4

ANSWERS TO QUESTIONS 4-1. Acceptance sampling is an inspection procedure in which samples of inputs, in-process products, or final products are taken and then compared to a quality standard to see if the entire lot the sample came from should be accepted or rejected. Process control is a statistical procedure for testing the production process itself to see if it is functioning properly. In attribute control charts (such as p-charts and c-charts), the measures of quality are discrete values reflecting a simple decision criterion such as good or bad. The quality measures used in variable-control charts (such as x -charts and R-charts) are continuous variables reflecting measurements such as weight, time, or volume. An R-chart reflects the process variability, whereas an x -chart indicates the tendency toward a mean value; thus, the two complement each other. That is, it is assumed that process average and variability must be in control for the process to be in control. When they are used together, the control limits are computed as x A2 R . In general, larger sample sizes are needed for attribute charts because more observations are required to develop a usable quality measure. Alternatively, variable measures require smaller sample sizes because each sample observation provides usable information such as weight or length. Thus, after only a few sample observations it is possible to compute a range or a sample average. Basically, a trial-and-error approach is required to determine a value for n that (along with c) achieves the desired operating characteristics. Tables and computer programs can make the determination much easier. A pattern test is used to determine if sample values from a process display a consistent pattern that is the result of a nonrandom cause, even though control charts may show the process to be in control. Width is determined by the size of the z value used; the smaller the value of z, the narrower the control limits. The AQL is the fraction of defective items in a lot that is deemed acceptable, and producers risk () is the probability of rejecting a lot that has an acceptable AQL. Alternatively, LTPD is an upper limit of the percentage of defective items that a consumer will accept in a lot, and the consumers risk () is the probability of accepting a lot in which the fraction of defective items exceeds the LTPD. 4-9.

STATISTICAL QUALITY CONTROL

A c-chart is used when it is not possible to determine a proportion defective (for a p-chart), for example, when counting the number of blemishes on a sheet of material. In a p-chart it must be possible to distinguish between individual defective and nondefective items.

4-2.

4-10. Tolerances are product-design specifications required by the customer, whereas control limits are the upper and lower bands of a control chart indicating when a process is out of control. 4-11. In a single sampling plan, only one sample from a lot is taken, and it must be of sufficient size to determine if the whole lot should be rejected. In a double-sampling plan a smaller sample is taken first; if the quality is very good, the lot is accepted, and if it is very bad, it is rejected. However, if the initial sample is inconclusive, a second sample is taken, and the lot is either accepted or rejected based on the combined results of the two samples. A multiple-sampling plan requires the smallest sample size of the sampling plans. An initial small sample is taken; if the results are inconclusive, additional samples are taken until the lot is either accepted or rejected. 4-12. Management usually selects 3-sigma limits because if the process is in control they want a high probability that the sample will fall within the control limits. With wider limits management is less likely to erroneously conclude that the process is out of control when points outside the control limits are due to normal, random variations. 4-13. Acceptance sampling accepts some level of defects because it is assumed that some defects are inevitable. However, TQM believes that defects can be eliminated altogether. In addition, acceptance sampling is based on inspection of the final product, however, according to TQM philosophy, that is too late and all the materials and labor that went into the defects are wasted. 4-14. When the product is a safety threat or life threatening or when inspection is very cheap compared to the consequences of a defect. 4-15. Process control charts could be used to monitor service time in a restaurant, bank, hospital, store, etc. 4-16. For example, in a fast food restaurant a control chart could be used to measure service times, defective menu items, out of stock menu items, customer complaints, cleanliness, and, order errors, among other things.

4-3.

4-4.

4-5.

4-6.

4-7.

4-8.

