This action might not be possible to undo. Are you sure you want to continue?

**Several exercises in this chapter differ from those in the 4th edition. An “*” following the
**

exercise number indicates that the description has changed (e.g., new values). A second

exercise number in parentheses indicates that the exercise number has changed. For

example, “2-16* (2-9)” means that exercise 2-16 was 2-9 in the 4th edition, and that the

description also differs from the 4th edition (in this case, asking for a time series plot

instead of a digidot plot). New exercises are denoted with an “☺”.

2-1*.

(a)

x = ∑ xi n = (16.05 + 16.03 +

n

i =1

(b)

n

( )

n

∑ x − ∑ xi

i =1

s=

2

i

+ 16.07 ) 12 = 16.029 oz

2

n

i =1

n −1

=

(16.052 +

+ 16.07 2 ) − (16.05 +

12 − 1

+ 16.07) 2 12

= 0.0202 oz

**MTB > Stat > Basic Statistics > Display Descriptive Statistics
**

Descriptive Statistics: Ex2-1

Variable

Ex2-1

Variable

Ex2-1

N N*

Mean

12

0 16.029

Maximum

16.070

SE Mean

0.00583

StDev

0.0202

Minimum

16.000

Q1

16.013

Median

16.025

Q3

16.048

2-2.

(a)

x = ∑ xi n = ( 50.001 + 49.998 +

n

i =1

(b)

n

s=

( )

n

2

∑ xi − ∑ xi

i =1

i =1

n −1

+ 50.004 ) 8 = 50.002 mm

2

n

=

(50.0012 +

+ 50.0042 ) − (50.001 +

8 −1

+ 50.004) 2 8

= 0.003 mm

**MTB > Stat > Basic Statistics > Display Descriptive Statistics
**

Descriptive Statistics: Ex2-2

Variable

Ex2-2

Variable

Ex2-2

N N*

Mean

8

0 50.002

Maximum

50.006

SE Mean

0.00122

StDev

0.00344

Minimum

49.996

Q1

49.999

Median

50.003

Q3

50.005

2-1

**Chapter 2 Exercise Solutions
**

2-3.

(a)

x = ∑ xi n = ( 953 + 955 +

n

i =1

(b)

n

s=

( )

n

∑ x − ∑ xi

i =1

2

i

i =1

n −1

+ 959 ) 9 = 952.9 °F

2

n

=

(9532 +

+ 9592 ) − (953 +

9 −1

+ 959) 2 9

= 3.7 °F

**MTB > Stat > Basic Statistics > Display Descriptive Statistics
**

Descriptive Statistics: Ex2-3

Variable

Ex2-3

Variable

Ex2-3

N N*

Mean

9

0 952.89

Maximum

959.00

SE Mean

1.24

StDev

3.72

Minimum

948.00

Q1

949.50

Median

953.00

Q3

956.00

2-4.

(a)

In ranked order, the data are {948, 949, 950, 951, 953, 954, 955, 957, 959}. The sample

median is the middle value.

(b)

Since the median is the value dividing the ranked sample observations in half, it remains

the same regardless of the size of the largest measurement.

2-5.

MTB > Stat > Basic Statistics > Display Descriptive Statistics

Descriptive Statistics: Ex2-5

Variable

Ex2-5

Variable

Ex2-5

N N*

Mean

8

0 121.25

Maximum

156.00

SE Mean

8.00

StDev

22.63

Minimum

96.00

Q1

102.50

Median

117.00

Q3

144.50

2-2

**Chapter 2 Exercise Solutions
**

2-6.

(a), (d)

MTB > Stat > Basic Statistics > Display Descriptive Statistics

Descriptive Statistics: Ex2-6

Variable

Ex2-6

Variable

Ex2-6

N N*

Mean

40

0 129.98

Maximum

160.00

SE Mean

1.41

StDev

8.91

Minimum

118.00

Q1

124.00

Median

128.00

Q3

135.25

(b)

Use √n = √40 ≅ 7 bins

MTB > Graph > Histogram > Simple

Histogram of Time to Failure (Ex2-6)

20

Frequency

15

10

5

0

112

120

128

136

Hours

144

152

160

(c)

MTB > Graph > Stem-and-Leaf

Stem-and-Leaf Display: Ex2-6

Stem-and-leaf of Ex2-6

Leaf Unit = 1.0

2

11 89

5

12 011

8

12 233

17 12 444455555

19 12 67

(5) 12 88999

16 13 0111

12 13 33

10 13

10 13 677

7

13

7

14 001

4

14 22

HI 151, 160

N

= 40

2-3

**Chapter 2 Exercise Solutions
**

2-7.

Use

n = 90 ≅ 9 bins

MTB > Graph > Histogram > Simple

**Histogram of Process Yield (Ex2-7)
**

18

16

14

Frequency

12

10

8

6

4

2

0

84

88

92

96

Yield

2-4

**Chapter 2 Exercise Solutions
**

2-8.

(a)

Stem-and-Leaf Plot

2

12o|68

6

13*|3134

12

13o|776978

28

14*|3133101332423404

(15)

14o|585669589889695

37

15*|3324223422112232

21

15o|568987666

12

16*|144011

6

16o|85996

1

17*|0

Stem Freq|Leaf

(b)

Use

n = 80 ≅ 9 bins

**MTB > Graph > Histogram > Simple
**

Histogram of Viscosity Data (Ex 2-8)

20

Frequency

15

10

5

0

13

14

15

Viscosity

16

17

**Note that the histogram has 10 bins. The number of bins can be changed by editing the
**

X scale. However, if 9 bins are specified, MINITAB generates an 8-bin histogram.

Constructing a 9-bin histogram requires manual specification of the bin cut points.

Recall that this formula is an approximation, and therefore either 8 or 10 bins should

suffice for assessing the distribution of the data.

2-5

**Chapter 2 Exercise Solutions
**

2-8(c) continued

MTB > %hbins 12.5 17 .5 c7

Row

1

2

3

4

5

6

7

8

9

10

11

Intervals

12.25 to 12.75

12.75 to 13.25

13.25 to 13.75

13.75 to 14.25

14.25 to 14.75

14.75 to 15.25

15.25 to 15.75

15.75 to 16.25

16.25 to 16.75

16.75 to 17.25

Totals

Frequencies

1

2

7

9

16

18

12

7

4

4

80

Percents

1.25

2.50

8.75

11.25

20.00

22.50

15.00

8.75

5.00

5.00

100.00

(d)

MTB > Graph > Stem-and-Leaf

Stem-and-Leaf Display: Ex2-8

Stem-and-leaf of Ex2-8 N = 80

Leaf Unit = 0.10

2

12 68

6

13 1334

12

13 677789

28

14 0011122333333444

(15) 14 555566688889999

37

15 1122222222333344

21

15 566667889

12

16 011144

6

16 56899

1

17 0

**median observation rank is (0.5)(80) + 0.5 = 40.5
**

x0.50 = (14.9 + 14.9)/2 = 14.9

Q1 observation rank is (0.25)(80) + 0.5 = 20.5

Q1 = (14.3 + 14.3)/2 = 14.3

Q3 observation rank is (0.75)(80) + 0.5 = 60.5

Q3 = (15.6 + 15.5)/2 = 15.55

(d)

10th percentile observation rank = (0.10)(80) + 0.5 = 8.5

x0.10 = (13.7 + 13.7)/2 = 13.7

90th percentile observation rank is (0.90)(80) + 0.5 = 72.5

x0.90 = (16.4 + 16.1)/2 = 16.25

2-6

**Chapter 2 Exercise Solutions
**

2-9 ☺.

MTB > Graph > Probability Plot > Single

Probability Plot of Liquid Detergent (Ex2-1)

Normal

99

Mean

StDev

N

AD

P-Value

95

90

16.03

0.02021

12

0.297

0.532

Percent

80

70

60

50

40

30

20

10

5

1

15.98

16.00

16.02

16.04

Fluid Ounces

16.06

16.08

**When plotted on a normal probability plot, the data points tend to fall along a straight
**

line, indicating that a normal distribution adequately describes the volume of detergent.

2-10 ☺.

MTB > Graph > Probability Plot > Single

Probability Plot of Furnace Temperatures (Ex2-3)

Normal

99

Mean

StDev

N

AD

P-Value

95

90

952.9

3.723

9

0.166

0.908

Percent

80

70

60

50

40

30

20

10

5

1

945.0

947.5

950.0 952.5

955.0

957.5

Temperature (deg F)

960.0

962.5

**When plotted on a normal probability plot, the data points tend to fall along a straight
**

line, indicating that a normal distribution adequately describes the furnace temperatures.

2-7

**Chapter 2 Exercise Solutions
**

2-11 ☺.

MTB > Graph > Probability Plot > Single

Probability Plot of Failure Times (Ex2-6)

Normal

99

Mean

StDev

N

AD

P-Value

95

90

130.0

8.914

40

1.259

<0.005

Percent

80

70

60

50

40

30

20

10

5

1

110

120

130

140

150

160

Hours

**When plotted on a normal probability plot, the data points do not fall along a straight line,
**

indicating that the normal distribution does not reasonably describe the failure times.

2-12 ☺.

MTB > Graph > Probability Plot > Single

Probability Plot of Process Yield Data (Ex2-7)

Normal

99.9

Mean

StDev

N

AD

P-Value

99

95

Percent

90

89.48

4.158

90

0.956

0.015

80

70

60

50

40

30

20

10

5

1

0.1

80

85

90

Yield

95

100

105

**When plotted on a normal probability plot, the data points do not fall along a straight line,
**

indicating that the normal distribution does not reasonably describe process yield.

2-8

**Chapter 2 Exercise Solutions
**

2-13 ☺.

MTB > Graph > Probability Plot > Single

**(In the dialog box, select Distribution to choose the distributions)
**

Probability Plot of Viscosity Data (Ex2-8)

Normal

99.9

Mean

StDev

N

AD

P-Value

99

Percent

95

90

14.90

0.9804

80

0.249

0.740

80

70

60

50

40

30

20

10

5

1

0.1

12

13

14

15

Viscosity

16

17

18

**Probability Plot of Viscosity Data (Ex2-8)
**

Lognormal

99.9

Loc

Scale

N

AD

P-Value

99

Percent

95

90

2.699

0.06595

80

0.216

0.841

80

70

60

50

40

30

20

10

5

1

0.1

12

13

14

15

Viscosity

16

17

18

19

2-9

the plot points on the Weibull probability plot are not straight—particularly in the tails— indicating it is not a reasonable model.010 10 5 3 2 1 0. 2-10 .Chapter 2 Exercise Solutions 2-13 continued Probability Plot of Viscosity Data (Ex2-8) Weibull Percent 99.10 15.1 10 11 12 13 14 Viscosity 15 16 17 18 Both the normal and lognormal distributions appear to be reasonable models for the data.032 <0. with no bends or curves.9 99 Shape Scale N AD P-Value 90 80 70 60 50 40 30 20 16. However. the plot points tend to fall along a straight line.36 80 1.

Chapter 2 Exercise Solutions 2-14 ☺.8537 20 0. MTB > Graph > Probability Plot > Single (In the dialog box.163 Percent 80 70 60 50 40 30 20 10 5 1 1000 10000 Cycles to Failure 100000 2-11 .549 0. select Distribution to choose the distributions) Probability Plot of Cycles to Failure (Ex2-14) Normal 99 Mean StDev N AD P-Value 95 90 8700 6157 20 0.521 0.776 0.137 Percent 80 70 60 50 40 30 20 10 5 1 -5000 0 5000 10000 15000 Cycles to Failure 20000 25000 Probability Plot of Cycles to Failure (Ex2-14) Lognormal 99 Loc Scale N AD P-Value 95 90 8.

336 >0.464 9624 20 0.Chapter 2 Exercise Solutions 2-14 continued Probability Plot of Cycles to Failure (Ex2-14) Weibull 99 Shape Scale N AD P-Value Percent 90 80 70 60 50 40 1.250 30 20 10 5 3 2 1 1000 10000 Cycles to Failure Plotted points do not tend to fall on a straight line on any of the probability plots. though the Weibull distribution appears to best fit the data in the tails. 2-12 .

ppm 100 150 Probability Plot of Concentration (Ex2-15) Lognormal 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 0.Chapter 2 Exercise Solutions 2-15 ☺.0 2-13 . ppm 100.426 <0.9347 1.56 40 8.201 0.005 Loc Scale N AD P-Value 0.0 Concentration.651 40 0.1 1. MTB > Graph > Probability Plot > Single (In the dialog box.0 10.470 22.873 Percent 80 70 60 50 40 30 20 10 5 1 -50 0 50 Concentration. select Distribution to choose the distributions) Probability Plot of Concentration (Ex2-15) Normal 99 95 90 Mean StDev N AD P-Value 9.

ppm 100.Chapter 2 Exercise Solutions 2-15 continued Probability Plot of Concentration (Ex2-15) Weibull 99 Shape Scale N AD P-Value Percent 90 80 70 60 50 40 0. 2-14 .000 10.091 30 20 10 5 3 2 1 0.010 0.782 40 0.100 1.637 0.6132 5.001 0. Plotted points on the normal and Weibull probability plots tend to fall off a straight line.000 Concentration.000 The lognormal distribution appears to be a reasonable model for the concentration data.

MTB > Graph > Time Series Plot > Single (or Stat > Time Series > Time Series Plot) Time Series Plot of Viscosity Data (Ex2-8) 17 Ex2-8 16 15 14 13 12 1 8 16 24 32 40 48 56 Time Order of Collection 64 72 80 From visual examination. 2-15 .Chapter 2 Exercise Solutions 2-16* (2-9). shifts or obvious patterns in the data. indicating that time is not an important source of variability. as evidenced by potentially cyclic behavior. MTB > Graph > Time Series Plot > Single (or Stat > Time Series > Time Series Plot) Time Series Plot of Yield Data (Ex2-7) 100 Ex2-7 95 90 85 1 9 18 27 36 45 54 63 Time Order of Collection 72 81 90 Time may be an important source of variability. there are no trends. 2-17* (2-10).

MTB > Graph > Time Series Plot > Single (or Stat > Time Series > Time Series Plot) Time Series Plot of Concentration Data (Ex2-15) 140 120 100 Ex2-15 80 60 40 20 0 4 8 12 16 20 24 28 Time Order of Collection 32 36 40 Although most of the readings are between 0 and 20. The order in which the data were collected may be an important source of variability. 35).476 Maximum 98. 2-19 (2-11). MTB > Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Ex2-7 Variable Ex2-7 Variable Ex2-7 N N* Mean 90 0 89. there are two unusually large readings (9.250 Q3 93.100 Median 89.125 2-16 . as well as occasional “spikes” around 20.158 Minimum 82.000 SE Mean 0.438 StDev 4.600 Q1 86.Chapter 2 Exercise Solutions 2-18 ☺.

those ending in 1.03 16. for example. 4.00 2-17 .02 16. The stem-and-leaf plot suggests that certain values may occur more frequently than others.Chapter 2 Exercise Solutions 2-20 (2-12).10 2 82 69 6 83 0167 14 84 01112569 20 85 011144 30 86 1114444667 38 87 33335667 43 88 22368 (6) 89 114667 41 90 0011345666 31 91 1247 27 92 144 24 93 11227 19 94 11133467 11 95 1236 7 96 1348 3 97 38 1 98 0 N = 90 Neither the stem-and-leaf plot nor the frequency histogram reveals much about an underlying distribution or a central tendency in the data. and 7.01 16. The data appear to be fairly well scattered.07 16. MTB > Graph > Stem-and-Leaf Stem-and-Leaf Display: Ex2-7 Stem-and-leaf of Ex2-7 Leaf Unit = 0. 6. MTB > Graph > Boxplot > Simple Boxplot of Detergent Data (Ex2-1) 16.05 16. 2-21 (2-13).04 16.06 Fluid Ounces 16.

2} + Pr{3. x = 3 3 / 36. x = 4 4 / 36. x = 9 3 / 36.92 − 7 2 11 = 0. x = 6 6 / 36.3} + Pr{2. 4. x: {the sum of two up dice faces} sample space: {2. 8.9950 2-23 (2-15).0000 49. 5. x = 12 0.9975 49.1} = 1 × 1 + 1 × 1 + 1 × 1 = 3 6 6 6 6 6 6 36 . x = 10 2 / 36. x = 5 5 / 36. 3. 9.1} = 1 × 1 + 1 × 1 = 2 6 6 6 6 36 Pr{x = 4} = Pr{1. 10. x = 2 2 / 36. 12} Pr{x = 2} = Pr{1.1} = 1 × 1 = 1 6 6 36 Pr{x = 3} = Pr{1. x = ∑ xi p( xi ) = 2 (1 36 ) + 3 ( 2 36 ) + 11 i =1 n ∑ xi p( xi ) − ⎡⎢ ∑ xi p ( xi ) ⎤⎥ ⎣ i =1 ⎦ S = i =1 n −1 n + 12 (1 36 ) = 7 2 n = 5.0050 mm 50. otherwise 2-24 (2-16). ( ) ( ( ) ) ( ) ( ) ⎧1/ 36. x = 8 4 / 36..38 10 2-18 . 11. x = 7 p ( x) = ⎨ ⎩5 / 36. 7..Chapter 2 Exercise Solutions 2-22 (2-14). 6. MTB > Graph > Boxplot > Simple Boxplot of Bearing Bore Diameters (Ex2-2) 50.0025 50.0075 50. 2} + Pr{2. x = 11 1/ 36.

02 (0. +∞ For f(x) to be a probability distribution.01. x ~ POI(0.02). ⎧(1 + 3k ) / 3.01 (0. 2-32) σ2 = 1/λ2 = 1 (Eqn.9802 = 0.Chapter 2 Exercise Solutions 2-25 (2-17). e −0.9900 = 0.5 k = 0. 2-33) 2-27 (2-19). 2-26 (2-18). x ~ POI(0.02.01).0100 Pr{x ≥ 1} = 1 − Pr{x = 0} = 1 − p(0) = 1 − 0! Cutting the rate at which defects occur reduces the probability of one or more defects by approximately one-half.02)1 = 0.5 + 5k ) / 3. µ = 1/λ = 1 (Eqn.0196 1! (b) Pr{x ≥ 1} = 1 − Pr{x = 0} = 1 − p(0) = 1 − e −0.5 + 5k ) =1 3 10k = 0.0198 to 0. This is a Poisson distribution with parameter λ = 0. x = 2 p ( x) = ⎨ ⎩(0.02 (0. x = 1 (1 + 2k ) / 3. from 0.02)0 = 1 − 0. use F ( x) = ∑ p( xi ) = 1 i =1 (1 + 3k ) + (1 + 2k ) + (0.0198 0! (c) This is a Poisson distribution with parameter λ = 0. otherwise (a) ∞ To solve for k.05 2-19 .01)0 = 1 − 0. (a) Pr{x = 1} = p(1) = e −0. x = 3 0. ∫ f ( x)dx must equal unity.0100. −∞ ∞ −x −x ∞ ∫ ke dx = [−ke ]0 = − k[0 − 1] = k ⇒ 1 0 This is an exponential distribution with parameter λ=1.

118 Approximately 11.Chapter 2 Exercise Solutions 2-27 continued (b) ⎡1 + 3(0.15 + 1.10/calculator. x = 1 ⎪ 3 ⎪ 1.15 ⎧ = 0.867 3 3 3 ⎣ ⎦ ⎣ ⎦ ⎣ 3 µ = ∑ xi p( xi ) = 1× ⎢ i =1 3 σ 2 = ∑ xi2 p( xi ) − µ 2 = 12 (0. x = 3 ⎪ 3 ⎩ 2-28 (2-20).15 + 1.75 = 1.250) − 1.5 + 5(0.125: Pr{x ≤ 1} = F (1) = 1 − e −0.8% will fail during the first year.867 2 = 0. 0 < r < 1.05) ⎤ ⎡1 + 2(0.05) ⎤ + 2× ⎢ + 3× ⎢ ⎥ ⎥ ⎥⎦ = 1. (a) This is an exponential distribution with parameter λ = 0. p ( x) = kr x . cost = $50/calculator Sale profit = $25/calculator Net profit = $[-50(1 + 0.90/calculator.615 i =1 (c) 1. (b) Mfg. 2-20 .383.1 + 0.125(1) = 0.000.05) ⎤ ⎡ 0. x = 2 3 ⎪ ⎪1. The effect of warranty replacements is to decrease profit by $5.383) + 22 (0. 2.1 ⎪ F ( x) = ⎨ = 0.118) + 75]/calculator = $19.750.… ∞ F ( x) = ∑ kr x = 1 by definition i =0 k ⎡⎣1 (1 − r ) ⎤⎦ = 1 k = 1− r 2-29 (2-21).1.367) + 32 (0. x = 0.

01) 25 = 1 − 0. ⎛ 25 ⎞ Pr{x ≥ 1} = 1 − Pr{x < 1} = 1 − Pr{x = 0} = 1 − ⎜ ⎟ (0.04)0 (1 − 0.22 ⎝0⎠ This decision rule means that 22% of the samples will have one or more nonconforming units.777821 2-32* (2-24).75 = 0.02)1 (1 − 0.125 12 11.04) Stop process if x ≥ 1. This exercise may also be solved using Excel or MINITAB: (1) Excel Function BINOMDIST(x.75 4 x2 2 12 − 47 x 11.02) 49 + ⎝0⎠ ⎝1⎠ ⎛ 50 ⎞ + ⎜ ⎟ (0.78 = 0.04) 25 = 1 − 0.04} = Pr{x ≤ 2} = ∑ ⎜ ⎟(0. n.01)0 (1 − 0.875 − 11. x ~ BIN(25.02)50 + ⎜ ⎟ (0.921 ⎝4⎠ 2-21 .01 and n = 25. This is a binomial distribution with parameter p = 0.02) 46 = 0. This is a binomial distribution with parameter p = 0.02)(50− x ) x =0 ⎝ x ⎠ ⎛ 50 ⎞ ⎛ 50 ⎞ = ⎜ ⎟ (0.75) dx = 11. This is a somewhat difficult operating situation.Chapter 2 Exercise Solutions 2-30 (2-22).02) 4 (1 − 0. and the process will be stopped to look for a cause.01 x P( X <= x ) 0 0. p.36 = 0. 4 ⎛ 50 ⎞ Pr{ pˆ ≤ 0.75 2-31* (2-23). 12 Pr{x < 12} = F (12) = ∫ −∞ 12 f ( x) dx = ∫ 4( x − 11.02)0 (1 − 0.64 ⎝0⎠ 2-33* (2-25). ⎛ 25 ⎞ Pr{x ≥ 1} = 1 − Pr{x < 1} = 1 − Pr{x = 0} = 1 − ⎜ ⎟ (0. 0. The process is stopped if x ≥ 1.02) x (1 − 0.75 = 11. TRUE) (2) MTB > Calc > Probability Distributions > Binomial Cumulative Distribution Function Binomial with n = 25 and p = 0.02 and n = 50.

99) + ⎜ ⎟ (0.0100 Pr{ pˆ > kσ + p} = 1 − Pr{ pˆ ≤ kσ + p} = 1 − Pr{x ≤ n(kσ + p)} k=1 1 − Pr{x ≤ n(kσ + p )} = 1 − Pr{x ≤ 100(1(0.99)97 ⎥ = 1− ∑ ⎜ ⎟ ⎟ x =0 ⎝ x ⎠ ⎝ 3 ⎠ ⎣ ⎦ = 1 − [0.99)100− x = 1 − ⎢0.0100) + 0.01(1 − 0.01) (1 − 0.01) 100 = 0.99) ⎥ ⎝ 1 ⎠ ⎝ 2 ⎠ ⎣⎝ 0 ⎠ ⎦ = 1 − [0. σ = 0.99) + ⎜ ⎟ (0.0100) + 0.01) (0.982] = 0.003 2-22 .Chapter 2 Exercise Solutions 2-34* (2-26).01)} = 1 − Pr{x ≤ 3} 3 ⎛ 100 ⎞ ⎡ ⎤ ⎛100 ⎞ (0.01) x (0.01) (0.921 + ⎜ (0.01) x =0 ⎝ x ⎠ ⎡⎛100 ⎞ ⎤ ⎛ 100 ⎞ ⎛100 ⎞ 0 100 1 99 2 98 = 1 − ⎢⎜ ⎟ (0.01 and n = 100.01) (0.99)100− x = 1 − ⎢0.01)} = 1 − Pr{x ≤ 2} 2 ⎛ 100 ⎞ 100 − x x = 1− ∑ ⎜ ⎟(0.921] = 0.01)} = 1 − Pr{x ≤ 4} 4 ⎛ 100 ⎞ ⎡ ⎤ ⎛100 ⎞ (0.992] = 0.01) x (0.01)3 (0.018 k=3 1 − Pr{x ≤ n(kσ + p )} = 1 − Pr{x ≤ 100(3(0.01) 4 (0.99)96 ⎥ = 1− ∑ ⎜ ⎟ ⎟ x =0 ⎝ x ⎠ ⎝ 4 ⎠ ⎣ ⎦ = 1 − [0.079 k=2 1 − Pr{x ≤ n(kσ + p )} = 1 − Pr{x ≤ 100(2(0. This is a binomial distribution with parameter p = 0.0100) + 0.982 + ⎜ (0.

(c) For N = 150. ⎛5⎞ Pr{acceptance} = p (0) = ⎜ ⎟ (0. 2-23 . D. M = 2.08)0 (1 − 0. The binomial approximation is not satisfactory in this case. This is a hypergeometric distribution with N = 25 and n = 5.Chapter 2 Exercise Solutions 2-35* (2-27). so the binomial approximation would be a satisfactory approximation the hypergeometric in this case. though close to the exact solution for x = 0. n/N = 5/150 = 0. violates the rule-ofthumb that n/N = 5/25 = 0. p = D/N = 2/25 = 0. and n = 5 x P( X <= x ) 0 0.633 (53.20 be less than the suggested 0.1. without replacement. (a) Given D = 2 and x = 0: ⎛ 2 ⎞ ⎛ 25 − 2 ⎞ ⎜ ⎟⎜ ⎟ 0 5 − 0 ⎠ (1)(33.130) ⎛ 25 ⎞ ⎜ ⎟ ⎝5⎠ This exercise may also be solved using Excel or MINITAB: (1) Excel Function HYPGEOMDIST(x. n.08)5 = 0. 649) Pr{Acceptance} = p (0) = ⎝ ⎠ ⎝ = = 0.033 ≤ 0.633333 (b) For the binomial approximation to the hypergeometric.08 and n = 5. N) (2) MTB > Calc > Probability Distributions > Hypergeometric Cumulative Distribution Function Hypergeometric with N = 25.1.659 ⎝0⎠ This approximation.

057 p(0) = ⎝ ⎠ ⎝ ⎠ = (3. 400) ⎛ 25 ⎞ ⎜ ⎟ ⎝ 11 ⎠ Let sample size n = 11. n = 5.038 p(0) = ⎝ ⎠ ⎝ ⎠ = (4. ⎛ 5 ⎞ ⎛ 25 − 5 ⎞ ⎛ 5 ⎞ ⎛ 20 ⎞ ⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟ 0⎠⎝ n − 0 ⎠ ⎝ 0⎠⎝ n ⎠ ⎝ = p(0) = ⎛ 25 ⎞ ⎛ 25 ⎞ ⎜ ⎟ ⎜ ⎟ ⎝n⎠ ⎝n⎠ try n = 10 ⎛ 5 ⎞ ⎛ 20 ⎞ ⎜ ⎟⎜ ⎟ 0 10 (1)(184. ⎛ 3 ⎞⎛ 30 − 3 ⎞ ⎜ ⎟⎜ ⎟ 1 ⎠⎝ 5 − 1 ⎠ (3)(17.Chapter 2 Exercise Solutions 2-35 continued (d) Find n to satisfy Pr{x ≥ 1 | D ≥ 5} ≥ 0. 457.433 ⎛ 30 ⎞ ⎜ ⎟ ⎝5⎠ 2-24 . 756) = 0.95.550) ⎝ Pr{x = 1} = p (1) = = = 0.567 = 0.506) ⎛ 30 ⎞ ⎜ ⎟ ⎝5⎠ ⎛ 3 ⎞ ⎛ 27 ⎞ ⎜ ⎟⎜ ⎟ 0 5 Pr{x ≥ 1} = 1 − Pr{x = 0} = 1 − p(0) = 1 − ⎝ ⎠ ⎝ ⎠ = 1 − 0. 2-36 (2-28). 760) ⎛ 25 ⎞ ⎜ ⎟ ⎝ 10 ⎠ try n = 11 ⎛ 5 ⎞ ⎛ 20 ⎞ ⎜ ⎟⎜ ⎟ 0 11 (1)(167.960) = 0. or equivalently Pr{x = 0 | D = 5} < 0. and D = 3.05.369 (142. This is a hypergeometric distribution with N = 30. 268.

020.00001 0! 2-40 (2-32).Chapter 2 Exercise Solutions 2-37 (2-29).020)50−1 = 1 − 0.1.01)1 = 0.364) = 0.020)50−0 = (1)(1)(0. the binomial distribution can be used to approximate the hypergeometric. This is a Poisson distribution with λ = 0.020)0 (1 − 0.01 (0.364 − ⎜ ⎟ (0.364 − 0.905 = 0.095 Pr{x ≥ 1} = 1 − Pr{x = 0} = 1 − p(0) = 1 − 0! This exercise may also be solved using Excel or MINITAB: (1) Excel Function POISSON(λ.01 errors/bill. x.264 ⎝1⎠ 2-38 (2-30).99999 = 0.020)1 (1 − 0.1 x P( X <= x ) 0 0.904837 2-39 (2-31). n = 50 pages.00001 (0.1 ≤ 0. and D = 10 errors. This is a hypergeometric distribution with N = 500 pages.1 (0. e −0.0099 Pr{x = 1} = p (1) = 1 2-25 . ⎛ 50 ⎞ Pr{x = 0} = p(0) = ⎜ ⎟ (0.364 ⎝0⎠ Pr{x ≥ 2} = 1 − Pr{x ≤ 1} = 1 − [Pr{x = 0} + Pr{x = 1}] = 1 − p(0) − p(1) ⎛ 50 ⎞ = 1 − 0.00001 stones/bottle. TRUE) (2) MTB > Calc > Probability Distributions > Poisson Cumulative Distribution Function Poisson with mean = 0. This is a Poisson distribution with λ = 0.372 = 0.1 defects/unit.1)0 = 1 − 0. with p = D/N = 10/500 = 0. e −0. Pr{x ≥ 1} = 1 − Pr{x = 0} = 1 − e −0. Checking n/N = 50/500 = 0.00001)0 = 1 − 0. This is a Poisson distribution with λ = 0.

000 How many fail the minimum specification. σ.012 0.6) = 1 − 0.9862 ⎝0⎠ ⎝1⎠ ⎝2⎠ 2-43* (2-35). the number that exceed 48 lb.6} 5 ⎭ ⎩ = 1 − Φ (1.01) (1 − 0.055 So.010 0.… ∞ d ∞ 1 µ = ∑ t ⎡⎣ p(1 − p)t −1 ⎤⎦ = p ⎡ ∑ q t ⎤ = ⎢ ⎥ t =1 dq ⎣ t =1 ⎦ p 2-42 (2-34).000) × (0.01) 3 1 − ⎝ ⎠ Pr{x > 51} = Pr{r = 0} + Pr{r = 1} + Pr{r = 2} ⎛ 50 ⎞ ⎛ 50 ⎞ ⎛ 50 ⎞ = ⎜ ⎟ 0. x ~ N (40. ⎛ 51 − 1⎞ 3 51−3 Pr{x = 51} = p(51) = ⎜ = (1225)(0.158655 How many exceed 48 lb.617290) = 0.000001)(0.945 = 0. and x = 1 + (5000/100) = 51. the number that fail the minimum specification are (50.9949 ⎜ ⎟ 0. r = 3 welds.? 35 − 40 ⎫ ⎧ Pr{x ≤ 35} = Pr ⎨ z ≤ ⎬ = Pr{z ≤ −1} = Φ (−1) = 0.0110.0008 ⎟ (0.159 5 ⎭ ⎩ So.3.? 48 − 40 ⎫ ⎧ Pr{x > 48} = 1 − Pr{x ≤ 48} = 1 − Pr ⎨ z ≤ ⎬ = 1 − Pr{z ≤ 1. 2. 52). Pr(t ) = p(1 − p)t −1 .159) = 7950. 2-26 .000) × (0.9950 + ⎜ ⎟ 0. n = 50.9948 = 0. t = 1.Chapter 2 Exercise Solutions 2-41 (2-33).01.055) = 2750. This exercise may also be solved using Excel or MINITAB: (1) Excel Function NORMDIST(X. TRUE) (2) MTB > Calc > Probability Distributions > Normal Cumulative Distribution Function Normal with mean = 40 and standard deviation = 5 x P( X <= x ) 35 0. µ. LSL = 35 lb. This is a Pascal distribution with Pr{defective weld} = p = 0. is (50.

