6.

1 Table of Hole Diameter Data

sample Carbon-fiber composite material

Number 1 2 3 4 5 Ri
1 -30 50 -20 10 30 8 80
2 0 50 -60 -20 30 0 110
3 -50 10 20 30 20 6 80
4 -10 -10 30 -20 50 8 70
5 20 -40 50 20 10 12 90
6 0 0 40 -40 20 4 80
7 0 0 20 -20 -10 -2 40
8 70 -30 30 -10 0 12 100
9 0 0 20 -20 10 2 40
10 10 20 30 10 50 24 40
11 40 0 20 0 20 16 40
12 30 20 30 10 40 26 30
13 30 -30 0 10 10 4 60
14 30 -10 50 -10 -30 6 80
15 10 -10 50 40 0 18 60
16 0 0 30 -10 0 4 40
17 20 20 30 30 -20 16 50
18 10 -20 50 30 10 16 70
19 50 -10 40 20 0 20 60
20 50 0 0 30 10 18 50
and 218 1270
and 10.9 63.5

(a) set up and R charts on the process. Is the process in statistical control?
From the problem, it is known that n =5, therefore we use Appendix table VI that D 3=0 and D4=2.114.
LC = =63.5
LCL = = 63.5 (0) = 0

UCL = = 63.5 (2.114) =134.239

For chart, it is used A2= 0.577, therefore
UCL = +A2 = 10.9 + 0.577 (63.5) = 47.5395

LC = =10.9
LCL = -A2 = 10.9 - 0.577 (63.5) = -25.7395
 R Chart of C1; ...; C5
140
UCL=134.3

120

100

Sample Range
80
_
R=63.5
60

40

20

0 LCL=0

1 3 5 7 9 11 13 15 17 19
Sample

6.1.1 R chart
Xbar Chart of C1; ...; C5
50
UCL=47.53

40

30
Sample Mean

20
_
_
10 X=10.9

-10

-20
LCL=-25.73
-30
1 3 5 7 9 11 13 15 17 19
Sample

6.1.2 chart

From the chart and R chart, there aren’t no indication of out-of-control condition. Since they both
exhibit control, we would conclude that the process in in control.
(b) Estimate the process standard deviation using the range method

= = 27.367

(c) if specification are at nominal ±100, what can you say about the capability of this process? Calculate
the PCR Cp.

in ten-thousandths of an inch = 10.9 x 10,000 = 109000 inch

USL and LSL= 109,100 and 108,900 or we can say 10.91 and 10.89 (in ten-thousandths of inch).
The process capability:
 = = = = 1.218

= x100% = x 100% = 0.821 x 100% ≈ 82%.

The process uses up about 82% of the specification band.

6.7

number
of
sample x1 x2 x3 x4 x5 x6 x7 x8 x9 x10
1 2.5 0.5 2 -1 1 -1 0.5 1.5 0.5 -1.5 0.5 1.333333
2 0 0 0.5 1 1.5 1 -1 1 1.5 -1 0.45 0.926463
3 1.5 1 1 -1 0 -1.5 -1 -1 1 -1 -0.1 1.125463
4 0 0.5 -2 0 -1 1.5 -1.5 0 -2 -1.5 -0.6 1.173788
5 0 0 0 -0.5 0.5 1 -0.5 -0.5 0 0 0 0.471405
6 1 -0.5 0 0 0 0.5 -1 1 -2 1 0 0.971825
7 1 -1 -1 -1 0 1.5 0 1 0 0 0.05 0.895979
8 0 -1.5 -0.5 1.5 0 0 0 -1 0.5 -0.5 -0.15 0.818196
9 -2 -1.5 1.5 1.5 0 0 0.5 1 0 1 0.2 1.183216
10 -0.5 3.5 0 -1 -1.5 -1.5 -1 -1 1 0.5 -0.15 1.528434
11 0 1.5 0 0 2 -1.5 0.5 -0.5 2 -1 0.3 1.206464
12 0 -2 -0.5 0 -0.5 2 1.5 0 0.5 -1 0 1.154701
13 -1 -0.5 -0.5 -1 0 0.5 0.5 -1.5 -1 -1 -0.55 0.68516
14 0.5 1 -1 -0.5 -2 -1 -1.5 0 1.5 1.5 -0.15 1.248332
15 1 0 1.5 1.5 1 -1 0 1 -2 -1.5 0.15 1.270389
-0.05 15.99315
-0.00333 1.06621

(a) Set up and s control charts on this process. Does the process exhibit statistical control?
If necessary, construct revised control limits.
 The parameters for s chart can be written as:
UCL = B4 = 1.716 x 1.06621= 1.829616
Center line = = 1.06621
LCL = B3 = 0.284 x 1.06621 = 0.302804
 The parameters for chart can be written as:

UCL = = = 1.036551

Center line = = -0.00333

 LCL = = = -1.04322

S Chart of 1; ...; 10
2.0

UCL=1.830

1.5
Sample StDev

_
S=1.066
1.0

0.5

LCL=0.302

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Sample

6.7.1. s chart

Xbar Chart of 1; ...; 10

1.0 UCL=1.040

0.5
Sample Mean

_
_
0.0 X=-0.003

-0.5

-1.0 LCL=-1.047

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Sample

6.7.2 chart

The process exhibit statistical control.