14

15

SOLUTIONS TO PROBLEMS 4-1. 453 = 0151 ; p= . (30)(100)

(0151)(0.849) p(1 p ) . = 0.0358 = n 100 UCL = p + z p(1 p ) n (0151)(0.849) . 100

= 0151 + 3 . = 0.258 LCL = p z

0.24 0.22 0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 2

UCL = 0.225

P = 0.153

LCL = 0.081

10

12

14

16

18

20

p(1 p ) n

Sample number

(0151)(0849) . . = 0151 3 . 100 = 0.044

0.28 0.26 0.24 0.22 0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02

In general, the proportion of defectives increases from sample 6 to sample 20, where it is eventually above the upper control limit. This indicates the process is moving out of control. p = 0.053 ; = (0.053)(0.947) p(1 p ) = = 0.016 n 200 UCL = p + 3 = 0.053 + 3(0.016) = 0101 .

4-3.
UCL = 0.258

P = 0.151

LCL = p 3 = 0.053 3( 0.016 ) = 0.005 The process does not seem to be out of control.

LCL = 0.044

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Sample number

The process does not seem to be out of control, although the decreasing number of defects from sample 8 to sample 17 should probably be investigated to see why the steady improvement occurred; likewise, the steadily increasing number of defects from sample 17 to sample 25 should probably be investigated to see why the quality deteriorated. p = 0153 ; = . p(1 p ) = 100 UCL = p + z = 0153 + 2(0.036) . = 0.225 LCL = p z = 0153 2(0.036) . = 0.081 0153(0847) . . = 0.036 100

0.12 0.11 0.10 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01

UCL = 0.101

P = 0.053

LCL = 0.005 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Sample number

4-2.

16

4-4.

UCL = p + z

p(1 p ) n 0.02(0.98) n

0.03 = 0.02 + z n=

2 0.02(0.98) 0.01 n = (28)2 n = 784

4-5.

a.

c=

742 = 24.73 30

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

UCL = 10.52

c = 4.3

LCL = 0 10 12 14 16 18 20

UCL = c + z c = 24.73 + 3 24.73 = 39.65 LCL = c z c = 24.73 3 24.73 = 9.81 4-7.

48 44 40 36 32 28 24 20 16 12 8 4 UCL = 39.65

Sample number

The process was not strictly out of control; however, from sample 10 to sample 20, the sample values were above the average and exhibited increasingly nonrandom behavior. c= 256 = 10.67 24

UCL = c + z c = 10.67 + 2 10.67 = 17.20 LCL = c z c

c = 24.73

= 10.67 2 10.67 = 414 .

24 22 20 18 16 14 12 10 8 6 4 2 2 4 6

LCL = 9.81

UCL = 17.20

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Sample number

b.

The process appears to be out of control. Nonrandom factors that might cause the process to move out of control could include (among other things) problems with the telephone order system, inexperienced operators taking orders, computer system problems, or shipping problems and delays. c= 86 = 4.3 20

c = 10.67

LCL = 4.14 8 10 12 14 16 18 20 22 24 Sample number

4-6.

a.

UCL = c + z c = 4.3 + 3 4.3 = 1052 . LCL = c z c = 4.3 3 4.3 = 192 or 0.0 (since the control chart . cannot go below zero) 4-8.

The process does not appear to be out of control, but sample 21 is close to the UCL and the process should be investigated.

= 9 in; = 0.6 in; n = 10 a. UCL = + z n 0.6 =9+3 10 = 9.57

   

   

17

LCL = z

b. c.

= 8.43 Yes, it appears to be. The control limits become narrower, but increasing the sample size will not effect the results in part b.
14 13 12 11 10 9 8 7 6 5 4 3 2 1 2

  n  0.6  =93  10 

a.

k 20 . From Table 4.1, D3 = 0.0 and D4 = 211 UCL = D4 R = 2.11(6.24) = 1317 . LCL = D3 R = 0.0(6.24) = 0
UCL = 13.17

R=

R = 12.7 = 6.24

4-9. Sample 1 2 3 4 5 6 a. R= R 0.67 0.69 0.93 0.52 0.64 0.71 Sample 7 8 9 10 11 12 R 0.45 0.17 0.32 0.99 0.65 0.15

R = 6.24

k 12 . From Table 4.1 in the text, D3 = 0 and D4 = 211 UCL = D4 R = 2.11(0.57) = 121 . LCL = D3 R = 0(057) = 0 .