0228.022). Φ ( z ) = 0.00621 = 0.9991 4 µ = −4(−1.29 Process variance must be reduced to 0. Find µ such that Pr{x < 32} = 0. 352) Pr{x > 1000} = 1 − Pr{x ≤ 1000} 1000 − 900 ⎫ ⎧ = 1 − Pr ⎨ x ≤ ⎬ 35 ⎩ ⎭ = 1 − Φ (2.98758 ⎝ 0.015 z −3.8571) = 1 − 0. 42 ). 0. is currently centered between the specification limits (target = 5 V).95 − 5 = = 0.29 z= LSL − µ σ .05 − 5 ⎞ ⎛ 4. Need Pr{x ≤ LSL} = 0.5) − Φ (−2.0005 z = Φ −1 (0.95 − 5 ⎞ = Φ⎜ ⎟−Φ⎜ ⎟ = Φ (2.9991 32 − µ = −1.001.0005) = −3. LSL = 4.02 ⎠ 2-45* (2-37). x ~ N ( µ .001 / 2 = 0. USL = 5.99379 − 0. Φ −1 (0. 2-46 (2-38). so σ = LSL − µ 4. x ~ N(5. with mean 5 V.95 V.Chapter 2 Exercise Solutions 2-44* (2-36). x ~ N(900.5) = 0.0005.0152 to have at least 999 of 1000 conform to specification.02 ⎠ ⎝ 0.0228) = −1.9991) + 32 = 40.0021 2-27 . Shifting the process mean in either direction would increase the number of nonconformities produced.0 2-47 (2-39).9979 = 0. The process. Assume that the process remains centered between the specification limits at 5 V. Desire Pr{Conformance} = 1 / 1000 = 0.05 V Pr{Conformance} = Pr{LSL ≤ x ≤ USL} = Pr{x ≤ USL} − Pr{x ≤ LSL} ⎛ 5.

x ~ N(5000.0000 = 0. then choose process 1 2-28 .500 ⎫ 5. x2 ~ N(7500. 000 1.000 h sales = $10/unit. 000 − 7. 000 − 7. profit = $10 × Pr{good} + $5 × Pr{bad} – c For Process 1 proportion defective = p1 = 1 − Pr{LSL ≤ x1 ≤ USL} = 1 − Pr{x1 ≤ USL} + Pr{x1 ≤ LSL} 10. 502).0000) + 5 (0. defect = $5/unit.0124) + 5 (0.5758 50 LSL = 50(−2.5) + Φ (−2.Chapter 2 Exercise Solutions 2-48 (2-40).9380 – c1 For Process 2 proportion defective = p2 = 1 − Pr{LSL ≤ x2 ≤ USL} = 1 − Pr{x2 ≤ USL} + Pr{x2 ≤ LSL} 10. Find LSL such that Pr{x < LSL} = 0.500 ⎫ ⎧ ⎧ = 1 − Pr ⎨ z2 ≤ ⎬ + Pr ⎨ z2 ≤ ⎬ 500 500 ⎩ ⎭ ⎩ ⎭ = 1 − Φ (5) + Φ (−5) = 1 − 1.5) = 1 − 0. σ22 = 5002).005) = −2. σ12 = 10002). USL = 10.9938 + 0. 000 − 7.5758 LSL − 5000 = −2.005 Φ −1 (0.0000 profit for process 2 = 10 (1 – 0.500 ⎫ 5.000 h.0124 profit for process 1 = 10 (1 – 0. x1 ~ N(7500. 000 ⎩ ⎭ ⎩ ⎭ = 1 − Φ (2.0000 + 0. 000 − 7.0124) – c1 = 9. LSL = 5.5758) + 5000 = 4871 2-49 (2-41).500 ⎫ ⎧ ⎧ = 1 − Pr ⎨ z1 ≤ ⎬ + Pr ⎨ z1 ≤ ⎬ 1.0062 = 0.0000) – c2 = 10 – c2 If c2 > c1 + 0.0620.

Chapter 2 Exercise Solutions 2-50 (2-42). 2 ⎣ ⎝ C0 + C2 ⎠ ⎦ 2-29 . Proportion less than lower specification: 6−µ⎫ ⎧ pl = Pr{x < 6} = Pr ⎨ z ≤ ⎬ = Φ (6 − µ ) 1 ⎭ ⎩ Proportion greater than upper specification: 8−µ ⎫ ⎧ pu = Pr{x > 8} = 1 − Pr{x ≤ 8} = 1 − Pr ⎨ z ≤ ⎬ = 1 − Φ (8 − µ ) 1 ⎭ ⎩ Profit = +C0 pwithin − C1 pl − C2 pu = C0 [Φ (8 − µ ) − Φ (6 − µ )] − C1[Φ (6 − µ )] − C2 [1 − Φ (8 − µ )] = (C0 + C2 )[Φ (8 − µ )] − (C0 + C1 )[Φ (6 − µ )] − C2 d d ⎡ 8− µ 1 ⎤ [Φ (8 − µ )] = exp(−t 2 / 2)dt ⎥ ∫ ⎢ dµ d µ ⎣ −∞ 2π ⎦ Set s = 8 – µ and use chain rule d d ⎡s 1 1 ⎤ ds =− [Φ (8 − µ )] = ⎢ ∫ exp(−t 2 / 2)dt ⎥ exp −1/ 2 × (8 − µ ) 2 dµ ds ⎣ −∞ 2π 2π ⎦ dµ ( ) d (Profit) ⎡ 1 ⎤ ⎡ 1 ⎤ = −(C0 + C2 ) ⎢ exp −1/ 2 × (8 − µ ) 2 ⎥ + (C0 + C1 ) ⎢ exp −1/ 2 × (6 − µ ) 2 ⎥ dµ ⎣ 2π ⎦ ⎣ 2π ⎦ ( ) ( ) Setting equal to zero 2 C0 + C1 exp −1/ 2 × (8 − µ ) = = exp(2µ − 14) C0 + C2 exp −1/ 2 × (8 − µ ) 2 ( ( So µ = ) ) ⎤ 1 ⎡ ⎛ C0 + C1 ⎞ ⎢ln ⎜ ⎟ + 14 ⎥ maximizes the expected profit.

p( x) = ∞ ∞ ⎛ e− λ λ x ⎝ x! µ = E[ x] = ∑ xi p( xi ) = ∑ x ⎜ i =1 x =0 e− λ λ x . ⎛n⎞ For the binomial distribution.... p ( x) = ⎜ ⎟ p x (1 − p) n − x .Chapter 2 Exercise Solutions 2-51 (2-43). x = 0. For the Poisson distribution. x = 0.… x! ⎞ − λ ∞ λ ( x −1) =e− λ λ eλ = λ ⎟=e λ∑ x = 0 ( x − 1)! ⎠ ( ) σ 2 = E[( x − µ ) 2 ] = E ( x 2 ) − [ E ( x)]2 ⎛ e− λ λ x ⎞ 2 E ( x ) = ∑ x p( xi ) = ∑ x ⎜ ⎟=λ +λ i =1 x =0 ⎝ x! ⎠ 2 ∞ ∞ 2 i 2 σ 2 = (λ 2 + λ ) − [ λ ] = λ 2 2-30 . n ⎝ x⎠ ∞ n ⎡⎛ n ⎞ ⎤ n −1 µ = E ( x) = ∑ xi p ( xi ) = ∑ x ⎢⎜ ⎟ p x (1 − p ) n − x ⎥ = n ⎡⎣ p + (1 − p ) ⎤⎦ p = np i =1 x =0 ⎣⎝ x ⎠ ⎦ σ 2 = E[( x − µ ) 2 ] = E ( x 2 ) − [ E ( x)]2 ∞ n ⎡⎛ n ⎞ n− x ⎤ E ( x 2 ) = ∑ xi2 p ( xi ) = ∑ x 2 ⎢⎜ ⎟ p x (1 − p ) ⎥ =np + (np ) 2 − np 2 i =1 x =0 ⎣⎝ x ⎠ ⎦ σ 2 = ⎡⎣ np + (np) 2 − np 2 ⎤⎦ − [ np ] = np(1 − p) 2 2-52 (2-44)..1.1.

setting u = x 2 and dv = λ exp(−λ x) uv − ∫ vdu = ⎡⎣ x exp(−λ x) ⎤⎦ 2 +∞ 0 +∞ + 2 ∫ x exp(−λ x)dx = (0 − 0) + 0 σ2 = 2 λ 2 − 1 λ 2 = 2 λ2 1 λ2 2-31 .Chapter 2 Exercise Solutions 2-53 (2-45). setting u = x and dv = λ exp(−λ x) +∞ +∞ 1 1 uv − ∫ vdu = ⎡⎣ − x exp ( −λ x ) ⎤⎦ 0 + ∫ exp ( −λ x ) dx = 0 + = λ 0 λ For the variance: ⎛1⎞ σ = E[( x − µ ) ] = E ( x ) − [ E ( x) ] = E ( x ) − ⎜ ⎟ ⎝λ⎠ 2 2 2 2 +∞ +∞ −∞ 0 2 2 E ( x 2 ) = ∫ x 2 f ( x)dx = ∫ x 2 λ exp(−λ x)dx Integrate by parts. f ( x ) = λ e − λ x . For the exponential distribution. x ≥ 0 For the mean: +∞ +∞ 0 0 µ = ∫ xf ( x)dx = ∫ x ( λ e − λ x )dx Integrate by parts.

002 15 Zα/2 = Z0.78 σ n 0. x = 127 psi.05 Test H0: µ = 8.25 Z0 = = = 6.251 MTB > Stat > Basic Statistics > 1-Sample Z > Summarized data One-Sample Z Test of mu = 8.002 cm (a) µ0 = 8.2535 The assumed standard deviation = 0.96 Reject H0: µ = 8.78 P 0.00000] = 0 (c) ⎞ ≤ µ ≤ x + Z ⎛σ ⎞ x − Zα / 2 ⎛⎜ σ ⎟ ⎟ α /2 ⎜ n⎠ n⎠ ⎝ ⎝ ( 8.00052 (8.25101) Z -6. H1: µ ≠ 8.2535 cm.828 σ n 2 8 Zα = Z0.05/2 = Z0. n = 15.249 ≤ µ ≤ 8. and conclude that the mean bearing ID is not equal to 8. Reject H0 if |Z0| > Zα/2.25 cm. x − µ0 127 − 125 Z0 = = = 2.645 Reject H0: µ = 125.25 − 1. H1: µ > 125.25. x − µ0 8. α = 0.002 15 ) 8.Chapter 3 Exercise Solutions 3-1.002 ) ( 15 ≤ µ ≤ 8.05 = 1.05 Test H0: µ = 125 vs.000 3-2. 3-1 . σ = 2 psi (a) µ0 = 125.25 + 1.025 = 1.002 N Mean SE Mean 95% CI 15 8.96 0.25000 0.96 0. x = 8.25. 8.25 vs. n = 8.2535 vs not = 8.2535 − 8.78)] = 2[1 − 1.25. (b) P-value = 2[1 − Φ(Z0)] = 2[1 − Φ(6. α = 0. σ = 0.24899. Reject H0 if Z0 > Zα. and conclude that the mean tensile strength exceeds 125 psi.

not simply that the mean does not equal the hypothesized value.0 − 25 t0 = = = 1. σ). α = 0. MTB > Stat > Basic Statistics > 1-Sample t > Samples in columns One-Sample T: Ex3-3 Test of mu = 25 vs > 25 Variable Ex3-3 N 10 Mean 26.Chapter 3 Exercise Solutions 3-2 continued (b) P-value = 1 − Φ(Z0) = 1 − Φ(2. s = 1. n−1 = t0.05.95 P 0. A one-sided hypothesis test lets us do this.5138 95% Lower Bound 25. x ~ N(µ. µ0 = 25.8 ≤ µ MTB > Stat > Basic Statistics > 1-Sample Z > Summarized data One-Sample Z Test of mu = 125 vs > 125 The assumed standard deviation = 2 95% Lower N Mean SE Mean Bound Z 8 127. we usually are interested in whether some minimum requirement is met.99766 = 0.837 2.645 ( 2 8 ) ≤ µ x − Zα σ 125. x − µ0 26.0581 T 1.62 10 tα. n = 10 (a) x = 26.0000 StDev 1.00234 (c) In strength tests.828) = 1 − 0. and conclude that the mean life exceeds 25 h.0. Reject H0 if t0 > tα. H1: µ > 25.952 S n 1.833 Reject H0: µ = 25. 10−1 = 1.62.707 125.042 3-2 .05 Test H0: µ = 25 vs.000 0.83 P 0. (d) ( n) ≤ µ 127 − 1.6248 SE Mean 0.002 3-3.

94 MTB > Stat > Basic Statistics > 1-Sample t > Samples in columns One-Sample T: Ex3-3 Test of mu = 25 vs not = 25 Variable N Mean StDev Ex3-3 10 26.62 10 ) 25. 3-3 .6248 SE Mean 0.625 10 0.62 n ≤ µ ≤ x + tα / 2.0 − 1.0581.114 0.986 Percent 80 70 60 50 40 30 20 10 5 1 20 22 24 26 28 Lifetime.833 1.5138 90% CI (25.Chapter 3 Exercise Solutions 3-3 continued (b) α = 0. 26.833 1. so the assumption that battery life is normally distributed is appropriate. Hours 30 32 The plotted points fall approximately along a straight line.95 P 0.95% CI 99 Mean StDev N AD P-Value 95 90 26 1.06 ≤ µ ≤ 26.10 x − tα / 2.0 + 1.0000 1.9419) T 1.n −1 S ( 26.083 (c) MTB > Graph > Probability Plot > Single Probability Plot of Battery Service Life (Ex3-3) Normal .n −1 S ) n ( 10 ≤ µ ≤ 26.

n −1 S ( 13. α = 0.39216 ≤ µ ≤ 13. 9 = 2.39618 − 3. x = 13.3962 0. x ~ N(µ. 13.3922.09 ) ( n) 10 ) ≤ µ ≤ 13. α = 0.00391 10 tα/2.0039 0. H1: µ ≠ 13.05.3962 0.2498 ( 0.39618 − 13.4.05 Test H0: µ = 13. and conclude that the mean thickness differs from 13.0012 99% CI (13. (a) x ~ N(µ.62 h.025. σ).089 S n 0.4 × 1000 Å.0 h. Reject H0 if |t0| > tα/2.4 Variable N Mean StDev SE Mean Ex3-5 10 13. 13.00391 P 0.00391 95% CI (13.4 vs.013 n ≤ µ ≤ x + tα / 2. s = 0.01 ( x − tα / 2.05.09 P 0.0039 0. x − µ0 13.4 vs not = 13.4 t0 = = = −3.n −1 S ( 26. MTB > Stat > Basic Statistics > 1-Sample t > Samples in columns One-Sample T: Ex3-5 Test of mu = 13.3990) T -3.n −1 S 10 ) 13. x = 26. n−1 = t0.833 ( ) 10 ) ≤ µ n ≤µ x − tα . n−1 = t0.013 3-4 .40020 MTB > Stat > Basic Statistics > 1-Sample t > Samples in columns One-Sample T: Ex3-5 Test of mu = 13. 3-5.Chapter 3 Exercise Solutions 3-4. tα. s = 1.9 = 1.833 1. n = 10.62 25.4 × 1000 Å.0 − 1.39618 + 3.4 vs not = 13.262 Reject H0: µ = 13.3934.4002) T -3.2498 0. σ).4.00391 µ0 = 13.06 ≤ µ The manufacturer might be interested in a lower confidence interval on mean battery life when establishing a warranty policy.0012 (b) α = 0.39618 × 1000 Å.4 Variable N Mean StDev SE Mean Ex3-5 10 13. n = 10.

015 − 12 t0 = = = 1.400 13. α = 0. n−1 = t0.005.395 13. 9 = 3.015. 3-6.0150 StDev 0. σ).0096 99% Lower Bound 11. MTB > Stat > Basic Statistics > 1-Sample t > Samples in columns One-Sample T: Ex3-6 Test of mu = 12 vs > 12 Variable Ex3-6 N 10 Mean 12.95% CI 99 Mean StDev N AD P-Value 95 90 13.Chapter 3 Exercise Solutions 3-5 continued (c) MTB > Graph > Probability Plot > Single Probability Plot of Photoresist Thickness (Ex3-5) Normal .076 3-5 . and conclude that there is not enough evidence that the mean fill volume exceeds 12 oz. H1: µ > 12.405 Thickness.410 The plotted points form a reverse-“S” shape. x − µ0 12.01 n = 10.385 13.0303 SE Mean 0. x = 12.5655 S n 0.40 0.380 13.390 13. x1000 Angstroms 13.711 Percent 80 70 60 50 40 30 20 10 5 1 13.030 Test H0: µ = 12 vs. s = 0.250 Do not reject H0: µ = 12. so the assumption that battery life is normally distributed is not appropriate.9880 T 1. (a) x ~ N(µ.57 P 0.003909 10 0. Reject H0 if t0 > tα.237 0.0303 10 tα/2. instead of a straight line. µ0 = 12.

15 The plotted points fall approximately along a straight line.62 ( S 10 ) 11.262 ( S 10 ) ≤ µ ≤ 12. 9 = 2. n−1 = t0.025.03028 10 0.0096 95% CI (11.993 ≤ µ ≤ 12. so the assumption that fill volume is normally distributed is appropriate.00 12.9600. α = 0. σ = 4 lb.05 Fill Volume.90 11.9600 4 ⎣ ) ) n ⎤ = total width ⎦ n ⎤ =1 ⎦ n = 246 3-6 .57 P 0. 12. find n ( ( 2 ⎡ Zα / 2 σ ⎣ 2 ⎡1.037 MTB > Stat > Basic Statistics > 1-Sample t > Samples in columns One-Sample T: Ex3-6 Test of mu = 12 vs not = 12 Variable N Mean StDev Ex3-6 10 12.025 = 1.262 x − tα / 2.015 + 2.274 0.0150 0.152 (c) MTB > Graph > Probability Plot > Single Probability Plot of Fill Volume (Ex3-6) Normal . total confidence interval width = 1 lb.95% CI 99 Mean StDev N AD P-Value 95 90 12.Chapter 3 Exercise Solutions 3-6 continued (b) α = 0.582 Percent 80 70 60 50 40 30 20 10 5 1 11.0303 SE Mean 0. Zα/2 = Z0. 3-7.0367) T 1. ounces 12.05.02 0.95 12.10 12.9933.015 − 2.n −1 S n ≤ µ ≤ x + tα / 2.n −1 S ( ) ( n) 12.05 tα/2.

504600 0.2498 95% CI (9.025.999 16 tα/2. 15 = 2.00 P 0. σ = 0. n = 16.0001 N Mean SE Mean 95% CI 25 0.9990 SE Mean 0.5025.5025 vs. Reject H0 if |t0| > tα/2. and conclude that the mean rod diameter differs from 0.Chapter 3 Exercise Solutions 3-8.5025. x −µ 0 = 0. α = 0.7270.05 n = 25.999 V (a) µ0 = 12.131 Reject H0: µ = 12.5025 The assumed standard deviation = 0. s = 0.5046 + 1.2594 0.0001 n ≤ µ ≤ x + Zα / 2 σ 25 ) 0. x − µ0 10. x = 10. 0. (a) x ~ N(µ. x = 0.971 S n 0.259 V.5046 − 0. x ~ N(µ.5025 = 105 Z0 = σ n 0.50464 3-9.504561.000 3-7 . MTB > Stat > Basic Statistics > 1-Sample Z > Summarized data One-Sample Z Test of mu = 0.05 Test H0: µ = 12 vs. 10.0001 ) ( n) 25 ) ≤ µ ≤ 0.000020 (0.5025 vs not = 0. MTB > Stat > Basic Statistics > 1-Sample t > Samples in columns One-Sample T: Ex3-9 Test of mu = 12 vs not = 12 Variable N Mean StDev Ex3-9 16 10.0001 25 Zα/2 = Z0.504639) Z 105.5025. µ0 = 0.50456 ≤ µ ≤ 0.7917) T -6.97 P 0. H1: µ ≠ 12.025 = 1.05/2 = Z0.5025.259 − 12 t0 = = = −6. Reject H0 if |Z0| > Zα/2. α = 0. σ). and conclude that the mean output voltage differs from 12V.5046 in.0001 in Test H0: µ = 0.5046 − 1. H1: µ ≠ 0.960 0.96 Reject H0: µ = 0.960 ( 0. n−1 = t0. σ).000 (b) P-value = 2[1 − Φ(Z0)] = 2[1 − Φ(105)] = 2[1 − 1] = 0 (c) ( x − Zα / 2 σ ( 0.

16−1 = 6. n-1 or χ20 < χ 21-α/2. (16 − 1)0.131 0.546 Since the 95% confidence interval on σ contains the hypothesized value. H1: σ 2 ≠ 1. α = 0. n−1 = χ20.488 χ21−α/2.262 2 0.05 Test H0: σ 2 = 1 vs.9992 2 ≤σ ≤ 27.9992 = 14. n-1.n −1 S ( 10.16−1 = 27. and conclude that there is insufficient evidence that the variance differs from 1.025.975. cannot be rejected. H0: σ 2 = 1.738 ≤ σ ≤ 1. σ 02 = 1.545 ≤ σ ≤ 2. Reject H0 if χ 20 > χ 2α/2.n −1 ≤σ2 ≤ (n − 1) S 2 χ2 1−α / 2.970 1 σ 02 χ2α/2. n −1 (16 − 1)0.391 2 0.792 (c) σ 02 = 1.259 + 2. the null hypothesis.262 Do not reject H0: σ 2 = 1.n −1 S 16 ) 9.131( 0. 3-8 .999 (16 − 1)0.999 ) ( n) 16 ) ≤ µ ≤ 10. χ 02 = (n − 1) S 2 = (d) (n − 1) S 2 χα2 / 2.999 n ≤ µ ≤ x + tα / 2.Chapter 3 Exercise Solutions 3-9 continued (b) ( x − tα / 2.259 − 2.488 6. n−1 = χ20.727 ≤ µ ≤ 10.

259 0.492793 16 M inimum 1st Q uartile M edian 3rd Q uartile M aximum 12 8.546 Mean Median 9.95.998 0.533 10.75 10.23 0.05.00 10.999 0.150 12.000 95% C onfidence Interv al for M ean 9.2609 α = 0. χ12−α .140 11.792 95% C onfidence Interv al for M edian 9.370 9.15 σ2 ≤ (n − 1) S 2 χ12−α .00 (e) 2 = 7.062 σ ≤ 1.50 10.2609 2 σ ≤ 2.116487 -0.n −1 (16 − 1)0.75 11.727 10.436 2 3-9 .430 10.738 1.767 M ean S tD ev V ariance S kew ness Kurtosis N 10.n −1 = χ 0.25 10.50 9.Chapter 3 Exercise Solutions 3-9 (d) continued MTB > Stat > Basic Statistics > Graphical Summary Summary for Output Voltage (Ex3-9) A nderson-D arling N ormality Test 8 9 10 11 A -S quared P -V alue 0.945 95% C onfidence Interv al for S tD ev 95% Confidence Intervals 0.9992 σ ≤ 7.

767 Percent 80 70 60 50 40 30 20 10 5 1 7 8 9 10 11 Output Voltage 12 13 14 From visual examination of the plot.05/2 = Z0. σ2 = 0.0152 20 σ 1 n1 + σ 2 n2 Zα/2 = Z0.010 l. (b) P-value = 2[1 − Φ(Z0)] = 2[1 − Φ(−7.07) − 0 Z0 = = = −7. n2 = 20.025 = 1. x2 = 2. 3-10. the assumption of a normal distribution for output voltage seems appropriate. (a) α = 0. n1 = 25.04 − 2.04 l.015 l.682)] = 2[1 − 1.Chapter 3 Exercise Solutions 3-9 continued (f) MTB > Graph > Probability Plot > Single Probability Plot of Output Voltage (Ex3-9) Normal .95% CI 99 Mean StDev N AD P-Value 95 90 10.682 2 2 0.26 0.00000] = 0 3-10 .9990 16 0.05. ∆ 0 = 0 Test H0: µ1 – µ2 = 0 versus H0: µ1 – µ2 ≠ 0.230 0. Reject H0 if Z0 > Zα/2 or Z0 < –Zα/2. x1 = 2.96 Reject H0: µ1 – µ2 = 0. σ1 = 0.96 −Zα/2 = −1.0102 25 + 0.07 l. ( x1 − x2 ) − ∆ 0 (2. and conclude that there is a difference in mean net contents between machine 1 and machine 2.

n + n 1 2 −2 S p 1 n1 + 1 n2 (1. n1+n2−2 = t0.376.n + n 1 2 −2 S p 1 n1 + 1 n2 ≤ ( µ1 − µ 2 ) ≤ ( x1 − x2 ) + tα / 2. (b) The practical implication of this test is that it does not matter which technician measures parts. 13 = 2.1604(0.383 − 1.mu (Ex3-11T2) Estimate for difference: 0.044 Difference = mu (Ex3-11T1) . 3-11 .127969.1252 = = 0.022 25 + 0.9600 0. n2 = 8. the readings will be the same.Chapter 3 Exercise Solutions 3-10 continued (c) ( x1 − x2 ) − Zα / 2 2 (2. We can conclude that there is no difference in measurements obtained by the two technicians.9600 0.383 − 1.917 DF = 13 Do not reject H0: µ1 – µ2 = 0.1604(0.115.376 0. we would have been concerned that the technicians obtained different measurements.010 20 −0.05.376) − 2. and an investigation should be undertaken to understand why.141183) T-Test of difference = 0 (vs not =): T-Value = 0.115 0. S1 = 0.11 Both use Pooled StDev = 0.07) − 1.043 Ex3-11T2 8 1.006607 95% CI for difference: (-0.120) 1 7 + 1 8 ≤ ( µ1 − µ 2 ) ≤ (1.04 − 2. Ex3-11T2 Two-sample T for Ex3-11T1 vs Ex3-11T2 N Mean StDev SE Mean Ex3-11T1 7 1. S2 = 0.120 n1 + n2 − 2 7+8−2 ( x1 − x2 ) − tα / 2.127 ≤ ( µ1 − µ 2 ) ≤ 0.125 0. x2 = 1.010 σ 12 25 n1 + σ 22 + 0.038 ≤ ( µ1 − µ 2 ) ≤ −0. 3-11. x1 = 1. (c) n1 = 7.1204 P-Value = 0. (a) MTB > Stat > Basic Statistics > 2-Sample t > Samples in different columns Two-Sample T-Test and CI: Ex3-11T1. We can conclude that the machines do not fill to the same volume.141 The confidence interval for the difference contains zero.07) + 1. tα/2.1152 + (8 − 1)0.04 − 2.025.383.015 2 20 The confidence interval for the difference does not contain zero.1604 Sp = (n1 − 1) S12 + (n2 − 1) S22 (7 − 1)0.125 α = 0. 0.383 0.376) + 2. If the null hypothesis had been rejected.120) 1 7 + 1 8 −0. and conclude that there is not sufficient evidence of a difference between measurements obtained by the two technicians.015 2 n2 σ 12 ≤ ( µ1 − µ 2 ) ≤ ( x1 − x2 ) + Zα / 2 n1 + σ 22 n2 2 ≤ ( µ1 − µ 2 ) ≤ (2.

01 0.975.05/ 2.Chapter 3 Exercise Solutions 3-11 continued (d) α = 0.n −1.3 1.8−1 = F0.8−1 = F0.6.85 0.n 1 2 −1 2 or F0 < F1−α / 2.n −1.20 0.7 = 0. 1 2 F0 = S S = 0.5 1. we would have been concerned about the difference in measurement variability between the technicians. Reject H 0 if F0 > Fα / 2. 3-12 .7 −1.025.920 0. If the null hypothesis is rejected. Ex3-11T2 F-Test Test Statistic P-Value Ex3-11T1 0.25 95% Bonferroni Confidence Intervals for StDevs 0.854 Lev ene's Test Test Statistic P-Value Ex3-11T2 0.n −1 = F0.05 Test H 0 : σ 12 = σ 22 versus H1 : σ 12 ≠ σ 22 .2 1.15 0.4 Data 1.30 Ex3-11T1 Ex3-11T2 1. and an investigation should be undertaken to understand why.8464 Fα / 2.7 −1.n −1.176 MTB > Stat > Basic Statistics > 2 Variances > Summarized data Test for Equal Variances for Ex3-11T1.7 = 5. and conclude that there is no difference in variability of measurements obtained by the two technicians.6 Do not reject H0.125 = 0.n −1.10 0.n 1 2 −1 = F1−0.n −1.115 0.05/ 2.119 2 1 1 2 2 2 2 F1−α / 2.6.

2 1.6899 α = 0.125 2 2 = 16. S2 = 0. x2 = 1. n1 −1 2 1 S2 S σ σ 12 0. χα2 / 2.1148 7 0.n2 −1 = χ 0.6955) ≤ ≤ 0.065 (g) MTB > Graph > Probability Plot > Multiple Probability Plot of Surface Finish by Technician (Ex3-11T1.0128 1.821 σ 22 (f) n2 = 8.n −1 ≤σ2 ≤ (n − 1) S 2 χ12−α / 2.n −1 = χ 0. Fα / 2.025.7 = 1.1954) (5.0128.n −1.376.1152 0.693 70 60 50 40 30 20 10 5 1 1.05 F1−α / 2.8 The normality assumption seems reasonable for these readings.6 = 0.943 0.975.1252 (8 − 1)0.6 1. Ex3-11T2) Normal .6 = 5.7.n 2 −1.1954.975.007 ≤ σ ≤ 0.0 1.1252 2 ≤σ ≤ 16.6955 S12 σ S F Fα / 2.383 1. n2 −1.Chapter 3 Exercise Solutions 3-11 continued (e) α = 0.7 2 (n − 1) S 2 χα2 / 2.025.4 Data 1.n −1 ≤ ≤ 2 1−α / 2.142 0.165 ≤ σ 12 ≤ 4.376 Percent 80 StDev N AD P 0.7.n −1 (8 − 1)0.1252 σ 22 0. 3-13 .05. n1 −1 = F0.1152 (0.1249 8 0. χ12−α / 2.1252 0. n1 −1 2 2 1 1 2 2 2 2 = F0.95% CI 99 Variable Ex3-11T1 Ex3-11T2 95 90 Mean 1.235 0.n 2 −1.6899 2 0.

97 1.46 (a) Assume σ 12 = σ 22 MTB > Stat > Basic Statistics > 2-Sample t > Samples in different columns Two-Sample T-Test and CI: Ex3-13SQ. and conclude that there is no difference between the quenching processes. From Eqn. for σ 12 ≠ σ 22 and both unknown.ν S12 n1 + S 22 n2 3-13. 3.6 − 149. S1 = 4. S2 = 5. x1 = 147.1009(5. Saltwater quench: n1 = 10.22) 1 10 + 1 10 ≤ ( µ1 − µ 2 ) ≤ (147.77 Both use Pooled StDev = 5.1 3-14 .97 2 + (10 − 1)5.ν S12 n1 + S 22 n2 ≤ ( µ1 − µ2 ) ≤ ( x1 − x2 ) + tα / 2. (b) α = 0. x2 = 149.7 Difference = mu (Ex3-13SQ) .40 5.80000 95% CI for difference: (-6.97 Oil quench: n2 = 10.22) 1 10 + 1 10 −6. Ex3-13OQ Two-sample T for Ex3-13SQ vs Ex3-13OQ N Mean StDev SE Mean Ex3-13SQ 10 147.n + n 1 2 −2 S p 1 n1 + 1 n2 ≤ ( µ1 − µ 2 ) ≤ ( x1 − x2 ) + tα / 2.Chapter 3 Exercise Solutions 3-12.mu (Ex3-13OQ) Estimate for difference: -1. tα/2.4) − 2.6 − 149.025.10615) T-Test of difference = 0 (vs not =): T-Value = -0.05. the test statistic is 2 2 2 + S n S n x − x 1 1 2 2 1 2 with degrees of freedom ν = −2 t0* = 2 2 2 2 2 S1 n1 + S 22 n2 S1 n1 S 2 n2 + ( n1 + 1) ( n2 + 1) ( ( ) ) ( ) A 100(1-α)% confidence interval on the difference in means would be: ( x1 − x2 ) − tα / 2. n1+n2−2 = t0.4.46 1. 3-54 and 3-55.70615.22 10 + 10 − 2 n1 + n2 − 2 ( x1 − x2 ) − tα / 2.n + n 1 2 −2 S p 1 n1 + 1 n2 (147.6 Ex3-13OQ 10 149.6.1009(5.2217 P-Value = 0.60 4.4) + 2.451 DF = 18 Do not reject H0.7 ≤ ( µ1 − µ 2 ) ≤ 3.1009 Sp = (n1 − 1) S12 + (n2 − 1) S 22 (10 − 1)4. 18 = 2.462 = = 5.

n1 −1 2 1 2 α / 2.462 0. 3-15 .Chapter 3 Exercise Solutions 3-13 continued (c) α = 0. n1 −1 = F0. (d) MTB > Graph > Probability Plot > Multiple Probability Plot of Quench Hardness (Ex3-13SQ.9.906 90 Percent 80 70 60 50 40 30 20 10 5 1 130 140 150 Hardness 160 170 The normal distribution assumptions for both the saltwater and oil quench methods seem reasonable.6 4.05 F1−α / 2.97 2 4.0260 S12 σ 12 S ≤ ≤ F F − α n − n − 1 / 2.2484) (4.169 0.2484. 2 1. 1 1 σ 22 S S22 σ 12 4.21 ≤ σ 12 ≤ 3.9.025. the assumption of equal variances seems reasonable. Fα / 2.461 10 0.34 σ 22 Since the confidence interval includes the ratio of 1.95% CI 99 Variable Ex3-13SQ Ex3-13OQ 95 AD P Mean StDev N 147. n1 −1 2 2 −1. Ex3-13OQ) Normal .462 5.218 0.9 = 4.9 2 −1.779 149.97 2 ≤ ≤ (0.0260) σ 22 5.971 10 0. n2 −1.4 5.n 0.975.n =F = 0.

pˆ = x/n = 18/200 = 0. P-value = 2[1 − Φ|Z0|] = 2[1 − Φ|−0.6382] = 0. (b) α = 0.10) Zα/2 = Z0.05. x = 18.3536 np0 (1 − p0 ) 20(1 − 0. 0.09 − 1.3536|] = 2[1 − 0.123 3-16 .09) 200 ≤ p ≤ 0. use the normal approximation to the binomial for x < np0.057 ≤ p ≤ 0.09) 200 0.05 = 1.645 pˆ − Zα / 2 pˆ (1 − pˆ ) n ≤ p ≤ pˆ + Zα / 2 pˆ (1 − pˆ ) n 0.05/2 = Z0.10/2 = Z0.09 (a) p0 = 0.7236 MTB > Stat > Basic Statistics > 1 Proportion > Summarized data Test and CI for One Proportion Test of p = 0. Test H0: p = 0.1 vs p not = 0.47 P-Value 0.10.5) − np0 (18 + 0. n = 200.10.129662) Z-Value -0.645 0.96 Do not reject H0. α = 0. Zα/2 = Z0.10) = 20 Since (x = 18) < (np0 = 20).050338. Z0 = ( x + 0.5) − 20 = = −0.09(1 − 0.10.637 Note that MINITAB uses an exact method.Chapter 3 Exercise Solutions 3-14. not an approximation. Reject H0 if |Z0| > Zα/2.645 0.1 Sample X N Sample p 95% CI 1 18 200 0.090000 (0.09 + 1.10.10 versus H1: p ≠ 0.025 = 1. np0 = 200(0.09(1 − 0. and conclude that the sample process fraction nonconforming does not differ from 0.

pˆ = x/n = 65/500 = 0.155 3-17 .08) = 40 Since (x = 65) > (np0 = 40).5) − np0 (65 − 0. np0 = 500(0. MTB > Stat > Basic Statistics > 1 Proportion > Summarized data Test and CI for One Proportion Test of p = 0.05 = 1.130000 (0.08 versus H1: p ≠ 0.08.05/2 = Z0.13 + 1.08) Zα/2 = Z0.99997] = 0.025 = 1.645 p ≤ pˆ + Zα pˆ (1 − pˆ ) n p ≤ 0. not an approximation. Reject H0 if |Z0| > Zα/2.13) 500 p ≤ 0.5) − 40 Z0 = = = 4.08.12 P-Value 0.645 0. (b) P-value = 2[1 − Φ|Z0|] = 2[1 − Φ|4.Chapter 3 Exercise Solutions 3-15.05. n = 500. Test H0: p = 0. and conclude the sample process fraction nonconforming differs from 0. Zα = Z0. x = 65.159478) Z-Value 4. ( x − 0. α = 0.130 (a) p0 = 0.000 Note that MINITAB uses an exact method.13(1 − 0.08. 0.0387 np0 (1 − p0 ) 40(1 − 0.08 vs p not = 0.00006 (c) α = 0.08 Sample X N Sample p 95% CI 1 65 500 0.96 Reject H0.05. use the normal approximation to the binomial for x > np0.0387|] = 2[1 − 0.100522.