(b) Set up an R chart, and compare with the s chart in part (a)
 number of
x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 R
sample
1 2.5 0.5 2 -1 1 -1 0.5 1.5 0.5 -1.5 4
2 0 0 0.5 1 1.5 1 -1 1 1.5 -1 2.5
3 1.5 1 1 -1 0 -1.5 -1 -1 1 -1 3
4 0 0.5 -2 0 -1 1.5 -1.5 0 -2 -1.5 3.5
5 0 0 0 -0.5 0.5 1 -0.5 -0.5 0 0 1.5
6 1 -0.5 0 0 0 0.5 -1 1 -2 1 3
7 1 -1 -1 -1 0 1.5 0 1 0 0 2.5
8 0 -1.5 -0.5 1.5 0 0 0 -1 0.5 -0.5 3
9 -2 -1.5 1.5 1.5 0 0 0.5 1 0 1 3.5
10 -0.5 3.5 0 -1 -1.5 -1.5 -1 -1 1 0.5 5
11 0 1.5 0 0 2 -1.5 0.5 -0.5 2 -1 3.5
12 0 -2 -0.5 0 -0.5 2 1.5 0 0.5 -1 4
13 -1 -0.5 -0.5 -1 0 0.5 0.5 -1.5 -1 -1 2
14 0.5 1 -1 -0.5 -2 -1 -1.5 0 1.5 1.5 3.5
15 1 0 1.5 1.5 1 -1 0 1 -2 -1.5 3.5

48
3.2

n =10, therefore we use Appendix table VI that D3=0.223 and D4=1.777. =3.2
LCL = = 3.2 (0.223) = 0.7136
UCL = = 3.2 (1.777) = 5.6864

R Chart of 1; ...; 10
6
UCL=5.686

4
Sample Range

_
R=3.2
3

1
LCL=0.714

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Sample

6.7.3 R chart
6.13

sample
x1 x2 x3 x4 x5 R
number
1 83 81.2 78.7 75.7 77 79.1 7.3
 2 88.6 78.3 78.8 71 84.2 80.2 17.6
3 85.7 75.8 84.3 75.2 81 80.4 10.5
4 80.8 74.4 82.5 74.1 75.7 77.5 8.4
5 83.4 78.4 82.6 78.2 78.9 80.3 5.2
6 75.3 79.9 87.3 89.7 81.8 82.8 14.4
7 74.5 78 80.8 73.4 79.7 77.3 7.4
8 79.2 84.4 81.5 86 74.5 81.1 11.5
9 80.5 86.2 76.2 64.1 80.2 77.4 22.1
10 75.7 75.2 71.1 82.1 74.3 75.7 11
11 80 81.5 78.4 73.8 78.1 78.4 7.7
12 80.6 81.8 79.3 73.8 81.7 79.4 8
13 82.7 81.3 79.1 82 79.5 80.9 3.6
14 79.2 74.9 78.6 77.7 75.3 77.1 4.3
15 85.5 82.1 82.8 73.4 71.7 79.1 13.8
16 78.8 79.6 80.2 79.1 80.8 79.7 2
17 82.1 78.2 75.5 78.2 82.1 79.2 6.6
18 84.5 76.9 83.5 81.2 79.2 81.1 7.6
19 79 77.8 81.2 84.4 81.6 80.8 6.6
20 84.5 73.1 78.6 78.7 80.6 79.1 11.4
and 1586.7 187.0
and 79.3 9.4

(a) Establish and R control charts for compressive strength using these data. Is the process in
statistical control?
 For R chart (use D3 = 0 and D4= 2.114 from Apendix table VI because n=5):
UCL = = 9.4 (2.114) =19.8716
Center line = = 9.4
LCL = = 9.4 (0) = 0
 For chart (use A2 =0.577 from Apendix table VI because n=5):
UCL = +A2 = 79.3+ 0.577 (9.4) = 84.7568
Center line = =79.3
LCL = -A2 = 79.3 - 0.577 (9.4) = 73.9092
Using minitab we can see the the R Chart and chart as below:
 R Chart of C1; ...; C5
25
1

20 UCL=19.77

Sample Range
15

_
10
R=9.35

0 LCL=0

1 3 5 7 9 11 13 15 17 19
Sample

6.13.1 R Chart for 20 subgroups

Xbar Chart of C1; ...; C5

85.0 UCL=84.97

82.5
Sample Mean

80.0 _
_
X=79.33

77.5

75.0

LCL=73.69

1 3 5 7 9 11 13 15 17 19
Sample

6.13.2 Chart for 20 subgroups

From the R Chart shown above, there is an out-of-control condition observed. The value which is
out-of-control from UCL is R= 22.1 whereas UCL is 19.8716. however, the chart exhibit control.

(b) After establishing the control charts in part (a), 15 new subgroups were collected and the
compressive strength are shown in table 6E.8. Plot the and R values against the control units
from part (a) and draw conclusions.

sample
number x1 x2 x3 x4 x5 xbar R
1 83 81.2 78.7 75.7 77 79.12 7.3
2 88.6 78.3 78.8 71 84.2 80.18 17.6
3 85.7 75.8 84.3 75.2 81 80.4 10.5
4 80.8 74.4 82.5 74.1 75.7 77.5 8.4
5 83.4 78.4 82.6 78.2 78.9 80.3 5.2
6 75.3 79.9 87.3 89.7 81.8 82.8 14.4
7 74.5 78 80.8 73.4 79.7 77.28 7.4
 8 79.2 84.4 81.5 86 74.5 81.12 11.5
9 80.5 86.2 76.2 64.1 80.2 77.44 22.1
10 75.7 75.2 71.1 82.1 74.3 75.68 11
11 80 81.5 78.4 73.8 78.1 78.36 7.7
12 80.6 81.8 79.3 73.8 81.7 79.44 8
13 82.7 81.3 79.1 82 79.5 80.92 3.6
14 79.2 74.9 78.6 77.7 75.3 77.14 4.3
15 85.5 82.1 82.8 73.4 71.7 79.1 13.8
16 78.8 79.6 80.2 79.1 80.8 79.7 2
17 82.1 78.2 75.5 78.2 82.1 79.22 6.6
18 84.5 76.9 83.5 81.2 79.2 81.06 7.6
19 79 77.8 81.2 84.4 81.6 80.8 6.6
20 84.5 73.1 78.6 78.7 80.6 79.1 11.4
21 68.9 81.5 78.2 80.8 81.5 78.18 12.6
22 69.8 68.6 80.4 84.3 83.9 77.4 15.7
23 78.5 85.2 78.4 80.3 81.7 80.82 6.8
24 76.9 86.1 86.9 94.4 83.9 85.64 17.5
25 93.6 81.6 87.8 79.6 71 82.72 22.6
26 65.5 86.8 72.4 82.6 71.4 75.74 21.3
27 78.1 65.7 83.7 93.7 93.4 82.92 28
28 74.9 72.6 81.6 87.2 72.7 77.8 14.6
29 78.1 77.1 67 75.7 76.8 74.94 11.1
30 78.7 85.4 77.7 90.7 76.7 81.84 14
31 85 60.2 68.5 71.1 82.4 73.44 24.8
32 86.4 79.2 79.8 86 75.4 81.36 11
33 78.5 99 78.3 71.4 81.8 81.8 27.6
34 68.8 62 82 77.5 76.1 73.28 20
35 83 83.7 73.1 82.2 95.3 83.46 22.2
and 2778 456.8
and 79.37143 13.05143

