Report Example: Basic EMP Study Description of Output

Date: 7/31/2018 Analyzed by: Bill

Gage: My Gage USL: 3
Characteristic: Thickness LSL: -3
Print out of information entered by the user
Operators (o): 3 Process Average:
Parts (p): 10 Process Sigma: 2.5
Trials (n): 3 Meas. Increment: 0.01

Operator-Part Control Charts

X̅ Chart for Operator-Part Averages

2.5 The X̅ chart is a plot of the subgroup averages for

2 the operator-part number combinations. The
1.5
Average

1 first subgroup is made up the results that

0.5 UCL
0 Avg Operator "a" got for part 1. This operator ran
-0.5 LCL this part three different times (the number of
-1
-1.5 trials).
-2
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
The average and control limits are calculated and
a b c added to the chart. The control limits on this
Operator-Part Number chart depend on the average range from the
range chart (see below).
R Chart for Operator-Part Ranges

1.2
1
0.8 UCL
Range

0.6 The R chart is a plot of the range of values within

0.4 Avg each operator-part number subgroup. Each
0.2 range value is a measure of the repeatability of
0
the test method. The average range and control
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
limits are calculated and added to the chart.
a b c
Operator-Part Number

Control Chart Calculations

X̅ Chart X̿ LCL = X̿ - A2R̅ UCL = X̿ + A2R̅
0.001 -0.348 0.351

R Chart R̅ LCL = D3R̅ UCL = D4R̅

The control chart calculations are given.
0.342 - 0.880

where A2, D3, and D4 are control chart constants depending on subgroup size.

A2 D3 D4
1.023 - 2.574

X̅ Chart Analysis
The X̅ chart shows the average value for each operator for each part. The X̅ chart is analyzed. The control limits on
The control limits on the X̅ chart are based on the average range. this chart are based on the average range from
The average range is representative of measurement error. the range chart. This average range represents
The X̅ chart control limits represent the variation obscured by measurement error. measurement variability. If the test method is
good, the measurement variability should be
The relative utility of the measurement system increases: small. So, the average range should be small and
* The more out of control points there on are on the X̅ chart. the control limits should be tight around the
* The further the out of control points are away from the control limits. average. The more out of control points the
better.
22 out of 30 points are out of control on the chart.

R Chart Analysis
The R chart shows the results for the repeated measurements for each operator for each part. The R chart is analyzed. This checks the
It is a check of the consistency of the measurement process between the operators. consistency between the operators. There
should be no out of control points. If there are,
There is 1 out of control point on the R chart; the ranges are not consistent. the reason should be found and eliminated and
The reason for the out of control point should be corrected and the study repeated. the study repeated.

There are 54.7 degrees of freedom associated with the average range.
The study should contain sufficient data (degrees
It is recommended to have at least 10 degrees of freedom.
of freedom). This is checked here.
ANOM Charts for Bias and Repeatability

Page 1
Main Effects (0.05 ANOME) Chart

0.200
0.150
0.100
0.050 UCL
Average

0.000 Avg The Analysis of Main Effects (ANOME) Chart

-0.050 LCL
-0.100 compares the overall averages for the operator.
-0.150 The average for each operator is plotted. The
-0.200
-0.250 overall average is plotted along with the ANOME
-0.300 upper and lower limits on the chart.
a (0.190) b (0.068) c (-0.254)
Operator

Mean Range (0.05 ANOMR) Chart

0.550
0.500
0.450 UCL
0.400 The Analysis of Mean Ranges (ANOMR) Chart
Range

0.350 Avg compares the average range between operators.

0.300
0.250 The average range for each operator is plotted.
LCL The overall average and the ANOMR upper and
0.200
0.150 lower limits are added to the chart.
a (0.184) b (0.513) c (0.328)
Operator

ANOM Calculations
Main Effects X̿ LCL = X̿ - ANOME0.05R̅ UCL = X̿ + ANOME0.05R̅
0.001 -0.070 0.073

Mean Range R̅ LCL = LMR0.05R̅ UCL = UMR0.05R̅

The ANOME and ANOMR calculations are given.
0.342 0.234 0.455

where ANOME, LMR, and UMR are scaling factors that depend on the amount of data.
ANOME0.05 LMR0.05 UMR0.05
0.209 0.685 1.331

Main Effects Chart Analysis

This chart plots the average part values for each operator.
The purpose of the chart is to check for operator bias. The main effects chart is analyzed. There are
Points beyond the control limits are indications that bias exists. differences (bias) in the operators if some of the
points are beyond the limits.
There is evidence of detectable bias between the operators.
Review the ANOME chart for the differences.