066667 Difference = p (1) .067) + 1.06 n1 + n2 200 + 300 Z0 = pˆ1 − pˆ 2 pˆ (1 − pˆ ) (1 n1 + 1 n2 ) 0.0580079.025 = 1.645 0.06) (1 200 + 1 300 ) = −0.067 (b) Use α = 0.96 Zα/2 = Z0. 0.052 ≤ ( p1 − p2 ) ≤ 0.05 n2 = 300.05/2 = Z0.067) 300 −0.05.06(1 − 0.0246745) Test for difference = 0 (vs not = 0): Z = -0.050 − 0.p (2) Estimate for difference: -0.067(1 − 0.05 − 0. x1 = 10.7842 −Zα/2 = −1. pˆ 2 = x2/n2 = 20/300 = 0.96 Do not reject H0.05(1 − 0.05) 200 + 0.77 P-Value = 0. (a) n1 = 200.067) − 1.018 3-18 .0166667 95% CI for difference: (-0.645 0. MTB > Stat > Basic Statistics > 2 Proportions > Summarized data Test and CI for Two Proportions Sample X N Sample p 1 10 200 0.Chapter 3 Exercise Solutions 3-16.050000 2 20 300 0. pˆ1 = x1/n1 = 10/200 = 0.05 − 0. Test H0: p1 = p2 versus H1: p1 ≠ p2.442 (c) ( pˆ 1 − pˆ 2 ) − Z α / 2 pˆ 1 (1 − pˆ 1 ) + pˆ 2 (1 − pˆ 2 ) n1 n2 ≤ ( p1 − p2 ) ≤ ( pˆ 1 − pˆ 2 ) + Z α / 2 (0.067 = 0.05(1 − 0.067(1 − 0.067) 300 pˆ 1 (1 − pˆ 1 ) + pˆ 2 (1 − pˆ 2 ) n1 n2 ≤ ( p1 − p2 ) ≤ (0. x2 = 20. Conclude there is no strong evidence to indicate a difference between the fraction nonconforming for the two processes.05) 200 + 0. Reject H0 if Z0 > Zα/2 or Z0 < –Zα/2 pˆ = x1 + x2 10 + 20 = = 0.

10.0987 F0 = 1.60576 5.18 = 2.n2 − 2 = F0.8232 and > 0.975. x2 = 8.85 − 8. S 22 = 6.n2 −1 Fα / 2.* before: n1 = 10.69405 F-Test (normal distribution) Test statistic = 1.n1+n2−2.24710 2 8 1.79 + (8 − 1) 6.05.05 Reject H 0 if F0 > Fα / 2.48596 5. 3-19 .85.025. 10+8−2 = 1.9.34856 T-Test of difference = 0 (vs >): T-Value = 1.082 DF = 16 The mean impurity after installation of the new purification unit is not less than before. F1−α / 2.n1 −1. so do not reject H 0 MTB > Stat > Basic Statistics > 2 Variances > Summarized data Test for Equal Variances 95% Bonferroni confidence intervals for standard deviations Sample N Lower StDev Upper 1 10 1.9.70449 2. tα.7 = 0.18 (a) Test H 0 : σ 12 = σ 22 versus H1 : σ 12 ≠ σ 22 .8232.05.77000 95% lower bound for difference: -0.79 6.922 The impurity variances before and after installation are the same.08.n1 −1.18 = 1.n1+n2−2 = t0.mu (2) Estimate for difference: 1.55525 2.83 2 8 8.79 after: n2 = 8.Chapter 3 Exercise Solutions 3-17.49 0.n2 − 2 or F0 < F1−α / 2. S12 = 6.88 Difference = mu (1) . (b) Test H0: µ1 = µ2 versus H1: µ1 > µ2.0987 < 4.n1 −1.5582 P-Value = 0. x1 = 9. p-value = 0.08 2.n1 −1.461 1 n1 + 1 n2 2.554 10 + 8 − 2 x1 − x2 9.746 SP = t0 = SP ( n1 − 1) S12 + ( n2 − 1) S22 n1 + n2 − 2 = (10 − 1) 6.554 1 10 + 1 8 MTB > Stat > Basic Statistics > 2-Sample t > Summarized data Two-Sample T-Test and CI Sample N Mean StDev SE Mean 1 10 9.n2 −1 = F0. α = 0.46 Both use Pooled StDev = 2. at α = 0.08 = = 1.2383.85 2.2383 F0 = S12 S 22 = 6.7 = 4.61 0. Reject H0 if t0 > tα.

69978 T-Test of difference = -5 (vs <): T-Value = -0. so H1: µ1 + 5 < µ2. σ1 = σ2 = 3. n2 = 16. n1 = 16.05 = −1. x1 = 175. test H0: µ1 − µ2 = − 5 versus H1: µ1 − µ2 < − 5.8 psi. The P-values are close due to the sample sizes. the Z-value is the same as the t-value.0000 P-Value = 0.4714 Z0 = 32 16 + 32 16 σ 12 n1 + σ 22 n2 (Z0 = −0.8 − 181. P-value = Φ(Z0) = Φ(−0.3 psi.00 0.75 2 16 181.Chapter 3 Exercise Solutions 3-18.50000 95% upper bound for difference: -3.3187 MTB > Stat > Basic Statistics > 2-Sample t > Summarized data Two-Sample T-Test and CI Sample N Mean StDev SE Mean 1 16 175.30 3. The mean strength of Design 2 does not exceed Design 1 by 5 psi.3) − (−5) = −0. so do not reject H0.80 3. So test a difference ∆0 = −5. Reject H0 if Z0 < − Zα .4714) > −1.00 0.645 x x − ( 1 2 ) − ∆ 0 = (175.4714) = 0. 3-20 . ∆0 = −5 −Zα = −Z0.75 Difference = mu (1) .mu (2) Estimate for difference: -5.47 Both use Pooled StDev = 3.320 DF = 30 Note: For equal variances and sample sizes.645. x2 = 181.0 psi Want to demonstrate that µ2 is greater than µ1 by at least 5 psi.

000379 95% CI for mean difference: (-0.10 (|t0| = 1.000417) T-Test of mean difference = 0 (vs not = 0): T-Value = -1. n1 + n2 − 2. There is no strong evidence to indicate that the two calipers differ in their mean measurements.10) < 2.8188.005.8188 d=1 ( n∑ j =1 n xMicrometer. j ) = 1 ⎡⎣( 0.001250.22 = 2. Test H0: µd = 0 versus H1: µd ≠ 0.295 3-21 .000468 Difference 12 -0. MTB > Stat > Basic Statistics > Paired t > Samples in Columns Paired T-Test and CI: Ex3-19MC.151 − 0. 0. so do not reject H0.000835 0.001311 12 = −1. n1 + n2 − 2 = t0. j − xVernier.152 ) ⎤⎦ = −0.150 − 0.000417 12 ⎛ n ⎞ − d ⎜∑dj ⎟ ∑ j =1 ⎝ j =1 ⎠ S d2 = ( n − 1) n 2 j t0 = d (S d ) 2 n = 0. Ex3-19VC Paired T for Ex3-19MC .151583 0.000417 0.151167 0.000417 0.000241 Ex3-19VC 12 0.10 P-Value = 0. tα/2. Reject H0 if |t0| > tα/2.151) + " + ( 0.001311 0.Chapter 3 Exercise Solutions 3-19.Ex3-19VC N Mean StDev SE Mean Ex3-19MC 12 0.001621 0.0013112 ( ) n = −0.

500 SE Mean 2.52 32. (n − 1) S 2 χα2 / 2.254 T 1.025. χ21-α/2.571 95% Lower Bound 149.05 Test H0: µ = 150 versus H1: µ > 150.7 − 150 ) 11.975.19 = 8. There is insufficient evidence to indicate that the mean strength is greater than 150 psi.4389 ( ) ( ) (t0 = 1. (a) The alternative hypothesis H1: µ > 150 is preferable to H1: µ < 150 we desire a true mean weld strength greater than 150 psi. n-1 = χ20. µ > 150.52 ⎤⎦ 1 = 42.083 3-21.80 1. so do not reject H0.85 ≤ σ 2 ≤ (20 − 1)1.1170 χ 02 = ⎡⎣(n − 1) S 2 ⎤⎦ σ 02 = ⎡⎣(20 − 1)1. n-1 = χ20. In order to achieve this result.19 = 1.85. s = 11. The standard deviation of the fill volume is not less than 1ml. α = 0. n −1.14 ≤ σ ≤ 2.5. tα. (b) χ2α/2. χ21-α.n −1 ≤ σ 2 ≤ (n − 1) S 2 χ12−α / 2. x = 752. α = 0.5. Reject H0 if t0 > tα. x = 153.7291.95. n-1 = χ20.19 3-22 .6 ml.Chapter 3 Exercise Solutions 3-20. (b) n = 20. s = 1.91 1.7291.91. Reject H0 if χ20 < χ21-α.19 = 32. so do not reject H0.75 > 10.75 χ20 = 42.n −1 (20 − 1)1. H0 must be rejected in favor of the alternative H1.4389) < 1.19 = 10.30 ≤ σ 2 ≤ 4. t0 = ( x − µ ) S n = (153.05.1170.44 P 0.52 8. n-1. MTB > Stat > Basic Statistics > 1-Sample t > Summarized data One-Sample T Test of mu = 150 vs > 150 N 20 Mean 153.05 (a) Test H0: σ2 = 1 versus H1: σ2 < 1.7.700 StDev 11.5 20 = 1. n −1 = t0. n = 20.

511 0.00 753.37 0.00 751.25 753.538 20 0. so the assumption that battery life is normally distributed is not appropriate.27 95% C onfidence Interv al for M edian 752.172 M ean S tD ev V ariance S kew ness Kurtosis N 752.8 753.95% CI 99 Mean StDev N AD P-Value 95 90 752.172 Percent 80 70 60 50 40 30 20 10 5 1 748 750 752 754 Fill Volume. ml 756 758 The plotted points do not fall approximately along a straight line. 3-23 .Chapter 3 Exercise Solutions 3-21 (b) continued MTB > Stat > Basic Statistics > Graphical Summary Summary for Pinot Gris Fill Volume.00 753.2 (c) MTB > Graph > Probability Plot > SIngle Probability Plot of Pinot Gris Fill Volume (Ex3-21) Normal . ml (Ex3-21) A nderson-D arling N ormality Test 749 750 751 752 753 754 755 A -S quared P -V alue 0.4 752.6 1.00 95% C onfidence Interv al for S tD ev 95% Confidence Intervals 1.55 1.0 752.00 756.191843 20 M inimum 1st Q uartile M edian 3rd Q uartile M aximum 756 750.25 Mean Median 752.54 2.17 2.83 753.51 0.281321 0.00 95% C onfidence Interv al for M ean 751.

6667 From Figure 3-7.9 0.2933) = 0. Check: β = Φ Zα / 2 − δ n σ − Φ − Zα / 2 − δ n σ = Φ 1. the operating characteristic curve for two-sided at α = 0. µ0 = 15. So.3733) − Φ (−5. 3-46.0. β. σ2 = 9.10? δ = µ1 − µ2 = 20 − 15 = 5 d = δ σ = 5 9 = 1. µ1 = 20.915181 3-23.96 − 5 4 3 − Φ −1.0848 − 0.Chapter 3 Exercise Solutions 3-22. What n is needed such that the Type II error. Let µ1 = µ0 + δ. ( β ≈ Φ Zα / 2 − δ n σ ( ) Φ ( β ) ≈ Φ −1 Zα / 2 − δ n σ ) − Z β ≈ Zα / 2 − δ n σ n ≈ ⎡⎣( Zα / 2 + Z β )σ δ ⎤⎦ 2 3-24 . α = 0. β = Φ Zα / 2 − δ n σ − Φ − Zα / 2 − δ n σ ( ) ( ) ( ) If δ > 0. Test H0: µ = 15 versus H1: µ ≠ 15.05 Assumed standard deviation = 3 Sample Target Difference Size Power Actual Power 5 4 0. then Φ − Zα / 2 − δ n σ is likely to be small compared with β. From Eqn. n = 4.96 − 5 4 3 ( ) ( ) ( ) ( ) = Φ (−1.05. is less than or equal to 0.0000 = 0.0848 MTB > Stat > Power and Sample Size > 1-Sample Z Power and Sample Size 1-Sample Z Test Testing mean = null (versus not = null) Calculating power for mean = null + difference Alpha = 0.05.

n1 . x1 − x2 Maximize: Z 0 = Subject to: n1 + n2 = N . 3-25. x1 . Assume µ1 = 2µ2 and let Q = ( x1 − x2 ) . x2 . σ 1 n1 + σ 2 n2 2 2 Since ( x1 − x2 ) is fixed. reject H 0 if Z 0 > Zα / 2 3-25 . Given x ~ N . x1 independent of x2 . n2 .Chapter 3 Exercise Solutions 3-24. E (Q) = E ( x1 − 2 x2 ) = µ1 − 2µ2 = 0 var(Q) = var( x1 − 2 x2 ) = var( x1 ) + var(2 x2 ) = var( x1 ) + 22 var( x2 ) = Z0 = var( x1 ) var( x2 ) +4 n1 n2 Q−0 x1 − 2 x2 = SD(Q) σ 12 n1 + 4 σ 22 n2 And. an equivalent statement is Minimize: L = σ 12 n1 + σ 22 n2 = σ 12 n1 + σ 22 N − n1 σ 22 ⎞ dL ⎡ −1 2 dL ⎛ σ 12 −1 + n1 σ 1 + ( N − n1 ) σ 22 ⎤ ⎜ ⎟= ⎦ dn1 ⎝ n1 N − n1 ⎠ dn1 ⎣ = −1n1−2σ 12 + (−1)(−1) ( N − n1 ) σ 22 = 0 −2 =− σ 12 n12 + σ 22 ( N − n1 ) 2 =0 n1 σ 1 = n2 σ 2 Allocate N between n1 and n2 according to the ratio of the standard deviations.

Reject H0: λ = λ0 if |Z0| > Zα/2 (b) x ~ Poi(λ).15 100 = −1. x = x/n = ~ N(λ. λ/n).110 Test H0: λ = 0. x = 3.05. Reject H0 if |Z0| > Zα/2. so reject H0. 3-27. n = 1000. Reject H0 if |Z0| > Zα. n = 5.645.5 ) 0.6 Test H0: λ = 0. so do not reject H0. n = 100. Z 0 = ( x − λ ) λ0 / n . x ~ Poi(λ).15 versus H1: λ ≠ 0. so do not reject H0. xn. x = x/N = 11/100 = 0. x = 688. if n is large.3162) < 1.05. Zα/2 = Z0.01.0328) < 2.6 − 0.15 ) 0.Chapter 3 Exercise Solutions 3-26.3162 (Z0 = 0.005 = 2. …. Zα = Z0. Reject H0 if Z0 > Zα. at α = 0. at α = 0.05 = 1.645 Z 0 = ( x − λ0 ) λ0 n = ( 0. x = 11.110 − 0. x ~ Poi(λ). i =1 Using the normal approximation to the Poisson. x = x/N = 688/1000 = 0. x2. 3-26 .96.688 Test H0: λ = 1 versus H1: λ ≠ 1. x = x/N = 3/5 = 0. Each xi ~ POI(λ). Zα/2 = Z0.5 5 = 0.5. (a) Wish to test H0: λ = λ0 versus H1: λ ≠ λ0.96 Z 0 = ( x − λ0 ) λ0 n = ( 0.5758 Z 0 = ( x − λ0 ) λ0 n = ( 0.688 − 1) 1 1000 = −9. n ∑x i ~ POI( nλ ) .5758.5 versus H1: λ > 0. at α = 0.025 = 1.15.8663) > 1.0328 (|Z0| = 1. 3-28. Select random sample of n observations x1.8663 (|Z0| = 9.

7132 (F0.Chapter 3 Exercise Solutions 3-29.0 4.80 Pooled StDev = 0.05.15 = 3.5231 (----------*---------) 200 6 3. so flow rate does not affect etch uniformity at a significance level α = 0. Boxplots of data MTB > Graph > Boxplot > One Y.00 3.6823) > (F0 = 3. the P-value is just slightly greater than 0.05.053 15 7. so there is some evidence that gas flow rate affects the etch uniformity.5 125 160 C2F6 Flow (SCCM) 200 Gas flow rate of 125 SCCM gives smallest mean percentage uniformity. (a) MTB > Stat > ANOVA > One-Way One-way ANOVA: Ex3-29Obs versus Ex3-29Flow Source Ex3-29Flow Error Total S = 0. 3-27 .34% R-Sq(adj) = 23.4167 0.5 4.9333 0.7600 (---------*----------) 160 6 4.32% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev -----+---------+---------+---------+---125 6 3.59).648 1.509 17 11.05.59 0.0 3. With Groups Boxplot of Etch Uniformity by C2F6 Flow Etch Uniformity (%) 5.8214 (----------*---------) -----+---------+---------+---------+---3.0 2.278 R-Sq = 32.3167 0.630 0.7132 DF SS MS F P 2 3.60 4.824 3. (b) MTB > Stat > ANOVA > One-Way > Graphs. However.2.20 4.5 3.

4 Residuals are satisfactory.0 4.0 0.4 3. 3-28 .8 Fitted Value 4.5 -1.5 Residual 1.0 -0.2 4. Normal plot of residuals Normal Probability Plot of the Residuals (response is Etch Uniformity (Ex3-29Obs)) 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 -2 -1 0 Residual 1 2 The normality assumption is reasonable.0 3.Chapter 3 Exercise Solutions 3-29 continued (c) MTB > Stat > ANOVA > One-Way > Graphs.6 3.2 3.5 0. (d) MTB > Stat > ANOVA > One-Way > Graphs. Residuals versus fits Residuals Versus the Fitted Values (response is Etch Uniformity (Ex3-29Obs)) 1.

0 3 .5087 6 = 0. 125 and 200.5 4 .3 3 .3 S c a le d t D is t r ib u t io n (2 0 0 ) (1 2 5 ) 3 .6 3 .8 M e a n E t c h U n if o r m it y The graph does not indicate a large difference between the mean etch uniformity of the three different flow rates.Chapter 3 Exercise Solutions 3-30. The statistically significant difference between the mean uniformities can be seen by centering the t distribution between.9% MSE n = 0.9 (1 6 0 ) 4 . 3-29 . say.2 4 . Flow Rate 125 160 200 scale factor = Mean Etch Uniformity 3. and noting that 160 would fall beyond the tail of the curve.4% 3.3% 4.

05.5 No difference due to rodding level at α = 0. Normal plot of residuals One-way ANOVA: Ex3-31Str versus Ex3-31Rod Source Ex3-31Rod Error Total S = 71.0 (-----------*----------) 15 3 1586.0 10.53 DF SS MS F P 3 28633 9544 1.87 0.7 77.9 (-----------*-----------) 25 3 1500.0 52.214 8 40933 5117 11 69567 R-Sq = 41. (b) Boxplot of Compressive Strength by Rodding Level 1750 Compressive Strength 1700 1650 1600 1550 1500 1450 1400 10 15 20 25 Rodding Level Level 25 exhibits considerably less variability than the other three levels.7 (-----------*-----------) 20 3 1606. 3-30 . (a) MTB > Stat > ANOVA > One-Way > Graphs> Boxplots of data.7 107.09% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev ----+---------+---------+---------+----10 3 1500.16% R-Sq(adj) = 19.Chapter 3 Exercise Solutions 3-31.0 (-----------*----------) ----+---------+---------+---------+----1440 1520 1600 1680 Pooled StDev = 71.

Chapter 3 Exercise Solutions 3-31 continued (c) Normal Probability Plot of the Residuals (response is Compressive Strength (Ex3-31Str)) 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 -150 -100 -50 0 Residual 50 100 150 The normal distribution assumption for compressive strength is reasonable. 3-31 .

Chapter 3 Exercise Solutions 3-32. 25 ) (1 5 ) (2 0 ) 1418 1459 1500 1541 1582 1623 1664 M e a n C o m p re s s iv e S tr e n g th There is no difference due to rodding level. 3-32 . Rodding Level 10 15 20 25 scale factor = Mean Compressive Strength 1500 1587 1607 1500 MSE n = 5117 3 = 41 S c a le d t D is tr ib u tio n (10 .

333 0.89% R-Sq(adj) = 5.75 42.324 Temperature level does not significantly affect mean baked anode density. (b) Normal Probability Plot of the Residuals (response is Baked Density (Ex3-33Den)) 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 -0.097 0.152 1.2 0.457 0.6 0.8 Normality assumption is reasonable.700 0.194 (----------*----------) 550 6 41.0 Residual 0.583 0.8 -0.4 -0.50 41.2 0.Chapter 3 Exercise Solutions 3-33.258 2.339 (----------*----------) 575 6 41.4 0.3238 SS MS F P 0.553 R-Sq = 17.497 (----------*----------) --------+---------+---------+---------+41.25 41.6 -0. 3-33 .00 Pooled StDev = 0.57% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev --------+---------+---------+---------+500 6 41.45 0.141 (----------*----------) 525 6 41. Normal plot of residuals One-way ANOVA: Ex3-33Den versus Ex3-33T Source DF Ex3-33T 3 Error 20 Total 23 S = 0.450 0.105 2. (a) MTB > Stat > ANOVA > One-Way > Graphs> Boxplots of data.

MTB > Stat > ANOVA > One-Way > Graphs> Residuals versus the Variables Residuals Versus Firing Temperature (Ex3-33T) (response is Baked Density (Ex3-33Den)) 0.00 Baked Desnity 41. Recommend 500°C for smallest variability.50 -0. 3-34 .75 41.25 41.75 40.50 Residual 0.50 41.50 500 525 550 Firing Temperature.Chapter 3 Exercise Solutions 3-33 continued (c) Boxplot of Baked Density by Firing Temperature 42. select the temperature with the smallest variance.00 -0.00 40. which probably also incurs the smallest cost (lowest temperature). More uniform anodes are produced at lower temperatures. 3-34. deg C 575 Since statistically there is no evidence to indicate that the means are different.25 -0.25 0. so does variability. 500°C (see Boxplot).75 500 510 520 530 540 550 Temperature (deg C) 560 570 580 As firing temperature increases.

000 1.0 77.309 (---*---) 0.711 DF SS MS F P 5 1133.51 0.304 (----*---) 1.85 0.65% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev ----+---------+---------+---------+----0.0 84.754 (---*---) ----+---------+---------+---------+----63.40 1. (a) MTB > Stat > ANOVA > One-Way > Graphs> Boxplots of data One-way ANOVA: Ex3-35Rad versus Ex3-35Dia Source Ex3-35Dia Error Total S = 2.750 2.35 23 1265.Chapter 3 Exercise Solutions 3-35.0 Pooled StDev = 2.99 4 62. at α = 0.062 (---*---) 0.4 orifice diameters.711 Orifice size does affect mean % radon release.559 (---*---) 1.71 4 75.99 and 1.99 Smallest % radon released at 1.750 2. 3-35 .68 30.37 4 82. Boxplot of Radon Released by Orifice Diameter Radon Released.000 3.000 18 132.51 4 77.0 70.40 4 65.38 226.63 R-Sq = 89.55% R-Sq(adj) = 86.02 4 71.37 0.25 7. % (Ex3-35Rad) 85 80 75 70 65 60 0.02 Orifice Diameter 1.05.000 2.826 (---*---) 1.750 3.71 1.

0 Residual 2.5 0.Chapter 3 Exercise Solutions 3-35 continued (b) MTB > Stat > ANOVA > One-Way > Graphs> Normal plot of residuals.0 -2. 3-36 . Residuals Versus Orifice Diameter (Ex3-35Dia) (response is Radon Released (Ex3-35Rad)) 5. 1.5 0.0 Residual 2.40.37 0.02.5 -5. Residuals versus the Variables Normal Probability Plot of the Residuals (response is Radon Released ( Ex3-35Rad)) 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 -5.99).0 0.0 Residuals violate the normality distribution.5 -5.0 -2.99 The assumption of equal variance at each factor level appears to be violated.0 Residual 2. 1.0 60 65 70 75 Fitted Value--Radon Released 80 85 Variability in residuals does not appear to depend on the magnitude of predicted (or fitted) values. with larger variances at the larger diameters (1.71 1.5 5.0 -2.5 0.40 Orifice Diameter (Ex3-35Dia) 1.51 0. Residuals Versus the Fitted Values (response is Radon Released (Ex3-35Rad)) 5. Residuals versus fits.02 1.

4066 − 0. Boxplot of Uniformity by Wafer Position Film Thickness Uniformity (Ex3-36Un) 6 5 4 3 2 1 1 (b) σˆτ2 = 2 3 Wafer Position (Ex3-36Pos) 4 MSfactor − MSE 5.6522 2 σˆ uniformity = σˆτ2 + σˆ 2 = 0. 3. Boxplots of data One-way ANOVA: Ex3-36Un versus Ex3-36Pos Source DF SS MS F P Ex3-36Pos 3 16.5 3.66% R-Sq(adj) = 66.0 Pooled StDev = 0.6522 = = 0.3962 + 0.652 Total 11 21. (a) MTB > Stat > ANOVA > One-Way > Graphs.4636 (------*------) 2 3 1.3570 (------*------) --------+---------+---------+---------+1.407 8.217 0.7733 0.0 4.008 Error 8 5.8076 There is a statistically significant difference in wafer position.3067 1.9267 0.53% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev --------+---------+---------+---------+1 3 4.5 6.Chapter 3 Exercise Solutions 3-36.4366 (------*------) 4 3 1.3962 n 12 (c) σˆ 2 = MSE = 0.6522 = 1.437 S = 0.220 5.8076 R-Sq = 75.0484 3-37 .3853 (------*------) 3 3 1. and 4.29 0. 1 is different from 2.3167 0.

0 4.0 Residual 0.0 3. Residuals versus the Variables Normal Probability Plot of the Residuals (response is Uniformity (Ex3-36Un)) 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 -2 -1 0 Residual 1 2 Normality assumption is probably not unreasonable. Residuals versus fits. but there are two very unusual observations – the outliers at either end of the plot – therefore model adequacy is questionable. 3-38 .5 1.5 0.Chapter 3 Exercise Solutions 3-36 continued (d) MTB > Stat > ANOVA > One-Way > Graphs> Normal plot of residuals.5 Fitted Value--Film Thickness Uniformity 4.0 1.5 The variability in residuals does appear to depend on the magnitude of predicted (or fitted) values.5 -1. Residuals Versus the Fitted Values (response is Uniformity ( Ex3-36Un)) 1.5 2.5 0.0 2.0 -0.5 1.5 1.0 -0.0 Residual 0.0 -1.5 1 2 3 Wafer Position (Ex3-36Pos) 4 Both outlier residuals are from wafer position 1.5 -1.5 3.0 -1. Residuals Versus Wafer Position (Ex3-36Pos) (response is Film Thickness Uniformity (Ex3-36Un)) 1.

**Chapter 4 Exercise Solutions
**

Several exercises in this chapter differ from those in the 4th edition. An “*” following the

exercise number indicates that the description has changed. New exercises are denoted

with an “☺”. A second exercise number in parentheses indicates that the exercise

number has changed.

4-1.

“Chance” or “common” causes of variability represent the inherent, natural variability of

a process - its background noise. Variation resulting from “assignable” or “special”

causes represents generally large, unsatisfactory disturbances to the usual process

performance. Assignable cause variation can usually be traced, perhaps to a change in

material, equipment, or operator method.

A Shewhart control chart can be used to monitor a process and to identify occurrences of

assignable causes. There is a high probability that an assignable cause has occurred when

a plot point is outside the chart's control limits. By promptly identifying these

occurrences and acting to permanently remove their causes from the process, we can

reduce process variability in the long run.

4-2.

The control chart is mathematically equivalent to a series of statistical hypothesis tests. If

a plot point is within control limits, say for the average x , the null hypothesis that the

mean is some value is not rejected. However, if the plot point is outside the control

limits, then the hypothesis that the process mean is at some level is rejected. A control

chart shows, graphically, the results of many sequential hypothesis tests.

**NOTE TO INSTRUCTOR FROM THE AUTHOR (D.C. Montgomery):
**

There has been some debate as to whether a control chart is really equivalent to

hypothesis testing. Deming (see Out of the Crisis, MIT Center for Advanced

Engineering Study, Cambridge, MA, pp. 369) writes that:

“Some books teach that use of a control chart is test of hypothesis: the process is

in control, or it is not. Such errors may derail self-study”.

Deming also warns against using statistical theory to study control chart behavior (falsealarm probability, OC-curves, average run lengths, and normal curve probabilities.

Wheeler (see “Shewhart’s Charts: Myths, Facts, and Competitors”, ASQC Quality

Congress Transactions (1992), Milwaukee, WI, pp. 533–538) also shares some of these

concerns:

“While one may mathematically model the control chart, and while such a model

may be useful in comparing different statistical procedures on a theoretical basis,

these models do not justify any procedure in practice, and their exact

probabilities, risks, and power curves do not actually apply in practice.”

4-1

**Chapter 4 Exercise Solutions
**

4-2 continued

On the other hand, Shewhart, the inventor of the control chart, did not share these views

in total. From Shewhart (Statistical Method from the Viewpoint of Quality Control

(1939), U.S. Department of Agriculture Graduate School, Washington DC, p. 40, 46):

“As a background for the development of the operation of statistical

control, the formal mathematical theory of testing a statistical hypothesis

is of outstanding importance, but it would seem that we must continually

keep in mind the fundamental difference between the formal theory of

testing a statistical hypothesis and the empirical theory of testing a

hypothesis employed in the operation of statistical control. In the latter,

one must also test the hypothesis that the sample of data was obtained

under conditions that may be considered random. …

The mathematical theory of distribution characterizing the formal and

mathematical concept of a state of statistical control constitutes an

unlimited storehouse of helpful suggestions from which practical criteria

of control must be chosen, and the general theory of testing statistical

hypotheses must serve as a background to guide the choice of methods of

making a running quality report that will give the maximum service as

time goes on.”

Thus Shewhart does not discount the role of hypothesis testing and other aspects of

statistical theory. However, as we have noted in the text, the purposes of the control

chart are more general than those of hypothesis tests. The real value of a control chart is

monitoring stability over time. Also, from Shewhart’s 1939 book, (p. 36):

“The control limits as most often used in my own work have been set so that after

a state of statistical control has been reached, one will look for assignable causes

when they are not present not more than approximately three times in 1000

samples, when the distribution of the statistic used in the criterion is normal.”

Clearly, Shewhart understood the value of statistical theory in assessing control chart

performance.

My view is that the proper application of statistical theory to control charts can provide

useful information about how the charts will perform. This, in turn, will guide decisions

about what methods to use in practice. If you are going to apply a control chart

procedure to a process with unknown characteristics, it is prudent to know how it will

work in a more idealized setting. In general, before recommending a procedure for use in

practice, it should be demonstrated that there is some underlying model for which it

performs well. The study by Champ and Woodall (1987), cited in the text, that shows the

ARL performance of various sensitizing rules for control charts is a good example. This

is the basis of the recommendation against the routine use of these rules to enhance the

ability of the Shewhart chart to detect small process shifts.

4-2

**Chapter 4 Exercise Solutions
**

4-3.

Relative to the control chart, the type I error represents the probability of concluding the

process is out of control when it isn't, meaning a plot point is outside the control limits

when in fact the process is still in control. In process operation, high frequencies of false

alarms could lead could to excessive investigation costs, unnecessary process adjustment

(and increased variability), and lack of credibility for SPC methods.

The type II error represents the probability of concluding the process is in control, when

actually it is not; this results from a plot point within the control limits even though the

process mean has shifted out of control. The effect on process operations of failing to

detect an out-of-control shift would be an increase in non-conforming product and

associated costs.

4-4.

The statement that a process is in a state of statistical control means that assignable or

special causes of variation have been removed; characteristic parameters like the mean,

standard deviation, and probability distribution are constant; and process behavior is

predictable. One implication is that any improvement in process capability (i.e., in terms

of non-conforming product) will require a change in material, equipment, method, etc.

4-5.

No. The fact that a process operates in a state of statistical control does not mean that

nearly all product meets specifications. It simply means that process behavior (mean and

variation) is statistically predictable. We may very well predict that, say, 50% of the

product will not meet specification limits! Capability is the term, which refers to the

ability to meet product specifications, and a process must be in control in order to

calculate capability.

4-6.

The logic behind the use of 3-sigma limits on Shewhart control charts is that they give

good results in practice. Narrower limits will result in more investigations for assignable

causes, and perhaps more false alarms. Wider limits will result in fewer investigations,

but perhaps fewer process shifts will be promptly identified.

Sometimes probability limits are used - particularly when the underlying distribution of

the plotted statistic is known. If the underlying distribution is unknown, care should be

exercised in selecting the width of the control limits. Historically, however, 3-sigma

limits have been very successful in practice.

4-3

**Chapter 4 Exercise Solutions
**

4-7.

Warning limits on control charts are limits that are inside the control limits. When

warning limits are used, control limits are referred to as action limits. Warning limits,

say at 2-sigma, can be used to increase chart sensitivity and to signal process changes

more quickly than the 3-sigma action limits. The Western Electric rule, which addresses

this type of shift is to consider a process to be out of control if 2 of 3 plot points are

between 2 sigma and 3 sigma of the chart centerline.

4-8.

The concept of a rational subgroup is used to maximize the chance for detecting variation

between subgroups. Subgroup samples can be structured to identify process shifts. If it

is expected that a process will shift and stay at the new level until a corrective action,

then sampling consecutive (or nearly) units maximizes the variability between subgroups

and minimizes the variability within a subgroup. This maximizes the probability of

detecting a shift.

4-9.

I would want assignable causes to occur between subgroups and would prefer to select

samples as close to consecutive as possible. In most SPC applications, process changes

will not be self-correcting, but will require action to return the process to its usual

performance level. The probability of detecting a change (and therefore initiating a

corrective action) will be maximized by taking observations in a sample as close together

as possible.

4-10.

This sampling strategy will very likely underestimate the size of the true process

variability. Similar raw materials and operating conditions will tend to make any fivepiece sample alike, while variability caused by changes in batches or equipment may

remain undetected. An out-of-control signal on the R chart will be interpreted to be the

result of differences between cavities. Because true process variability will be

underestimated, there will likely be more false alarms on the x chart than there should

be.

4-4

**Chapter 4 Exercise Solutions
**

4-11.

(a)

No.

(b)

The problem is that the process may shift to an out-of-control state and back to an incontrol state in less than one-half hour. Each subgroup should be a random sample of all

parts produced in the last 2½ hours.