R = 6.87 = 0.57

LCL = 0 4 6 8 10 12 14 16 18 20

Sample number

1.20 1.10 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10

UCL = 1.21

b. 4-11.

The temperature is within the control limits.

R =2 a. From Table 4.1, D3 = 0 and D4 = 2.28 UCL = D4 R = 2.28(2) = 4.56

R = 0.57

LCL = D3 R = 0(2) = 0 b. Sample 1 2 3 4 5 6 7 8 9 10 R 2.1 4.5 7.0 3.0 5.0 4.3 6.5 2.0 3.2 4.0

LCL = 0 2 4 6 8 10 12 Sample number

b. 4-10.

The process variability is within the control limits.

Sample 1 2 3 4 5 6 7 8 9 10

R 8.5 5.8 8.1 6.4 7.1 6.0 9.9 2.5 1.6 9.4

Sample 11 12 13 14 15 16 17 18 19 20

R 8.1 4.4 5.8 3.9 6.2 5.4 3.6 10.9 7.5 3.6

18

The process appears to be in control from both the x and R charts, although sample 6 is close to the LCL and perhaps should be investigated.
5.0 4.0 3.0 2.0 1.0 LCL = 0 1 2 3 4 5 6 7 8 9 10 UCL = 4.56

4-13. Sample 1 2 3 4 5 6 7 8 9 10 x 43.9 39.7 37.2 40.4 39.0 41.8 39.4 40.7 41.6 39.0 Sample 11 12 13 14 15 16 17 18 19 20 x 39.2 39.7 42.9 37.8 36.6 37.6 39.9 40.7 38.2 39.0

R=2

Sample number

The process clearly seems to be out of control. There are three of the sample points above the UCL, and all other sample values are above the center line for R , indicating nonrandom variations. 4-12. Sample 1 2 3 4 5 6 x= x 8.89 8.88 8.99 9.19 9.04 8.71 = Sample 7 8 9 10 11 12 x 9.05 9.16 8.97 9.06 9.09 9.01

x = 39.7 From Table 4.1, A2 = 058 . . UCL = x + A2 R = 39.7 + 058(6.24) . = 4332 . LCL = x A2 R = 39.7 058(6.24) . = 36.08
46 45 44 43 42 41 40 39 38 37 36 35 2 4 6

108.04 = 9.00 12 12 From Table 4.1, A2 = 058 . . UCL = x + A2 R = 9.00 + 0.58(0.57) = 9.33 LCL = x A2 R = 9 0.58(0.57) = 8.67
10.0 9.8 9.6 9.4 9.2 9.0 8.8 8.6 8.4 8.2

UCL = 43.32

x = 39.7

LCL = 36.08 8 10 12 14 16 18 20

Sample number

The process appears to be in control, with sample 1 seeming to be an aberration, however, the process should still be checked.
UCL = 9.33 x = 9.00

LCL = 8.67

10 11 12

Sample number

19

4-14. Sample 1 2 3 4 5 6 7 8 9 10 x 37.0 35.8 35.3 36.1 36.1 34.3 36.3 35.6 37.7 34.5

4-16.
Sample 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Above/ Below A B A A B B B B A A A A A A A Up/ Down D U U D U D D U U D U D D D Zone C B A A B C C B B A C A A A C Sample 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Above/ Below B B B B A A B B A A A B B B B Up/ Down D D U U U U D U U U U D D U D Zone B A B B C B A B C C A C A B B

x = 359 . From Table 4.1, A2 = 0.73 . UCL = x + A2 R . = 359 + 0.73(2) = 37.36 LCL = x A2 R . = 359 0.73(2) = 34.44
40 39 38 37 36 35 34 33 32 31 1 2 3 4 5 6 7 8 9 10 LCL = 34.44 UCL = 37.36 x = 35.9

The zone pattern test is violated for samples 10 through 14 (2 out of 3 consecutive points in zone A but within limits). 4-17.
Sample 1 2 3 4 5 6 7 8 9 10 Above/ Below A B A A A B A B B A Up/ Down D U D U D U D D U Zone C C C C C C B B A B Sample 11 12 13 14 15 16 17 18 19 20 Above/ Below A B B B A B B A A B Up/ Down D D U D U D D U D D Zone C C C B C C B A C B