Mean Range Chart Analysis

This charts plot the average range values for each operator.
The purpose of the chart is to see if the test-retest error is the same for each operator. The mean range is analyzed. There are
Points beyond the control limits are indications that differences in repeatability exist. differences in repeatability if some of the points
are beyond the limits.
There is evidence of differences in the test-retest error between the operators.
Review the ANOMR chart for the differences.

Repeatability (Test-Retest Error)

d2 pe = R̅ /d2
The repeatability (standard deviation) of the test
1.693 0.20181138
method is determined from the average range.
where d2 is a control chart constant depending on subgroup size.

Probable Error (PE) and Measurement Increment

PE 0.136 Probable Error (0.675pe)
0.2(PE) 0.0272 Smallest Effective Measurement Increment
2(PE) 0.272 Largest Effective Measurement Increment The probable error (PE) is calculated to
determine if the measurement increment is
PE is the minimum medium error of the measurement process. adequate. The measurement increment should
50% of the measurements will fall within +/- one PE. be between 0.2PE and 2PE. If the measurement
PE defines the effective resolution of the measurement process. increment is too small, round up and rerun the
The resolution should be between 0.2(PE) and 2(PE). analysis. If it is too large, look at ways to
decrease the measurement increment.
The measurement increment (0.01) is less than 0.2(PE),
increase the measurement increment so it is between 0.2PE and 2PE.

Variance Components

Page 2
 Component Variance % of Total Estimates: Sigma
Repeatability 0.0407 0.7% pe2 Repeatability (pure error) variance 0.202 The variance components are calculated and the
Reproducibility 0.0514 0.8% o 2 Reproducibility variance 0.227 % of total variance for the various components is
given.
R&R 0.0922 1.5% e2 Combined R&R variance 0.304
Product 6.158 98.5% p 2 Product variance 2.481 If the process sigma was entered, it is used to
Total 6.250 x2 Total variance 2.500 determine the total variance. If not, the parts
used in the study are used to determine the total
Product variance estimated from the process sigma entered. variance.

o2 = so2 - (o/n o p)pe2 where so2 = variance of operator averages. The equations used in the calculations are given.
p2 = x2 - e2

Intraclass Correlation Coefficients

Intraclass Correlation Coefficient (Repeatability) = 0.9934
rpe = p2/(p2+pe2)
Two interclass coefficients are calculated: one
based on the repeatability and one based on
Intraclass Correlation Coefficient (Repeatability & Reproducibility) = 0.9853
both the repeatability and reproducibility
re = p2/(p2+e2)

Type of Class Monitor

Based on Repeatability: This is a First Class Monitor
Based on Repeatability and Reproducibility: This is a First Class Monitor

Type of Reduction of Process Chance of Detecting ± 3 Ability to Track Process

r
Monitor Signala Std. Error Shiftsb Improvementsc

0.8 to 1.0 First Class Less than 10% >99% with Rule 1 Up to Cp80 = 2.22
The type of class monitor is determined for each
intraclass coefficient. The row for the class
0.5 to 0.8 Second Class From 10% to 30% >88% with Rule 1 Up to Cp50 = 3.51
monitor for the intraclass coefficient based on
repeatability is highlighted in yellow. If one or
Third Class >91% with Rules 1, 2, 3, & more specification is entered, the various Cp
0.2 to 0.5 From 30% to 55% Up to Cp20 = 4.439
Monitor 4 values are determined.

Fourth Class
0.0 to 0.2 Greater than 55% Rapidly Vanishes Unable to Track
Monitor
Table reproduced with permission from EMP III Evaluating the Measurement System
by Donald J. Wheeler, Copyright 2006 SPC Press.
a
A signal occurring on a control chart is reduced in strength by 1 - square root of ro.
b
The probability that the measurement process can detect a significant shift.
Rule 1: Point beyond the control limits.
Rule 2: 2 out of 3 consecutive points on the same side of the average are > 1 sigma from the average.
Rule 3: 4 out of 5 consecutive points on the same side of the average are > 2 sigma from the average.
Rule 4: 8 consecutive points on the same side of the average.
c
The process capability where the measurement process will move down to a lower class.