4-12.

No. The problem is that with a slow, prolonged trend upwards, the sample average will

tend to be the value of the 3rd sample --- the highs and lows will average out. Assume

that the trend must last 2½ hours in order for a shift of detectable size to occur. Then a

better sampling scheme would be to simply select 5 consecutive parts every 2½ hours.

4-13.

No. If time order of the data is not preserved, it will be impossible to separate the

presence of assignable causes from underlying process variability.

4-14.

An operating characteristic curve for a control chart illustrates the tradeoffs between

sample size n and the process shift that is to be detected. Generally, larger sample sizes

are needed to increase the probability of detecting small changes to the process. If a large

shift is to be detected, then smaller sample sizes can be used.

4-15.

The costs of sampling, excessive defective units, and searches for assignable causes

impact selection of the control chart parameters of sample size n, sampling frequency h,

and control limit width. The larger n and h, the larger will be the cost of sampling. This

sampling cost must be weighed against the cost of producing non-conforming product.

4-16.

Type I and II error probabilities contain information on statistical performance; an ARL

results from their selection. ARL is more meaningful in the sense of the operations

information that is conveyed and could be considered a measure of the process

performance of the sampling plan.

4-5

**Chapter 4 Exercise Solutions
**

4-17.

Evidence of runs, trends or cycles? NO. There are no runs of 5 points or cycles. So, we

can say that the plot point pattern appears to be random.

4-18.

Evidence of runs, trends or cycles? YES, there is one "low - high - low - high" pattern

(Samples 13 – 17), which might be part of a cycle. So, we can say that the pattern does

not appear random.

4-19.

Evidence of runs, trends or cycles? YES, there is a "low - high - low - high - low" wave

(all samples), which might be a cycle. So, we can say that the pattern does not appear

random.

4-20.

Three points exceed the 2-sigma warning limits - points #3, 11, and 20.

4-21.

Check:

• Any point outside the 3-sigma control limits? NO.

• 2 of 3 beyond 2 sigma of centerline? NO.

• 4 of 5 at 1 sigma or beyond of centerline? YES. Points #17, 18, 19, and 20 are

outside the lower 1-sigma area.

• 8 consecutive points on one side of centerline? NO.

One out-of-control criteria is satisfied.

4-22.

Four points exceed the 2-sigma warning limits - points #6, 12, 16, and 18.

4-23.

Check:

• Any point outside the 3-sigma control limits? NO. (Point #12 is within the lower

3-sigma control limit.)

• 2 of 3 beyond 2 sigma of centerline? YES, points #16, 17, and 18.

• 4 of 5 at 1 sigma or beyond of centerline? YES, points #5, 6, 7, 8, and 9.

• 8 consecutive points on one side of centerline? NO.

Two out-of-control criteria are satisfied.

4-6

**Chapter 4 Exercise Solutions
**

4-24.

The pattern in Figure (a) matches the control chart in Figure (2).

The pattern in Figure (b) matches the control chart in Figure (4).

The pattern in Figure (c) matches the control chart in Figure (5).

The pattern in Figure (d) matches the control chart in Figure (1).

The pattern in Figure (e) matches the control chart in Figure (3).

4-25 (4-30).

Many possible solutions.

MTB > Stat > Quality Tools > Cause-and-Effect

**Cause-and-Effect Diagram for Late Arrival
**

Driv e

Family

"Turtle"

Children/School

Route

Put out pet

Accident

Children/Homework

**Find badge, keys
**

Fix breakfast

A rriv e late to

O ffice

Errands

Fix lunch

Eat breakfast

Carpool

Read paper

Dress

Gas

Shower

Get up late

Activ ities

Coffee

Stops

4-7

**Chapter 4 Exercise Solutions
**

4-26 (4-31).

Many possible solutions.

MTB > Stat > Quality Tools > Cause-and-Effect

**Cause-and-Effect Diagram for Car Accident
**

Driv er

Car

A sleep

Tires

Drunk

Brakes

M isjudgment

S uspension

**Talking on cell phone
**

D istracted

S teering

S tate of Repair

Raining

Blocked

P oor v isibility

Windy

W eather

Out-of-contr

ol car strikes

tree

Icy /snow -cov ered

Road

4-8

**Chapter 4 Exercise Solutions
**

4-27 (4-32).

Many possible solutions.

MTB > Stat > Quality Tools > Cause-and-Effect

**Cause-and-Effect Diagram for Damaged Glassware
**

Glassware

Glassware

Pack aging

**Deliv ery Serv ice
**

Handling

Crushed

Not enough padding

Strength flaw

Dropped

Severe transport vibration

Dropped

Crushed

Internal Handling

Weak box

Broken at start

Glassware

Damaged

Droppped

Carelessly packed

Manufacturer Handling

4-9

Many possible solutions.Chapter 4 Exercise Solutions 4-28☺. MTB > Stat > Quality Tools > Cause-and-Effect Cause-and-Effect Diagram for Coffee-making Process Personnel Method Machine Worn-out C offee drinkers C leanliness E spresso drinkers Brew temperature Insufficient training A ge of brew Brew method C onsistently Bad C offee Ty pe of filter C offee grind A mount of w ater C offee roast C offee beans Water temperature Env ironment A mount of beans Measurement Water source Material 4-10 .

Example of a check sheet to collect data on personal opportunities for improvement. Many possible solutions. The next step would be to determine the causes of “Being Rude” and to work on eliminating those causes.9 g in et C m om itm t en 19 13.8 81.0 To reduce total count of defects. beginning and end of process are shown below. 4-11 . Defect Overeating Being Rude Not meeting commitments Missing class Etc.Chapter 4 Exercise Solutions 4-29☺. Yellow is nonvalue-added activity.3 100. Snooze No Check time Awake Yes 6:30am ? Get out of bed Arrive at work … 4-31☺. 1 2 0 2 10 11 4 2 4 6 3 1 9 2 3 4 0 9 2 2 Month/Day 5 6 7 1 0 1 7 10 11 1 0 1 7 9 4 TOTAL 18 21 15 13 16 19 … … … … … 31 TOTAL 1 6 9 76 7 19 2 37 17 19 138 Pareto Chart of Personal Opportunities for Improvement 140 100 120 Count 80 60 60 40 40 20 20 0 Defect g in Be Count Percent Cum % Percent 80 100 de Ru M 76 55.1 55. Many possible solutions.1 g in iss s as Cl e tM No 37 26. including defect categories and counts. green is value-added activity.7 r he Ot 0 6 4. “Being Rude” represents the greatest opportunity to make an improvement.8 95.

0027) = 1 − 0.0027)5 = 1 − 0.0027) = 0.986573 m = 10 ⎛10 ⎞ 0 10 ⎟ (0.0526 ⎝0⎠ α1 = 1 − Pr{0 of 20 beyond} = 1 − ⎜ Cumulative Distribution Function Binomial with n = 20 and p = 0.0027) = 0.0027) (1 − 0.0027 x P( X <= x ) 0 0.0027 x P( X <= x ) 0 0.0027 x P( X <= x ) 0 0.0267 ⎝0⎠ α1 = 1 − Pr{0 of 10 beyond} = 1 − ⎜ Cumulative Distribution Function Binomial with n = 10 and p = 0.947363 m = 30 ⎛ 30 ⎞ 0 30 ⎟ (0.0134 ⎝0⎠ MTB > Calc > Probability Distributions > Binomial.922093 m = 50 ⎛ 50 ⎞ 0 50 ⎟ (0.0027) = 0.873556 Although the probability that a single point plots beyond the control limits is 0.0027)0 (1 − 0.0779 0 ⎝ ⎠ α1 = 1 − Pr{0 of 30 beyond} = 1 − ⎜ Cumulative Distribution Function Binomial with n = 30 and p = 0.Chapter 4 Exercise Solutions 4-32☺.0027 x P( X <= x ) 0 0.9733 = 0. the probability that at least one of the points is beyond the limits also increases.1025 ⎝0⎠ α1 = 1 − Pr{0 of 50 beyond} = 1 − ⎜ Cumulative Distribution Function Binomial with n = 50 and p = 0.0027) (1 − 0. m=5 α1 = Pr{at least 1 out-of-control} = Pr{1 of 5 beyond} + Pr{2 of 5 beyond} + " + Pr{5 of 5 beyond} ⎛5⎞ = 1 − Pr{0 of 5 beyond} = 1 − ⎜ ⎟ (0.973326 m = 20 ⎛ 20 ⎞ 0 20 ⎟ (0.0027) (1 − 0. as the number of samples increases (m). 4-12 . Cumulative Probability Cumulative Distribution Function Binomial with n = 5 and p = 0.9866 = 0.0027) (1 − 0.0027.0027 x P( X <= x ) 0 0.

In fact. The lack of independence in the sample statistics will affect the estimates of the process population parameters. sampling frequencies are often designed to increase the likelihood of detecting a special or assignable cause. When the process mean µ and variance σ2 are unknown. they must be estimated by sample means x and standard deviations s. However. 4-13 .Chapter 4 Exercise Solutions 4-33☺. the points used to estimate these sample statistics are not independent—they do not reflect a random sample from a population.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Sample No.00 + 0.71) = 31.2 x= 1 2 = = 34.577(4. The MINITAB convention for determining whether a point is out of control is: (1) if a plot point is within the control limits. it is in control.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Sample No.71) = 36.71 LCL = 31.00 R x-bar 35. MINITAB uses pooled standard deviation to estimate standard deviation for control chart limits and capability estimates.96 UCL = 36.00 − 0.577(4. or (2) if a plot point is on or beyond the limits.00 R chart for Bearing ID (all samples in calculations) X-bar Chart for Bearing ID (all samples in calculations) 41.0 12 39.71 LCL R = D3 R = 0(4.71 m 24 UCL x = x + A2 R = 34.29 UCL R = D4 R = 2. (a) for n = 5.5 + 34.00 LCL x = x − A2 R = 34.0 2 29.0 CL = 4. This can be changed under Tools > Options > Control Charts and Quality Tools > Define Tests. 4. 5-1. MINITAB defines some sensitizing rules for control charts differently than the standard rules. In particular. New exercises are denoted with an “☺”.Chapter 5 Exercise Solutions Notes: 1.2 + " + 34.29 4 31.114.577. Several exercises in this chapter differ from those in the 4th edition.0 UCL = 9. An “*” indicates that the description has changed.00 m 24 R + R2 + " + Rm 3 + 4 + " + 2 R= 1 = = 4.0 6 33. it is out of control.115(4. A2 = 0. 5-1 .0 12 10 15 37. 3. not 8.71) = 9. This can be changed in dialog boxes or under Tools>Options>Control Charts and Quality Tools>Estimating Standard Deviation.72 CL x = x = 34. D4 = 2. D3 = 0 x + x + " + xm 34. 2.71) = 0. a run of n consecutive points on one side of the center line is defined as 9 points.72 8 CL = 34.96 CL R = R = 4.0 LCL = 0 0 27. A second exercise number in parentheses indicates that the exercise number has changed.

50 31.0 12 10 UCL = 9. The new process parameter estimates are: x = 33.00050 5-2 .0 LCL = 31. σˆ x = R / d 2 = 4.Chapter 5 Exercise Solutions 5-1 (a) continued The process is not in statistical control.CL R = 4.29) = 0 + 1 − 0. 15. 12 and Sample No.65. Assuming an assignable cause is found for these two out-of-control points.65 R x-bar 35. R = 4.25 8 CL = 33.CL x = 33.326 = 1.0 CL = 4.5 / 2.65. the two samples can be excluded from the control limit calculations.65 ⎞ ⎤ = Φ⎜ ⎟ + ⎢1 − Φ ⎜ ⎟⎥ ⎝ 1.07) + 1 − Φ (3.93 UCL x = 36.0 12 39.0 6 33. 15 excluded) 41. (b) pˆ = Pr{x < LSL} + Pr{x > USL} = Pr{x < 20} + Pr{x > 40} = Pr{x < 20} + [1 − Pr{x < 40}] ⎛ 20 − 33.25.5.0 4 2 27.52 15 37. x is beyond the upper control limit for both Sample No.0 UCL =36. 15 excluded) R chart for Bearing ID (samples 12.06 29.65 ⎞ ⎡ ⎛ 40 − 33.93 ⎠ ⎣ ⎝ 1.99950 = 0.52.00 x-bar Chart for Bearing ID (samples 12.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Sample No.06 UCL R = 9.5. LCL R = 0.93 ⎠ ⎦ = Φ (−7. LCL x = 31. 0 LCL = 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Sample No.

23 * 42 Exp. With xi = (observed voltage on unit i – 350) × 10: USLT = +50. LSLT = –50 USL − LSL +50 − (−50) Cˆ P = = = 5.00 PPM Total 0. 6σˆ 6(3. σˆ X = R / d 2 = 6.25.44 4.0 U C L=14.00 PPM > USL 0. x = 10.36 4. Overall Performance PPM < LSL 0.25 / 2.Chapter 5 Exercise Solutions 5-2. (b) n = 4.88 12. so the process is capable. or cycles.0 7.49 CPL 6. Actual specs are 350 ± 5 V.49 .035) MTB > Stat > Quality Tools > Capability Analysis > Normal Process Capability Analysis of Ex5-2V LSL LSL Target USL Within Ov erall Process Data -50.5 LC L=5.36 5.00 -14 0 14 28 5.00 PPM > USL 0. Within Performance PPM < LSL 0.25 4 0 LC L=0 2 4 6 8 10 Sample 12 14 16 18 20 The process is in statistical control with no out-of-control signals.23 4. runs. trends.03545 3. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Ex5-2V Sample M ean 15.49 Overall Capability Pp PPL PPU Ppk Cpm -42 Observed Performance PPM < LSL 0. R = 6.77 5.00000 10.00 PPM > USL 0.0 2 4 6 8 10 Sample 12 14 16 18 20 16 Sample Range U C L=14.32500 80 3.12282 CPU Cpk CCpk 4.26 12 8 _ R=6.035 .00000 * USL Sample Mean Sample N StDev(Within) StDev(Overall) Potential (Within) Capability Cp 5.059 = 3.00 PPM Total 0.62 50.00 -28 Exp.00 PPM Total 0.5 _ _ X=10.33 10.33.00 5-3 .34 6.

runs.9 0 -20 LC L=-25.1 0 5 10 Ex5-2V 15 20 A normal probability plot of the transformed output voltage shows the distribution is close to normal.Chapter 5 Exercise Solutions 5-2 continued (c) MTB > Stat > Basic Statistics > Normality Test Probability Plot of Ex5-2V Normal 99. trends. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Ex5-3Dia U C L=47.113 80 0.3 100 _ R=63.5 50 LC L=0 0 2 4 6 8 10 Sample 12 14 16 18 20 The process is in statistical control with no out-of-control signals. or cycles.53 Sample M ean 40 20 _ _ X=10. 5-3. 5-4 .73 2 4 6 8 10 Sample 12 14 16 18 20 150 Sample Range U C L=134.33 3.704 0.064 Percent 80 70 60 50 40 30 20 10 5 1 0.9 Mean StDev N AD P-Value 99 95 90 10.

48 5-5 .17 * 90 Exp.00000 * 100. Within Performance PPM < LSL 24.00 -60 Exp.90000 100 27. so the process is capable.00000 10.326 = 27.30 PPM > USL 549.35 1.46 1.3) MTB > Stat > Quality Tools > Capability Analysis > Normal Process Capability Analysis of Ex5-3Dia LSL USL Within Ov erall Process Data LSL Target USL Sample Mean Sample N StDev(Within) StDev(Overall) -100.09 1.Chapter 5 Exercise Solutions 5-3 continued (b) σˆ x = R / d 2 = 63.81 PPM > USL 213.3 (c) USL = +100.17 1.5 / 2.32 1.22 Overall Capability Pp PPL PPU Ppk Cpm -90 Observed Performance PPM < LSL 0. Cˆ P = 6σˆ x 6(27.22 .30009 Potential (Within) Capability Cp 1.79 PPM Total 574. LSL = –100 USL − LSL +100 − (−100) = = 1.00 PPM Total 0.09 -30 0 30 60 1.09 1.29384 1. Overall Performance PPM < LSL 5.22 CPL CPU Cpk CCpk 25.67 PPM Total 219.00 PPM > USL 0.

Chapter 5 Exercise Solutions 5-4. at points: 15 If graph is updated with new data. Test Failed at points: 22 TEST 5. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Thickness (Ex5-4Th) Sample M ean 0.0024 U C L=0.0018 0.00 standard deviations from center line.0640 U C L=0. 5-6 .00092 0.00 standard deviations from center line.0630 0.062952 0. Test Failed at points: 22 Test Results for R Chart of Ex5-4Th TEST 1.0000 LC L=0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 Test Results for Xbar Chart of Ex5-4Th TEST 1.0006 0.0635 _ _ X=0. One Test Failed * WARNING * * point more than 3.0620 LC L=0. One point more than 3.062011 1 2 4 6 8 10 12 14 Sample 16 18 20 22 24 1 Sample Range 0.0012 _ R=0.0625 0.063893 0. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). the results above may no longer be correct.002368 0.

22) and recalculate control limits.000823 0. 14 removed from control limits calculations) 0.000486) = [0.063787 0. 0. Assuming assignable causes are found.693 = 0.06295 ± 3(0. remove the out-of-control points (samples 15.0000 LCL=0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 (b) σˆ x = R / d 2 = 0.0625 LCL=0.062945 0. sample 14 is also out-of-control on the x chart.0012 _ R=0. 22.0635 _ _ X=0.0024 Sample Range UCL=0.0006 0. With the revised limits.002118 0.061492.000486 (c) Natural tolerance limits are: x ± 3σˆ x = 0. Removing all three samples from calculation.0640 1 Sample Mean UCL=0.Chapter 5 Exercise Solutions 5-4 continued The process is out-of-control.0630 0.064408] 5-7 .0018 0.000823 /1.062104 0. the new control limits are: Xbar-R Chart of Thickness (Ex5-4Th) (Samples 15. failing tests on both the x and the R charts.0620 1 2 4 6 8 10 12 14 Sample 16 18 20 22 24 1 0.

06150 Target * USL 0.88 5-8 .028 Cˆ P = 6σˆ x 6(0. 15.0624 0.70 PPM Total 2157.0640 0.0628 0.52 Exp.0015) = = 1.0644 Observed Performance PPM < LSL 15151.0616 0.0620 0. Within Performance PPM < LSL 1467. 15.0015 − (−0. Overall Performance PPM < LSL 3419.61 PPM > USL 689. and excluding samples 14.06450 Sample Mean 0.0632 0.Chapter 5 Exercise Solutions 5-4 continued (d) Assuming that printed circuit board thickness is normally distributed.90 * 0.55 PPM Total 5233.52 PPM > USL 0.000486) MTB > Stat > Quality Tools > Capability Analysis > Normal Process Capability Analysis of Thickness (Ex5-4Th_w/o) (Estimated without Samples 14.90 0.33 PPM > USL 1814.00 PPM Total 15151.07 Cpk 0.31 Exp.99 CPU 1.00049 StDev(Overall) 0. 22) LSL USL Within Ov erall Process Data LSL 0.03 CPL 0.99 CCpk 1.94 0.0636 0.03 Overall Capability Pp PPL PPU Ppk Cpm 0.97 0. and 22 from the process capability estimation: USL − LSL +0.00053 Potential (Within) Capability Cp 1.06295 Sample N 66 StDev(Within) 0.

5 LC L=0. There is no difference in interpretation from the x − s chart.0 Sample StDev U C L=1.5 -1. (b) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R (Ex5-5Vol) Xbar-R Chart of Fill Volume (Ex5-5Vol) Sample M ean 1.983 0.0 LC L=-1.0 1 2 3 4 5 6 7 8 Sample 9 10 11 12 13 14 15 The process is in statistical control. trends. trends.0 -0.5 LC L=0.5 _ _ X=-0.003 0. Xbar-S Chart of Fill Volume (Ex5-5Vol) U C L=1.Chapter 5 Exercise Solutions 5-5. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-S (Ex5-5Vol) Under “Options.5 -1. Estimate” select Sbar as method to estimate standard deviation.990 1 2 3 4 5 6 7 8 Sample 9 10 11 12 13 14 15 Sample Range 6.302 1 2 3 4 5 6 7 8 Sample 9 10 11 12 13 14 15 The process is in statistical control.686 4. or cycles. with no out-of-control signals.830 1. or cycles. 5-9 .043 1 2 3 4 5 6 7 8 Sample 9 10 11 12 13 14 15 2.0 -0.0 0. runs.0 0.2 3.5 _ S =1.0 U C L=0.037 Sample M ean 1.0 U C L=5.066 1. runs.0 LC L=-0.5 _ _ X=-0.714 0.003 0. with no out-of-control signals.0 1.5 _ R=3.

07 ) = 0. n = 15. and “By Variables” is the sample ID column (Ex5-5Sample).Chapter 5 Exercise Solutions 5-5 continued (c) Let α = 0.136 1.5 CL = 1.0662 (15 − 1) ( 4.33 0.0662 (15 − 1) ( χ12−(0. To construct an s2 control chart. and there would be no difference in interpretation from either the x − s or the x−R chart.15 −1 ) = 1. Copy results from the session window by holding down the keyboard “Alt” key. Otherwise there are no runs.010 / 2).542 2.n −1 = 1.010 / 2.330 MINITAB’s control chart options do not include an s2 or variance chart. and then copying & pasting to an empty worksheet column (results in Ex5-5Variance).066. CL = s 2 = 1. 5-10 .5 LCL = 0.542 LCL = s 2 (n − 1) χ12−(α / 2). s = 1.066 (15 − 1) ( 31.0662 (15 − 1) ( χ 0. or cycles. Graph > Time Series Plot > Simple Control limits can be added using: Time/Scale > Reference Lines > Y positions Control Chart for Ex5-5Variance UCL = 2.15−1 ) = 1. If the limits had been calculated using α = 0. selecting only the variance column.32 ) = 2. trends.136 2 2 UCL = s 2 (n − 1) χα2 / 2.0027 (not tabulated in textbook). “Variables” is column with sample data (Ex5-5Vol).0 1 2 3 4 5 6 7 8 9 Sample 10 11 12 13 14 15 Sample 5 signals out of control below the lower control limit.010. first calculate the sample standard deviations and then create a time series plot. sample 5 would be within the limits.n −1 = 1.0 0. In “Statistics” select “Variance”.0 1.0662 = 1.5 s^2 (Variance) 2. Results are displayed in the session window. To obtain sample standard deviations: Stat > Basic Statistics > Store Descriptive Statistics.

5420 16.268.268 16.204 5-11 . runs.326 = 0.0 LC L=15.25 0. σˆ x = R / d 2 = 0.2 16.475 / 2.4 _ _ X=16.Chapter 5 Exercise Solutions 5-6.9940 2 4 6 8 10 Sample 12 14 16 18 20 U C L=1.00 0. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Net Weight (Ex5-6Wt) 16.6 Sample M ean U C L=16. R = 0. trends.004 Sample Range 1.50 0.75 _ R=0.00 LC L=0 2 4 6 8 10 Sample 12 14 16 18 20 The process is in statistical control with no out-of-control signals. x = 16.475 0. or cycles.475. (b) n = 5.

the histogram has one mode.50 15.75 16. Points on the normal probability plot generally fall along a straight line.1 15.2014 100 1.2 Ex5-6Wt 16.0 16.25 Ex5-6Wt 16.Chapter 5 Exercise Solutions 5-6 continued (c) MTB > Graph > Histogram > Single (Ex5-6Wt) Histogram of Net Weight (Ex5-6Wt) 20 Frequency 15 10 5 0 15.005 80 70 60 50 40 30 20 10 5 1 0.27 0.00 16.8 16.95% CI 99. 5-12 .50 16. and is approximately symmetrical with a bell shape.4 16.9 Mean StDev N AD P-Value 99 Percent 95 90 16.257 <0.75 17.6 MTB > Graph > Probability Plot > Single (Ex5-6Wt) Probability Plot of Net Weight (Ex5-6Wt) Normal .00 Visual examination indicates that fill weights approximate a normal distribution .

so the process is not capable of meeting Cˆ P 6σˆ x 6(0.7 − 16.20196 Potential (Within) Capability Cp 0.71 0.70000 Sample Mean 16. MTB > Stat > Quality Tools > Capability Analysis > Normal Under “Estimate” select Rbar as method to estimate standard deviation.5) = = 0.Chapter 5 Exercise Solutions 5-6 continued (d) USL − LSL +0.20421 StDev(Overall) 0.82 CPL 0.61 16.96 5-13 .71 CCpk 0. Overall Performance PPM < LSL 2458. Process Capability Analysis of Net Weight (Ex5-6Wt) LSL USL Within Ov erall Process Data LSL 15.0027 0.26800 Sample N 100 StDev(Within) 0.83 0.73 PPM Total 18673.2 16.23 PPM > USL 16215.93 CPU 0.5 − (−0.23 PPM > USL 16215.8 Observed Performance PPM < LSL 0.70000 Target * USL 16.268 ⎞ pˆ lower = Pr{x < LSL} = Pr{x < 15.73 PPM Total 18673.78) = 0.00 16. "Overall" Performance PPM < LSL 2458.41 PPM Total 19902.4 0.204 ⎝ ⎠ The MINITAB process capability analysis also reports Exp.71 Cpk 0.204) specifications.00 PPM Total 0.00 PPM > USL 0. Within Performance PPM < LSL 2706.6 Exp.96 (e) ⎛ 15.71 * 16.82 .0 Exp.82 Overall Capability Pp PPL PPU Ppk Cpm 15.7} = Φ ⎜ ⎟ = Φ (−2.20 PPM > USL 17196.94 0.

703 3.5 0.33 10.0 U C L=14.Chapter 5 Exercise Solutions 5-7.91 Sample M ean 40 20 _ _ X=10.73 12.9 0 -20 LC L=-25.5 LC L=5. trends. trends. 5-8.5 _ _ X=10.0 2 4 6 8 10 Sample 12 14 16 18 20 U C L=6. or cycles. or cycles.0 7.11 2 4 6 8 10 Sample 12 14 16 18 20 60 Sample StDev U C L=52. runs.0 LC L=0 2 4 6 8 10 Sample 12 14 16 18 20 The process is in statistical control with no out-of-control signals.125 Sample StDev 6.5 _ S =2.71 45 30 _ S =25.0 4. 5-14 . MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-S (Ex5-2Vl) Xbar-S Chart of Output Voltage (Ex5-2V) Sample M ean 15.92 5. runs.0 1. MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-S (Ex5-3Dia) Xbar-S Chart of Deviations from Nominal Diameter (Ex5-3Dia) U C L=46.23 15 LC L=0 0 2 4 6 8 10 Sample 12 14 16 18 20 The process is in statistical control with no out-of-control signals.

Chapter 5 Exercise Solutions 5-9☺.00 73.024 0. (b) The control limits on the x charts in Example 5-3 were calculated using S to estimate σ. 5-15 .01458 Sample M ean 74.02324 0.036 _ R=0. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R (Ex5-9ID) Xbar-R Chart of Inner Diameter (Ex5-9ID) U C L=74.012 0.01 _ _ X=74. runs. and in general.99 LC L=73.000 LC L=0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 The process is in statistical control with no out-of-control signals.04914 Sample Range 0. trends. They will not always be the same. or cycles.98777 2 4 6 8 10 12 14 Sample 16 18 20 22 24 U C L=0. in this exercise R was used to estimate σ. the x control limits based on S will be slightly different than limits based on R.048 0.00118 74.

00 Exp.040 Observed Performance PPM < LSL 0. MTB > Stat > Quality Tools > Capability Analysis > Normal Under “Estimate” select Rbar as method to estimate standard deviation.668 Cˆ P = 6σˆ x 6(0.59 * 73.05}] ⎛ 73.71 CPU 1.950 73.63 1.Chapter 5 Exercise Solutions 5-9 continued (c) σˆ x = R / d 2 = 0.05} = Pr{x < 73.00 PPM Total 0. Overall Performance PPM < LSL 0.63 CCpk 1.02324 / 2.010 74.00118 ⎞ ⎡ ⎛ 74.67 CPL 1.15 PPM > USL 0.00 PPM > USL 0.95} + Pr{x > 74.59 1.95 = = 1.009991 ⎝ ⎠ ⎣ ⎝ ⎠⎦ = Φ (−5.05 − 73.16 pˆ = Pr{x < LSL} + Pr{x > USL} = Pr{x < 73.05000 Sample Mean 74.009991) specifications.965 73.995 74. so the process is not capable of meeting USL − LSL 74.886) = 0 +1−1 =0 5-16 .28 PPM > USL 0. Within Performance PPM < LSL 0.67 1.66 Exp.67 Overall Capability Pp PPL PPU Ppk Cpm 1.009991 0.89 PPM Total 1.51 PPM Total 0.01022 Potential (Within) Capability Cp 1.95 − 74.123) + 1 − Φ (4.00118 ⎞ ⎤ = Φ⎜ ⎟ + ⎢1 − Φ ⎜ ⎟⎥ 0.326 = 0.00999 StDev(Overall) 0.009991 . Process Capability Analysis of Inner Diameter (Ex5-9ID) LSL USL Within Ov erall Process Data LSL 73.980 73.95000 Target * USL 74.025 74.63 Cpk 1.05 − 74.00118 Sample N 125 StDev(Within) 0.95} + [1 − Pr{x < 74.

01458 74. 39.99 LC L=73.048 Sample Range U C L=74. 38.04914 0. Test Failed at points: 37.00118 74.02324 0. 38. MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R (Ex5-10ID) Xbar-R Chart of Inner Diameter (Ex5-10ID) 1 1 Sample M ean 74. until the x -value from the 37th sample is plotted.02 1 5 5 _ _ X=74.036 _ R=0. 5-17 . increasing the process mean.000 LC L=0 4 8 12 16 20 Sample 24 28 32 36 40 Test Results for Xbar Chart of Ex5-10ID TEST 1. Test Failed at points: 38.98777 4 8 12 16 20 Sample 24 28 32 36 40 U C L=0. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). One point more than 3.Chapter 5 Exercise Solutions 5-10☺. 37. 40 TEST 6.012 0. an assignable cause has likely occurred.00 73.00 standard deviations from center line.024 0. 39 TEST 5. Since this point and the three subsequent points plot above the upper control limit. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL).01 0. 39. Test Failed at points: 35. 40 The control charts indicate that the process is in control.

and A = 0.727 S UCL S = B6σ x = 1. m = 50 samples (a) 50 ∑ xi 50 ∑ Ri 2000 200 = 40.669(10) = 16. reduce variability such that the natural process tolerance limits are closer to. n = 10.669. ∑ xi = 2000.0) = = 1. or 0.483(4) = 41. 44.51 centerline = c4σ x = 0. 50 50 i =1 i =1 n = 6 items/sample.949(10) = 70. σˆ x ≈ 1. R = i =1 = =4 m m 50 50 UCL x = x + A2 R = 40 + 0.949(10) = 89. to reduce scrap and rework costs. so the process is not capable.949.276 centerline x = µ = 80 UCL x = µ + Aσ x = 80 + 0.579 ⎠ or 0.000005 ⎝ 1.0005%. ⎝ 1.016 LCL R = D3 R = 0(4) = 0 (b) natural tolerance limits: x ± 3σˆ x = x ± 3 ( R / d 2 ) = 40 ± 3(4 / 2.579) (d) ⎛ 36 − 40 ⎞ pˆ scrap = Pr{x < LSL} = Pr{x < 36} = Φ ⎜ ⎟ = Φ (−2.056 . B5 = 0.932 x= i =1 = LCL x = x − A2 R = 40 − 0. σ x = 10 in-lb.579 ⎠ ⎛ 47 − 40 ⎞ pˆ rework = Pr{x > USL} = 1 − Pr{x < USL} = 1 − Φ ⎜ ⎟ = 1 − Φ (4.76 5-12* (5-10). say.57%.253 . Second.736] (c) USL .Chapter 5 Exercise Solutions 5-11 (5-9).264. B6 = 1.LSL +5. (e) First.433) = 1 − 0.69 LCL S = B5σ x = 0.0 − (−5.483(4) = 38.533) = 0.999995 = 0. µ = 80 in-lb. not 40.0057 . 5-18 . Cˆ P = 6σˆ x 6(1.004(4) = 8. center the process at 41.9727(10) = 9.068 UCL R = D4 R = 2.276(10) = 2.49 LCL x = µ − Aσ x = 80 − 0. ∑ Ri = 200.534) = [35.

Chapter 5 Exercise Solutions

5-13* (5-11).

50

50

i =1

i =1

**n = 4 items/subgroup; ∑ xi = 1000; ∑ Si = 72; m = 50 subgroups
**

(a)

50

x=

∑ xi

i =1

m

=

1000

= 20

50

50

∑ Si

72

= 1.44

50

m

UCL x = x + A3 S = 20 + 1.628(1.44) = 22.34

S=

i =1

=

LCL x = x − A3 S = 20 − 1.628(1.44) = 17.66

UCL S = B4 S = 2.266(1.44) = 3.26

LCL S = B3 S = 0(1.44) = 0

(b)

⎛S ⎞

⎛ 1.44 ⎞

natural process tolerance limits: x ± 3σˆ x = x ± 3 ⎜ ⎟ = 20 ± 3 ⎜

⎟ = [15.3, 24.7]

⎝ 0.9213 ⎠

⎝ c4 ⎠

(c)

USL - LSL

+4.0 − (−4.0)

=

= 0.85 , so the process is not capable.

Cˆ P =

6σˆ x

6(1.44 / 0.9213)

(d)

⎛ 23 − 20 ⎞

pˆ rework = Pr{x > USL} = 1 − Pr{x ≤ USL} = 1 − Φ ⎜

⎟ = 1 − Φ (1.919) = 1 − 0.9725 = 0.0275

⎝ 1.44 / 0.9213 ⎠

or 2.75%.

⎛ 15 − 20 ⎞

pˆ scrap = Pr{x < LSL} = Φ ⎜

⎟ = Φ (−3.199) = 0.00069 , or 0.069%

⎝ 1.44 / 0.9213 ⎠

Total = 2.88% + 0.069% = 2.949%

(e)

⎛ 23 − 19 ⎞

pˆ rework = 1 − Φ ⎜

⎟ = 1 − Φ (2.56) = 1 − 0.99477 = 0.00523 , or 0.523%

⎝ 1.44 / 0.9213 ⎠

⎛ 15 − 19 ⎞

pˆ scrap = Φ ⎜

⎟ = Φ (−2.56) = 0.00523 , or 0.523%

⎝ 1.44 / 0.9213 ⎠

Total = 0.523% + 0.523% = 1.046%

Centering the process would reduce rework, but increase scrap. A cost analysis is needed

to make the final decision. An alternative would be to work to improve the process by

reducing variability.

5-19

Chapter 5 Exercise Solutions

5-14 (5-12).