 For R chart (use D3 = 0 and D4= 2.114 from Apendix table VI because n=5):
UCL = = 13.05143 (2.114) = 27.60
Center line = = 13.05143
LCL = = 13.05143 (0) = 0
 For chart (use A2 =0.577 from Apendix table VI because n=5):
UCL = +A2 = 79.37143 + 0.577 (13.05143) = 86.90
Center line = =79.37143
LCL = -A2 = 79.37143 - 0.577 (13.05143) = 71.84
 Using minitab we can see the the R Chart and chart as below:

R Chart of 1; ...; 5
30
1 1
UCL=27.60
25

20

Sample Range
15 _
R=13.05

10

0 LCL=0

1 4 7 10 13 16 19 22 25 28 31 34
Sample

6.13.3 R Chart for 35 subgroups

Xbar Chart of 1; ...; 5

88
UCL=86.90
86

84

82
Sample Mean

_
_
80
X=79.37
78

76

74

72 LCL=71.84

70
1 4 7 10 13 16 19 22 25 28 31 34
Sample

6.13.4 Chart for 35 subgroups

Conclusions:

6.19 Control charts for and R are maintained for an important quality characteristic. The sample size is
n= 7; and R are computed for each sample. After 35 samples, we have found that:

=7805 and =1200

(a) Set up and R charts using these data

Control limits for the R Chart (since n=7, the value of D3 is 0.076 and D4 is 1.924 ):

UCL = D4 = 1.924 x 34.286 =65.966

Center line = = = = 34.286

LCL = D3 = 0.076 x 34.286 =2.605

Control limits for the Chart (since n=7, the value of A2=0.419 and is = 223):

UCL = +A2 = 223 + 0.419 (34.286) =237.3658

Center line = =223

LCL = -A2 = 223 - 0.419 (34.286) = 208.6342

(b) Assuming that both charts exhibit control, estimate the process mean and standard deviation

The process mean = µ ≈ = 223

The estimator of is = = = 12.679

(c) If the quality characteristic is normally distributed and if the spesifications are 220± 35, can the
process meet the specifications? Estimate the fraction nonconforming.

USL = 220 +35 = 255 and LSL = 220-35=185

P= P{x<185} + P{x>255}

=Φ +1- Φ

= Φ(-2.996) +1 – Φ(2.5237)

= 0.001368 + 1 - 0.994194 = 0.007174

It is about 0.7174 percent [7174 parts per millon(ppm)] of the products produced will be
aoutside of the specifications.

(d) Assuming the variance to remain constant, state where the process mean should be located to
minimize the fraction nonconforming. What would be the value of fraction nonconforming
under these conditions?
6.25 Suppose that the following the construction of the and R control charts in exercise 6.23, the
process engineers decided to change the subgroup size to n = 2. Table 6E.11 contains 10 new
subgroups of tickness data. Plot this data on the control charts from exercise 6.23 (a) based on the
new subgroup size. Is the process in statistical control?

subgroup x1 x2 x3 x4 x bar R
1 459 449 435 450 448.25 24
2 443 440 442 442 441.75 3
3 457 444 449 444 448.5 13
4 469 463 453 438 455.75 31
5 443 457 445 454 449.75 14
6 444 456 456 457 453.25 13
7 445 449 450 445 447.25 5
8 446 455 449 452 450.5 9
9 444 452 457 440 448.25 17
10 432 463 463 443 450.25 31
11 445 452 453 438 447 15
12 456 457 436 457 451.5 21
13 459 445 441 447 448 18
14 441 465 438 450 448.5 27
15 460 453 457 438 452 22
16 453 444 451 435 445.75 18
17 451 460 450 457 454.5 10
18 422 431 437 429 429.75 15
19 444 446 448 467 451.25 23
20 450 450 454 454 452 4
and 8973.75 333
and 448.6875 16.65

From exercise 6.23, we know that

n old =4 , = 16.65

and from Appendix Table VI we have

d2 (old) =2.059 d2 (new)= 1.128,

therefore, the new control limits on the chart are found :

UCL = = 448.69 + 1.88(0.548)16.65= 465.84

 LCL = = 448.69 - 1.88(0.548)16.65= 431.54

For the R chart, the new parameters are :

UCL = D4 = 3.267 (0.548)16.65 = 29.808

CL = = = 9.1242

LCL = max {0, D3 }= 0

From here, we can make the chart to se whether the process in statistical control or not.

6.31.Specifications on a cigar lighter detent are 0.3220 and 0.3200 in. Samples of size 5 are taken every
45 min with the results shown in Table 6E.13 (measured as deviations from 0.3210 in 0.0001 in).

