device, cost of repair, or other factors)
12
Gage Repeatability and Reproducibility (GR&R)
– Vernier Caliper (% Study Var & %Tolerance)
Interpreting the Output – % Contribution vs. % Study vs. % Tolerance
Process tolerance = 0.4
% Study Variation
M (Measurement variation) • Study variation of that source
Study Var %Study Var %Tolerance
Source StdDev (SD) (6 × SD) (%SV) (SV/Toler) divided by the study variation
Total Gage R&R 0.028391 0.170345 28.08 42.59 of the total variation.
Repeatability 0.027657 0.165941 27.35 41.49
Reproducibility 0.006413 0.038481 6.34 9.62 • That is, for total Gage R&R:
OPERATORS_Ve 0.006413 0.038481 6.34 9.62
Part-To-Part 0.097039 0.582232 95.98 145.56 %Study Var (%SV) = [(6*M)
Total Variation 0.101107 0.606640 100.00 151.66 / (6*T)] * 100
P (Part variation) T (Total variation)
Number of Distinct Categories = 4 • (In terms of % Study
Variation (28.08% which is
%Tolerance
between 10% and 30%),
• The %Tolerance (SV/Toler) column only appears when a tolerance
this Measurement System
is entered.
is Marginally Acceptable
depending on application, the
• Each value of (6*SD) is divided by tolerance. That is, for total
cost of the measuring device,
Gage R&R: %Tolerance (SV/Toler) = [(6*M) /Tolerance ] * 100,
cost of repair, or other
e.g. 0.170345 / 0.4 = 42.59%.
factors)
• Dividing by tolerance is to compare measurement system variation
to customer specifications.

• (In terms of % Tolerance (42.59% which is greater than 30%),

this Measurement System is Unacceptable) 13
Gage Repeatability and Reproducibility (GR&R)
– Vernier Caliper (ndc)
Interpreting the Output – % Contribution vs. % Study vs. % Tolerance
Study Var %Study Var %Tolerance
Source StdDev (SD) (6 × SD) (%SV) (SV/Toler)
Total Gage R&R 0.028391 0.170345 28.08 42.59
Repeatability 0.027657 0.165941 27.35 41.49
Reproducibility 0.006413 0.038481 6.34 9.62
OPERATORS_Ve 0.006413 0.038481 6.34 9.62
Part-To-Part 0.097039 0.582232 95.98 145.56
Total Variation 0.101107 0.606640 100.00 151.66

Number of Distinct Categories = 4

ndc

The number of distinct categories of parts that the measurement • Ndc < 5 (Measurement
system is currently able to distinguish. The capability of a system is not acceptable.
measurement system (telling the differences between parts) This Measurement System
increases when the number of distinct categories (ndc) is not able to distinguish
increases. parts sufficiently)

Number of Distinct Categories = 2 * (P/M)

Marginally Acceptable

Acceptable Not acceptable

ndc=1 ndc=3 ndc=5 14

Gage Repeatability and Reproducibility (GR&R)
– Romer Arm (Interpretation)
Gage R&R Interpreting the Output

Variance Components • % Contribution = 0.81%

(< 1% -> acceptable)
• % Study Variation = 9.01%
Source VarComp %Contribution
(of VarComp) (< 10% -> acceptable)
Total Gage R&R 0.0000138 0.81 • % Tolerance = 5.57%
Repeatability 0.0000121 0.71 (< 10% -> acceptable)
Reproducibility 0.0000017 0.10
• Ndc = 15
OPERATORS_Ro 0.0000017 0.10
Part-To-Part 0.0016837 99.19 (>4 -> acceptable)
Total Variation 0.0016975 100.00

Process tolerance = 0.4

Gage Evaluation
Source StdDev (SD) Study Var %Study Var %Tolerance
(6 × SD) (%SV) (SV/Toler)
Total Gage R&R 0.0037136 0.022282 9.01 5.57
Repeatability 0.0034823 0.020894 8.45 5.22
Reproducibility 0.0012903 0.007742 3.13 1.94
OPERATORS_Ro 0.0012903 0.007742 3.13 1.94
Part-To-Part 0.0410325 0.246195 99.59 61.55
Total Variation 0.0412002 0.247201 100.00 61.80

Number of Distinct Categories = 15

15
Comparison of MSA studies between tools
Items Vernier Caliper Romer Arm Criteria (AIAG guidelines)

%Linearity 11.20% 0.00% < 10% (acceptable)

%Bias (average) 7.20% 1.50%
%Bias (1st) 54.5% (14.5mm) 1.5% (50mm)
< 10% (acceptable)
%Bias (2nd) 12.8% (20.5mm) 1.5% (75mm)
%Bias (3rd) 63.2% (25mm) 1.4% (100mm)
< 1% (acceptable)
GR&R (%
7.88% 0.81% 1-9% (marginally acceptable)
contribution)
> 9% (unacceptable)
GR&R (% study
28.08% 9.01% < 10% (acceptable)
var)
10-30% (marginally acceptable)
GR&R (%
42.59% 5.57% > 30% (unacceptable)
Tolerance)
➢ 4 acceptable
Ndc 4 15
(Able to distinguish data)

Good sample %Part-to-part

Yes Yes
selection > %GR&R (% contribution)

16
Analysis of Results for Variables Gage Study

• In this problem, bias, linearity, and gage R&R are conducted.

• Master standard is used to ensure that the tools are not drifted and give the values between
the acceptability criteria of the known reference values.
• Summary of acceptability of measurement systems:
• Romer Arm (and CMM) are the best tools as all measurement aspects are acceptable or
marginally acceptable.
• Vernier caliper has unacceptable %bias and %linearity. This maybe the reason that Vernier
caliper has unacceptable GRR (>30% %tolerance); and repeatability contributed more to
GR&R compared to reproducibility.
• When the tools are marginally acceptable, there should be some justifications and/or
corrective action in order to continue using the tool.

17
Learning Outcomes 2/3

• List possible sources of process variation in measurement

• State the reasons for performing Measurement System Analysis (MSA) in data collection
• List possible causes of poor measurement in Gage Repeatability and Reproducibility (GR&R)
• Explain the four (4) aspects of measurement system capability using gage study
• Explain the two (2) types of appraiser assessment using attribute agreement analysis
• Apply Variables Gage Study to handle variable data in Minitab
• Apply Attribute Agreement Analysis to handle attribute data in Minitab
• Interpret acceptability in measurement system based on AIAG guidelines
• Interpret degree of agreement in appraisers’ responses based on Kappa Statistics and p-
value

18
E339 Quality and Reliability Engineering
– Topic Flow
Introduction to Quality Control, Acceptance Sampling,
Measurement System Analysis Reliability Concepts and System Reliability
Statistical Process Control and Capability

Quality Control and Reliability Concepts 1

Reliability Overview

Geometric Statistical Process Acceptance Sampling Reliability Concepts 2

Dimensioning and Control (Variable) Plans (AQL)
Tolerancing

We are
here! Variables Gage Study Statistical Process Acceptance Sampling System Reliability 1
Control (Attribute) Plans (AOQL and LTPD)

Attribute Agreement Process Capability System Reliability 2

Analysis Analysis

19