(a)

MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R

Xbar-R Chart of Critical Dimension (Ex5-14ax1, ..., Ex5-14ax5)

U C L=154.45

Sample M ean

150

140

_

_

X=130.88

130

120

110

LC L=107.31

2

4

6

8

10

Sample

12

14

16

18

20

U C L=86.40

Sample Range

80

60

_

R=40.86

40

20

0

LC L=0

2

4

6

8

10

Sample

12

14

16

18

20

**The process is in statistical control with no out-of-control signals, runs, trends, or cycles.
**

(b)

MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R

**Under “Options, Estimate” use subgroups 1:20 to calculate control limits.
**

Xbar-R Chart of Critical Dimension (Ex5-14bx1, ..., Ex5-14bx5)

1

Sample M ean

180

1

1

1

1

1

1

1

1

1

160

U C L=154.45

140

_

_

X=130.88

120

LC L=107.31

100

3

6

9

12

15

Sample

18

21

24

27

30

U C L=86.40

Sample Range

80

60

_

R=40.86

40

20

0

LC L=0

3

6

9

12

15

Sample

18

21

24

27

30

Starting at Sample #21, the process average has shifted to above the UCL = 154.45.

5-20

Chapter 5 Exercise Solutions

5-14 continued

(c)

MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R

**Under “Options, Estimate” use subgroups 1:20 to calculate control limits.
**

Xbar-R Chart of Critical Dimension (Ex5-14cx1, ..., Ex5-14cx5)

1

Sample M ean

180

1

1

1

1

1

1

1

1

160

1

U C L=154.45

140

_

_

X=130.88

120

2

6

2

5

LC L=107.31

1

100

4

8

12

16

20

Sample

24

28

32

36

40

U C L=86.40

Sample Range

80

60

_

R=40.86

40

20

0

LC L=0

4

8

12

16

20

Sample

24

28

32

36

40

**The adjustment overcompensated for the upward shift. The process average is now
**

between x and the LCL, with a run of ten points below the centerline, and one sample

(#36) below the LCL.

5-21

Chapter 5 Exercise Solutions

5-15* (5-13).

(a)

MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R

Xbar-R Chart of Strength Test (Ex5-15aSt)

Sample M ean

85.0

U C L=84.58

82.5

_

_

X=79.53

80.0

77.5

75.0

LC L=74.49

2

4

6

8

10

Sample

12

14

16

18

20

Sample Range

20

U C L=18.49

15

_

R=8.75

10

5

0

LC L=0

2

4

6

8

10

Sample

12

14

16

18

20

**Yes, the process is in control—though we should watch for a possible cyclic pattern in
**

the averages.

5-22

Chapter 5 Exercise Solutions

5-15 continued

(b)

MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R

**Under “Options, Estimate” use subgroups 1:20 to calculate control limits.
**

Xbar-R Chart of Strength Test (Ex5-15bSt)

1

Sample M ean

85.0

8

82.5

8

U C L=84.58

8

_

_

X=79.53

80.0

77.5

75.0

LC L=74.49

1

3

6

9

12

15

18

Sample

21

24

27

30

30

33

1

1

Sample Range

1

1

1

1

1

1

20

U C L=18.49

2

_

R=8.75

10

0

LC L=0

3

6

9

12

15

18

Sample

21

24

27

30

33

**Test Results for R Chart of Ex5-15bSt
**

TEST

Test

TEST

Test

**1. One point more than 3.00 standard deviations from center line.
**

Failed at points: 25, 26, 27, 31, 33, 34, 35

2. 9 points in a row on same side of center line.

Failed at points: 32, 33, 34, 35

**A strongly cyclic pattern in the averages is now evident, but more importantly, there are
**

several out-of-control points on the range chart.

5-23

Chapter 5 Exercise Solutions

5-16 (5-14).

(a)

MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-S

Xbar-S Chart of Strength Test (Ex5-15aSt)

Original Data

Sample Mean

85.0

UCL=84.64

82.5

_

_

X=79.53

80.0

77.5

75.0

LCL=74.43

2

4

6

8

10

Sample

12

14

16

18

20

Sample StDev

8

UCL=7.468

6

_

S=3.575

4

2

0

LCL=0

2

4

6

8

10

Sample

12

14

16

18

20

**Under “Options, Estimate” use subgroups 1:20 to calculate control limits.
**

Xbar-S Chart of Strength Test (Ex5-15bSt)

Original plus New Data

1

Sample Mean

85.0

UCL=84.64

82.5

_

_

X=79.53

80.0

77.5

75.0

LCL=74.43

1

3

6

9

12

15

18

Sample

21

24

27

30

1

33

1

1

Sample StDev

10.0

1

1

7.5

1

1

1

1

UCL=7.47

5.0

_

S=3.57

2.5

LCL=0

0.0

3

6

9

12

15

18

Sample

21

24

27

30

33

**Test Results for Xbar Chart of Ex5-15bSt
**

TEST 1. One point more than 3.00 standard deviations from center line.

Test Failed at points: 24, 31, 34

**Test Results for S Chart of Ex5-15bSt
**

TEST 1. One point more than 3.00 standard deviations from center line.

Test Failed at points: 22, 25, 26, 27, 31, 33, 34, 35

5-24

Chapter 5 Exercise Solutions

5-16 continued

(b)

Yes, the s chart detects the change in process variability more quickly than the R chart

did, at sample #22 versus sample #24.

5-17 (5-15).

nold = 5; xold = 34.00; Rold = 4.7

(a)

for nnew = 3

⎡ d 2(new ) ⎤

⎡ 1.693 ⎤

UCL x = xold + A2(new ) ⎢

(4.7) = 37.50

⎥ Rold = 34 + 1.023 ⎢

⎣ 2.326 ⎥⎦

⎣⎢ d 2(old) ⎦⎥

⎡ d 2(new ) ⎤

⎡ 1.693 ⎤

LCL x = xold − A2(new ) ⎢

⎥ Rold = 34 − 1.023 ⎢

⎥ (4.7) = 30.50

⎣ 2.326 ⎦

⎢⎣ d 2(old) ⎥⎦

⎡ d 2(new ) ⎤

⎡ 1.693 ⎤

UCL R = D4(new) ⎢

⎥ Rold = 2.574 ⎢

⎥ (4.7) = 8.81

⎣ 2.326 ⎦

⎢⎣ d 2(old) ⎥⎦

⎡d

⎤

⎡ 1.693 ⎤

CL R = Rnew = ⎢ 2(new ) ⎥ Rold = ⎢

(4.7) = 3.42

⎣ 2.326 ⎥⎦

⎣⎢ d 2(old) ⎦⎥

⎡d

⎤

⎡ 1.693 ⎤

LCL R = D3(new) ⎢ 2(new ) ⎥ Rold = 0 ⎢

(4.7) = 0

⎣ 2.326 ⎥⎦

⎢⎣ d 2(old) ⎥⎦

(b)

The x control limits for n = 5 are “tighter” (31.29, 36.72) than those for n = 3 (30.50,

37.50). This means a 2σ shift in the mean would be detected more quickly with a sample

size of n = 5.

5-25

Chapter 5 Exercise Solutions

5-17 continued

(c)

for n = 8

⎡ d 2(new ) ⎤

⎡ 2.847 ⎤

UCL x = xold + A2(new ) ⎢

⎥ Rold = 34 + 0.373 ⎢

⎥ (4.7) = 36.15

⎣ 2.326 ⎦

⎣⎢ d 2(old) ⎦⎥

⎡ d 2(new ) ⎤

⎡ 2.847 ⎤

LCL x = xold − A2(new ) ⎢

(4.7) = 31.85

⎥ Rold = 34 − 0.373 ⎢

⎣ 2.326 ⎥⎦

⎢⎣ d 2(old) ⎥⎦

⎡ d 2(new ) ⎤

⎡ 2.847 ⎤

UCL R = D4(new) ⎢

(4.7) = 10.72

⎥ Rold = 1.864 ⎢

⎣ 2.326 ⎥⎦

⎣⎢ d 2(old) ⎦⎥

⎡d

⎤

⎡ 2.847 ⎤

CL R = Rnew = ⎢ 2(new ) ⎥ Rold = ⎢

(4.7) = 5.75

⎣ 2.326 ⎥⎦

⎢⎣ d 2(old) ⎥⎦

⎡d

⎤

⎡ 2.847 ⎤

LCL R = D3(new) ⎢ 2(new ) ⎥ Rold = 0.136 ⎢

⎥ (4.7) = 0.78

⎣ 2.326 ⎦

⎢⎣ d 2(old) ⎥⎦

(d)

The x control limits for n = 8 are even "tighter" (31.85, 36.15), increasing the ability of

the chart to quickly detect the 2σ shift in process mean.

5-18☺.

nold = 5, xold = 74.001, Rold = 0.023, nnew = 3

⎡ d 2(new ) ⎤

⎡ 1.693 ⎤

UCL x = xold + A2(new ) ⎢

⎥ Rold = 74.001 + 1.023 ⎢

⎥ (0.023) = 74.018

⎣ 2.326 ⎦

⎣⎢ d 2(old) ⎦⎥

⎡ d 2(new ) ⎤

⎡ 1.693 ⎤

LCL x = xold − A2(new ) ⎢

(0.023) = 73.984

⎥ Rold = 74.001 − 1.023 ⎢

⎣ 2.326 ⎥⎦

⎢⎣ d 2(old) ⎥⎦

⎡ d 2(new ) ⎤

⎡ 1.693 ⎤

UCL R = D4(new) ⎢

(0.023) = 0.043

⎥ Rold = 2.574 ⎢

⎣ 2.326 ⎥⎦

⎣⎢ d 2(old) ⎥⎦

⎡d

⎤

⎡ 1.693 ⎤

CL R = Rnew = ⎢ 2(new ) ⎥ Rold = ⎢

(0.023) = 0.017

⎣ 2.326 ⎥⎦

⎢⎣ d 2(old) ⎥⎦

⎡d

⎤

⎡ 1.693 ⎤

LCL R = D3(new) ⎢ 2(new ) ⎥ Rold = 0 ⎢

(0.023) = 0

⎣ 2.326 ⎥⎦

⎣⎢ d 2(old) ⎦⎥

5-26

Chapter 5 Exercise Solutions

5-19 (5-16).

35

35

i =1

i =1

**n = 7; ∑ xi = 7805; ∑ Ri = 1200; m = 35 samples
**

(a)

35

x=

∑ xi

i =1

m

=

7805

= 223

35

35

∑ Ri

1200

= 34.29

m

35

UCL x = x + A2 R = 223 + 0.419(34.29) = 237.37

R=

i =1

=

**LCL x = x − A2 R = 223 − 0.419(34.29) = 208.63
**

UCL R = D4 R = 1.924(34.29) = 65.97

LCL R = D3 R = 0.076(34.29) = 2.61

(b)

µˆ = x = 223; σˆ x = R / d 2 = 34.29 / 2.704 = 12.68

(c)

**USL − LSL +35 − (−35)
**

=

= 0.92 , the process is not capable of meeting

Cˆ P =

6σˆ x

6(12.68)

specifications.

**pˆ = Pr{x > USL} + Pr{x < LSL} = 1 − Pr{x < USL} + Pr{x < LSL} = 1 − Pr{x ≤ 255} + Pr{x ≤ 185}
**

⎛ 255 − 223 ⎞

⎛ 185 − 223 ⎞

= 1− Φ ⎜

⎟+ Φ⎜

⎟ = 1 − Φ (2.52) + Φ (−3.00) = 1 − 0.99413 + 0.00135 = 0.0072

⎝ 12.68 ⎠

⎝ 12.68 ⎠

(d)

The process mean should be located at the nominal dimension, 220, to minimize nonconforming units.

⎛ 255 − 220 ⎞

⎛ 185 − 220 ⎞

pˆ = 1 − Φ ⎜

⎟+Φ⎜

⎟ = 1 − Φ (2.76) + Φ (−2.76) = 1 − 0.99711 + 0.00289 = 0.00578

⎝ 12.68 ⎠

⎝ 12.68 ⎠

5-27

50.36 / 2.50 ⎞ ⎛ 25.155 ⎝ ⎠ ⎝ ⎠ = 0. 25 25 i =1 i =1 n = 5.155 0.577(0.90 − 26.36) = 26.90 − 26. ∑ xi = 662.40 ⎞ pˆ = 1 − Φ ⎜ ⎟+Φ⎜ ⎟ = 1 − Φ (3.00499 (c) ⎛ 26.58) + Φ (−3.36) = 0.40 ⎞ ⎛ 25.71 R= i =1 = LCL x = x − A2 R = 26.50 25 25 ∑ Ri 9.155 pˆ = Pr{x > USL} + Pr{x < LSL} = 1 − Pr{x ≤ USL} + Pr{x < LSL} ⎛ 26.00005 0.29 UCL R = D4 R = 2.36 m 25 UCL x = x + A2 R = 26.114(0.23) 0.50 + 0.00.99506 + 0.326 = 0.00062 = 0.155 0.50 − 0.00124 5-28 . ∑ Ri = 9.50 = 26.00 = 0.90 − 26.36) = 26.23) + Φ (−3.Chapter 5 Exercise Solutions 5-20 (5-17).99938 + 0.36) = 0 (b) σˆ x = R / d 2 = 0.577(0.87) = 1 − 0.155 ⎝ ⎠ ⎝ ⎠ = 1 − 0.90 − 26.76 LCL R = D3 R = 0(0.50 ⎞ = 1− Φ ⎜ ⎟+Φ⎜ ⎟ = 1 − Φ (2. m = 25 samples (a) 25 x= ∑ xi i =1 m = 662.

57926 − 0 = 0.0 + 1. n = 5.6 5 ⎝ ⎠ ⎝ ⎠ = 0. R = 9. LSL=90 σˆ x = R / d 2 = 9.5 / 0. USL − LSL +10 − (−10) Cˆ P = = = 0.0. X ~ N .0 − 1.99 is larger than the width of the tolerance band.5) = 3.13 LCL S = B3 S = 0(1.9400 = 1.8338 6σˆ x 6(3. USL=110.30.998) 5-29 .14 LCL x = x − A3 S = 20.30 / 2.5) = 22. x = 104.326 = 3. even if the mean is located at the nominal dimension.20) − Φ (−5.57926 5-22 (5-19).5) = 0 (c) Pr{in control} = Pr{LCL ≤ x ≤ UCL} = Pr{x ≤ UCL} − Pr{x ≤ LCL} ⎛ 22. x = 20.84134 ⎞⎤ ⎡ ⎛ ⎞ ⎛ ⎞⎤ ⎟ ⎥ = 1 − ⎢Φ ⎜ 209 − 188 ⎟ − Φ ⎜ 191 − 188 ⎟ ⎥ ⎟⎥ ⎢⎣ ⎝ 6 4 ⎠ ⎝ 6 4 ⎠ ⎥⎦ ⎠⎦ 5-23 (5-20).6 5 1.84134 = 0.427(1.5) = 17.79) 1.427(1.089(1.998 and 6σˆ x = 6(3. m = 50 samples (a) σˆ x = S / c4 = 1.14 − 22 ⎞ ⎛ 17.60 (b) UCL x = x + A3 S = 20. S = 1. not all of the output will meet specification. So.Chapter 5 Exercise Solutions 5-21 (5-18). 2(10) = 20. n = 5.998) = 23.86 UCL S = B4 S = 2. Pr{detect} = 1 − Pr{not detect} = 1 − [Pr{LCL ≤ x ≤ UCL}] = 1 − [Pr{x ≤ UCL} − Pr{x ≤ LCL}] ⎡ ⎛ UCL − µ ⎞ ⎛ LCL − µ x x new ⎟ − Φ ⎜ new = 1 − ⎢Φ ⎜ ⎢ ⎜ ⎟ ⎜ n n σ σ x x ⎠ ⎝ ⎣ ⎝ = 1 − Φ (7) + Φ (1) = 1 − 1 + 0.86 − 22 ⎞ = Φ⎜ ⎟ −Φ⎜ ⎟ = Φ (0.5. 100.

686(2.7979(2.5) = 0 (c) centerline S = c4σ x = 0. x (a) centerline x = µ = 10 UCL x = µ + Aσ x = 10 + 2. σ = 2. µ = 10.5) = 0 5-30 .5) = 4.5) = 9.52 LCL S = B5σ = 0(2. These are standard values.30 LCL x = µ − Aσ x = 10 − 2.5 . n = 2.70 (b) centerline R = d 2σ x = 1.82 UCL R = D2σ = 3.5) = 1.121(2.606(2.121(2.5) = 6.128(2.Chapter 5 Exercise Solutions 5-24* (5-21).5) = 15.5) = 2.99 UCL S = B6σ = 2.22 LCL R = D1σ = 0(2.

56) = 22. x = 20.Chapter 5 Exercise Solutions 5-25 (5-22).56) = 0.56) = 0 (b) σˆ x = R / d 2 = 4.37 UCL R = D4 R = 2. n = 5.326 = 1.56) = 17.85 .56.05938 5-31 . (d) Pr{not detect} = Pr{LCL ≤ x ≤ UCL} = Pr{x ≤ UCL} − Pr{x ≤ LCL} ⎛ UCL x − µnew ⎞ ⎛ LCL x − µnew ⎞ ⎛ 22.05938 − 0 = 0.63 LCL x = x − A2 R = 20 − 0.96 5 ⎠ ⎝ 1. R = 4.56) + Φ (−7.96) specifications.63 − 24 ⎞ ⎛ 17. m = 25 samples (a) UCL x = x + A2 R = 20 + 0.56) = 9.114(4. so the process is not capable of meeting Cˆ P = 6σˆ x 6(1.56 / 2.37 − 24 ⎞ = Φ ⎜⎜ ⎟⎟ − Φ ⎜⎜ ⎟⎟ = Φ ⎜ ⎟−Φ⎜ ⎟ ⎝ 1.64 LCL R = D3 R = 0(4.96 (c) USL − LSL +5 − (−5) = = 0.96 5 ⎠ ⎝ σˆ x n ⎠ ⎝ σˆ x n ⎠ = Φ (−1.577(4.577(4.

74 10 0 LCL=0 2 4 6 8 10 Sample 12 14 16 18 20 Test Results for Xbar Chart of Ex5-26Th TEST 1.Chapter 5 Exercise Solutions 5-26☺. MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Sample M ean Xbar-R Chart of TiW Thickness (Ex5-26Th) 460 U C L=460.49 430 1 2 4 6 8 10 Sample 12 14 16 18 20 Sample Range 40 UCL=38. so these limits can be used for future production.68 450 440 LCL=437.56 430 1 2 4 6 8 10 Sample 12 14 16 18 20 Sample Range 40 U C L=37. One point more than 3. One point more than 3.00 standard deviations from center line.00 standard deviations from center line. 5-32 .69 440 LC L=436. Excluding subgroup 18 from control limits calculations: Xbar-R Chart of TiW Thickness (Ex5-26Th) Excluding subgroup 18 from calculations UCL=461.98 30 20 _ R=16. Test Failed at points: 18 The process is out of control on the x chart at subgroup 18.88 Sample Mean 460 _ _ X=449. Test Failed at points: 18 No additional subgroups are beyond the control limits.18 30 20 _ R=16.65 10 0 LC L=0 2 4 6 8 10 Sample 12 14 16 18 20 Test Results for Xbar Chart of Ex5-26Th TEST 1.82 450 _ _ X=448.

5-33 .7 9.9 Mean StDev N AD P-Value 99 Percent 95 90 448.68 σˆ x = R / d 2 = 16.74 / 2.Chapter 5 Exercise Solutions 5-26 continued (b) Excluding subgroup 18: x = 449.059 = 8.672 80 70 60 50 40 30 20 10 5 1 0.13 (c) MTB > Stat > Basic Statistics > Normality Test Probability Plot of TiW Thickness (Ex5-26Th) Normal 99.1 420 430 440 450 Ex5-26Th 460 470 480 A normal probability plot of the TiW thickness measurements shows the distribution is close to normal.269 0.111 80 0.

13944 Overall Capability Pp PPL PPU Ppk Cpm 420 Observed Performance PPM < LSL 0. so the process is capable.13) MTB > Stat > Quality Tools > Capability Analysis > Normal Process Capability Analysis of TiW Thickness (Ex5-26Th) LSL USL W ithin O v erall Process Data LSL 420. Cp = 1.00000 Sample Mean 448.24.29 Cpk 1.Chapter 5 Exercise Solutions 5-26 continued (d) USL = +30.00000 Target * USL 480.00 PPM Total 0.68750 Sample N 80 StDev(Within) StDev(Overall) Potential (Within) Capability Cp 1. σ x is estimated using R .18 CPU 1.14 1.24 8.92 248.18 CCpk 1. Within Performance PPM < LSL 194.38 PPM > USL PPM Total 53.18 The Potential (Within) Capability. Cˆ P = 6σˆ x 6(8.05 * 480 Exp. 5-34 .08645 9. LSL = –30 USL − LSL +30 − (−30) = = 1.00 PPM > USL 0.30 440 450 460 470 1. or in other words. Overall Performance PPM < LSL 848.23 .24 CPL 1.17 PPM Total 1154.09 1.05 1. This is the same result as the manual calculation. is estimated from the within-subgroup variation.00 430 Exp.01 PPM > USL 306.

74 10 0 LCL=0 3 6 9 12 15 Sample 18 21 24 27 30 Test Results for Xbar Chart of Ex5-27Th TEST 1. MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of TiW Thickness (Ex5-27Th) Using previous limits with 10 new subgroups UCL=461.68 450 440 LCL=437.00 standard deviations from center line.Chapter 5 Exercise Solutions 5-27☺. 5-35 .49 430 1 3 6 9 12 15 Sample 18 21 24 27 30 Sample Range 40 UCL=38. One point more than 3. Test Failed at points: 18 The process continues to be in a state of statistical control.18 30 20 _ R=16.88 Sample Mean 460 _ _ X=449.

74) = 432.880 ⎢ ⎣ 2.059 ⎦ ⎣⎢ d 2(old) ⎦⎥ ⎡ d 2(new ) ⎤ ⎡ 1.92 ⎥ Rold = 449.74) = 9. nnew = 2 ⎡ d 2(new ) ⎤ ⎡ 1. Parameters.128 ⎤ UCL x = xold + A2(new ) ⎢ (16. and enter new parameter values.96 ⎥ Rold = 3.74) = 29.267 ⎢ ⎣ 2.128 ⎤ LCL x = xold − A2(new ) ⎢ (16.44 ⎥ Rold = 449. xold 449.96 Sample Range 30 20 10 _ R=9.74) = 0 ⎣ 2.128 ⎤ CL R = Rnew = ⎢ ⎥ Rold = ⎢ ⎥ (16.93 460 _ _ X=449. nold = 4. Limits derived from N=4 subgroups Sample Mean 470 UCL=466.Chapter 5 Exercise Solutions 5-28☺.68 + 1.68 450 440 LCL=432.13 MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Select Xbar-R options.128 ⎤ LCL R = D3(new) ⎢ ⎥ Rold = 0 ⎢ ⎥ (16.059 ⎥⎦ ⎣⎢ d 2(old) ⎦⎥ ⎡ d 2(new ) ⎤ ⎡ 1.68 − 1.059 ⎦ ⎢⎣ d 2(old) ⎥⎦ σˆ new = Rnew d 2(new ) = 9. 5-36 .128 = 8.059 ⎥⎦ ⎣⎢ d 2(old) ⎦⎥ ⎡ d 2(new ) ⎤ ⎡ 1.17 1.68.17 ⎣ 2. Xbar-R Chart of TiW Thickness (Ex5-28Th) New subgroups with N=2. R old = 16.74.059 ⎥⎦ ⎢⎣ d 2(old) ⎥⎦ ⎡ d 2(new ) ⎤ ⎡ 1.128 ⎤ UCL R = D4(new) ⎢ (16.17 0 LCL=0 1 2 3 4 5 6 7 8 9 10 Sample The process remains in statistical control.74) = 466.43 430 1 2 3 4 5 6 7 8 9 10 Sample UCL=29.880 ⎢ ⎣ 2.

68 450 440 LCL=437.22 Sample Mean 460 _ _ X=449.Chapter 5 Exercise Solutions 5-29☺. Xbar-S Chart of Thickness (Ex5-27Th) 10 subgroups of new data.00 standard deviations from center line.22 Sample Mean 460 _ _ X=449. One point more than 3. After finding assignable cause. so these limits can be used for future production.68 450 440 LCL=437. 5-37 . exclude subgroup 18 from control limits calculations: MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-S Xbar-S Chart of Thickness (Ex5-26Th) Excluding subgroup 18 from calculations UCL=462.70 5 0 LCL=0 2 4 6 8 10 Sample 12 14 16 18 20 Xbar-S Chart of Ex5-26Th Test Results for Xbar Chart of Ex5-26Th TEST 1.15 430 1 3 6 9 12 15 Sample 18 21 24 27 30 20 Sample StDev UCL=17.70 5 0 LCL=0 3 6 9 12 15 Sample 18 21 24 27 30 The process remains in statistical control. with prior limits UCL=462. Test Failed at points: 18 No additional subgroups are beyond the control limits. The process is out of control on the x chart at subgroup 18.44 15 10 _ S=7.44 15 10 _ S=7.15 430 1 2 4 6 8 10 Sample 12 14 16 18 20 20 Sample StDev UCL=17.

97 6 ⎠ = Φ (4.483(5) = 202.75) = 1 − 0.95994 β − risk = Pr{not detect} = Φ ⎜ 5-38 .85 Cˆ p = 6σˆ x 6(1. m = 30 samples (a) 30 x= ∑ xi = i =1 m 6000 = 200 30 30 ∑ Ri 150 =5 m 30 UCL x = x + A2 R = 200 + 0.59 UCL R = D4 R = 2.97 6 ⎠ ⎝ 1.59 − 199 ⎞ ⎟−Φ⎜ ⎟ ⎝ 1. ∑ Ri = 150. (c) ⎛ 202.004(5) = 10.42 − 199 ⎞ ⎛ 197.534 = 1.04006 = 0. Even though the process is centered at nominal.25) − Φ (−1.42 R= i =1 = LCL x = x − A2 R = 200 − 0. ∑ xi = 6000.483(5) = 197.97) The process is not capable of meeting specification.02 LCL R = D3 R = 0(5) = 0 (b) σˆ x = R / d 2 = 5 / 2.Chapter 5 Exercise Solutions 5-30 (5-23).97 USL − LSL +5 − (−5) = = 0. the variation is large relative to the tolerance. 30 30 i =1 i =1 n = 6.

R = 3.34) ] = 1 − [1 − 0.579) = 101.577(3. L = 3.95 UCL x = x + A2 R = 104.065 LCL x = x − A2 R = 104 − 0.579 / 2. n = 4.771 UCL R = D4 R = 2.579) = 0 (b) Without sample #4.05. excluding #4 and recalculating: x = 104. µ0 = 100.62 LNTL = x − 3σˆ x = 104 − 3(1.577(3.37 ] = 0.329 LCL x = x − A2 R = 104.326 = 1.579) = 7.95) = 8.37 5-32 (5-25).Chapter 5 Exercise Solutions 5-31 (5-24).579 UCL x = x + A2 R = 104 + 0.577(3.33) 4 ) ⎤ ⎣ ⎦ = 1 − ⎡Φ L − k n − Φ − L − k n ⎤ ⎣ ⎦ = 1 − [ Φ (5.05 − 0.114(3. So.95) = 0 Sample #4 is out of control on the Range chart.05 + 0.33 Pr{detecting shift on 1st sample} = 1 − Pr{not detecting shift on 1st sample} = 1− β ( ) ( ) = 1 − ⎡Φ ( 3 − (−1.33) 4 ) − Φ ( −3 − (−1.350 LCL R = D3 R = 0(3.566 LCL R = D3 R = 0(3.95) = 101. σˆ x = R / d 2 = 3.539) = 99. µ1 = 92 k = ( µ1 − µ0 ) σ = ( 92 − 100 ) 6 = −1.114(3.38 5-39 .579) = 106.577(3. σ = 6.95) = 106.66) − Φ (−0. (a) x = 104.935 UCL R = D4 R = 2.539 (c) UNTL = x + 3σˆ x = 104 + 3(1. R = 3.539) = 108.

0006 = 0.95) + Φ (−3.26 30 m ∑ Ri 144 = 4.064 ⎠ 5-40 .064 ⎛ 16 − 20.9744 + 0.667 ⎠ ⎝ 0.8) = 0 (b) σˆ x = R / d 2 = 4.0000 = 0.667 .9953 + 0.60) = 1 − 0.8 / 2.0195 ⎝ 2.8) = 17. if σˆ x = 0.Chapter 5 Exercise Solutions 5-32 continued (d) ⎛ 107 − 104 ⎞ ⎛ 99 − 104 ⎞ pˆ = 1 − Φ ⎜ ⎟+ Φ⎜ ⎟ = 1 − Φ (1.8) = 23.577(4.114(4.539 ⎠ ⎝ 1. ⎛ 107 − 103 ⎞ ⎛ 99 − 103 ⎞ pˆ = 1 − Φ ⎜ ⎟+ Φ⎜ ⎟ = 1 − Φ (2. 30 n = 5.0262 ⎝ 1.997) + Φ (−5.0000 ⎝ 0.8 = 20.326 = 2. m = 30 i =1 (a) m x= ∑ xi i =1 m = 607.8.0047 = 0.60) + Φ (−2.147 LCL R = D 3 R = 0(4.8) = 10.0094 ⎝ 1.8 30 m UCL x = x + A2 R = 20. ⎛ 107 − 103 ⎞ ⎛ 99 − 103 ⎞ pˆ = 1 − Φ ⎜ ⎟+ Φ⎜ ⎟ = 1 − Φ (5.667 ⎠ 5-33 (5-26).26 ⎞ pˆ = Pr{x < LSL} = Φ ⎜ ⎟ = Φ (−2. ∑ xi = 607.997) = 1 − 1.26 − 0.064) = 0.0000 + 0. then almost 100% of parts will be within specification.25) = 1 − 0.49 UCL R = D 4 R = 2. i =1 30 ∑ Ri = 144.26 + 0.539 ⎠ Next work on reducing the variability.03 R= i =1 = LCL x = x − A2 R = 20. first center the process at nominal.539 ⎠ ⎝ 1.539 ⎠ (e) To reduce the fraction nonconforming.577(4.

Test Failed at points: 12 Process is not in statistical control -. R Chart of Detent (Ex5-34Det) 20 1 15 Sample Range UCL=13.67 10 _ R=6.47 5 0 LCL=0 1 2 3 4 5 6 7 8 9 Sample 10 11 12 13 14 15 Test Results for R Chart of Ex5-34Det TEST 1.Chapter 5 Exercise Solutions 5-34 (5-27).sample #12 exceeds the upper control limit on the Range chart. 5-41 . (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > R Under “Options.00 standard deviations from center line. Estimate” select Rbar as method to estimate standard deviation. One point more than 3.

Estimate” omit subgroup 12 and select Rbar.42 (d) Assume the cigar lighter detent is normally distributed.64 5 0 LCL=0 1 2 3 4 5 6 7 8 9 Sample 10 11 12 13 14 15 Test Results for R Chart of Ex5-34Det TEST 1.93 10 _ R=5.00 standard deviations from center line. Test Failed at points: 12 (c) Without sample #12: σˆ x = R / d 2 = 5.64 / 2.3220 − 0. R Chart of Detent (Ex5-34Det) Sample 12 Excluded from Calculations 20 1 Sample Range 15 UCL=11.42 × 0.3200 Cˆ P = = = 1. One point more than 3.Chapter 5 Exercise Solutions 5-34 continued (b) Excluding Sample Number 12: MTB > Stat > Control Charts > Variables Charts for Subgroups > R Under “Options.0001) 5-42 .38 6σˆ x 6(2. Without sample #12: USL − LSL 0.326 = 2.

Test Failed at points: 2. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 18. 22. MTB > Stat > Control Charts > Variables Charts for Subgroups > R Under “Options. 20 TEST 6. 25 5-43 . 23 2. 9 points in a row on same side of center line.Chapter 5 Exercise Solutions 5-35 (5-28).0 6 -2. 18. 25 5.0 1 1 1 1 1 UCL=3. Failed at points: 12 2. and select Rbar. 19.00 standard deviations from center line. 20.0 2 4 6 8 10 20 Sample Range _ _ X=0.2 LCL=-3. One point more than 3. Failed at points: 24.93 2 10 2 5 0 _ R=5. 2. Failed at points: 1. Test Failed at points: 15. 13. 23. One point more than 3. 13-15 1 1 Sample Mean 5. 9 points in a row on same side of center line. 20.64 LCL=0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 Test Results for Xbar Chart of Ex5-35Det TEST Test TEST Test TEST 1. Xbar-R Chart of Ex5-35Det Limits based on Samples 1-11. 24 Test Results for R Chart of Ex5-35Det TEST Test TEST Test 1. Estimate” use subgroups 1:11 and 13:15. 13. 12. 3.5 6 6 2 2 0. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 16.5 1 -5.00 standard deviations from center line. Failed at points: 24. 17.05 1 12 14 Sample 16 18 20 22 24 1 15 UCL=11. 17.45 5 2.

12 and 13 1 1 Sample Mean 5. 2.5 1 -5.66 0. 3. No additional samples are out of control on the x chart. Removing sample 3 gives Xbar-R Chart of Ex5-35Det Limits based on first 15 samples. While the limits on the above charts may be used to monitor future production. 2.0 1 1 1 1 1 1 2. the fact that 6 of 15 samples were out of control and eliminated from calculations is an early indication of process instability.0 2 4 6 8 10 20 Sample Range UCL=1.26 1 12 14 Sample 16 18 20 22 24 1 15 1 1 10 1 1 2 2 UCL=9.5 5 6 6 2 5 2 -2.5 5 0 LCL=0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 Sample 14 is now out of control on the R chart. 5-44 .52 _ R=4. (a) Given the large number of points after sample 15 beyond both the x and R control limits on the charts above. 12. Note that samples 1. sample 3 is also out of control on the x chart. and 13 are out of control on the x chart.0 LCL=-3.94 _ _ X=-0. the process appears to be unstable. excluding 1. If these samples are removed and the limits recalculated.Chapter 5 Exercise Solutions 5-35 continued We are trying to establish trial control limits from the first 15 samples to monitor future production.

Test Failed at points: 1.23 0.5 1 -5. 13.08 1 12 14 Sample 16 18 20 22 24 1 UCL=15.00 standard deviations from center line.82 15 10 _ R=7. 12. 16.Chapter 5 Exercise Solutions 5-35 continued (b) Xbar-R Chart of Detent (Ex5-35Det) All Samples in Calculations 1 1 1 UCL=5. One point more than 3.55 Sample Mean 5.00 standard deviations from center line. 17 Test Results for R Chart of Ex5-35Det TEST 1. One point more than 3.0 -2.0 2.5 _ _ X=1. Test Failed at points: 12 5-45 .48 5 LCL=0 0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 With Test 1 only: Test Results for Xbar Chart of Ex5-35Det TEST 1.0 2 4 6 8 10 Sample Range 20 LCL=-3.