(a) Set up an R chart and examine the process for statistical control.

sample
x1 x2 x3 x4 x5 x bar R
number
1 1 9 6 9 6 6.2 8
2 9 4 3 0 3 3.8 9
3 0 9 0 3 2 2.8 9
4 1 1 0 2 1 1 2
5 -3 0 -1 0 -4 -1.6 4
6 -7 2 0 0 2 -0.6 9
7 -3 -1 -1 0 -2 -1.4 3
8 0 -2 -3 -3 -2 -2 3
9 2 0 -1 -3 -1 -0.6 5
10 0 2 -1 -1 2 0.4 3
11 -3 -2 -1 -1 2 -1 5
12 -16 2 0 -4 -1 -3.8 18
13 -6 -3 0 0 -8 -3.4 8
14 -3 -5 5 0 5 0.4 10
15 -1 -1 -1 -2 -1 -1.2 1
and -1 97
and -0.06667 6.466667
Control limits for the R Chart (since n=5, the value of D3 is 0 and D4 is 2.114 ):

UCL = D4 = 2.114 x 6.47 =13.67

 Center line = = = 6.47

LCL = D3 = 0 x 6.47 = 0

R Chart of C1; ...; C5

20
1

15
UCL=13.67
Sample Range

10

_
R=6.47
5

0 LCL=0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Sample

From the picture, there is a sample that is out of control (sample 12). Then, we can eliminate the
extrim data and recalculate the parameter.

(c) What parameters would you recommend for an R chart for on-line control?

The parameters recommended in this case are:

UCL = D4 new= 2.114 x5.643 =11.929

Center line = new= = 5.643

LCL = D3 new= 0 x5.643 = 0

The chart is given below:

 R Chart of C1; ...; C5

12 UCL=11.93

10

8
Sample Range

6 _
R=5.64

0 LCL=0

1 2 3 4 5 6 7 8 9 10 11 12 13 14
Sample

The R chart above indicate that there’s no out-of-control sample. Therefore, since the R chart
exhibit control, we would conclude that the parameters we would reccomend are UCL = 11.929,
Center line = 5.643 and LCL = 0

(d) Estimate the standard deviation of the process.

= = 2.426

(e) What is the process capability?

Cp= = = 1.374

P= = = 72.78%

6.37 Thirty samples each of size 7 have been collected to establish control over a process. The following

data were collected: and

(a) calculate trial control limits for the two charts

Trial control limit for chart:

UCL = +A2 = 90 +0.419(4)=91.676

CL = = =90
 LCL = -A2 = 90 - 0.419(4)=88.324

Trial control limit for R chart:

UCL = D4 = 1.92 4(4)=7.696

CL= = =4

LCL = = D3 =0.076(4) = 0.304

(b) On the assumption that the R chart is in control, estimate the process standard deviation.

= = 1.479

(c) Suppose an s chart were desired. What would be the appropriate control limits and center line?

Suppose that is given above

UCL = B6 =1.806 (1.479)= 2.671

CL = = 1.479

LCL =B5 = 0.113 (1.479)= 0.167

6.43. and R charts with n=4 and the control parameters below both exhibit control.

Chart R Chart
UCL = 815 UCL=46.98
Center Line = 800 Center Line=20.59
LCL= 785 LCL=0
What is the probability that a shift in the process mean to 790 will be detected in the first sample
following the shift?

First, calculate the probability of not detecting this shift that is:

Since N(µ,
̴ /n), then :

We suppose that L=3 (the usual three-sigma limits), k=(µ 1-µ0)/σ with σ can be estimated by =

= =10 →k=(790-800)/10= -1
 then we have:

= (5)- (-1)

= 1- 0.158655 = 0.841344

Last, the probability that such a shift will be detected in the first subsequent is 1-β= 1- 0.841344
= 0.158655

6.49 The following and s charts based on n=4 have shown statistical control:

Chart S Chart
UCL = 710 UCL=18.08
Center Line = 700 Center Line=7.979
LCL= 690 LCL=0
(a) Estimate the process parameters µ and σ

For µ, the estimate parameter is = 700

For σ, the estimate parameter is =8.66

(b) If the specifications are at 705 ± 15, and the process output is normally distributed, estimate the
fraction nonconforming.

= = = 0.577

= = =1.733x100% =173.3%. the process uses up more than 100%

of the tolerance band. In this case the process is very yield-sensitive, and a large number of
nonconforming units will be produced.

(c) For the Chart, find the probability of type I error, assuming σ is constant

(d) Suppose the process mean shifts to 693 and the standard deviation simultaneously shifts to 12.
Find the probability of detecting this shift on the Chart on the first subsequent sample.

(e) For the shift of part(d), find the average run length.

6.55. Michelson actually made 100 measurements on the velocity of light in five trials of 20 observations
each. The second set of 20 measurements is shown in table 6E.20.
 (a) plot these new measurement on the control charts constructed in exercise 6.56. are these new
measurements in statistical control? Give a practical interpretation of the control charts.

(b) is there evidence that the variability in the measurement has decreased between trial 1 and trial
2?

6.61.The vane heights for 20 of the castings from Fig. 6.25 are shown in table 6E.23. Construct the
“between/within “ control charts for these process data using a range chart to monitor the within-
castings vane height. Compare these to the control charts shown in fig.6.27

7.1 The data in table 7E.1 represent the results of inspecting all units of a personal computer produced
for the past ten days. Does the process appear to be in control?

non fraction
units
day conforming non- 1- stdev LCL UCL
inspected
units conforming

1 80 4 0.05 0.06 0.94 0.026552 -0.01966 0.139656

2 110 7 0.064 0.06 0.94 0.022643 -0.00793 0.12793
3 90 5 0.056 0.06 0.94 0.025033 -0.0151 0.1351
4 75 8 0.107 0.06 0.94 0.027423 -0.02227 0.142268
5 130 6 0.046 0.06 0.94 0.020829 -0.00249 0.122487
6 120 6 0.05 0.06 0.94 0.021679 -0.00504 0.125038
7 70 4 0.057 0.06 0.94 0.028385 -0.02516 0.145155
8 125 5 0.04 0.06 0.94 0.021241 -0.00372 0.123724
9 105 8 0.076 0.06 0.94 0.023176 -0.00953 0.129529
10 95 7 0.074 0.06 0.94 0.024366 -0.0131 0.133097
Total : 1000 Total: 60 Total: 0.62 0.132398
Because UCL and LCL can’t be negative, then the value of LCL = 0. From the calculation above, we
can se from the p-chart below that there’s no indication of out-of-control condition, therefore the
process appear to be in control.
 P Chart of nonconforming unit
0.16

0.14
UCL=0.1331
0.12

0.10

Proportion
0.08
_
0.06 P=0.06

0.04

0.02

0.00 LCL=0

1 2 3 4 5 6 7 8 9 10
Sample
Tests performed with unequal sample sizes

7.7. A control chart indicates that the current process fraction nonconforming is 0.02. If fifty items are
inspected each day, what is the probability of defecting a shift in the fraction nonconforming to
0.04 on the first day after the shift? By the end of the third day following the shift?