Watershed Specifications1 and Precision to Tolerance Ratio

Watershed USL = 3.005
Watershed LSL = -3.005
Watershed Tol. = 6.01

PE Used to
Precision to Tolerance Precision + Bias to If at least one specification is entered, the
% Mfg. Specs 2 Tighten Mfg. LSL 4 Mfg. USL 4
3 Ratio 5 Tolerance Ratio 6 watershed specifications and precision to
Specs tolerance ratio are determined. The watershed
85.0% 1 -2.86877732 2.86877732 4.53% 6.82%
specifications are used to determine
96.0% 2 -2.73255464 2.73255464 9.07% 13.64%
99.0% 3 -2.59633196 2.59633196 13.60% 20.46% manufacturing specifications that take into
99.9% 4 -2.46010927 2.46010927 18.13% 27.28% account measurement error (probable error).
1
Watershed specification limits take into account the measurement increment. The ratio is given for repeatability only and for
Watershed USL = USL + 0.5(measurement increment) both the repeatability and reproducibility
Watershed LSL = LSL - 0.5(measurement increment)
Watershed Tolerance = Watershed USL - Watershed LSL The precision to tolerance ratio describes the %
2
% Mfg Specs is the probability that an item, with a measured value that of the tolerance that is consumed by the
falls between the Mfg. LSL and Mfg. USL, conforms to specifications. probable error adjustment to the watershed
3
PE Used to Tighten Spec s is the number of PE units used to reduce the watershed specifications. specifications.
4
Mfg. LSL and Mfg. USL are the specifications based on the PE adjustments.
Example: 96%, Mfg. LSL = Watershed LSL + 2(PE) and Mfg. USL = Watershed USL - 2(PE)
5
Precision to Tolerance Ratio is the % of the watershed tolerance consumed

Page 3
 of the tolerance that is consumed by the
probable error adjustment to the watershed
specifications.

by the PE adjustment.
Example: For 96% Mfg. Specs, P/T = 4(PE)/Watershed Tolerance
6
Precision + Bias to Tolerance Ratio is the % of the watershed tolerance consumed
by the PE adjustment using both the repeatability and reproducibility.

Data
Optional Data Table
Run No. Operator Part Result Comment
1 a 1 0.29
31 a 1 0.41
61 a 1 0.64
2 a 2 -0.56
32 a 2 -0.68
62 a 2 -0.58
3 a 3 1.34
33 a 3 1.17
63 a 3 1.27
4 a 4 0.47
34 a 4 0.5
64 a 4 0.64
5 a 5 -0.8
35 a 5 -0.92
65 a 5 -0.84
6 a 6 0.02
36 a 6 -0.11
66 a 6 -0.21
7 a 7 0.59
37 a 7 0.75
67 a 7 0.66
8 a 8 -0.31
38 a 8 -0.2
68 a 8 -0.17
9 a 9 2.26
39 a 9 1.99
69 a 9 2.01
10 a 10 -1.36
40 a 10 -1.25
70 a 10 -1.31
11 b 1 0.08
41 b 1 0.25
71 b 1 0.07
12 b 2 -0.47
42 b 2 -1.22
72 b 2 -0.68
13 b 3 1.19
43 b 3 0.94
73 b 3 1.34
14 b 4 0.01
44 b 4 1.03
74 b 4 0.2
15 b 5 -0.56
45 b 5 -1.2
75 b 5 -1.28
16 b 6 -0.2
46 b 6 0.22
76 b 6 0.06
17 b 7 0.47
47 b 7 0.55
77 b 7 0.83
18 b 8 -0.63
48 b 8 0.08
78 b 8 -0.34
19 b 9 1.8
49 b 9 2.12
79 b 9 2.19

Page 4
20 b 10 -1.68
50 b 10 -1.62
80 b 10 -1.5
21 c 1 0.04
51 c 1 -0.11
81 c 1 -0.15
22 c 2 -1.38
52 c 2 -1.13
82 c 2 -0.96
23 c 3 0.88
53 c 3 1.09
83 c 3 0.67
24 c 4 0.14
54 c 4 0.2
84 c 4 0.11
25 c 5 -1.46
55 c 5 -1.07
85 c 5 -1.45
26 c 6 -0.29
56 c 6 -0.67
86 c 6 -0.49
27 c 7 0.02
57 c 7 0.01
87 c 7 0.21
28 c 8 -0.46
58 c 8 -0.56
88 c 8 -0.49
29 c 9 1.77
59 c 9 1.45
89 c 9 1.87
30 c 10 -1.49
60 c 10 -1.77
90 c 10 -2.16

Page 5