17 excluded from calculations 1 1 1 1 Sample Mean 5. 20 Test Results for R Chart of Ex5-35Det TEST 1. and 17 from calculations: Xbar-R Chart of Detent (Ex5-35Det) Samples 1. 12.0 UCL=4. 16. 16. 13. One point more than 3.5 _ _ X=0.0 2 4 6 8 10 Sample Range 20 1 12 14 Sample 16 18 20 22 24 1 15 UCL=14.85 5 0 LCL=0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 With Test 1 only: Test Results for Xbar Chart of Ex5-35Det TEST 1. 16.48 10 _ R=6. One point more than 3. 13.Chapter 5 Exercise Solutions 5-35 (b) continued Removing samples 1. 12.00 standard deviations from center line.96 1 -5.5 LCL=-2.0 -2. 17.99 0.00 standard deviations from center line.94 2. Test Failed at points: 12 5-46 . 12. Test Failed at points: 1. 13.

Chapter 5 Exercise Solutions 5-35 continued Sample 20 is now also out of control. Test Failed at points: 1.00 standard deviations from center line.5 _ _ X=0. Removing sample 20 from calculations. 18.66 2. and that the sources of variation need to be identified and removed.0 2 4 6 8 10 Sample Range 20 LCL=-3. Test Failed at points: 12 Sample 18 is now out-of-control. 20 excluded from calculations 1 1 Sample Mean 5.00 standard deviations from center line. for a total 7 of the 25 samples. Xbar-R Chart of Ex5-35Det Samples 1. 13. One point more than 3. with runs of points both above and below the centerline.11 1 12 14 Sample 16 18 20 22 24 1 15 UCL=14. 20 Test Results for R Chart of Ex5-35Det TEST 1. One point more than 3. 17.78 0. 5-47 .5 1 -5. 16. 16. 13.0 -2. 17. 12.74 5 LCL=0 0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 With Test 1 only: Test Results for Xbar Chart of Ex5-35Det TEST 1. This suggests that the process is inherently unstable.0 1 1 1 UCL=4. 12.24 10 _ R=6.

Then σˆ z = σˆ x2 + σˆ y2 = 0.0) = 0 (b) m ∑ xi 12.006) ⎝ 0. 30 30 i =1 i =1 n = 6.758 µˆ = x = i =1 = 5-48 .978 /10) / 2.400 / m ) d = (6.Chapter 5 Exercise Solutions 5-36 (5-29).346 5-37 (5-30).006 ⎝ σˆ z ⎠ ⎛ 0.09 − z ⎞ ⎜ ⎟ = −2.326 = 0. ∑ Ry .608.i = 18. (a) 20 10 i =1 i =1 n = 5.5121 ⎝ 0. ∑ Rx .0) = 90.870 = 429.42 + 0.608 / 20) / 2.534 = 17.500) + 0. ∑ Ri = 1350. mx = 20.500 ⎠ ⎛ 0.300 σˆ x = Rx / d 2 = ∑ Rx .0 / 2.006. m y = 10.18 R= i =1 = LCL R = D3 R = 0(45.500 ⎛ 0. m = 30 (a) m ∑ Ri 1350 = 45.0 30 m UCL R = D4 R = 2. ∑ xi = 12.004(45.i / mx i =1 σˆ y 10 y 2 i =1 y .09 − z ⎞ Φ⎜ ⎟ = 0.09} = 0.500 ⎠ z = +2.i = 6.09 = 1. Let z = x − y.0 30 m σˆ x = R / d 2 = 45.32 = 0.5121(0.870.326 = 0.978 ( = R / d = (∑ R 20 ) d = (18.i y 2 2 (b) Want Pr{(x − y) < 0.09 − z ⎞ Φ −1 ⎜ ⎟ = Φ (0.

758 ⎠ ⎝ 17.63) = 1 − 0. 30 30 i =1 i =1 n = 4.667 / 0.969 S= i =1 = LCL S = B3 S = 0(13. ∑ Si = 410.40 = 400 USL − LSL 480 − 400 = = 0.9213 = 14.870 = 429.758 ⎠ (d) To minimize fraction nonconforming the mean should be located at the nominal dimension (440) for a constant variance. LSL = 440 . ∑ xi = 12.0516 = 0.667) = 0 (b) m ∑ xi 12.751 Cˆ p 6σˆ x 6(17.266(13.0 30 m σˆ x = S / c4 = 13.9979 + 0.87) + Φ (−1.667 30 m UCL S = B4 S = 2.870. m = 30 (a) m ∑ Si 410 = 13.834 µˆ = x = i =1 = 5-49 .758) ⎛ 480 − 429 ⎞ ⎛ 400 − 429 ⎞ pˆ = 1 − Φ ⎜ ⎟+Φ⎜ ⎟ = 1 − Φ (2.0537 ⎝ 17.Chapter 5 Exercise Solutions 5-37 continued (c) USL = 440 + 40 = 480. 5-38 (5-31).667) = 30.

9535 − 0.56) ⎟ −Φ⎜ ⎟ = Φ⎜ σx σx ⎝ ⎠ ⎝ ⎠ ⎝ 8 5 ⎠ ⎝ 8 5 ⎠ = 0. 1 1 1 ARL1 = = = = 2. centerline x = 100. 5-50 .68) − Φ (−0.2877 = 0. µ = 100.6658)3 = 0. LCL x = 96.50 Cˆ P = = = = 0.6658 1 − [Pr{not detected}]3 = 1 − (0. µ = 98. UCL x = 104.807 ( = µ − k (σ UCL x = µ + kσ x = µ + k σ x LCL x = µ − kσ x x ) ( 4 ) = 111.005 / 2 = Z 0.6658 5-42 (5-35).7049 5-41 (5-34). k = 3.7678 6σˆ x 6 (1. σ x = 8 ( = µ − 2 (σ UCL x = µ + 2σ x = µ + 2 σ x LCL x = µ − 2σ x x ) ( 4 ) = 108 n ) = 100 − 2 ( 8 4 ) = 92 n = 100 + 2 8 (b) k = Zα / 2 = Z 0. σ x = 8 Pr{out-of-control signal by at least 3rd plot point} = 1 − Pr{not detected by 3rd sample} = 1 − [Pr{not detected}]3 Pr{not detected} = Pr{LCL x ≤ x ≤ UCL x } = Pr{x ≤ UCL x } − Pr{x ≤ LCL x } ⎛ 104 − 98 ⎞ ⎛ 96 − 98 ⎞ ⎛ UCL x − µ ⎞ ⎛ LCL x − µ ⎞ = Φ⎜ ⎟−Φ⎜ ⎟ = Φ (1.228 n ) = 100 − 2.0025 = 2.807 ( 8 4 ) = 88.807 8 5-40 (5-33). (a) n = 4.9213) 6 ( S c4 ) The process is not capable of meeting specifications. USL − LSL USL − LSL 202. n = 5.772 n = 100 + 2.992 1 − β 1 − Pr{not detect} 1 − 0.Chapter 5 Exercise Solutions 5-39 (5-32).50-197.000 0.

8000 0. 0. 0. β = Φ L − k n − Φ − L − k n ( ) ( ) for k = {0.3 1.0000 beta 0. β = {0.7734.5. 0.213 UCL S = B6σ x = 2. 0.0 1.75. 0.0013.2000 0.5.Chapter 5 Exercise Solutions 5-43 (5-36).970 ) 6 ( R d2 ) Process is capable of meeting specifications.0000. 0.5. 0. 0.0.2 n = 200 + 1.05/ 2 = Z 0.8 1.8 n ) = 200 − 1. 1. L = 3.0.0 2.9332. (b) n = 9. USL = 600 + 20 = 620.2266. σ x = 10 (a) centerline S = c4σ = 0. LSL = 600 − 20 = 580 (a) USL − LSL USL − LSL 620 − 580 Cˆ P = = = = 1.0668.0000 0.2000 1.0000} Operating Characteristic Curve for n = 9. 0.82 / 2.0 0.25.6000 0.4000 0. 0.0}.96 (10 4 ) = 190.96 10 5-44 (5-37).025 = 1.9213(10) = 9. n = 9.5 0. µ = 200. L = 3 1. 2. n = 4.111 6σˆ x 6 (17.5 2.96 ( = µ − k (σ UCL x = µ + kσ x = µ + k σ x LCL x = µ − kσ x x ) ( 4 ) = 209.5. 1.5 3.088(10) = 20.9974. 3.88 LCL S = B5σ x = 0(10) = 0 (b) k = Zα / 2 = Z 0. 1.0 k 5-51 . 2.

σˆ x = R / d 2 = 20. µ = 600.882(1.676 x= i =1 = LCL x = x − A2 R = 90 − 0.005 = 2.7 n = 600 + 2.419) = 0.576 ( = µ − k (σ UCL x = µ + kσ x = µ + k σ x LCL x = µ − kσ x x ) ( 9 ) = 610.1587 5-52 . ∑ Ri = 120. α =0.324 UCL R = D 4 R = 1. R = i =1 = =4 m m 30 30 UCL x = x + A2 R = 90 + 0. 30 30 i =1 i =1 n = 7.Chapter 5 Exercise Solutions 5-45 (5-38).059 = 10 Pr{detect shift on 1st sample} = Pr{x < LCL} + Pr{x > UCL} = Pr{x < LCL} + 1 − Pr{x ≤ UCL} ⎛ 785 − 790 ⎞ ⎛ 815 − 790 ⎞ ⎛ LCL − µnew ⎞ ⎛ UCL − µnew ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎟ +1− Φ ⎜ ⎟ = Φ⎜ σx σx ⎝ ⎠ ⎝ ⎠ ⎝ 10 4 ⎠ ⎝ 10 4 ⎠ = Φ (−1) + 1 − Φ (5) = 0.1587 + 1 − 1. n = 9.479 (c) S = c4σˆ x = 0. σ x = 12.01/ 2 = Z 0.419 UCL S = 1.419(4) = 91.167 5-46 (5-39). m = 30 (a) m ∑ xi m ∑ Ri 2700 120 = 90.59 / 2.01 k = Zα / 2 = Z 0.419) = 2.576 (12 9 ) = 589.304 (b) σˆ x = R / d 2 = 4 / 2.118(1.3 n ) = 600 − 2.9594(1.419(4) = 88.704 = 1. ∑ xi = 2700.696 LCL R = D 3 R = 0.479) = 1.0000 = 0.924(4) = 7.671 LCL S = 0.076(4) = 0.576 12 5-47 (5-40).

005 = 2.30 ARL1 = 1 − β 1 − Pr{not detect} Pr{detect} 0. (c) µnew = 357 ⎛ UCL − µnew Pr{not detect on 1st sample} = Pr{LCL ≤ x ≤ UCL} = Φ ⎜⎜ ⎝ σˆ x n ⎞ ⎛ LCL − µnew ⎟⎟ − Φ ⎜⎜ ⎠ ⎝ σˆ x n ⎞ ⎟⎟ ⎠ ⎛ 363 − 357 ⎞ ⎛ 357 − 357 ⎞ = Φ⎜ ⎟ −Φ⎜ ⎟ = Φ (6) − Φ (0) = 1.576 3 ) 9 = 362.1587 5-49 (5-42).5000 ⎝ 3 9 ⎠ ⎝ 3 9 ⎠ (d) α = 0. k = Zα / 2 = Z 0. (a) σˆ x = R / d 2 = 8.424 5-53 .0000 − 0.91/ 2.0026 ⎝ 3 9 ⎠ ⎝ 3 9 ⎠ (b) USL − LSL +6 − (−6) Cˆ P = = = 0.01/ 2 = Z 0.970 = 3.01.Chapter 5 Exercise Solutions 5-48 (5-41).576 3 ) ) ( n = 360 + 2.576 9 = 357.9987 = 0.0013 + 1 − 0.000 ⎛ LCL − x ⎞ ⎛ UCL − x ⎞ ⎟ +1− Φ ⎜ ⎟ ⎝ σx ⎠ ⎝ σx ⎠ α = Pr{x < LCL} + Pr{x > UCL} = Φ ⎜ ⎛ 357 − 360 ⎞ ⎛ 363 − 360 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ = Φ (−3) + 1 − Φ (3) = 0.576 ( UCL x = x + kσ x = x + k σˆ x ( LCL x = 360 − 2. 1 1 1 1 = = = = 6.667 6σˆ x 6(3) The process is not capable of producing all items within specification.5000 = 0.

000 (b) S = c4σˆ x = 0.9772 (f) Pr{detect by 3rd sample} = 1 − Pr{not detect by 3rd sample} = 1 − (Pr{not detect})3 = 1 − (1 − 0.0000 = 0.351 LCL S = B3 S = 0(3.8944 = 0.685) = 8.059 = 4. (e) Pr{detect on 1st sample} = Pr{x < LCL} + Pr{x > UCL} ⎛ LCL − µnew ⎞ ⎛ UCL − µnew ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ σx σx ⎝ ⎠ ⎝ ⎠ ⎛ 614 − 610 ⎞ ⎛ 626 − 610 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 4 4 ⎠ ⎝ 4 4 ⎠ = Φ (2) + 1 − Φ (8) = 0.25) = 0.Chapter 5 Exercise Solutions 5-50 (5-43).0000 5-54 . Also consider reducing the process variability.266(3.236 / 2.685) = 0 (c) ⎛ LSL − x ⎞ ⎛ USL − x ⎞ pˆ = Pr{x < LSL} + Pr{x > USL} = Φ ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ σˆ x ⎠ ⎝ σˆ x ⎠ ⎛ 595 − 620 ⎞ ⎛ 625 − 620 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ 4 4 ⎝ ⎠ ⎝ ⎠ = Φ (−6.25) + 1 − Φ (1.865 UCL S = B4 S = 2.9213(4) = 3.9772)3 = 1. try moving the center of the process from its current mean of 620 closer to the nominal dimension of 610.1056 (d) To reduce the fraction nonconforming.0000 + 1 − 0. (a) σˆ x = R / d 2 = 8.9772 + 1 − 1.

00. σˆ x = S / c4 = 1.0000 = 0.827 ⎠ = Φ (−1.827 ⎠ ⎝ 1.738 / 0.52 (c) pˆ = Pr{x < LSL} + Pr{x > USL} ⎛ LSL − x ⎞ ⎛ USL − x ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ σˆ x ⎠ ⎝ σˆ x ⎠ ⎛ 703 − 706 ⎞ ⎛ 709 − 706 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 1.48 LNTL = 706 − 3(1.1006 (d) Pr{detect on 1st sample} = Pr{x < LCL} + Pr{x > UCL} ⎛ LCL − µnew ⎞ ⎛ UCL − µnew ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ σx σx ⎝ ⎠ ⎝ ⎠ ⎛ 703. (a) µˆ = x = 706.Chapter 5 Exercise Solutions 5-51 (5-44).2 − 702 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 1.9920)3 = 1.8 − 702 ⎞ ⎛ 708.0503 + 1 − 0.9497 = 0.827 6 ⎠ ⎝ 1.642) = 0.827) = 711.9515 = 1.827 (b) UNTL = x + 3σˆ x = 706 + 3(1.31) = 0.9920 + 1 − 1.827 6 ⎠ = Φ (2.0000 5-55 .41) + 1 − Φ (8.642) + 1 − Φ (1.827) = 700.9920 (e) Pr{detect by 3rd sample} = 1 − Pr{not detect by 3rd sample} = 1 − (Pr{not detect})3 = 1 − (1 − 0.

3085 + 1 − 0.661 ⎠ ⎝ 8.15) + 1 − Φ (2.661 ⎠ = Φ (−1.3108 (e) ARL1 = 1 1 1 1 = = = = 3.1251 + 1 − 0.5) + 1 − Φ (2.new ⎝ ⎠ ⎝ ⎠ ⎛ 690 − 693 ⎞ ⎛ 710 − 693 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 12 4 ⎠ ⎝ 12 4 ⎠ = Φ (−0.0104 + 1 − 0.31) = 0.661 (b) pˆ = Pr{x < LSL} + Pr{x > USL} ⎛ LSL − x ⎞ ⎛ USL − x ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ σˆ x ⎠ ⎝ σˆ x ⎠ ⎛ 690 − 700 ⎞ ⎛ 720 − 700 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 8.31) = 0.661 4 ⎠ ⎝ 8.Chapter 5 Exercise Solutions 5-52 (5-45).9213 = 8.83) = 0.31) + 1 − Φ (2. σˆ x = S / c4 = 7.661 4 ⎠ = Φ (−2.0208 (d) Pr{detect on 1st sample} = Pr{x < LCL} + Pr{x > UCL} ⎛ LCL − µnew ⎞ ⎛ UCL − µnew ⎞ = Φ⎜ +1− Φ ⎜ ⎟ ⎟⎟ ⎜ σ ⎟ ⎜ σ x .9977 = 0.new x .9896 = 0.1355 (c) α = Pr{x < LCL} + Pr{x > UCL} ⎛ LCL − x ⎞ ⎛ UCL − x ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ σx ⎠ ⎝ σx ⎠ ⎛ 690 − 700 ⎞ ⎛ 710 − 700 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 8.9896 = 0.3108 5-56 . (a) µˆ = x = 700.979 / 0.22 1 − β 1 − Pr{not detect} Pr{detect} 0.

021055.1052.14 _ X=16.021055 ⎠ 5-57 .08 16.150 Visual examination of the normal probability indicates that the assumption of normally distributed coffee can weights is valid. MR2 = 0.1684 16.342 Percent 80 70 60 50 40 30 20 10 5 1 16.02044 25 0.1052 16.100 Ex5-53Wt 16.17 U C L=16.00 LC L=0 2 4 6 8 10 12 14 O bser vation 16 18 20 22 24 There may be a “sawtooth” pattern developing on the Individuals chart.0420 2 4 6 8 10 12 14 O bser vation 16 18 20 22 24 M oving Range 0. %underfilled = 100% × Pr{x < 16 oz} ⎛ 16 − 16.07760 0.9964) = 0.Chapter 5 Exercise Solutions 5-53 (5-46).02375 MTB > Stat > Basic Statistics > Normality Test Probability Plot of Weight (Ex5-53Wt) Normal 99 Mean StDev N AD P-Value 95 90 16.1052 ⎞ = 100% × Φ ⎜ ⎟ = 100% × Φ (−4.11 16.04 __ M R=0.08 U C L=0. σˆ x = 0.02375 0. MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Weight (Ex5-53Wt) Individual Value 16.397 0.075 16.050 16.06 0.125 16.02 0. x = 16.05 LC L=16.00003% ⎝ 0.11 0.

72 1 2 3 4 5 6 7 8 9 O bser vation 10 11 12 13 14 15 U C L=10.217 Percent 80 70 60 50 40 30 20 10 5 1 46 48 50 52 54 Ex5-54Har 56 58 60 Although the observations at the tails are not very close to the straight line.84954.82 Individual Value 60 55 _ X=53.0 7. indicating that it may be reasonable to assume that hardness is normally distributed.2667. MR2 = 3. MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Hardness (Ex5-54Har) U C L=61.50 M oving Range 10.05.Chapter 5 Exercise Solutions 5-54(5-47).27 50 45 LC L=44.465 0.0 1 2 3 4 5 6 7 8 9 O bser vation 10 11 12 13 14 15 x = 53. 5-58 . the p-value is greater than 0.0 __ M R=3. σˆ x = 2.27 2.5 5.21429 MTB > Stat > Basic Statistics > Normality Test Probability Plot of Hardness (Ex5-54Har) Normal 99 Mean StDev N AD P-Value 95 90 53.21 2.5 LC L=0 0.712 15 0.

MR2 = 148.2 120 LC L=0 0 2 4 6 8 10 12 O bser vation 14 16 18 20 The process appears to be in statistical control. with no out-of-control points.346. σˆ x = 131. (a) MTB > Stat > Basic Statistics > Normality Test Probability Plot of Viscosity (Ex5-55Vis) Normal 99 Mean StDev N AD P-Value 95 90 2929 129. (c) µˆ = x = 2928. (b) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Viscosity (Ex5-55Vis) Individual V alue 3400 U C L=3322.0 20 0.9.1 M oving Range 480 360 240 __ M R=148. or other patterns.9 2800 2600 LC L=2534. runs.Chapter 5 Exercise Solutions 5-55 (5-48).158 5-59 .9 2 4 6 8 10 12 O bser vation 14 16 18 20 U C L=484. trends.511 Percent 80 70 60 50 40 30 20 10 5 1 2600 2700 2800 2900 3000 Ex5-55Vis 3100 3200 3300 Viscosity measurements do appear to follow a normal distribution.9 3200 3000 _ X=2928.319 0.

9 3000 2800 2600 LCL=2534.9 3200 _ X=2928. However all of the new points on the I chart are above the center line.2 120 LCL=0 0 2 4 6 8 10 12 14 Observation 16 18 20 22 24 All points are inside the control limits.Chapter 5 Exercise Solutions 5-56 (5-49).9 2 4 6 8 10 12 14 Observation 16 18 20 22 24 UCL=484. MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Viscosity (Ex5-56Vis) With five next measurements 3400 Individual Value UCL=3322.1 Moving Range 480 360 240 __ MR=148. 5-60 . indicating that a shift in the mean may have occurred.

14 60 _ X=49.85 50 40 LC L=34.85 4. (a) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Oxide Thickness (Ex5-57aTh) Individual V alue U C L=65. 5-61 . MTB > Stat > Basic Statistics > Normality Test Probability Plot of Oxide Thickness (Ex5-57aTh) Normal 99 Mean StDev N AD P-Value 95 90 49.534 30 0.79 15 10 __ M R=5.480 Percent 80 70 60 50 40 30 20 10 5 1 40 45 50 Ex5-57aTh 55 60 The normality assumption is reasonable.75 5 0 LC L=0 3 6 9 12 15 18 O bser vation 21 24 27 30 The process is in statistical control.338 0.Chapter 5 Exercise Solutions 5-57 (5-50).55 30 3 6 9 12 15 18 O bser vation 21 24 27 30 M oving Range 20 U C L=18.

Chapter 5 Exercise Solutions 5-57 continued (b) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Oxide Thickness (Ex5-57bTh) With 10 new measurements and some sensitizing rules 70 1 Individual Value 5 60 6 6 2 2 UCL=65. One point more than 3.14 _ X=49.85 50 40 LCL=34.75 5 0 LCL=0 4 8 12 16 20 Observation 24 28 32 36 40 Test Results for I Chart of Ex5-57bTh TEST Test TEST Test TEST 1. 39. 9 points in a row on same side of center line. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 38. even without use of the sensitizing rules.79 15 10 __ MR=5.00 standard deviations from center line. Test Failed at points: 35. 39. There is a run above the centerline. and several 2 of 3 beyond 2 sigma on the x chart. it is clear that the process is out of control during this period of operation. 37. Failed at points: 38. 5-62 . 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). several 4 of 5 beyond 1 sigma.55 30 4 8 12 16 20 Observation 24 28 32 36 40 Moving Range 20 UCL=18. Test Failed at points: 34. 40 5. However. 40 We have turned on some of the sensitizing rules in MINITAB to illustrate their use. 40 TEST 6. Failed at points: 38 2. 39.

14 2 _ X=49.Chapter 5 Exercise Solutions 5-57 continued (c) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Oxide Thickness (Ex5-57cTh) 10 + 20 New Measurements.85 50 40 LCL=34. with Sensitizing Rules On 70 1 Individual Value 5 60 6 6 2 UCL=65.79 15 10 __ MR=5.55 30 1 6 12 18 24 30 Observation 36 48 54 60 1 20 Moving Range 42 UCL=18. 5-63 .75 5 0 LCL=0 1 6 12 18 24 30 Observation 36 42 48 54 60 The process has been returned to a state of statistical control.

One point more than 3. with two Western Electric rule violations.26 30 20 __ M R=11.71 10 0 LC L=0 3 6 9 12 15 18 O bser vation 21 24 27 30 Test Results for I Chart of Ex5-58C TEST 5.88 5 100 80 _ X=73. Test Failed at points: 11 Test Results for MR Chart of Ex5-58C TEST 1.59 40 3 6 9 12 15 18 O bser vation 24 27 30 1 40 M oving Range 21 U C L=38. (b) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Concentration (Ex5-58C) Individual V alue U C L=104.Chapter 5 Exercise Solutions 5-58 (5-51).73 60 LC L=42. Test Failed at points: 17 The process is not in control. in particular due to the curve of the larger values and three distant values. 5-64 .00 standard deviations from center line. (a) The normality assumption is a little bothersome for the concentration data. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).

0 4.5 4.9 4. again due to the curve of the larger values and three distant values.Chapter 5 Exercise Solutions 5-58 continued (c) MTB > Stat > Basic Statistics > Normality Test Probability Plot of ln(Concentration) (Ex5-58lnC) Normal 99 Mean StDev N AD P-Value 95 90 4.327 Percent 80 70 60 50 40 30 20 10 5 1 3.408 0.4 Ex5-58lnC 4.1567 30 0.3 4.6 4.2 4.7 The normality assumption is still troubling for the natural log of concentration. 5-65 .1 4.288 0.

Chapter 5 Exercise Solutions 5-58 continued (d) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of ln(Concentration) (Ex5-58lnC) Individual Value 4.12 0.1577 0. One point more than 3.48 0.6 4. There does not appear to be much difference between the two control charts (actual and natural log).8 U C L=4.2 4.00 standard deviations from center line. Test Failed at points: 11 Test Results for MR Chart of Ex5-58lnC TEST 1.8689 3 6 9 12 15 18 O bser vation 21 24 27 30 1 U C L=0.5154 M oving Range 0. 5-66 .24 __ M R=0.4 _ X=4.00 LC L=0 3 6 9 12 15 18 O bser vation 21 24 27 30 Test Results for I Chart of Ex5-58lnC TEST 5. with the same to Western Electric Rules violations. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).0 LC L=3.2884 4.36 0. Test Failed at points: 17 The process is still not in control.7079 5 4.

9 20 0. use all data for control limits calculations.1 1000 _ X=909 750 LC L=582. For now.672 0. consider the process to be in a state of statistical process control. Test Failed at points: 8 The out-of-control signal on the moving range chart indicates a significantly large difference between successive measurements (7 and 8). There may also be an early indication of less variability in the later measurements. MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Velocity of Light (Ex5-59Vel) Individual V alue 1250 U C L=1235. Since neither of these measurements seems unusual.067 Percent 80 70 60 50 40 30 20 10 5 1 600 700 800 900 Ex5-59Vel 1000 1100 1200 Velocity of light measurements are approximately normally distributed.9 500 2 4 6 8 10 12 O bser vation 14 16 18 20 1 M oving Range 400 U C L=400.Chapter 5 Exercise Solutions 5-59☺. One point more than 3. 5-67 . MTB > Stat > Basic Statistics > Normality Test Probability Plot of Velocity of Light (Ex5-59Vel) Normal 99 Mean StDev N AD P-Value 95 90 909 104.6 100 0 LC L=0 2 4 6 8 10 12 O bser vation 14 16 18 20 I-MR Chart of Ex5-59Vel Test Results for MR Chart of Ex5-59Vel TEST 1.00 standard deviations from center line.7 300 200 __ M R=122.

5-68 . 9 points in a row on same side of center line. meaning the method is producing gradually smaller measurements.1 1000 _ X=909 2 2 2 750 2 2 LCL=582. Failed at points: 8 2. Estimate to specify which subgroups to use in calculations I-MR Chart of Velocity of Light (Ex5-60Vel) New measurements with old limits Individual Value 1250 UCL=1235. 37.00 standard deviations from center line.9 500 4 8 12 16 20 Observation 24 28 32 36 40 1 UCL=400. 9 points in a row on same side of center line. 38. (a) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR Select I-MR Options. Test Failed at points: 36. however the method of measuring is producing varying results—this is a chart of the measurement process.7 Moving Range 400 300 200 100 2 __ MR=122. One point more than 3. 39. 40 Test Results for MR Chart of Ex5-60Vel TEST Test TEST Test 1. (b) Early measurements exhibit more variability than the later measurements. There is a distinct downward trend in measurements. 37 The velocity of light in air is not changing. Failed at points: 36.Chapter 5 Exercise Solutions 5-60☺.6 2 0 LCL=0 4 8 12 16 20 Observation 24 28 32 36 40 I-MR Chart of Ex5-60Vel Test Results for I Chart of Ex5-60Vel TEST 2. which is reflected in the number of observations below the centerline of the moving range chart.

8 Ex5-61lnUn 3.005 Percent 80 70 60 50 40 30 20 10 5 1 0 5 10 15 20 Ex5-61Un 25 30 35 The data are not normally distributed.158 <0.4 2. so a compressive transform such as natural log or square-root may be appropriate.6 The distribution of the natural-log transformed uniformity measurements is approximately normally distributed. as well as the Anderson-Darling test p-value.3493 30 0.626 0.07 5. 5-69 . Probability Plot of ln(Uniformity) (Ex5-61lnUn) Normal 99 Mean StDev N AD P-Value 95 90 2.093 Percent 80 70 60 50 40 30 20 10 5 1 2.546 30 1.0 2. as evidenced by the “S”.shaped curve to the plot points on a normal probability plot.Chapter 5 Exercise Solutions 5-61☺. The data are skewed right. (a) MTB > Stat > Basic Statistics > Normality Test Probability Plot of Uniformity Determinations (Ex5-61Un) Normal 99 Mean StDev N AD P-Value 95 90 15.653 0.2 3.

Chapter 5 Exercise Solutions 5-61 continued (b) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of ln (Uniformity) (Ex5-61lnUn) U C L=3.5 2.586 Individual V alue 3.00 0. 5-70 .5 3 6 9 12 15 18 O bser vation 21 24 27 30 U C L=1.351 0.00 LC L=0 3 6 9 12 15 18 O bser vation 21 24 27 30 The etching process appears to be in statistical control.720 1.0 LC L=1.0 _ X=2.5 3.146 M oving Range 1.50 __ M R=0.25 0.653 2.75 0.

174 <0.81 0.85 0.Chapter 5 Exercise Solutions 5-62 (5-52).87 Purity is not normally distributed.80 0.84 Ex5-62Pur 0.83 0.01847 20 1. 5-71 .824 0. (a) MTB > Stat > Basic Statistics > Normality Test Probability Plot of Batch Purity (Ex5-62Pur) Normal 99 Mean StDev N AD P-Value 95 90 0.82 0.005 Percent 80 70 60 50 40 30 20 10 5 1 0.86 0.79 0.78 0.

824 . Test Failed at points: 11. One point more than 3.82 6 0. σˆ x = 0. Test Failed at points: 18 TEST 5.024 __ M R=0.000 LC L=0 2 4 6 8 10 12 O bser vation 14 16 18 20 Test Results for I Chart of Ex5-62Pur TEST 1.00 standard deviations from center line. 20 The process is not in statistical control. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL).0135 without sample 18: µˆ = 0.78 2 4 6 8 10 12 O bser vation 14 16 18 20 U C L=0.824 0.78341 0.8216 .0133 5-72 .012 0.80 LC L=0.86459 5 0. Test Failed at points: 19 TEST 6.04987 M oving Range 0.01526 0.036 0. σˆ x = 0.84 _ X=0.Chapter 5 Exercise Solutions 5-62 continued (b) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Purity (Ex5-62Pur) 1 Individual V alue U C L=0. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).86 6 0.048 0. (c) all data: µˆ = 0.

1681 16. control limits for both are essentially the same. MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR Select “Estimate” to change the method of estimating sigma I-MR Chart of Can Weight (Ex5-53Wt) Individual V alue 16.11 16.08 U C L=0.05 LC L=16.Chapter 5 Exercise Solutions 5-63 (5-53).14 _ X=16.06 0.0423 2 4 6 8 10 12 14 O bser vation 16 18 20 22 24 M oving Range 0.1052 16.17 U C L=16.08 16. 5-73 .07726 0.02 0.04 __ M R=0.02365 0.00 LC L=0 2 4 6 8 10 12 14 O bser vation 16 18 20 22 24 There is no difference between this chart and the one in Exercise 5-53.

5 0. 5-74 . with no practical difference for this set of observations.27 50 LC L=45.Chapter 5 Exercise Solutions 5-64 (5-54). MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR Select “Estimate” to change the method of estimating sigma I-MR Chart of Hardness-Coded (Ex5-54Har) U C L=61.0 __ M R=2.0 LC L=0 1 2 3 4 5 6 7 8 9 O bser vation 10 11 12 13 14 15 The median moving range method gives slightly tighter control limits for both the Individual and Moving Range charts.13 Individual V alue 60 55 _ X=53.5 5.96 2.0 U C L=9.41 45 1 2 3 4 5 6 7 8 9 O bser vation 10 11 12 13 14 15 M oving Range 10.66 7.

7 3200 3000 _ X=2928. 5-75 .2 M oving Range 480 360 240 __ M R=153. with no practical meaning for this set of observations.Chapter 5 Exercise Solutions 5-65 (5-55). MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR Select “Estimate” to change the method of estimating sigma I-MR Chart of Polymer Viscosity (Ex5-55Vis) Individual Value 3400 U C L=3337.9 2800 2600 LC L=2520.1 2 4 6 8 10 12 O bser vation 14 16 18 20 U C L=502.7 120 0 LC L=0 2 4 6 8 10 12 O bser vation 14 16 18 20 The median moving range method gives slightly wider control limits for both the Individual and Moving Range charts.

One point more than 3. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL).07 2 6 _ X=51. Remove observation 38 and recalculate control limits.05 50 40 LCL=36. 5-76 . Failed at points: 38. 40 Test Results for MR Chart of Ex5-57cTh TEST 1.00 standard deviations from center line.Chapter 5 Exercise Solutions 5-66 (5-56). 9 points in a row on same side of center line. 40 TEST 6. Test Failed at points: 41 Recall that observations on the Moving Range chart are correlated with those on the Individuals chart—that is. 39. 39. Test Failed at points: 35.00 standard deviations from center line. 40 5.65 5 0 LCL=0 1 6 12 18 24 30 Observation 36 42 48 54 60 Test Results for I Chart of Ex5-57cTh TEST Test TEST Test TEST 1. 37. 39. MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR (a) I-MR Chart of Oxide Thickness (Ex5-57cTh) All 60 Observations--Average Moving Range Method 70 1 Individual Value 5 60 2 6 UCL=66.46 15 10 __ MR=5. Test Failed at points: 34. Failed at points: 38 2. 38. One point more than 3. the out-of-control signal on the MR chart for observation 41 is reflected by the shift between observations 40 and 41 on the Individuals chart. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).03 1 6 12 18 24 30 Observation 36 48 54 60 1 20 Moving Range 42 UCL=18.