From the problem above it is known that , n=50,

P{D=x} =

7.13.

(a) set up a control chart for the number nonconforming in samples of n=100

sample fraction
sample sample number of
number size nonconforming nonconforming
1 100 10 0.1
2 100 15 0.15
3 100 31 0.31
4 100 18 0.18
5 100 24 0.24
6 100 12 0.12
7 100 23 0.23
8 100 15 0.15
9 100 8 0.08
10 100 8 0.08
Total: 164 = 0.164
  The np Control Chart:

= 0.164, n=100

=n +3 = 100(0.164) +3 = 27.508 ~28

UCL

Center line = n = 100 x 0.164=16.4

LCL= n -3 = 100(0.164) -3 = 5.292 ~ 5

(b) For the chart established in part (a), what is the probability of detecting a shift in the
process fraction nonconforming to 0.30 on the first sample after the shift has occured?

δ= the magnitude of the process shift = 0.3 – 0.164 =0.136

L= 3 (if it is assumed that three-sigma limits are used)

n = 100

7.19 A fraction nonconforming control chart has center line 0.01, UCL=0.0399, LCL=0, and n=100. If
three-sigma limits are used, find the smallest sample size that would yield a positive lower control
limit.

The sample size must be

= 891

Thus, if n ≥ 892 units, the control chart will have a positive lower control limit.

7.25 A fraction nonconforming control chart is to be established with a center line of 0.01 and two-
sigma control limits.

(a) How large should the sample size be if the lower control limit is to be nonzero?

The sample size must be

= 396

Thus, if n ≥ 397 units, the control chart will have a positive lower control limit.

(b) how large should the sample size be if we wish the probability of detecting a shift to 0.04 to be
0.5?
 → n= = = 1.63 ≈2

7.31.Consider an np chart with k-sigma control limits. Derive a general formula for determining the
minimum sample size to ensure that the chart has a positive lower control limit.

Based on the LCL for fraction nonconforming, the minimum sample size to ensure that the chart
has a positive lower control limit can be obtained:

LCL = p - L >0

, therefore the LCL for np chart should be:

LCL = np - L >0

7.37 Consider the data in Exercise 7.35. Suppose a new inspection unit is defined as 2500 m of wire.

Based on exercise 7.35 we can get the information below:

sample number of sample number of

number nonconformities number nonconformities
1 1 12 6
2 1 13 9
3 3 14 11
4 7 15 15
5 8 16 8
6 10 17 3
7 5 18 6
8 13 19 7
9 0 20 4
10 19 21 9
11 24 22 20
Total: 22 Total : 189

= = 8.59
 C Chart of C1
25 1

1
20 1

UCL=17.38

Sample Count
15

10 _
C=8.59

0 LCL=0

1 3 5 7 9 11 13 15 17 19 21
Sample

(a) What are the center line and control limits for a control chart for monitoring future production
based on the total number of nonconformities in the new inspection unit?

The center line = = 8.59

UCL = +3 = 8.59 +3 =17.38

LCL = - 3 = 8.59 -3 = - 0.2, because these calculations yield a negative value for the
LCL, then set LCL=0.

(b) What are the center line and control limits for a control chart for average nonconformities per
unit used to monitor future production?

sample number of sample

average
number nonconformities size
1 1 2500 0.0004
2 1 2500 0.0004
3 3 2500 0.0012
4 7 2500 0.0028
5 8 2500 0.0032
6 10 2500 0.004
7 5 2500 0.002
8 13 2500 0.0052
9 0 2500 0
10 19 2500 0.0076
11 24 2500 0.0096
12 6 2500 0.0024
13 9 2500 0.0036
14 11 2500 0.0044
15 15 2500 0.006
16 8 2500 0.0032
 17 3 2500 0.0012
18 6 2500 0.0024
19 7 2500 0.0028
20 4 2500 0.0016
21 9 2500 0.0036
22 20 2500 0.008
189 0.0756

= = =0.0034, therefore, the parameters for the control chart:

The center line = =0.0034

UCL = +3 = 0.0034 +3 =0.0069

LCL = -3 = 0.0034 -3 ≈0

7.43. Find 0.900 and 0.100 probability limits for a c chart when the process average is equal to sixteen
nonconformities.

7.49. A textile mill wishes to establish a control procedure on flaws in towels it manufactures. Using an
inspection unit of 50 units, past inspection data show that 100 previous inspection units had 850
total flaws. What type of control chart is appropriate? Design the control chart such that it has two-
sided probability control limits of α=0.06, approximately. Give the center line and control limits.

The appropriate control chart for this case is control charts for nonconformities because each
product contain more than one nonconformities. We can use the u Chart

From the problem above, it is known that:

n=50, then the past inspection data shows 100 sample number, number of nonconformities= 850,
α=0.06→β= 1-0.06 = 0.94

= =8.5

7.55. A production line assembles electric clocks. The average number of nonconformities per clock is
estimated to be 0.75. The quality engineer wishes to establish a c chart for this operation, using an
inspection unit of six clocks. Find the three-sigma limits for this chart.

It is known that =0.75

 Inspection unit= 6

UCL = +3 = 0.75+3 = 3.348

LCL = - 3 = 0.75-3 =-1.848, because these calculations yield a negative value for the LCL,
then set LCL=0.