Samples 7 through 10 appear to violate a zone pattern test (four out of five points in zone B or beyond) so there may be a nonrandom pattern.
Sample number

4-18.
Sample 1 2 3 4 5 6 7 8 9 10 Above/ Below A B A B A B A A B Up/ Down D D U D U D U U D Zone A C A C C B C C B C Sample 11 12 13 14 15 16 17 18 19 20 Above/ Below B B B B B B B A B B Up/ Down U U U D D U U U D U Zone C C A B A B C C B C

The process may be out of control (sample 6 and 9), although overall the x -chart does not reflect the magnitude of the out of control situation, as does the R-chart in Problem 4-11. The process should be investigated to determine a cause for the out of control samples. 4-15. Sample 1 2 3 4 5 6 7 8 9 10 11 12 Above/Below B B B A A B A A B A A A Up/Down D U U D D U U D U U D Zone B B C B C A C B C C C C

There are several instances where zone pattern test rules are violated; samples 1 to 3, and sample 13 through 16. Nonrandom patterns may exist.

There are no discernible nonrandom patterns.

20

4-19. Sample 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Above/ Below B B B A A A B A A A B B B B B Up/ Down

U D U U U D U D D D U U D D

Sample 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Above/ Below B B B B B A A A A A A A A A A

Up/ Down D D U U U U U U D U D U U D U 4-22.

5 4 3 2 1 0

10 11 12

Sample number

The process appears to be in control. p= 718 = 0.36 2000 p(1 p ) 100 = 0.36 + (3)( 0.048) = 0.504

There are eight consecutive points on one side of the center line on two occasionssamples 11-20 and samples 21-30 indicating nonrandom patterns may exist. 4-20. c= 146 = 7.3 20 = 7.3 + (3) 7.3 = 15.41 LCL = c z c = 7.3 (3)(2.70) 0
16 14 12 10 8 6 4 2 0

UCL = p + 3

UCL = c + z c

= 7.3 + (3)(2.70)

3 8

LCL = p (3)( 0.048) = 0.36 0.144 = 0.216

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

10

12

14

16

18

20

Sample number

The process appears to be out of control.

2 4 6 8 10 12 14 16 18 20

Sample number

4-23.

c=

326 = 16.3 20

The process appears to be in control however the process may be moving toward an out-of-control situation. 4-21. x = 325, R = 386, A2 (n = 5) = 0.58 . . UCL = x + A2 R . . = 3.25 + (058)(386) . = 548 LCL = x A2 R . = 3.25 (0.58)(386) . = 101

UCL = c + z c = 16.3 + (3)(4.03) = 28.412 LCL = c z c = 16.3 (3)(4.03) = 4188 . The process is in-control.

21

4-24.

x = 7.28, R = 4.25, A2 (n = 5) = 058 . UCL = x + A2 R . = 7.28 + (058)(4.25) = 9.75 LCL = x A2 R = 7.28 2.47 = 4.81
10 9 8 7 6 5 4 3 2 1 0

18 16 14 12 10 8 6 4 2 0

10

12

14

16

18

20

Sample number

The process appears to be in control. 4-27. x = 317, R = 325, A2 (n = 5) = 058 . . . UCL = x + A2 R . . . = 317 + (058)(325) . = 504
2 4 6 8 10 12 14 16 18 20

LCL = x A2 R . . = 317 189 . = 129

6 5 4 3 2 1 0

Sample number

While the process appears to be in control the mean of 7.28 appears to be significantly lower than the objective of 8 chips per cookie that management has established. Thus, the company should adjust their process to increase the number of chips and construct a new control chart. 4-25. p= 105
(30 )(16)

= 0.22

UCL = p + z

p(1 p ) n = 0.22 + (3)(0.0756) = 0.4452 LCL = 0.22 (3)(0.0756)

0 Although the process appears to be in control, the average proportion of defects among leaving patients, p = 0.22 , seems high and the hospital should probably adopt some quality improvement measures. 4-26. c= 202 = 101 . 20 4-28.