77 50 40 LCL=35. Test Failed at points: 41 5-77 .00 standard deviations from center line.60 2 6 _ X=50. One point more than 3.22 15 10 5 __ MR=5. Failed at points: 38 2. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). Failed at points: 38. 40 TEST 6. 40 5. 39. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL).58 0 LCL=0 1 6 12 18 24 30 Observation 36 42 48 54 60 Test Results for I Chart of Ex5-57cTh TEST Test TEST Test TEST 1.94 1 6 12 18 24 30 Observation 36 42 48 54 60 1 20 Moving Range UCL=18.Average Moving Range Method 70 1 Individual Value 5 60 2 6 UCL=65. Test Failed at points: 34. 9 points in a row on same side of center line.Chapter 5 Exercise Solutions 5-66 (a) continued Excluding observation 38 from calculations: I-MR Chart of Oxide Thickness (Ex5-57cTh) Less Observation 38 -. 40 Test Results for MR Chart of Ex5-57cTh TEST 1. 39. Test Failed at points: 35. 37. One point more than 3.00 standard deviations from center line. 39. 38.

One point more than 3. 38. 40 5. 40 TEST 6.05 50 40 LCL=36. 40 Test Results for MR Chart of Ex5-57cTh TEST 1. Failed at points: 38 2.56 0 LCL=0 1 6 12 18 24 30 Observation 36 42 48 54 60 Test Results for I Chart of Ex5-57cTh TEST Test TEST Test TEST 1. Failed at points: 38.Chapter 5 Exercise Solutions 5-66 continued (b) I-MR Chart of Oxide Thickness (Ex5-57cTh) All 60 Observations -. One point more than 3. Test Failed at points: 34. 39. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). Test Failed at points: 35. 9 points in a row on same side of center line. Test Failed at points: 41 5-78 .00 standard deviations from center line. 39.00 standard deviations from center line. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 39.83 2 6 _ X=51.27 1 6 12 18 24 30 Observation 36 42 48 54 60 1 20 Moving Range UCL=18. 37.16 15 10 5 __ MR=5.Median Moving Range Method 70 1 Individual Value 5 60 2 6 UCL=65.

4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).00 standard deviations from center line. One point more than 3.93 1 6 12 18 24 30 Observation 36 42 48 54 60 1 20 Moving Range UCL=17. 40 TEST 6. Failed at points: 38. 40 Test Results for MR Chart of Ex5-57cTh TEST 1.77 50 40 LCL=36. 38 2. 9 points in a row on same side of center line. 38.61 6 _ X=50. 39. Test Failed at points: 34.00 15 10 5 __ MR=5.00 standard deviations from center line. 40 5. 33. Test Failed at points: 35. 39. One point more than 3. 37. 39. 5-79 .Median Moving Range Method 70 1 1 Individual Value 5 60 2 6 2 UCL=64. Test Failed at points: 41 (c) The control limits estimated by the median moving range are tighter and detect the shift in process level at an earlier sample.Chapter 5 Exercise Solutions 5-66 (b) continued Excluding observation 38 from calculations: I-MR Chart of Oxide Thickness (Ex5-57cTh) Excluding Observation 38 from Calculations -. Failed at points: 33.20 0 LCL=0 1 6 12 18 24 30 Observation 36 42 48 54 60 Test Results for I Chart of Ex5-57cTh TEST Test TEST Test TEST 1.

630 σˆ x = S / c4 = 1.018 12 _ X=10.079 2 4 6 8 10 12 14 O bser vation 16 18 20 22 24 U C L=4.128 = 1.342 / 0.Chapter 5 Exercise Solutions 5-67 (5-57).157 (b) MTB > Stat > Basic Statistics > Descriptive Statistics Descriptive Statistics: Ex5-67Meas Variable Ex5-67Meas Total Count 25 Mean 10.682 5-80 .305 /1.549 StDev 1.342 Median 10.7979 = 1.549 10 8 LC L=7. (a) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Measurements (Ex5-67Meas) Individual V alue 14 U C L=14.305 1 0 LC L=0 2 4 6 8 10 12 14 O bser vation 16 18 20 22 24 σˆ x = R / d 2 = 1.262 M oving Range 4 3 2 __ M R=1.

961 12 _ X=10.549 10 8 LCL=7.128 = 1.693 = 1. This tends to be true for unstable processes. 5-81 .435 (e) As the span of the moving range is increased.049 /1.283 1 LCL=0 0 2 4 6 8 10 12 14 Observation 16 18 20 22 24 σˆ x = R / d 2 = 1.186 / 3.192 Moving Range 4 3 2 __ MR=1.283 /1.210 Average MR4 Chart: σˆ x = R / d 2 = 2.735 = 1.137 (d) Average MR3 Chart: σˆ x = R / d 2 = 2.598 / 2.689 = 1. σ gets larger as the span increases.262 Average MR19 Chart: σˆ x = R / d 2 = 5.137 2 4 6 8 10 12 14 Observation 16 18 20 22 24 UCL=4. For this example.36 / 3.Chapter 5 Exercise Solutions 5-67 continued (c) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Measurements (Ex5-67Meas) Median Moving Range Method--Span = 2 Individual Value 14 UCL=13. there are fewer observations to estimate the standard deviation. and the estimate becomes less reliable.059 = 1.406 Average MR20 Chart: σˆ x = R / d 2 = 5.

5-82 .1233 0.70 LCL=5.04 0.0143831 Between/Within 0.Chapter 5 Exercise Solutions 5-68 (5-58).. For this example.00 LCL=0 2 4 6 8 10 Ex5-68Cast 12 14 16 18 20 I-MR-R/S Standard Deviations of Ex5-68v1.0358361 The Individuals and Moving Range charts for the subgroup means are identical.00 LCL=0 2 4 6 8 10 12 14 16 18 20 Sample Range 0.8569 5.08 UCL=0.6563 MR of Subgroup Mean 2 4 6 8 10 12 14 16 18 20 UCL=0.10 0.07074 _ R=0. Ex5-68v5 Standard Deviations Between 0..82 _ X=5.76 5. .. Estimate” and choose R-bar method to estimate standard deviation I-MR-R (Between/Within) Chart of Vane Heights (Ex5-68v1...7566 5.0328230 Within 0. the R chart tells the same story—same data pattern and no out-of-control points.0377 0.. the control schemes are identical. . When compared to the s chart for all data.03345 0. Ex5-68v5) Subgroup Mean UCL=5. MTB > Stat > Control Charts > Variables Charts for Subgroups > I-MR-R/S (Between/Within) Select “I-MR-R/S Options.05 __ MR=0.

1292 Sample Range 0. we may want to consider treating the averages as an Individual value and graphing “between/within” range charts..Chapter 5 Exercise Solutions 5-69 (5-59).. Test Failed at points: 7 (b) Though the R chart is in control.7931 11. 5-83 .73 LC L=11.12 0. One point more than 3.7579 11..00 standard deviations from center line.79 U C L=11. .06 0.0611 0.70 1 1 2 4 6 8 10 Sample 12 14 16 18 20 U C L=0. Ex5-69d5 Test Results for Xbar Chart of Ex5-69d1. 17 TEST 5. 7. This will lead to a understanding of the greatest source of variability.82 1 Sample M ean 1 11. Test Failed at points: 5.09 _ R=0. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). we might expect that the diameter of a single casting will not change much with location..7226 11. Ex5-69d5) 1 11. . between castings or within a casting...00 LC L=0 2 4 6 8 10 Sample 12 14 16 18 20 Xbar-R Chart of Ex5-69d1. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Casting Diameter (Ex5-69d1.. .. Ex5-69d5 TEST 1. 9. If no assignable cause can be found for these out-of-control points. Since these are high precision castings.. plot points on the x chart bounce below and above the control limits.76 _ _ X=11.03 0. 13.

..10 _ R=0. . It describes the measurement variability within a sample (variability in diameter of a single casting). Ex5-69d5 Standard Deviations Between 0. .00 LCL=0 2 4 6 8 10 Sample 12 14 16 18 20 I-MR-R/S (Between/Within) Chart of Ex5-69d1.7 LCL=11.1292 0. we should continue to monitor “within” to look for wear. Estimate” and choose R-bar method to estimate standard deviation MR-R/S (Between/Within) Chart of Casting Diameter (Ex5-69d1.8685 11. Ex5-69d5 I-MR-R/S Standard Deviations of Ex5-69d1.0611 0.05 0.7579 11. 5-84 . etc..8 _ X=11.0437327 (d) We are taking several diameter measurements on a single precision casting.9 Subgroup Mean UCL=11...10 __ MR=0.1360 0.05 LCL=0 0.. (e) The “within” chart is the usual R chart (n > 1).Chapter 5 Exercise Solutions 5-69 continued (c) MTB > Stat > Control Charts > Variables Charts for Subgroups > I-MR-R/S (Between/Within) Select “I-MR-R/S Options..0349679 Within 0. Ex5-69d5 11.. Though the nature of this process leads us to believe that the diameter at any location on a single casting does not change much.6472 MR of Subgroup Mean 2 4 6 8 10 12 14 16 18 20 UCL=0. . damage..00 2 4 6 8 10 12 14 16 18 20 Sample Range UCL=0..0262640 Between/Within 0. in the wax mold.0416 0.

Then a chart for individual x and moving range would provide information on lot-to-lot variability. then only one location. 5-85 . Individuals X and Moving Range charts should be used for process monitoring. (2) moving range is the difference between successive wafers. and (3) sample range is the difference within a wafer ( Rww ) . The data from each wafer could also be used to monitor between-wafer variability by maintaining an individuals X and moving range chart for each of the five fixed positions. However. then either five X − R charts or some multivariate technique is needed. Alternatively. If the positions are essentially identical. (c) Both between-wafer and total process variability could be estimated from measurements at one point on five consecutive wafers. (e) The simplest scheme would be to randomly select one wafer from each lot and treat the average of all measurements on that wafer as one observation. (a) Both total process variability and the overall process average could be estimated from a single measurement on one wafer from each lot. needs to be monitored. within-wafer variability could be monitored with a standard X − R control chart. (d) Within-wafer variability can still be monitored with randomly selected test sites.Chapter 5 Exercise Solutions 5-70 (5-60). with one X − R chart. If it is necessary to separately monitor the variation at each location. no information will be obtained about the pattern of variability within a wafer. a multivariate process control technique could be used. (b) Assuming that each wafer is processed separately. The Minitab “between/within” control charts do this in three graphs: (1) wafer mean ( xww ) is an “individual value”.

04515 200 1. with many repeated values.15 2. s = 0.9 Mean StDev N AD P-Value 99 Percent 95 90 2.005 80 70 60 50 40 30 20 10 5 1 0.04515 UNTL = x + 3s = 2.939 5-86 .20 2.074 0.333 <0. the plot points do fall along a straight line.90 1.074 − 3(0.209 LNTL = x − 3s = 2.1 1. The natural tolerance limits (± 3 sigma above and below mean) are: x = 2.95 2.10 Ex5-71All 2.05 2.25 Although the p-value is very small.074. (a) MTB > Stat > Basic Statistics > Normality Test Probability Plot of Critical Dimensions (Ex5-71All) Normal 99.Chapter 5 Exercise Solutions 5-71 (5-61).04515) = 2.00 2.04515) = 1.074 + 3(0. The wafer critical dimension is approximately normally distributed.

construct an R chart for each sample of 5 wafer positions (two wafers per lot number). for a total of 40 subgroups. indicating that within-wafer variability is also in control.. MTB > Stat > Control Charts > Variables Charts for Subgroups > R R Chart of Critical Dimension Within Wafer (Ex5-71p1.Chapter 5 Exercise Solutions 5-71 continued (b) To evaluate within-wafer variability...14 Sample Range 0.16 UCL=0.08 _ R=0.1480 0.00 LCL=0 4 8 12 16 20 24 28 32 Sample (Lot Number-Wafer Order) 36 40 The Range chart is in control. Ex5-71p5) 0.12 0.10 0.07 0.06 0. 5-87 .02 0. .04 0.

1066 0. Estimate” and choose R-bar method to estimate standard deviation I-MR-R/S (Between/Within) Chart of Crit Dim (Ex5-71p1.10 0.05 __ MR=0.16 UCL=0..0255911 Within 0.0326 LCL=0 0. .00 LCL=1.0735 6 6 2.1603 2.08 0. indicating that between-wafer variability is also in-control. set up Individuals and Moving Range charts where the x statistic is the average wafer measurement and the moving range is calculated between two wafer averages.00 LCL=0 4 8 12 16 20 Sample 24 28 32 36 40 I-MR-R/S Standard Deviations of Ex5-71p1. 5-88 .07 0..00 4 8 12 16 20 24 28 32 36 40 Sample Range 0. . Ex5-71p5 Standard Deviations Between 0.9868 MR of Subgroup Mean 4 8 12 16 20 24 28 32 36 40 UCL=0. MTB > Stat > Control Charts > Variables Charts for Subgroups > I-MR-R/S (Between/Within) Select “I-MR-R/S Options.08 _ X=2.0395043 Both “between” control charts (Individuals and Moving Range) are in control.Chapter 5 Exercise Solutions 5-71 continued (c) To evaluate variability between wafers... Ex5-71p5) Subgroup Mean Variability between wafers 2.1480 _ R=0..0300946 Between/Within 0. The “within” chart (Range) is not required to evaluate variability between wafers.16 UCL=2..

0459 0.096 0.00 LCL=0 2 4 6 8 10 12 14 16 18 20 Sample Range UCL=0. MTB > Stat > Control Charts > Variables Charts for Subgroups > I-MR-R/S (Between/Within) I-MR-R/S (Between/Within) Chart of Ex5-71All Subgroup Mean Lot-to-Lot Variability 2.2 UCL=2.0503081 All three control charts are in control.10 0. indicating that the lot-to-lot variability is also in-control.1 _ X=2.08 __ MR=0.0311891 Between/Within 0.05 LCL=0.Chapter 5 Exercise Solutions 5-71 continued (d) To evaluate lot-to-lot variability.0 LCL=1.15 _ R=0.1706 0.0214 2 4 6 8 10 Ex5-7Lot All 12 14 16 18 20 I-MR-R/S Standard Deviations of Ex5-71All Standard Deviations Between 0.1500 0.16 UCL=0. and (3) range within a lot—the Minitab “between/within” control charts.0735 2.1956 2.0394733 Within 0. three charts are needed: (1) lot average. (2) moving range between lot averages.9515 2 4 6 8 10 12 14 16 18 20 MR of Subgroup Mean 0. 5-89 .

it is considered to be out of control.0585) = 0. not 8.0585(1 − 0.16 1 0.1289 Proportion 0. ∑ Di = 117. For these solutions. 3. a run of n consecutive points on one side of the center line is defined as 9 points.0585 0.02 0. 2. MINITAB defines some sensitizing rules for control charts differently than the standard rules.04 0. Also fewer special cause tests are available for attributes control charts.Chapter 6 Exercise Solutions Notes: 1. New exercises are denoted with an “☺”.00 standard deviations from center line. it is considered to be in control.0585 + 3 = 0. In particular.0585 − 0. p = i =1 ∑ Di i =1 mn = 117 = 0. 6-1. This can be changed under Tools > Options > Control Charts and Quality Tools > Define Tests.0585 − 3 = 0.0585) = 0.14 UCL=0.0585 20(100) UCL p = p + 3 p (1 − p ) 0. If a plot point is within the control limits.10 0. m = 20. m m n = 100.00 LCL=0 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-1Num TEST 1. One point more than 3.0585(1 − 0. Test Failed at points: 12 6-1 . If a plot point is on or beyond the control limits.06 0.08 _ P=0.12 0.0704 ⇒ 0 n 100 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Assemblies (Ex6-1Num) 0.1289 n 100 LCL p = p − 3 p (1 − p ) 0. we follow the MINITAB convention for determining whether a point is out of control.

0537 − 0.0537) = 0.0676 ⇒ 0 100 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Assemblies (Ex6-1Num) Sample 12 removed from calculations 0.0537 0.0537) = 0.Chapter 6 Exercise Solutions 6-1 continued Sample 12 is out-of-control.1213 Proportion 0.1213 100 LCL p = 0. Test Failed at points: 12 6-2 .06 0.08 _ P=0.14 UCL=0.02 0. so remove from control limit calculation: m m n = 100.10 0.04 0.16 1 0.00 standard deviations from center line.0537(1 − 0.0537 19(100) UCL p = 0. One point more than 3.00 LCL=0 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-1Num TEST 1. ∑ Di = 102.0537 − 3 0.12 0.0537 + 3 0.0537(1 − 0. p = i =1 ∑ Di i =1 mn = 102 = 0. m = 19.

Chapter 6 Exercise Solutions 6-2. 17 6-3 . ∑ Di = 69.0230 − 3 = 0. Test Failed at points: 9. p = ∑ Di i =1 mn i =1 = 69 = 0. m m n = 150.0230 + 3 = 0.0230(1 − 0.0597 n 150 LCL p = p − 3 p (1 − p ) 0.00 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-2Num TEST 1.0230 − 0.00 standard deviations from center line. One point more than 3.023 0.0230) = 0.0367 ⇒ 0 n 150 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Switches (Ex6-2Num) 1 0.04 _ P=0. m = 20.06 UCL=0.02 LCL=0 0.0230 20(150) UCL p = p + 3 p (1 − p ) 0.10 0.08 Proportion 1 0.0230) = 0.0597 0.0230(1 − 0.

0163 0.04 0.02 _ P=0. 17 6-4 .08 Proportion 1 0.06 1 UCL=0.0473 0.10 0.Chapter 6 Exercise Solutions 6-2 continued Re-calculate control limits without samples 9 and 17: MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Switches (Ex6-2Num) Samples 9 and 17 excluded from calculations 1 0.00 standard deviations from center line. Test Failed at points: 1. 9. One point more than 3.00 LCL=0 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-2Num TEST 1.

00 standard deviations from center line. m = 17. 17 excluded from calculations 1 0.Chapter 6 Exercise Solutions 6-2 continued Also remove sample 1 from control limits calculation: m m n = 150.0430 150 LCL p = 0.08 Proportion 1 0.0141 + 3 0. 17 6-5 . 9.0141) = 0.0289 ⇒ 0 150 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Switches (Ex6-2Num) Samples 1.0141) = 0.10 0.06 1 UCL=0.0141 − 0.04 0. ∑ Di = 36. One point more than 3.0141(1 − 0.0141 17(150) UCL p = 0. Test Failed at points: 1.0141(1 − 0.0141 − 3 0.02 _ P=0. 9.00 LCL=0 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-2Num TEST 1. p = i =1 ∑ Di i =1 mn = 36 = 0.0141 0.0430 0.

∑ ni = 1000. The Fraction Nonconforming for Day 5 should be 0. ∑ Di = 60.16 0.10 0.06 − 0.14 UCL=0.06 0.00 LCL=0 1 2 3 4 5 6 Sample 7 8 9 10 Tests performed with unequal sample sizes The process appears to be in statistical control.06 UCLi = p + 3 p (1 − p ) ni and LCLi = max{0.1397 LCL1 = p − 3 p (1 − p ) n1 = 0.06 0.06) 80 = 0. p − 3 p (1 − p ) ni } As an example. for n = 80: UCL1 = p + 3 p (1 − p ) n1 = 0. 6-6 .02 0.Chapter 6 Exercise Solutions 6-3.06 − 3 0.06(1 − 0. p = ∑ Di ∑ ni = 60 1000 = 0.06(1 − 0.06 + 3 0. m m m m i =1 i =1 i =1 i =1 m = 10.06) 80 = 0.08 _ P=0.046.04 0.12 0. NOTE: There is an error in the table in the textbook.1331 Proportion 0.0797 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Units (Ex6-3Num) 0.

0167(1 − 0.0167 − 3 0. 6-7 .0167 UCL = p + 3 p (1 − p ) n = 0. ∑ Di = 50.0167) 150 = 0. (1 − p) 2 n> L p (1 − 0.0167 + 3 0. (a) m m i =1 i =1 n = 150. (b) Using Equation 6-12.0167(1 − 0.03 0.01 LCL=0 0.9 Select n = 530.0167) 2 (3) > 0.0167) 150 = 0.0314 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Units (Ex6-4Num) 0.0167 > 529.05 UCL=0.04802 Proportion 0.01667 0.04 0.02 _ P=0.00 2 4 6 8 10 12 Sample 14 16 18 20 The process appears to be in statistical control.0480 LCL = p − 3 p (1 − p ) n = 0. p = ∑ Di mn = 50 20(150) = 0. m = 20.Chapter 6 Exercise Solutions 6-4.0167 − 0.

1228(1 − 0. 6-8 .1031 1 1 1 0. 3.075 1 1 0.1228) 2500 = 0. 2. Instead.100 LCL=0.1425 LCL = p − 3 p (1 − p ) n = 0. the process should be investigated for causes of the wild swings in p.125 _ P=0. 5.050 1 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-5Num TEST 1. 12. 17.200 1 1 0.1425 0. 19.1228 0.1228) 2500 = 0.Chapter 6 Exercise Solutions 6-5.1228 + 3 0. One point more than 3.1228 − 3 0.1031 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Belts (Ex6-5Num) 0. 11.1228(1 − 0.00 standard deviations from center line. 16.150 UCL=0. (a) UCL = p + 3 p (1 − p ) n = 0. 15. 20 (b) So many subgroups are out of control (11 of 20) that the data should not be used to establish control limits for future production. Test Failed at points: 1.175 1 1 1 Proportion 0.

98 8 6 __ NP=4 4 2 LCL=0 0 1 2 3 4 5 6 Ex6-6Day 7 8 9 10 Test Results for NP Chart of Ex6-6Num TEST 1.976 LCL = np − 3 np (1 − p ) = 4 − 3 4(1 − 0.008) = 9. Test Failed at points: 6 6-9 .008) = 4 − 5.976 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > NP NP Chart of Number of Nonconforming Units (Ex6-6Num) 1 12 Sample Count 10 UCL=9. One point more than 3.00 standard deviations from center line.Chapter 6 Exercise Solutions 6-6. UCL = np + 3 np (1 − p ) = 4 + 3 4(1 − 0.

11 2 0 LCL=0 1 2 3 4 5 6 Ex6-6Day 7 8 9 10 Test Results for NP Chart of Ex6-6Num TEST 1. Test Failed at points: 6 Recommend using control limits from second chart (calculated less sample 6). 6-10 .6 continued Recalculate control limits without sample 6: NP Chart of Number of Nonconforming Units (Ex6-6Num) Day 6 excluded from control limits calculations 1 12 Sample Count 10 UCL=8.Chapter 6 Exercise Solutions 6. One point more than 3.39 8 6 4 __ NP=3.00 standard deviations from center line.

2) = 1 – 0.02 − 3 0. use the Poisson approximation to the binomial with λ = npnew = 50(0.2) + POI(0.0170. n = 250. Pr{detected by 3rd sample} = 1 – Pr{detected after 3rd} = 1 – (1 – 0. 0.02) 50 = 0.0440 = 0.0044) 250 = 0. The data from the shipment do not indicate statistical control.0126 ⇒ 0 No.04) = 2.0794 LCL = p − 3 p (1 − p ) n = 0.0170 10 m = 10.857 + 0.278)3 = 0.02.0440. the UCL. ( pˆ 6 = 0.0044 − 3 0. 6-11 .278 where POI(⋅) is the cumulative Poisson distribution.02 − 0. ∑ pˆ i = 0.0044 − 0.02(1 − 0.00.0044 + 3 0.0044(1 − 0. n = 50 UCL = p + 3 p (1 − p ) n = 0. From the 6th sample.02(1 − 0.02 + 3 0.0594 ⇒ 0 Since pnew = 0. p = UCL = p − 3 p (1 − p ) n = 0.0044) 250 = 0.Chapter 6 Exercise Solutions 6-7.02) 50 = 0.04 < 0.0044(1 − 0.0044 i =1 10 UCL = p + 3 p (1 − p ) n = 0. Pr{detect|shift} = 1 – Pr{not detect|shift} =1–β = 1 – [Pr{D < nUCL | λ} – Pr{D ≤ nLCL | λ}] = 1 – Pr{D < 50(0.0794) | 2} + Pr{D ≤ 50(0) | 2} = 1 – POI(3.624 6-8.1 and n = 50 is "large". p = 0.020) > 0.135 = 0.

10 − 0.000000 0.6) | p} − Pr{D ≤ 0 | p} p 0.480098 0.20 0.21 0.215 0.05 0.10 0.10 > 81 6-12 .000000 0.000000 β 0. (1 − p ) 2 n> L p (1 − 0.996172 0.2125) | p} − Pr{D ≤ 64(0) | p} = Pr{D < 13. p = 0.480098 0. n = 64 UCL = p + 3 p (1 − p ) n = 0.44154 0.503553 Assuming L = 3 sigma control limits.999999 0.10 − 3 0.1125 ⇒ 0 β = Pr{D < nUCL | p} − Pr{D ≤ nLCL | p} = Pr{D < 64(0.Chapter 6 Exercise Solutions 6-9.10.10) 2 (3) > 0.22 0.212 Pr{D ≤ 13|p} 0.44154 0.598077 0.503553 Pr{D ≤ 0|p} 0.10(1 − 0.001179 0.000000 0.598077 0.519279 0.994993 0.037524 0.10(1 − 0.000000 0.10) 64 = 0.962475 0.10) 64 = 0.519279 0.2125 LCL = p − 3 p (1 − p ) n = 0.10 + 3 0.

16) = 27.16) = 5.25 So.5 − 20 ⎞ = 1− Φ ⎜ + Φ⎜ ⎟ ⎜ 20(1 − 0. Pr{detect shift on 1st sample} = 1 − β = 1 − [Pr{D < UCL | p} − Pr{D ≤ LCL | p}] ⎛ UCL + 1/ 2 − np ⎞ ⎛ LCL − 1/ 2 − np ⎞ = 1− Φ ⎜ +Φ⎜ ⎟ ⎟ ⎜ ⎜ np (1 − p) ⎟⎠ np (1 − p ) ⎟⎠ ⎝ ⎝ ⎛ 27 + 0.50. n = 100. so use normal approximation to binomial distribution. n = 48 is the minimum sample size for a positive LCL.00 (a) npnew = 20.Chapter 6 Exercise Solutions 6-10.10 ⎠ 6-13 .0 > 15.0873 (b) Assuming L = 3 sigma control limits.0.970 + 0.10) = 81 ⎝δ ⎠ ⎝ 0.875) + Φ (−3.000 = 0.10 = 0. (1 − p ) 2 n> L p (1 − 0.10)(1 − 0. desire Pr{detect} = 0.2) ⎟ ⎜ 20(1 − 0.030 Pr{detect by at least 3rd} = 1 – Pr{detected after 3rd} = 1 – (1 – 0.10. p = 0.16 UCL = np + 3 np(1 − p ) = 16 + 3 16(1 − 0.030)3 = 0. 6-11.16) 2 (3) > 0. assume k = 3 sigma control limits new − p = 0.5 − 20 ⎞ ⎛ 5 − 0.00 LCL = np − 3 np (1 − p ) = 16 − 3 16(1 − 0.10 δ =p new 2 2 ⎛k⎞ ⎛ 3 ⎞ n = ⎜ ⎟ p(1 − p) = ⎜ ⎟ (0.875) = 1 − 0. np = 16. p = 16 100 = 0.20.16 > 47.20 − 0.2) ⎟⎟ ⎝ ⎠ ⎝ ⎠ = 1 − Φ (1. p = 0.

07494 − 0 = 0. (c) npnew = 100(0.37) = 0.20 0 − 0.161 − 0.080 < 0. UCL = 0.080) = 8 − 8.99997 6-14 .080) = 8 UCL = np + 3 np(1 − p ) = 8 + 3 8(1 − 0.996] = 0.08(1 − 0.1 and n =100 is large. p = 0.08(1 − 0.004 where POI(⋅) is the cumulative Poisson distribution. Pr{type II error} = β = Pr{ pˆ < UCL | pnew } − Pr{ pˆ ≤ LCL | pnew } ⎛ UCL − pnew ⎞ ⎛ LCL − pnew ⎞ = Φ⎜ −Φ⎜ ⎟ ⎜ p (1 − p ) n ⎟ ⎜ p (1 − p ) n ⎟⎟ ⎝ ⎠ ⎝ ⎠ ⎛ ⎞ ⎛ ⎞ 0.8)] = 0 + [1 – 0.Chapter 6 Exercise Solutions 6-12. n = 100. LCL = 0 (a) np = 100(0. Pr{type I error} = α = Pr{D < LCL | p} + Pr{D > UCL | p} = Pr{D < LCL | p} + [1 – Pr{D ≤ UCL | p}] = Pr{D < 0 | 8} + [1 – Pr{D ≤ 16 | 8}] = 0 + [1 – POI(16.44) − Φ (−7.08.07494)4 = 0. so use Poisson approximation to the binomial.161.08) 100 ⎟⎟ ⎝ ⎠ ⎝ ⎠ = Φ (−1.1388 ⇒ 0 (b) p = 0.20 = Φ⎜ −Φ⎜ ⎟ ⎜ 0.07494 (d) Pr{detect shift by at most 4th sample} = 1 – Pr{not detect by 4th} = 1 – (0.14 LCL = np − 3 np (1 − p ) = 8 − 3 8(1 − 0. so use the normal approximation to the binomial.080) = 16.20) = 20 > 15.08) 100 ⎟ ⎜ 0.

07) 400 = 0.07) 400 = 0.20) 2 (3) > 0.032 (b) npnew = 400(0.533) = 1 − 0.20) = 400 ⎝δ ⎠ ⎝ 0.297) = 0.07. k = 3 sigma control limits. Pr{detect on 1st sample} = 1 − Pr{not detect on 1st sample} = 1− β = 1 − [Pr{ pˆ < UCL | p} − Pr{ pˆ ≤ LCL | p}] ⎛ UCL − p ⎞ ⎛ LCL − p ⎞ = 1− Φ ⎜ +Φ⎜ ⎟ ⎜ p (1 − p) n ⎟ ⎜ p(1 − p ) n ⎟⎟ ⎝ ⎠ ⎝ ⎠ ⎛ ⎛ 0.07 − 3 0.000 = 0.Chapter 6 Exercise Solutions 6-13.297)(0. n = 400 UCL = p + 3 p(1 − p) n = 0. so use the normal approximation to the binomial.297 (c) Pr{detect on 1st or 2nd sample} = Pr{detect on 1st} + Pr{not on 1st}×Pr{detect on 2nd} = 0.032 − 0. δ =pnew − p = 0.20 = 0.07(1 − 0. p = 0.1(1 − 0.1 ⎞ 0.1(1 − 0. (a) p = 0.06 2 2 ⎛k⎞ ⎛ 3 ⎞ n = ⎜ ⎟ p(1 − p) = ⎜ ⎟ (0.20 > 36 For Pr{detect} = 0.10) = > 40.26.1) 400 ⎟ ⎜ 0.26 − 0.1 ⎞ = 1− Φ ⎜ + Φ⎜ ⎟ ⎜ 0.06 ⎠ 6-15 .506 6-14.703 + 0.108 − 0.20)(1 − 0.297 + (1 – 0.108 LCL = p − 3 p(1 − p ) n = 0.1) 400 ⎟⎟ ⎝ ⎠ ⎝ ⎠ = 1 − Φ (0.20 and L = 3 sigma control limits (1 − p ) 2 n> L p (1 − 0.07(1 − 0.07 + 3 0.533) + Φ (−4.50 after a shift to pnew = 0.

4 + 3 16. (a) m = 10.29 5 1 2 3 4 5 6 Sample 7 8 9 10 Test Results for NP Chart of Ex6-15Num TEST 1. Test Failed at points: 3 6-16 . np = 16.51 Sample Count 25 20 __ NP=16.Chapter 6 Exercise Solutions 6-15. i =1 10 p = ∑ Di i =1 ( mn ) = 164 [10(100)] = 0.00 standard deviations from center line.51 LCL = np − 3 np (1 − p ) = 16. 10 ∑ Di = 164.164) = 27.4 UCL = np + 3 np (1 − p ) = 16.4 − 3 16. n = 100.292 MTB > Stat > Control Charts > Attributes Charts > NP NP Chart of Number Nonconforming (Ex6-15Num) 1 30 UCL=27. One point more than 3.164.4(1 − 0.4 15 10 LCL=5.164) = 5.4(1 − 0.

Chapter 6 Exercise Solutions 6-15 continued Recalculate control limits less sample 3: NP Chart of Number Nonconforming (Ex6-15Num) Sample 3 excluded from calculations 1 30 UCL=25. One point more than 3.13 1 2 3 4 5 6 Sample 7 8 9 10 Test Results for NP Chart of Ex6-15Num TEST 1.42 Sample Count 25 20 __ NP=14.78 15 10 5 LCL=4.00 standard deviations from center line. Test Failed at points: 3 6-17 .

Pr{detect on 1st sample | p} = 1 − Pr{not detect | p} = 1− β = 1 − [Pr{ pˆ < UCL | p} − Pr{ pˆ ≤ LCL | p}] = 1 − Pr{D < nUCL | np} + Pr{D ≤ nLCL | np} = 1 − Pr{D < 200(0.16) + POI(0.03) 200 = 0.0662 LCL p = p − 3 p (1 − p ) n = 0.9943 6-18 .08.0662) | 200(0.03(1 − 0.16) = 1 − 0.30 is not too far from 0. the normal approximation to the binomial can be used.2745 + 0.187) + (0. Since (pnew = 0.50.3) ⎟⎠ ⎝ ⎝ ⎠ = 1 − Φ (−0.13 − 0.03 + 3 0.03 − 0.5 − 30 ⎞ ⎛ 4.08)} + Pr{D ≤ 200(0) | 200(0.3) ⎟⎟ 30(1 − 0.000 = 0. Pr{detect on 1st} = 1 − Pr{not detect on 1st} = 1− β = 1 − [Pr{D < UCL | p} − Pr{D ≤ LCL | p}] ⎛ UCL + 1/ 2 − np ⎞ ⎛ LCL − 1/ 2 − np ⎞ = 1− Φ ⎜ + Φ⎜ ⎟ ⎟ ⎜ ⎜ np (1 − p ) ⎟⎠ np (1 − p ) ⎟⎠ ⎝ ⎝ ⎛ 25.5 − 30 ⎞ = 1− Φ ⎜ +Φ⎜ ⎟ ⎜ ⎜ 30(1 − 0. and n = 100 > 10.30.03 − 3 0.7255 where POI(⋅) is the cumulative Poisson distribution.0362 ⇒ 0 (b) pnew = 0. Pr{detect by at least 4th} = 1 – Pr{detect after 4th} = 1 – (1 – 0.10 and n is large.7255)4 = 0.000) = 0.813 6-16. Since p = 0.7544) = 1 − (0. (a) UCL p = p + 3 p (1 − p ) n = 0.08)} = 1 − POI(13.03) 200 = 0.42 + 0.08) < 0.03(1 − 0.Chapter 6 Exercise Solutions 6-15 continued (b) pnew = 0.8903) + Φ (−5. use the Poisson approximation to the binomial.