7.61.(a) Set up a c chart for the total number of errors. Is the process in control?

Number of
day record 1 record 2 record 3 record 4 record 5
nonconformities
1 8 7 1 11 17 44
2 11 1 11 2 9 34
3 1 1 8 2 5 17
4 3 2 5 1 4 15
5 3 2 13 6 5 29
6 6 3 3 3 1 16
7 8 8 2 1 5 24
8 4 10 2 6 4 26
9 1 6 1 3 2 13
10 15 1 3 2 8 29
11 1 7 13 5 1 27
12 6 7 9 3 1 26
13 7 6 3 3 1 20
14 2 9 3 8 7 29
15 6 14 7 1 8 36
16 2 9 4 2 1 18
17 11 1 1 3 2 18
18 5 5 19 1 3 33
19 6 15 5 6 6 38
20 2 7 9 2 8 28
21 7 5 6 14 10 42
22 4 3 8 1 2 18
23 4 1 4 20 5 34
24 15 2 7 10 17 51
25 2 15 3 11 2 33
Total : 698

= =27.92

UCL = +3 = 27.92 +3 = 43.77

LCL = - 3 = 27.92 -3 = 12.068

 C Chart of C1
1
50

1
UCL=43.77
40
Sample Count

30 _
C=27.92

20

LCL=12.07
10
1 3 5 7 9 11 13 15 17 19 21 23 25
Sample

From the c chart above, it is indicate that the process is out of control because sample 24 is plotted
outside the UCL.

(c) Set up a t chart for the total number of errors, assuming a geometric distribution with a=1. Is
the process in control?

(d) discuss the findings from parts (a) and (b). Is the poisson distribution a good model for the
customer error data? Is there evidence of this in the data?

9.1. A machine is used to fill cans with motor oil additive. A single sample can is selected every hour
and the wight of the can is obtained. Since the filling process is automated, it has very stable
variability, and long experience indicates that σ=0.05 oz. The individual observations for 24 hours
of operation are shown in table 9E.1

(a) Assuming that the process target is 8.02 oz, set up tabular cusum for this process. Design the
cusum using the standardized values h=4.77 and k=1/2.

µ0= 8.02, n= 1, σ=0.05, h=4.77 , k=1/2

=max , with =0

=max , with =0

Calculation (for i=1, for the rest is ilustrated from the table):

=max

= max

= max
 =0

=max

= max

= max

=0

a b
sample
x
number
xi-8.52 Ci+ N+ 7.52-xi Ci- N-

1 8 -0.52 0 0 -0.48 0 0
2 8.01 -0.51 0 0 -0.49 0 0
3 8.02 -0.5 0 0 -0.5 0 0
4 8.01 -0.51 0 0 -0.49 0 0
5 8 -0.52 0 0 -0.48 0 0
6 8.01 -0.51 0 0 -0.49 0 0
7 8.06 -0.46 0 0 -0.54 0 0
8 8.07 -0.45 0 0 -0.55 0 0
9 8.01 -0.51 0 0 -0.49 0 0
10 8.04 -0.48 0 0 -0.52 0 0
11 8.02 -0.5 0 0 -0.5 0 0
12 8.01 -0.51 0 0 -0.49 0 0
13 8.05 -0.47 0 0 -0.53 0 0
14 8.04 -0.48 0 0 -0.52 0 0
15 8.03 -0.49 0 0 -0.51 0 0
16 8.05 -0.47 0 0 -0.53 0 0
17 8.06 -0.46 0 0 -0.54 0 0
18 8.04 -0.48 0 0 -0.52 0 0
19 8.05 -0.47 0 0 -0.53 0 0
20 8.06 -0.46 0 0 -0.54 0 0
21 8.04 -0.48 0 0 -0.52 0 0
22 8.02 -0.5 0 0 -0.5 0 0
23 8.03 -0.49 0 0 -0.51 0 0
24 8.05 -0.47 0 0 -0.53 0 0

(b) Does the value of σ=0.05 seem reasonable for this process?

If σ=0.05 , then H= 5(0.05) = 0.25 It seems to be reasonable because there’s no out-of-control

process.
9.7. Set up a tabular cusum scheme for the flow width data used in example 6.1 (see Tables 6.1 and 6.2).
when the procedure is applied to all 45 samples, does the cusum react more quickly than the
chart to the shift in the process mean? Use σ=0.14 in setting up the cusum, and design the
procedure to quickly detect a shift of about 1σ.

