Metrology and Tolerancing for

Manufacture
(Context of Quality Engineering)
Level: 3rd Year (CADCAM Section) _Sem. 1

Major: Mechanical Engineering

Chapter.4.

ACCURACY AND REPEATABILITY IN PRECISION MACHINES

AND INSTRUMENTS

A. INTRODUCTION

▪ Factors that affect measurement system variation can be studied using the Gauge
Repeatability& Reproducibility technique (GR&R). WHEN USING THE GR&R SYSTEM,
MEASUREMENT SYSTEM VARIATION CAN BE CHARACTERIZED BY LOCATION (STABILITY,
BIAS, LINEARITY) AND WIDTH OR SPREAD (REPEATABILITY AND REPRODUCIBILITY).
Sources of variations in a production process/measuring system may include; tool wear,
product degradation, operator error and performance, measurement system, material
and/or environmental variations, etc. PRIMARILY, THERE EXIST THREE MAJOR SOURCES
OF VARIATION IN A PRODUCTION PROCESS, I.E., REPEATABILITY, REPRODUCIBILITY, PART-
TO-PART AND TOTAL. For example, you have a known standard that is exactly 5.00g.
You weigh it multiple times and get the following readings: 5.01g, 4.99g, 4.97g, 5.03g
and 5.01g. The differences in the measurements are due to measurement system
variation. If, however, you weigh different parts that come off your manufacturing line,
are the differences due to measurement system variation or due to actual differences in
the parts themselves? Use Minitab's measurement system analysis tools to determine
the contribution of different sources of variation. IN A PROCESS, IF THE MEASUREMENT
SYSTEM VARIATION IS LARGE COMPARED TO PART-TO-PART VARIATION, THE
MEASUREMENTS MAY NOT PROVIDE USEFUL INFORMATION.

▪ Like any other process, a MEASUREMENT SYSTEM IS SUBJECT TO BOTH COMMON

(PREDICTABLE BUT NOT CONTROLLABLE VARIATION) AND SPECIAL CAUSE VARIATION
(UNPREDICTABLE BUT EASILY CONTROLLABLE/CORRIGIBLE). To control the measurement
system variation, you must first identify the sources of the variation (see Fig. 1), then
you must either eliminate or reduce the various causes.

Fig.2
ACCURACY IS THE MEASURE OF HOW CLOSELY A MEASUREMENT RESEMBLES THE ACTUAL OR
ACCEPTED VALUE. HOW CLOSELY TWO MEASUREMENTS OF THE SAME THING DIFFER IS REFERRED
TO AS PRECISION.

B. GAGE R&R STUDY

▪ GAGE R&R MEASURES THE AMOUNT OF VARIABILITY IN MEASUREMENTS CAUSED BY THE

MEASUREMENT SYSTEM ITSELF. Gage R&R is very important when new workers are
assigned, new tools are used, or any significant process change. THE TOTAL VARIATION
IS COMPOSED BY THE GAGE R&R AND THE PART-TO-PART VARIATIONS.

𝝈𝑻𝒐𝒕𝒂𝒍 = √𝝈𝟐𝑮𝒂𝒈𝒆 𝑹&𝑹 + 𝝈𝟐𝑷𝒂𝒓𝒕−𝒕𝒐−𝒑𝒂𝒓𝒕

𝑻𝒐𝒍 = (𝑼𝑺𝑳 − 𝑳𝑺𝑳) ≥ 𝟔 ∗ 𝝈𝑻𝒐𝒕𝒂𝒍

▪ MEASUREMENT SYSTEM ERRORS CAN BE CLASSIFIED INTO TWO CATEGORIES: ACCURACY

AND PRECISION. Within any measurement system, you can have one or both of these
components (see Fig. 2)

a) ACCURACY, which describes the difference between the mean observation and the
part's true/accepted/target value. In a measuring system, THE ACCURACY IS BROKEN
INTO THREE COMPONENTS.

o LINEARITY − It is an indicator of the consistency of measurements over the

entire range of measurements. It tells us how well the instrument measurement
corresponds to reality.
o BIAS − is the difference between the observed average measurement and the
master/reference/true/target value.
o STABILITY − a measure of the capability of the measuring system to produce the
same values over time when measuring THE SAME SAMPLE, i.e., ABSENCE OF
SPECIAL CAUSE VARIATION (only random variation is present).

Fig. 2

b) PRECISION, which describes the observed variation when you measure the same part
repeatedly with the SAME CONDITIONS (i.e., machine, process, instrument, operator,
shift, day/week, etc.). This is composed of two metrics: REPEATABILITY (SAME
CONDITIONS) and REPRODUCIBILITY (IF DIFFERENT CONDITIONS ARE EXPERIENCED).