10 11 12

Sample number

UCL = c + z c = 101 + (3)(318) . . = 19.6 LCL = c z c = 101 (3)(318) . . = 0.57

The process appears out of control for sample 10, however, this could be an aberration since there are no other apparent nonrandom patterns or out-of-control points. Thus, this point should probably be thrown out and a new control chart developed with the remaining eleven samples. AQL = 0.02; LTPD = 0.08; = 0.05; = 010 . a. Using POM for Windows, n = 99 and c = 4 . results in = 0.049 and = 0.095 .

This

22

b.
.9509

CASE: Quality Control at Rainwater Brewery This is basically a discussion question; therefore, the student responses might vary. In general, an acceptance sampling plan is probably not appropriate for Rainwater Brewery. The owners have stated in the case description that the chances of a batch being spoiledand, thus, an unhealthy batchis very unlikely. However, even a very slight risk of a contaminated batch of 1,000 bottles might be too much, given the health consequences of a spoiled batch. Testing only a small sample of even a few bottles would indicate if the batch was, in fact, bad, so a simple testing procedure such as opening and testing 5 to 10 bottles might be prudent, without having to go to the trouble of developing a detailed acceptance sampling plan. Most of the quality control efforts should focus on process control procedures at the various stages of the brewing process. Obvious candidates are x - and R-charts to monitor temperature, specific gravity, and pH during the fermentation and aging stages. Some type of process control testing of the final bottled product probably is warranted also. One practical possibility might be a sequential sampling plan with the acceptance and rejection limits used as the limits of a process control chart. In other words, bottles could be tested sequentially for taste and color, and if the number of bottles sampled that were found to be off the norm exceeded the normal upper limit for off bottles, the owners could adjust the process rather than rejecting all 1,000 bottles in a batch. Finally, quality control methods can also be used at the beginning of the brewing process for checking materials such as bottles and caps and ingredients such as yeast, hops, and grain. Bottles and caps that are not completely clean and sterile can result in a spoiled batch, and poor-quality ingredients can obviously contribute to an off brew. It is feasible to consider acceptance sampling plans for these items. CASE: Quality Control at Grass, Unlimited c= 249 defects = 415 . 60 samples z = 2.00

Probability of Acceptance

.0948 0.02 AQL 0.08 LTPD Percent Defective

c. 4-29.

Using POM for Windows, AOQL = 0.0257 .

AQL = 0.04; LTPD = 010; = 0.05; = 010 . Using POM for . . Windows software, n = 136 and c = 9 , resulting in = 0.04 and = 0.117 .

4-30. No, a sampling plan with n = 150 and c = 4 will result in = 0.95 and = 0 . A sampling plan that would approximately achieve the desired results is n = 208 and c 18 , resulting in = 0.046 and = 0.08 . CASE SOLUTION: Improving Service Time at Daves Burgers x = 170, R = 140, A2 (n = 6) = 0.48 . . UCL = x + A2 R . . = 170 + (0.48)(140) = 2.37 LCL = x A2 R . . = 170 (0.48)(140) . = 103
3 2.5 2 1.5 1 0.5 0

UCL = c + Z c = 415 + 2 415 = 8.22 . . LCL = c Z c = 415 2 415 = 0.076 . . The chart exceeds the control limits for samples 12 and 14, however, the chart appears to have been in control prior to sample 12. Although, the reasons for the out-of-control occurrences must be investigated, the chart based on samples 1 through 11 could be used, or, several observations could be collected once the process is brought back in control and used with samples 1 through 12. Thus, this chart could be implemented for continued use. Other examples of control charts that could be used include p-charts for the number of errors in a sample of orders, or the number of customer complaints for a sample survey of customers. A c-chart could be used for the number of defects found (for cleanliness) during an inspection of facilities.

1 2 3 4

5 6 7 8

9 10 11 12 13 14 15

Sample number