sample wafers
number 1 2 3 4 5 xbar=xi
1 1.3235 1.4128 1.6744 1.4573 1.6914 1.5119
2 1.4314 1.3592 1.6075 1.4666 1.6109 1.4951
3 1.4284 1.4871 1.4932 1.4324 1.5674 1.4817
4 1.5028 1.6352 1.3841 1.2831 1.5507 1.4712
5 1.5604 1.2735 1.5265 1.4363 1.6441 1.4882
6 1.5955 1.5451 1.3574 1.3281 1.4198 1.4492
7 1.6274 1.5064 1.8366 1.4177 1.5144 1.5805
8 1.419 1.4303 1.6637 1.6067 1.5519 1.5343
9 1.3884 1.7277 1.5355 1.5176 1.3688 1.5076
10 1.4039 1.6697 1.5089 1.4627 1.522 1.5134
11 1.4158 1.7667 1.4278 1.5928 1.4181 1.5242
12 1.5821 1.3355 1.5777 1.3908 1.7559 1.5284
13 1.2856 1.4106 1.4447 1.6398 1.1928 1.3947
14 1.4951 1.4036 1.5893 1.6458 1.4969 1.5261
15 1.3589 1.2863 1.5996 1.2497 1.5471 1.4083
16 1.5747 1.5301 1.5171 1.1839 1.8662 1.5344
17 1.368 1.7269 1.3957 1.5014 1.4449 1.4874
18 1.4163 1.3864 1.3057 1.621 1.5573 1.4573
19 1.5796 1.4185 1.6541 1.5116 1.7247 1.5777
20 1.7106 1.4412 1.2361 1.382 1.7601 1.5060
21 1.4371 1.5051 1.3485 1.567 1.488 1.4691
22 1.4738 1.5936 1.6583 1.4973 1.472 1.5390
23 1.5917 1.4333 1.5551 1.5295 1.6866 1.5592
24 1.6399 1.5243 1.5705 1.5563 1.553 1.5688
25 1.5797 1.3663 1.624 1.3732 1.6887 1.5264
26 1.4483 1.5458 1.4538 1.4303 1.6206 1.4998
27 1.5435 1.6899 1.583 1.3358 1.4187 1.5142
28 1.5175 1.3446 1.4723 1.6657 1.6661 1.5332
29 1.5454 1.0931 1.4072 1.5039 1.5264 1.4152
30 1.4418 1.5059 1.5124 1.462 1.6263 1.5097
31 1.4301 1.2725 1.5945 1.5397 1.5252 1.4724
32 1.4981 1.4506 1.6174 1.5837 1.4962 1.5292
33 1.3009 1.506 1.6231 1.5831 1.6454 1.5317
34 1.4132 1.4603 1.5808 1.7111 1.7313 1.5793
35 1.3817 1.3135 1.4953 1.4894 1.4596 1.4279
36 1.5765 1.7014 1.4026 1.2773 1.4541 1.4824
 37 1.4936 1.4373 1.5139 1.4808 1.5293 1.4910
38 1.5729 1.6738 1.5048 1.5651 1.7473 1.6128
39 1.8089 1.5513 1.825 1.4389 1.6558 1.6560
40 1.6236 1.5393 1.6738 1.8698 1.5036 1.6420
41 1.412 1.7931 1.7345 1.6391 1.7791 1.6716
42 1.7372 1.5663 1.491 1.7809 1.5504 1.6252
43 1.5971 1.7394 1.6832 1.6677 1.7974 1.6970
44 1.4295 1.6536 1.9134 1.7272 1.437 1.6321
45 1.6217 1.822 1.7915 1.6744 1.9404 1.7700

9.13.Consider the velocity of light data introduced in exercises 6.56 and 6.55. use only the 20
obseervations in Exercise 6.56 to set up a cusum with target value 734.5. plot all 40 observations
from both exercises 6.56 and 6.55 on this cusum. What conclusions can you draw?
Cusum with target value 734.5 for 20 observations:

sample a b
x
number xi-735 Ci+ N+ 734-xi Ci- N-
1 850 115 115 1 -116 0 0
2 1000 265 380 2 -266 0 0
3 740 5 385 3 -6 0 0
4 980 245 630 4 -246 0 0
5 900 165 795 5 -166 0 0
6 930 195 990 6 -196 0 0
7 1070 335 1325 7 -336 0 0
8 650 -85 1240 8 84 84 1
9 930 195 1435 9 -196 0 0
10 760 25 1460 10 -26 0 0
11 850 115 1575 11 -116 0 0
12 810 75 1650 12 -76 0 0
13 950 215 1865 13 -216 0 0
14 1000 265 2130 14 -266 0 0
15 980 245 2375 15 -246 0 0
16 1000 265 2640 16 -266 0 0
17 980 245 2885 17 -246 0 0
18 960 225 3110 18 -226 0 0
19 880 145 3255 19 -146 0 0
20 960 225 3480 20 -226 0 0

Plot of Cusum with target value 734.5 for all observations:

 CUSUM Chart of C1

4000

3000

Cumulative Sum 2000

1000

UCL=359
0 0
LCL=-359

1 5 9 13 17 21 25 29 33 37
Sample

It indicates that the process is out of control.

9.19.Rework Exercise 9.4 using an EWMA control chart with λ= 0.1 and L=2.7. Compare your results to
those obtained within the cusum.

Based on the exercise 9.4, the tabular cusum for target value=1050 and σ=25 is described below:

sample a b
x
number
xi-1062.5 Ci+ N+ 1037.5-xi Ci- N-
1 1045 -17.5 0 0 -7.5 0 0
2 1055 -7.5 0 0 -17.5 0 0
3 1037 -25.5 0 0 0.5 0.5 1
4 1064 1.5 1.5 1 -26.5 0 0
5 1095 32.5 34 2 -57.5 0 0
6 1008 -54.5 0 0 29.5 29.5 1
7 1050 -12.5 0 0 -12.5 17 2
8 1087 24.5 24.5 1 -49.5 0 0
9 1125 62.5 87 2 -87.5 0 0
10 1146 83.5 170.5 3 -108.5 0 0
11 1139 76.5 247 4 -101.5 0 0
12 169 -893.5 0 0 868.5 868.5 1
13 1151 88.5 88.5 1 -113.5 755 2
14 1128 65.5 154 2 -90.5 664.5 3
15 1238 175.5 329.5 3 -200.5 464 4
16 1125 62.5 392 4 -87.5 376.5 5
17 1163 100.5 492.5 5 -125.5 251 6
18 1188 125.5 618 6 -150.5 100.5 7
19 1146 83.5 701.5 7 -108.5 0 0
 20 1167 104.5 806 8 -129.5 0 0

The calculations for EWMA control chart with λ = 0.1 and L=2.7 are summarized in the table below:

To ilustrate the calculation, consider the first observation is:

UCL= = =1056.75

Center line = µ0 =1050

LCL = = =1043.25

z1 = λx1 +(1- λ)z0 = 0.1(1045)+0.9(1050) = 1049.5

sample sample
x EWMA x EWMA
number number
1 1045 1049.5 11 1139 1058.9
2 1055 1050.5 12 169 961.9
3 1037 1048.7 13 1151 1060.1
4 1064 1051.4 14 1128 1057.8
5 1095 1054.5 15 1238 1068.8
6 1008 1045.8 16 1125 1057.5
7 1050 1050 17 1163 1061.3
8 1087 1053.7 18 1188 1063.8
9 1125 1057.5 19 1146 1059.6
10 1146 1059.6 20 1167 1061.7
 EWMA Chart of C1
1100

UCL=1083.4
1080
_
_
X=1066.3
1060

LCL=1049.2
EWMA

1040

1020

1000

1 3 5 7 9 11 13 15 17 19
Sample

9.25.Derive the variance of the exponentially weighted moving average zi

9.31.An EWMA control chart uses λ=0.4. How wide will the limit be on the Shewhart control chart,
expressed as a multiple of width of the steady-state EWMA limits?