Could you comment the gaging distribution as it is shown in Fig. 3?

Fig. 3

▪ Other than TYPE1 GAGE STUDY (bias and repeatability evaluation for repeated
measurement of one part in same conditions), there exist three others Gage R&R.

a) CROSSED GAGE R&R- WHEN EACH OPERATOR MEASURES EACH PART MANY TIMES
as in Fig. 4. It must have a balanced design with random factors. IT IS USED FOR NON-
DESTRUCTIVE TESTING.

Fig. 4

b) NESTED Gage R&R-WHEN ONLY ONE OPERATOR MEASURES EACH PART (see Fig.
5). IT IS USED FOR DESTRUCTIVE TESTING and it must have a balanced design with
random factors.

Fig. 5

c) EXPANDED GAGE R&R- WHEN WE NEED TO INCLUDE MORE FACTORS THAN

OPERATOR AND PART. TYPICALLY, CROSSED AND NESTED DEALS WITH ONLY TWO
FACTORS (OPERATOR AND PART). Design can be balanced or unbalanced.

▪ THERE EXIST THREE METHODS TO ANALYZE GAGE R&R: I) RANGE METHOD, II) AVERAGE
AND RANGE METHOD (X BAR AND R), iii) and ANALYSIS OF VARIANCE METHOD
(ANOVA).
 THE X AND R METHOD BREAKS DOWN THE OVERALL VARIATION INTO THREE CATEGORIES:
PART-TO-PART, REPEATABILITY, AND REPRODUCIBILITY. THE ANOVA METHOD BREAKS
DOWN REPRODUCIBILITY INTO OPERATOR AND OPERATOR-PART COMPONENTS. THE
ANOVA METHOD IS MORE ACCURATE THAN THE X AND R METHOD, IN PART, BECAUSE IT
ACCOUNTS FOR THE OPERATOR-PART INTERACTION.

C. APPLICATIONS (TYPE 1 GAGE STUDY)

▪ Case study 1: Electrode manufacturer

A electrode manufacturer wants to evaluate the measurement system which measures

the outer diameter of electrode shafts to determine whether the system accurately
measures the shaft diameter within 0.05mm tolerance. An operator measures a reference
shaft with a known diameter of 11.8mm 30 times. Using Minitab, analyze the variability
of the measuring instrument. What do you recommend?

Trial 1 2 3 4 5 6 7 8 9 10
Measurement
12.1 12.1 11.90 11.5 12.3 12 12 12 11.8 12.1
in coded units
Trial 11 12 13 14 15 16 17 18 19 20
Measurement
12.06 12.31 12.2 12.2 12 12 12 11.8 11.9 11.9
in coded units
Trial 21 22 23 24 25 26 27 28 29 30
Measurement
12.2 11.7 11.4 11.8 11.8 12.1 12.2 12.1 12.1 12.0
in coded units

▪ Case study 2: engine torque

A quality engineer wants to evaluate the measurement system for the torque developed
by and electrical engine. The USL=12.335N.m, the LSL=12.385N.m and the nominal
value is 12.306N.m. A reference engine is measured 50 times. Using Minitab, analyze the
variability of the measuring instrument. What do you recommend?

Type1 Gage Study2

Tria Meas. Tria Meas. Tria Meas. Tria Meas. Tria Meas.
l Data l Data l Data l Data l Data
1 12.3075 11 12.3096 21 12.3142 31 12.3102 41 12.3104
2 12.3018 12 12.3092 22 12.3014 32 12.3068 42 12.3076
3 12.3116 13 12.3036 23 12.3055 33 12.3116 43 12.3071
4 12.3074 14 12.319 24 12.3028 34 12.3029 44 12.3162
5 12.3077 15 12.3086 25 12.3096 35 12.3045 45 12.3117
6 12.3127 16 12.3154 26 12.3023 36 12.303 46 12.3104
7 12.3035 17 12.3009 27 12.2979 37 12.3068 47 12.3066
8 12.2985 18 12.3054 28 12.3091 38 12.3153 48 12.3019
9 12.3059 19 12.3132 29 12.3091 39 12.3072 49 12.3124
10 12.3127 20 12.2995 30 12.301 40 12.3097 50 12.3061
▪ Case study 3: Pump pressure

In a production system, we want to evaluate the measurement system of the outlet

pressure of a centrifugation pump using a mechanical manometer. The IT is ±0.2bars. An
operator measures a reference pump having an outlet pressure of 2.75 bars. The
measurement process is repeated 30 times. Using Minitab, analyze the variability of the
measuring instrument. What do you recommend?

Trial 1 2 3 4 5 6 7 8 9 10
Measurement
2.78 1.87 1.87 2.36 2.36 2.21 2.22 2.45 2.33 4.56
in coded units
Trial 11 12 13 14 15 16 17 18 19 20
Measurement
4.21 4.13 3.56 3.47 3.69 2.11 2.23 2.12 2.24 2.24
in coded units
Trial 21 22 23 24 25 26 27 28 29 30
Measurement
2.27 5.47 5.89 5.31 2.44 1.8 1.72 4.1 3.88 3.56
in coded units

▪ Case study 4: Blood sugar testing system

In a hospital a new blood sugar system is tested. We want to evaluate the new
measurement system using a reference test of 1.7g. The upper specification limit is 4g.
The operator measures the reference sample 30 times. Using Minitab, analyze the
variability of the new measuring instrument. What do you recommend?

Trial 1 2 3 4 5 6 7 8 9 10
Measurement
2.56 2.22 2.14 2.13 2.36 2.34 2.41 2.36 2.33 4.12
in coded units
Trial 11 12 13 14 15 16 17 18 19 20
Measurement
4.36 4.56 4.37 3.89 3.78 2.11 2.23 2.12 2.22 2.23
in coded units
Trial 21 22 23 24 25 26 27 28 29 30
Measurement
2.24 5.36 5.47 5.87 2.44 1.8 1.72 4.65 4.89 4
in coded units

▪ Case study 5: Vibration frequency testing system

A mechanism is tested for vibration frequency using a frequency meter. We want to

evaluate the new measurement system using a reference test of 3.0Hz. The lower
specification limit is 1.3 Hz. The operator measures the reference sample 30 times. Using
Minitab, analyze the variability of the new measuring instrument. What do you
recommend?

Trial 1 2 3 4 5 6 7 8 9 10
Measurement
256 222 215 256 212 232 212 245 233 425
in coded units
Trial 11 12 13 14 15 16 17 18 19 20
 Measurement
436 422 347 389 375 211 223 212 227 226
in coded units
Trial 21 22 23 24 25 26 27 28 29 30
Measurement
222 521 536 554 244 180 172 412 325 369
in coded units

D. APPLICATIONS (LINEARITY AND BIAS STUDY)

Part 1 2 3 4 5
Ref Value 7 9 11 13 15
1 7.7 10.1 10.8 12.6 14.1
2 7.5 8.9 10.7 12.7 14.3
3 7.4 9.2 10.9 12.8 14.5
4 7.5 10 10.9 12.7 14.3
5 7.7 8.8 11.0 12.8 14.4
6 7.3 8.9 11.1 12.8 14.5
7 7.5 8.9 11.0 12.8 14.5
8 7.5 8.9 11.1 12.7 14.5
9 7.4 8.9 11.4 12.8 14.6
10 7.4 9.0 11.3 12.5 14.2

A new measurement system will be used to monitor production for some quality characteristic.
Five parts are selected from the process, which represent the range of the quality characteristic
in production. The reference values of each part are determined. Then one operator measures
each part ten times. The obtained process variation is 7.55 from a Gage R&R study using the
ANOVA method. The data are shown in the Table above. Using Minitab, check for linearity
and the bias of the measuring system. What do you advise?

E. APPLICATIONS (CROSSED AND NESTED GAGE STUDY)

▪ Case study 1: Geometry deviation in a forging process (use ANOVA)

A testing engineer selected 10 parts that represent the expected range of the
geometry deviation in a forging process. Three technicians A, B and C measure
randomly ten parts three times. Assess the measurement system Gage R&R.

Operator Part Trial 1 Trial2 Trial3 Operator Part Trial 1 Trial2 Trial3

1 2.78 1.87 1.87 6 2.11 2.23 2.12

2 2.36 2.36 2.21 7 2.22 2.23 2.24
3 2.22 2.45 2.33 B 8 5.36 5.47 5.87
4 4.56 4.21 4.13 9 2.44 1.8 1.72
A 5 3.56 3.47 3.69 10 4.65 4.89 4
6 2.11 2.23 2.12 1 2.56 2.22 2.15
7 2.24 2.24 2.27 2 2.56 2.12 2.32
C
8 5.47 5.89 5.31 3 2.12 2.45 2.33
9 2.44 1.8 1.72 4 4.25 4.36 4.22
 10 4.1 3.88 3.56 5 3.47 3.89 3.75
1 2.56 2.22 2.14 6 2.11 2.23 2.12
2 2.13 2.36 2.34 7 2.27 2.26 2.22
B 3 2.41 2.36 2.33 8 5.21 5.36 5.54
4 4.12 4.36 4.56 9 2.44 1.8 1.72
5 4.37 3.89 3.78 10 4.12 3.25 3.69

▪ Case study 2: Wearing in CNC machine

An engineer wants to monitor the wearing in a CNC machine. He randomly selects
30 samples that represent the expected range of the process variation and gives 10
random samples to 3 machine operators (A, B and C) who are chosen in random. The 3
machine operators measure the wearing using an ultrasonic piezo system. The 10 sample
parts are measured twice for a total of 60 measurements. Each part is unique to operator,
so, no 2 operators measured the same part. Because the measurements are nested within
operator, the engineer performs a nested gage R&R (GRR) study to assess the variability
in measurements that may emanate from the measurement system. Assess the
measurement system Gage R&R and say your recommendations.