From the control limit of EWMA control chart for steady-state value, then the wide is:

= = =

10.1.Discuss how you would use a cusum in the short production run solution. What advantages would
it have relative to a Shewhart chart, such as a DNOM version of the chart?
10.7.Reconsider the data in Exercise 10.4 and 10.6. Suppose the process measurement are individual
data values, not subgroup averages.

head
sample
1 2 3 4
number
xbar R xbar R xbar R xbar R
1 53 2 54 1 56 2 55 3
2 51 1 55 2 54 4 54 4
3 54 2 52 5 53 3 57 2
4 55 3 54 3 52 1 51 5
5 54 1 50 2 51 2 53 1
6 53 2 51 1 54 2 52 2
7 51 1 53 2 58 5 54 1
8 52 2 54 4 51 2 55 2
9 50 2 52 3 52 1 51 3
10 51 1 55 1 53 3 53 5
11 52 3 57 2 52 4 55 1
12 51 2 55 1 54 2 58 2
13 54 4 58 2 51 1 53 1
14 53 1 54 4 50 3 54 2
15 55 2 52 3 54 2 52 6
16 54 4 51 1 53 2 58 5
17 53 3 50 2 57 1 53 1
18 52 1 49 1 52 1 49 2
19 51 2 53 3 51 2 50 3
20 52 4 52 2 50 3 52 2

(a) Use observations 1-20 in Exercise 10.4 to construct appropriate group of control charts.

 chart:

CL= 52.988

UCL = = 52.988+ = 58.727

LCL= = 52.988- = 47.248

 R chart:

CL=2.158

UCL = D4MR= 3.267x2.158= 7.05

 LCL=D3MR = 0 x 2.158 = 0

= = =52.988

MR= = 2.158

(b) Plot observations 21-30 from Exercise 10.6 on the charts from part (a). Discuss your findings.

head
1 2 3 4
sample x x x x
number bar R bar R bar R bar R
21 50 3 54 1 57 2 55 5
22 51 1 53 2 54 4 54 3
23 53 2 52 4 55 3 57 1
24 54 4 54 3 53 1 56 2
25 50 2 51 1 52 2 58 4
26 51 2 55 5 54 5 54 3
27 53 1 50 2 51 4 60 1
28 54 3 51 4 54 3 61 4
29 52 2 52 1 53 2 62 3
30 52 1 53 3 50 4 60 1

(c) Using observations 1-20, construct an individual chart using the average of the readings on all
four heads as an individual measurement and an s control chart using the individual
measurements on each head. Discuss how these charts function relative to the group control
chart.

10.13. Specification on a bearing diameter are established at 8.0 ± 0.01 cm. Sample of size n=8 are used,
and a control chart for s shows statistical control, with the best current estimate of the population
standard deviation S=0.001. If the fraction of nonconforming product that is barely acceptable is
0.135%, find the three-sigma limits on the modified control chart for this process.

10.19. Set up a moving center line EWMA control chart for the concentration data in Exercise 10.16.
Compare it to the residuals control chart in Exercise 10.16 part (c).

=0.1, λ=0.7055, =3.227. observations 8, 56 and 90 exceed control limits

10.25.
 (a) Discuss the use of the moving range method to estimate the process standard deviation when
the data are positively autocorrelated.

(b)Discuss the use of the sample variance s2 with positively autocorrelated data. Specifically, if the
observations at lag ρi, is s2 still an unbiased estimator for σ2?

(c) Does your answer in part (b) imply that s 2 would really be a good way (in practice) to estimate σ 2
in constructing a control chart for autocorrelated data?

11.1. A product has three quality characteristics. The nominal values of these quality characteristics and
their sample covariance matrix have been determined from the analyzis of 30 preliminary samples
of size n=10 as follows:

S=

The sample means for each quality characteristic for 15 additional samples of size n=10 are shown
in table 11E.1. Is the process in statistical control?

sample
number
1 3.1 3.7 3
2 3.3 3.9 3.1
3 2.6 3 2.4
4 2.8 3 2.5
5 3 3.3 2.8
6 4 4.6 3.5
7 3.8 4.2 3
8 3 3.3 2.7
9 2.4 3 2.2
10 2 2.6 1.8
11 3.2 3.9 3
12 3.7 4 3
13 4.1 4.7 3.2
14 3.8 4 2.9
15 3.2 3.6 2.8

11.7. rework exercise 11.6, assuming that the subgroup size is n=5

(a) find the phase II control limits assuming that α=0.005.

UCL= = =21.3
 (b) Compare the control limits from part (a) to the chi-square control limit. What is the magnitude
of the difference in the two control limits?

Chi-square control limit = = =18.548

(c) how many preliminary samples would have to be taken to ensure that the exact phase II control
limit is within 1% of the chi-square control limit?

11.13. consider all 30 observations on the first two process variables in Table 11.6. Calculate an estimate
of the sample covariance matrix using both estimators S1 and S2 discussed in Section 11.3.2. Are
the estimates very different? Discuss your findings.

11.19. Consider the p=9 process variables in Table 11.5.

(a) perform a PCA on the first 30 observations. Be sure to work with the standardized variables.

(b) how much variability is explained if only the first r=3 principal components are retained?

(c) Construct an appropriate set of pairwise plot of the first r=3 principal component scores.

(d) Now consider the last 10 observations. Obtain the principal component scores and plot them on
the chart in part (c). Does the process seem to be in control?