Measurements Measurements
Operators Trial Trial
in coded units in coded units
1 4.861 4.860 16 4.759 4.865
2 4.802 4.884 17 4.815 4.889
3 4.827 4.917 B 18 4.821 4.885
4 4.793 4.867 19 4.858 4.866
5 4.858 4.866 20 4.814 4.854
A
6 4.811 4.855 21 4.829 4.898
7 4.797 4.884 22 4.788 4.872
8 4.842 4.859 23 4.820 4.888
9 4.822 4.905 24 4.800 4.884
10 4.820 4.888 25 4.787 4.906
C
11 4.787 4.906 26 4.829 4.898
12 4.802 4.901 27 4.827 4.917
B 13 4.869 4.819 28 4.822 4.905
14 4.793 4.867 29 4.820 4.888
15 4.822 4.905 30 4.824 4.866

▪ Case study 3: Tensile stress monitoring

An engineer wants to control the tensile stress variation in crank mechanism. He

randomly selects 30 samples that represent the expected range of the process variation
and gives 10 random samples to 3 quality control technicians (A, B and C) who are
chosen in random. Each 10 sample parts are measured twice for a total of 60
measurements. Assess the measurement system Gage R&R and say your
recommendations.
 Measurements Measurements
Operator Trial Operator Trial
in coded units In coded units
1 862.485 875.890 16 862.473 875.900
2 862.423 875.830 17 862.446 875.660
3 862.483 875.550 B 18 862.474 875.920
4 862.471 875.880 19 862.454 875.800
5 862.454 875.660 20 862.479 875.840
A
6 862.475 875.640 21 862.480 875.760
7 862.428 875.440 22 862.490 875.720
8 862.497 875.880 23 862.454 875.660
9 862.470 875.770 24 862.419 875.730
10 862.454 875.890 25 862.468 875.840
C
11 862.468 875.790 26 862.480 875.750
12 862.459 875.850 27 862.483 875.760
B 13 862.450 875.650 28 862.470 875.750
14 862.471 875.710 29 862.454 875.660
15 862.470 875.960 30 862.474 875.730

▪ Case study 4: residual stresses

An engineer wants to control the residual stresses after a hardening process a

bearing hub. He randomly selects 30 samples that represent the expected range of the
process variation and gives 10 random samples to 3 quality control technicians (A, B
and C) who are chosen in random. Each 10 sample parts are measured fourfold for a
total of 120 measurements. Assess the measurement system Gage R&R and say your
recommendations.

Measurements Measurements
Operator Trial Operator Trial
in coded units in coded units
1 6.693 6.766 6.807 6.714 16 6.632 6.743 6.806 6.746
2 6.594 6.709 6.804 6.814 17 6.626 6.807 6.980 6.747
3 6.657 6.736 6.819 6.797 B 18 6.653 6.751 6.794 6.754
4 6.616 6.709 6.808 6.744 19 6.661 6.807 6.793 6.765
5 6.661 6.807 6.793 6.765 20 6.667 6.684 6.808 6.748
A
6 6.631 6.751 6.794 6.753 21 6.639 6.731 6.838 6.771
7 6.587 6.707 6.805 6.807 22 6.630 6.685 6.803 6.782
8 6.677 6.741 6.808 6.717 23 6.651 6.745 6.809 6.744
9 6.659 6.714 6.834 6.749 24 6.605 6.718 6.839 6.744
10 6.651 6.745 6.809 6.744 25 6.631 6.684 6.831 6.761
C
11 6.631 6.684 6.831 6.761 26 6.639 6.731 6.838 6.771
12 6.637 6.700 6.832 6.747 27 6.657 6.736 6.819 6.797
B 13 6.642 6.783 6.768 6.700 28 6.659 6.714 6.834 6.749
14 6.616 6.709 6.808 6.744 29 6.651 6.745 6.809 6.744
15 6.659 6.714 6.834 6.749 30 6.638 6.744 6.851 6.742
Note: Pls refer to the following Table for the GR&R metrics as established by the Automotive
Industry Action Group (AIAG) guidelines.