(a) m p = ∑ Di i =1 UCL LCL np np ( mn ) = 1200 [30(400)] = 0. np = 400(0.10) = 22 (b) npnew = 400 (0.Chapter 6 Exercise Solutions 6-17.15) = 60 > 15.10) = 40 = np + 3 np (1 − p ) = 40 + 3 40(1 − 0. so use the normal approximation to the binomial.15) ⎟⎟ ⎝ ⎠ ⎝ ⎠ = 1 − Φ (−0.39) = 1 − 0.210) + Φ (−5.000 = 0.15) ⎟ ⎜ 60(1 − 0.5 − 60 ⎞ = 1− Φ ⎜ + Φ⎜ ⎟ ⎜ 60(1 − 0.10) = 58 = np − 3 np (1 − p ) = 40 − 3 40(1 − 0.5 − 60 ⎞ ⎛ 22 − 0.583 6-19 . Pr{detect on 1st sample | p} = 1 − Pr{not detect on 1st sample | p} = 1− β = 1 − [Pr{D < UCL | np} − Pr{D ≤ LCL | np}] ⎛ UCL + 1/ 2 − np ⎞ ⎛ LCL − 1/ 2 − np ⎞ = 1− Φ ⎜ + Φ⎜ ⎟ ⎟ ⎜ ⎜ np (1 − p ) ⎟⎠ np (1 − p ) ⎟⎠ ⎝ ⎝ ⎛ 58 + 0.417 + 0.10.

10) = 0.1 ⎠ ⎝ UCL − p ⎠ 2 (b) Using the Poisson approximation to the binomial. 20) − POI(1.616 ≅ 2 6-20 . 6-19.20) = 20.381 ARL1 = 1/(1 –β) = 1/(1 – 0.381) = 1.19) | 20} − Pr{D ≤ 100(0.10) = 10. (c) pnew = 0.19) |10} = POI(0.583 ARL1 = 1/(1 –β) = 1/(0.004 where POI(⋅) is the cumulative Poisson distribution.000 = 0.715 ≅ 2 6-20.996 = 0. Pr{type II error} = β = Pr{D < nUCL | λ} − Pr{D ≤ nLCL | λ} = Pr{D < 100(0.19 − 0.381 where POI(⋅) is the cumulative Poisson distribution. 20) = 0.583) = 1. λ = npnew = 100(0.01) |10} + 1 − Pr{D ≤ 100(0.20. 1 – β = 0.381 − 0. λ = np = 100(0. from 6-17(b). This is a continuation of Exercise 6-17. not 6-18. β = 0.10) + 1 − POI(19. Pr{type I error} = Pr{ pˆ < LCL | p} + Pr{ pˆ > UCL | p} = Pr{D < nLCL | λ} + 1 − Pr{D ≤ nUCL | λ} = Pr{D < 100(0.01) | 20} = POI(18. NOTE: There is an error in the textbook. from 6-18(c).1) ⎜ ⎟ = 100 ⎝ 0.1(1 − 0.Chapter 6 Exercise Solutions 6-18. Using the Poisson approximation to the binomial. (a) UCL = p + 3 p (1 − p ) n 2 ⎛ ⎞ 3 3 ⎛ ⎞ n = p (1 − p ) ⎜ ⎟ = 0.000 + 1 − 0.

03 _ P=0. (a) For a p chart with variable sample size: p = ∑ i Di ∑ i ni = 83 / 3750 = 0.07 0. The first approach would be to plot pˆ i and use constant limits unless there is a different size sample or a plot point near a control limit.00 2 4 6 8 10 12 Sample 14 16 18 20 Tests performed with unequal sample sizes Process is in statistical control.02 0. UCLi] [0. 6-21 .0221) ni . In those cases.0500] ni 100 150 200 250 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Second Visit Required (Ex6-21Sec) 0.0221 ± 3 0. would be to construct standardized control limits with control limits at ± 3.0221) 0.05 UCL=0.0533] [0.02213 0.06 Proportion 0. 0.0662] [0. (b) There are two approaches for controlling future production.Chapter 6 Exercise Solutions 6-21.0221 and control limits are at p ± 3 p (1 − p ) / ni [LCLi. 0.05005 0.04 0.01 LCL=0 0. The second approach.0221(1 − 0. 0.0581] [0. calculate the exact control limits by p ± 3 p (1 − p ) / ni = 0. 0. preferred in many cases. and to plot Z i = ( pˆ i − 0.0216 / ni .

Use a sample size of n = 187.04 0. MTB > Stat > Control Charts > Attributes Charts > P P Chart of Second Visit Required (Ex6-21Sec) Limits based on average sample size (n=187) 0. however MINITAB accepts only integer values for n.05451 0. MTB > Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Ex6-21Req Variable Ex6-21Req N 20 Mean 187.06 UCL=0. 6-22 .00 LCL=0 2 4 6 8 10 12 Sample 14 16 18 20 Process is in statistical control.02219 0.03 _ P=0.05 Proportion 0. and carefully examine points near the control limits.5 Average sample size is 187.01 0.Chapter 6 Exercise Solutions 6-22.02 0.5.

Chapter 6 Exercise Solutions 6-23.0216 / ni MTB > Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Standardized Second Visit Data (Ex6-23zi) UCL=3 3 Individual Value 2 1 _ X=0 0 -1 -2 -3 LCL=-3 2 4 6 8 10 12 Observation 14 16 18 20 Process is in statistical control. zi = ( pˆ i − p ) p (1 − p ) ni = ( pˆ i − 0.0221) 0. 6-23 .

UCL200 = 0.03 0.00 2 4 6 8 10 12 Week 14 16 18 20 6-24 .07 Variable Ex6-24pi Ex6-24n100 Ex6-24n150 Ex6-24n200 Ex6-24n250 Ex6-24C L Ex6-24LCL 0.0662.0221.04 0. LCL = 0 UCL100 = 0. UCL150 = 0.Chapter 6 Exercise Solutions 6-24. UCL250 = 0.0500 MTB > Graph > Time Series Plot > Multiple Control Chart of Second Visit Data with Limits for Various Sample Sizes (Ex6-24pi) Proportion of Second Visits Required 0.0533.06 0.01 0.02 0. CL = 0.0581.05 0.

0895 = 0.0000 + 0.01.05 + L(0.64 |12} = 1 − POI(32. The np chart is inappropriate for varying sample sizes because the centerline (process center) would change with each ni.8349 np (1 − p ) = 20 − 2. n = 100 ⎛ 1− p ⎞ 2 n>⎜ ⎟L ⎝ p ⎠ ⎛ 1 − 0.01 ⎞ 2 >⎜ ⎟3 ⎝ 0. LCL = 0.12) = 1 − 1. UCL = 0.8349 20(1 − 0.05) = 20 UCL = np + 2.8349 20(1 − 0.05(1 − 0.36 LCL = np − 2.0809 = 0.05) = 7.0500.0809. LCL = 0. p = CL = 0.0399.64 (c) n = 400 is large and p = 0.03 on 1st sample} = 1 − Pr{not detect} = 1− β = 1 − [Pr{D < UCL | λ} − Pr{D ≤ LCL | λ}] = 1 − Pr{D < 32. use Poisson approximation to binomial.01 ⎠ > 891 ≥ 892 6-26.36 |12} + Pr{D ≤ 7. n = 400. 6-27.1.8349 (b) CL = np = 400(0. UCL = 0.05 < 0.05 + L 0.0109) L = 2. 6-25 .12) + POI(7.0895 where POI(·) is the cumulative Poisson distribution.Chapter 6 Exercise Solutions 6-25.05) 400 = 0. Pr{detect shift to 0.8349 np(1 − p) = 20 + 2. p = CL = 0.05) = 32.0191 (a) 0.

15(1 − 0. L = 2 (a) ⎛ 1− p ⎞ 2 n>⎜ ⎟L ⎝ p ⎠ ⎛ 1 − 0.15 0.03 ⎠ 6-26 . use normal approximation to binomial.05(1 − 0.01) = 44 ⎝δ ⎠ ⎝ 0. p = 0.00135 + 0.01)(1 − 0.15(1 − 0.Chapter 6 Exercise Solutions 6-28.15) 400 ⎟ ⎜ 0.04 – 0.19) = 1 − 0.01 ⎞ 2 >⎜ ⎟2 ⎝ 0.15 = 1− Φ ⎜ +Φ⎜ ⎟ ⎟ ⎜ 0.01.00) + Φ (−8.01 ⎠ > 396 ≥ 397 (b) δ = 0. Pr{detect on 1st sample after shift} = 1 − Pr{not detect} = 1− β = 1 − [Pr{ pˆ < UCL | p} − Pr{ pˆ ≤ LCL | p}] ⎛ UCL − p ⎞ ⎛ LCL − p ⎞ = 1− Φ ⎜ +Φ⎜ ⎟ ⎜ p (1 − p ) n ⎟ ⎜ p (1 − p ) n ⎟⎟ ⎝ ⎠ ⎝ ⎠ ⎞ ⎛ ⎞ ⎛ 0.01 = 0.05) 400 L = 4. λ = np = 400(0.03 2 2 ⎛L⎞ ⎛ 2 ⎞ n = ⎜ ⎟ p(1 − p) = ⎜ ⎟ (0.24 (b) p = 15.15) 400 ⎟ ⎝ ⎠ ⎝ ⎠ = 1 − Φ (−3. (a) UCL = p + L p (1 − p ) n 0.0962 = 0.000 = 0.99865 6-29.0962 − 0.0038 − 0.0500 + L 0.15) = 60 > 15.

6) = 0.948 = 0.005) |100(0. (a) Pr{type I error} = Pr{ pˆ < LCL | p} + Pr{ pˆ > UCL | p} = Pr{D < nLCL | np} + 1 − Pr{D ≤ nUCL | np} = Pr{D < 100(0.075) |100(0. (b) Pr{type II error} =β = Pr{D < nUCL | np} − Pr{D ≤ nLCL | np} = Pr{D < 100(0.002 = 0.04)} = POI(0.018 + 1 − 0. 6-27 .0050) |100(0.04)} + 1 − Pr{D ≤ 100(0. 6) − POI(0.070 where POI(⋅) is the cumulative Poisson distribution.06)} − Pr{D ≤ 100(0.06) = POI(7.744 − 0.Chapter 6 Exercise Solutions 6-30. 4) + 1 − POI(7.742 where POI(⋅) is the cumulative Poisson distribution.075) |100(0. 4) = 0.

0742: ARL1 = 1/(1 –β) = 1/(1 – 0.0000 Pr{D<=0.0008 0.9989 0.0000 1.070 = 14.0000 0.25 0.9000 0.15 0.1 0.0000 0.742) = 3.1000 0.0000 0.xls : worksheet Ex6-30 p 0 0.5978 0.0001 0.8666 0.05 0. LCL=0.0000 beta 0.0183 0.2 0.9881 0.6065 0.4526 0.07 0.0008 0.6321 0.2202 0.0498 0.2202 0. CL=4.5 1 2 3 4 5 6 7 8 9 10 12.5 |100 p} Excel : workbook Chap06.0000 0.3000 0.8000 0.7440 0.5|np} 1.0025 0. α = 0. β = 0.25 np 0 0.0750) |100 p} − Pr{D ≤ 100(0.0000 1.1 0.0000 OC Curve for n=100.4530 0.1353 0.3239 0.9306 0.0180 0.08 0.070: ARL0 = 1/α = 1/0.3935 0.0000 0.15 0.0000 0 0.2 0.6000 0.0003 0.8599 0.7000 Beta 0.0698 0.5|np} 1.005 0.3 p (d) from part (a).2000 0.09 0.0180 0.4000 0.Chapter 6 Exercise Solutions 6-30 continued (c) β = Pr{D < nUCL | np} − Pr{D ≤ nLCL | np} = Pr{D < 100(0.125 0.9383 0.04 0.0050) |100 p} = Pr{D < 7.06 0.02 0.03 0.0000 0.5000 0.0000 0.5 1.3679 0.0067 0.0009 0.3238 0.5987 0.29 ≅ 15 from part (b). UCL=7.8636 0.5 |100 p} − Pr{D ≤ 0.9489 0.861 ≅ 4 6-28 .0698 0.7415 0.05 0.0000 0.5 15 20 25 Pr{D<7.01 0.5.

062 LCL = p − 3 p (1 − p ) n = 0.05 0.02 0.04 0.06 0.062 Proportion 0.02 − 3 0.02 (a) UCL = p + 3 p (1 − p ) n = 0.07 UCL=0.00 LCL=0 1 2 3 4 5 6 Sample 7 8 9 10 Test Results for P Chart of Ex6-31Num TEST 1.03 _ P=0. p = 0.Chapter 6 Exercise Solutions 6-31.038 and σˆ p = 0.08 0.00 standard deviations from center line.02(1 − 0. p = 0.02) 100 = 0. n = 100.0191 6-32.02 + 3 0.02(1 − 0. Test Failed at points: 4 Sample 4 exceeds the upper control limit.02 0. One point more than 3.01 0.02) 100 ⇒ 0 (b) MTB > Stat > Control Charts > Attributes Charts > P P Chart of Number Nonconforming (Ex6-31Num) 0.09 1 0. LCL = np − k np (1 − p ) > 0 np > k np (1 − p ) ⎛ 1− p ⎞ n > k2 ⎜ ⎟ ⎝ p ⎠ 6-29 .

p = 0. m = 20.204 7 Sample Count 6 5 4 3 __ NP=2.0167 CL = np = 150(0.0167) = 2. n = 150.505(1 − 0.213 LCL = np − 3 np (1 − p ) = 2.505 + 3 2. 6-30 .505 UCL = np + 3 np (1 − p ) = 2. ∑ D = 50. results are the same as for the p chart.708 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > NP NP Chart of Numer of Nonconforming Units (Ex6-4Num) 8 UCL=7.Chapter 6 Exercise Solutions 6-33.505 − 4.0167) = 7.5 2 1 LCL=0 0 2 4 6 8 10 12 Sample 14 16 18 20 The process is in control.

17. 5. 19. 15. CL = np = 2500(0.1228) = 257.8 200 1 1 1 1 1 1 100 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for NP Chart of Ex6-5Num TEST 1.1228) = 307 UCL = np + 3 np (1 − p ) = 307 + 3 307(1 − 0. indicating that this data should not be used to establish control limits for future production. 11. 20 Like the p control chart.77 MTB > Stat > Control Charts > Attributes Charts > NP NP Chart of Number of Nonconforming Belts (Ex6-5Num) 500 1 1 1 400 Sample Count 1 1 UCL=356. many subgroups are out of control (11 of 20).Chapter 6 Exercise Solutions 6-34. 2.3 __ NP=307. 16.23 LCL = np − 3 np (1 − p ) = 307 − 3 307(1 − 0. 6-31 . Test Failed at points: 1.00 standard deviations from center line. 12.1 300 LCL=257. 3.1228) = 356. One point more than 3.

6-32 .06(1 − 0.Chapter 6 Exercise Solutions 6-35.06 zi = ( pˆ i − 0.0564 / ni MTB > Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Standardized Fraction Nonconforming (Ex6-35zi) 3 UCL=2. p = 0. results are the same as for the p chart.414 -3 1 2 3 4 5 6 Observation 7 8 9 10 The process is in control.494 Individual Value 2 1 _ X=0.06) 0.040 0 -1 -2 LCL=-2.06) 0.06) / ni = ( pˆ i − 0.

36 + 3 2.36 − 3 2.36 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > C C Chart of Number of Nonconformities on Plate (Ex6-36Num) 9 1 8 Sample Count 7 UCL=6.36 = 6.969 6 5 4 3 _ C=2. 6-33 . CL = c = 2. Test Failed at points: 13 No.36 2 1 0 LCL=0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 Test Results for C Chart of Ex6-36Num TEST 1.97 LCL = c − 3 c = 2. The plate process does not seem to be in statistical control.36 UCL = c + 3 c = 2. One point more than 3.Chapter 6 Exercise Solutions 6-36.00 standard deviations from center line.

1.2926] [0.4 UCL=1.249 Sample Count Per Unit 1.701 0.1881.2 1.153 0.1088.2 LCL=0. UCLi] [0. 1.7007 − 3 0.6 0. CL = u = 0.7007 UCLi = u + 3 u ni = 0.1527.4 0.7007 / ni ni 18 20 21 22 24 [LCLi.Chapter 6 Exercise Solutions 6-37.2622] [0.7007 / ni LCLi = u − 3 u ni = 0.2361] [0.1653.7007 + 3 0. 1.2487] [0.0 0. 1.1392.8 _ U=0.0 2 4 6 8 10 12 Sample 14 16 18 20 Tests performed with unequal sample sizes 6-34 .2133] MTB > Stat > Control Charts > Attributes Charts > U U Chart of Imperfections in Paper Rolls (Ex6-37Imp) 1. 1.

Chapter 6 Exercise Solutions 6-38.7007 + 3 0.6 0.7007. n = 20.72 0.151 0.4 0.55.1467 MTB > Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Ex6-37Rol Variable Ex6-37Rol N 20 Mean 20.55 = 1. however MINITAB accepts only integer values for n.7007 / 20.4 UCL=1.2 LCL=0.7007 / 20.55 UCL = u + 3 u n = 0. CL = u = 0. MTB > Stat > Control Charts > Attributes Charts > U U Chart of Imperfections in Paper Rolls (Ex6-37Imp) with average sample size n=20 1.8 _ U=0.2547 LCL = u − 3 u n = 0. and carefully examine points near the control limits.55 = 0.0 0.2 1.550 Average sample size is 20. Use a sample size of n = 20.289 Sample Count Per Unit 1.7007 − 3 0.0 2 4 6 8 10 12 Sample 14 16 18 20 6-35 .

zi = (ui − u ) u ni = (ui − 0.898 Individual Value 1 _ X=-0.004 0 -1 LCL=-1.Chapter 6 Exercise Solutions 6-39.906 -2 2 4 6 8 10 12 Observation 14 16 18 20 6-36 .7007) 0.7007 / ni MTB > Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Standardized Paper Roll Imperfections (Ex6-39zi) 2 UCL=1.

Chapter 6 Exercise Solutions 6-40.174 5 Sample Count 4 3 2 _ C=1.17 LCL ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > C C Chart of Cassette Deck Nonconformities (Ex6-40Num) UCL=5. Use these limits to control future production.5 + 3 1. 6-37 .5 1 0 LCL=0 2 4 6 8 10 Sample 12 14 16 18 Process is in statistical control. c chart based on # of nonconformities per cassette deck CL = c = 1.5 UCL = c + 3 c = 1.5 = 5.

59.00 standard deviations from center line. 22 6-38 . 11. One point more than 3. UCL = c + 3 c = 8.59 − 3 8. CL = c = 8. Test Failed at points: 10.59 = 17. LCL = c − 3 c = 8.38 15 10 _ C=8.59 + 3 8.59 5 0 LCL=0 2 4 6 8 10 12 Sample 14 16 18 20 22 Test Results for C Chart of Ex6-41Num TEST 1.Chapter 6 Exercise Solutions 6-41.59 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > C C Chart of Number of Nonconformities (Ex6-41Num) per 1000 meters telephone cable 25 1 20 1 1 Sample Count UCL=17.384.

17 + 3 6. 15. 15.Chapter 6 Exercise Solutions 6-41 continued Process is not in statistical control. UCL = c + 3 c = 6.17 5 LCL=0 0 2 4 6 8 10 12 Sample 14 16 18 20 22 Test Results for C Chart of Ex6-41Num TEST 1. 22 excluded from calculations 25 1 Sample Count 20 1 1 1 15 UCL=13. then re-calculate the control limits. LCL ⇒ 0 C Chart of Number of Nonconformities (Ex6-41Num) Samples 10. Exclude subgroups 10. three subgroups exceed the UCL. Test Failed at points: 10. 11 and 22.17 = 13. CL = c = 6. Subgroup 15 will then be out of control and should also be excluded. 11.62.62 10 _ C=6.17. 11. 22 6-39 .00 standard deviations from center line. One point more than 3.

CL = nc = 2.21 LCL = nc − 3 nc = 15.43 = 27.20 LCL = u − 3 u n = 15. total nonconformities 27 = = 6.42 − 3 15.5(6.43 + 3 15. (a) The new inspection unit is n = 2500/1000 = 2.42 CL = u = total inspection units (18 × 1000) / 2500 UCL = u + 3 u n = 15.43 = 3. A u chart of average nonconformities per inspection unit is appropriate. A c chart of the total number of nonconformities per inspection unit is appropriate.42 /1 = 3.35 LCL = nc − 3 nc = 6 − 3 6 ⇒ 0 (b) The sample is n =1 new inspection units.43 UCL = nc + 3 nc = 15. 6-40 . is the average number of nonconformities found in 2500m. A c chart of the total number of nonconformities per inspection unit is appropriate.65 The plot point. and since n = 1.35 LCL = u − 3 u n = 6 − 3 6 1 ⇒ 0 6-43.17) = 15.64 The plot point. c . CL = nc = 4(1.43 − 3 15. A u chart of average nonconformities per inspection unit is appropriate. (b) The sample is n =1 new inspection units.5 of the old unit.5) = 6 UCL = nc + 3 nc = 6 + 3 6 = 13. (a) The new inspection unit is n = 4 cassette decks.42 /1 = 27.00 CL = u = total inspection units (18 / 4) UCL = u + 3 u n = 6 + 3 6 1 = 13. total nonconformities 111 = = 15. this is the same as the total number of nonconformities. is the total number of nonconformities found while inspecting a sample 2500m in length.Chapter 6 Exercise Solutions 6-42. u .42 + 3 15.

since this chart was established for average nonconformities per unit. (If this was a c chart for total nonconformities in the sample.396 LCL = u − 3 u n = 0.422 4 = −0. the control limits would need revision.2 1. the same control limits may be used for future production with the new sample size.Chapter 6 Exercise Solutions 6-44.396 Sample Count Per Unit 1.211 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > U U Chart of Manual Transmission Subassemblies (Ex6-44Num) 1.0 LCL=0 2 4 6 8 10 Sample 12 14 16 (b) The process is in statistical control.422 − 3 0.0 0.) 6-41 .6 _ U=0.4 0.422 + 3 0.422 4 = 1.422 UCL = u + 3 u n = 0. (a) A u chart of average number of nonconformities per unit is appropriate. CL = u = ∑ ui m = ( ∑ xi / n ) m = (27 / 4) 16 = 6.8 0.422 0. However. (c) The new sample is n = 8/4 = 2 inspection units.2 0.4 UCL=1.75 16 = 0. with n = 4 transmissions in each inspection.

Chapter 6 Exercise Solutions 6-45.0774.9108. n = 4 CL = u = c / n = 16 / 4 = 4 UCL = u + 3 u n = 4 + 3 4 / 4 = 7 LCL = u − 3 u n = 4 − 3 4 / 4 = 1 6-42 . n = 4 CL = u = c / n = 4 / 4 = 1 UCL = u + 3 u n = 1 + 3 1/ 4 = 2. (a) CL = c = 9 UCL = c + 3 c = 9 + 3 9 = 18 LCL = c − 3 c = 9 − 3 9 = 0 (b) c = 16. LCL = 10 6-47. UCL = 21 Pr{x ≤ 10 | c = 16} = 0. c = 16 Pr{x ≤ 21| c = 16} = 0.5 LCL = u − 3 u n = 1 − 3 1/ 4 ⇒ 0 6-46. Use the cumulative Poisson tables. (a) CL = c = 4 UCL = c + 3 c = 4 + 3 4 = 10 LCL = c − 3 c = 4 − 3 4 ⇒ 0 (b) c = 4.

Find limits such that Pr{D ≤ UCL} = 0.980 and Pr{D < LCL} = 0.6} 0.055 0. As a comparison.4 − 1. the normal distribution gives: UCL = u + z0.972 0.00 LCL = c − z0.095 6-49. As a comparison.96 7.975 c = 7.905 LCL = u + z0. the normal distribution gives: UCL = u + z0.645 1.00.020 u n = 6 − 2.Chapter 6 Exercise Solutions 6-48.96 7.019 0.030 0.0 and n = 3.025 c = 7.70.6 = 2. UCL = 13 and LCL = 2.6 − 1.784 6-43 .645 1.054 6 3 = 3. c = u × n = 18. From the cumulative Poisson tables: x 9 10 26 27 Pr{D ≤ x | c = 18} 0.4 10 = 2.4 + 1.952 UCL = x/n = 20/10 = 2.976 for the c chart.032 0.015 0. As a comparison. Using the cumulative Poisson distribution with c = u n = 1.4 10 = 0.6 = 13.923 0. and LCL = x/n = 7/10 = 0.95 u n = 1.980 u n = 6 + 2.6 + 1.020. and LCL = x/n = 9/3 = 3.954 0.016 LCL = u + z0.20 6-50.054 6 3 = 8.062 0. Using the cumulative Poisson distribution: x 2 3 12 13 Pr{D ≤ x | c = 7. the normal distribution gives UCL = c + z0.05 u n = 1. u chart with u = 6.983 UCL = x/n = 27/3 = 9.4(10) = 14: x 7 8 19 20 Pr{D ≤ x | c = 14} 0.

0} = 0.0. So set UCL = 6. u chart with control limits based on each sample size: u = 7.Chapter 6 Exercise Solutions 6-51.005)(0.995.94 Sample Count Per Unit 14 12 10 8 _ U=7 6 4 2 0 LCL=0 1 2 3 4 5 6 Ex6-51Day 7 8 9 10 Tests performed with unequal sample sizes The process is in statistical control. Pr{x ≤ 6 | c = 2.00003 6-44 . UCLi = 7 + 3 7 / ni . 6-52. (a) From the cumulative Poisson table.005) = 0. LCLi = 7 − 3 7 / ni MTB > Stat > Control Charts > Attributes Charts > U U Chart of Total Number of Imperfections (Ex6-51Imp) 16 UCL=14. (b) Pr{two consecutive out-of-control points} = (0.

000 + [1 − 0.25 = 12.10.5 − 3 10.10.02 6-54. CL = c = ∑ D m = 924 /176 = 5.5} 3 0.Chapter 6 Exercise Solutions 6-53.88 8.5 + 1.5) + [1 − POI(20.22 LCL = nc − 3 nc = 10.25 UCL = c + 3 c = 5.22 |10. (b) Let new inspection unit n = 2 water heaters CL = nc = 2(5. approximately 8/day.5 − 1.997 ] = 0.78 |10. c = 850 /100 = 8.5}] = POI(0.5 = 20. the normal distribution gives UCL = c + z0.5 = 13.25 + 3 5.97 c = 8.5 UCL = nc + 3 nc = 10.25) = 10.12 LCL ⇒ 0 Plot the results after inspection of each water heater.003 6-45 . From the cumulative Poisson distribution: x Pr{D ≤ x | c = 8.030 13 0.5 = 0. For comparison.5 = 3.98 LCL = c + z0.03 c = 8. A c chart with one inspection unit equal to 50 manufacturing units is appropriate.949 14 0.78 (c) Pr{type I error} = Pr{D < LCL | c} + Pr{D > UCL | c} = Pr{D < 0.5 + 3 10.5) ] = 0.973 LCL = 3 and UCL = 13.5} + [1 − Pr{D ≤ 20.5 . (a) Plot the number of nonconformities per water heater on a c chart.88 8.

0 average number of nonconformities/unit. 2. 6-56.533 ± 3 0.017 where POI(⋅) is the cumulative Poisson distribution.63 0. u = 4.983 = 0. c = 16 / 30 = 0. ∑ Di = 16 in 30 refrigerators.00 An UCL = 9 will give a probability of 0.89 0.533}] = 0 + [1 − POI(2.09 0.24 0.723] (b) α = Pr{D < LCL | c} + Pr{D > UCL | c} = Pr{D < 0 | 0.99 1.02 0.533} + [1 − Pr{D ≤ 2.98 0.72 | 0.Chapter 6 Exercise Solutions 6-55.99 of concluding the process is in control. Use the cumulative Poisson distribution to determine the UCL: MTB : worksheet Chap06. 0.99.43 0. 6-46 . Desire α = 0. Use a c chart for nonconformities with an inspection unit n = 1 refrigerator.79 0.533) ] = 1 − 0.mtw Ex6-55X 0 1 2 3 4 5 6 7 8 9 10 11 Ex6-55alpha 0.533 = [0.00 1.95 0.533 (a) 3-sigma limits are c ± 3 c = 0. when in fact it is.

406 − 0. 2) − POI(0.533}] = 0 + [1 − POI(1. 0.1353 = 0.541 6-57.533 (a) c ± 2 c = 0.533)] = 1 − 0.Chapter 6 Exercise Solutions 6-56 continued (c) β = Pr{not detecting shift} = Pr{D < UCL | c} − Pr{D ≤ LCL | c} = Pr{D < 2. 2) − POI(0.0} − Pr{D ≤ 0 | 2.135 = 0.271 where POI(⋅) is the cumulative Poisson distribution.1004 where POI(⋅) is the cumulative Poisson distribution.533 = [0.72 | 2. 2) = 0.8996 = 0. (d) ARL1 = 1 1 = = 2. (c) β = Pr{D < UCL | c} − Pr{D ≤ LCL | c} = Pr{D < 1.0} = POI(2.271 6-47 . (d) ARL1 = 1 1 = = 1.533} + [1 − Pr{D ≤ 1.533 + 2 0.993 | 0.5414 where POI(⋅) is the cumulative Poisson distribution.1.993 | 2} − Pr{D ≤ 0 | 2} = POI(1. 2) = 0.372 ≈ 2 1 − β 1 − 0.993] (b) α = Pr{D < LCL | c } + Pr{D > UCL | c } = Pr{D < 0 | 0.6767 − 0.18 ≈ 2 1 − β 1 − 0. c = 0.

708 LCL ⇒ 0 6-59. u = average # nonconformities/calculator = 2 (a) c chart with c = u × n = 2(2) = 4 nonconformities/inspection unit CL = c = 4 UCL = c + k c = 4 + 3 4 = 10 LCL = c − k c = 4 − 3 4 ⇒ 0 (b) Type I error = α = Pr{D < LCL | c } + Pr{D > UCL | c } = Pr{D < 0 | 4} + [1 − Pr{D ≤ 10 | 4}] = 0 + [1 − POI(10. 6-60.5 average nonconformities/radio CL = c = u × n = 0. u = 0.997 = 0. L: sigma control limits nc − L nc > 0 nc > L nc n > L2 c 6-48 .5 + 3 4.86 LCL ⇒ 0 6-61. 1 inspection unit = 6 clocks.5 = 10. u = 0. c: nonconformities per unit.003 where POI(⋅) is the cumulative Poisson distribution.5 UCL = c + 3 c = 4.Chapter 6 Exercise Solutions 6-58.5(10) = 5 UCL = c + 3 c = 5 + 3 5 = 11.75(6) = 4. 4)] = 1 − 0.75 nonconformities/clock CL = c = u × n = 0. 1 inspection unit = 10 radios.

(b) Probability Plot of Transformed "Days-between-Homicides" (Ex6-62t27) Normal .95% CI 99 Mean StDev N AD P-Value 95 90 12.Chapter 6 Exercise Solutions 6-62.04 28 1.2777th root transformation makes the data more closely resemble a sample from a normal distribution.760 Percent 80 70 60 50 40 30 20 10 5 1 0 1 2 Ex6-62t27 3 4 The 0.572 <0.5635 28 0. (a) MTB > Graphs > Probability Plot > Single Probability Plot of Days-Between-Homicides (Ex6-62Bet) Normal . indicating that the normal distribution assumption is not reasonable.25 12.95% CI 99 Mean StDev N AD P-Value 95 90 1.238 0.005 Percent 80 70 60 50 40 30 20 10 5 1 -30 -20 -10 0 10 20 Ex6-62Bet 30 40 50 There is a huge curve in the plot points. 6-49 .806 0.

5 The 0.5 2.95% CI 99 Mean StDev N AD P-Value 95 90 1.4789 28 0.0 0.5 2. It is not very different from the transformed data in (b).0 1.0 3 6 9 12 15 18 Observation 21 24 27 6-50 .0 Individual Value 2.807 Percent 80 70 60 50 40 30 20 10 5 1 0.695 0.5 3.5 1.246 0.5 1.806 1.0 3.Chapter 6 Exercise Solutions 6-62 continued (c) Probability Plot of Transformed "Days-betwee-Homicides" (Ex6-62t25) Normal .25th root transformation makes the data more closely resemble a sample from a normal distribution.366 3.5 UCL=3.223 0.2777 root) (Ex6-62t27) 3.5 LCL=0.0 _ X=1.0 0. (d) MTB > Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Transformed Homicide Data (0.0 Ex6-62t25 2.

There are endless possibilities for collection of attributes data from nonmanufacturing processes.0 _ X=1.695 1. etc. etc. invoiced incorrectly. There is no difference in interpretation.25 root) (Ex6-62t25) UCL=3. etc. packaged incorrectly (wrong material.0 3 6 9 12 15 18 Observation 21 24 27 Both Individuals charts are similar. wrong package labeling).0 Individual Value 2.Chapter 6 Exercise Solutions 6-62 continued (e) I Chart of Transformed Homicide Data (0. with an identical pattern of points relative to the UCL.5 LCL=0. (hopefully caught internally with a verification step).025 3.). mean and LCL. One could track the number of orders filled incorrectly (wrong parts. If a change is made.5 2. wrong part labeling. 6-63.. errors in tax preparation. meaning that the days-between-homicides is approximately constant.5 1. law. the mean time between may get longer (or shorter) with plot points above the upper (or below the lower) control limit. too few/many parts. Consider a product distribution center (or any warehouse) with processes for filling and shipping orders. (f) The “process” is stable. say in population.365 0.0 0. Or consider an accounting firm—errors in statements. workforce. policy. 6-51 . which affects the rate at which homicides occur.

if you’re looking for data on time-between-events that must be obtained from external sources (for example.94 30 25 _ U=25. However. One point more than 3. There are ways to address these difficulties. it can usually be obtained reliably and consistently.77 94 94 94 94 94 94 94 94 94 94 94 94 95 95 95 M EB A R PR A Y UN JUL UG EP C T OV EC A N EB A R A J F J S F D A J A M N O M M Ex6-65MON Tests performed with unequal sample sizes Test Results for U Chart of Ex6-65NSCANB TEST 1. 6-52 . MTB > Stat > Control Charts > Attributes Charts > U U Chart of CAT Scans (Ex6-65NSCANB) Sample Count Per Unit 40 1 35 UCL=35. time-to-field failures). Also. If time-between-events data (say failure time) is being sought for internally generated data. The variable NYRSB can be thought of as an “inspection unit”. 6-65☺. it may be hard to determine with sufficient accuracy—both the “start” and the “end”. Collection of “start” time data may be facilitated by serializing or date coding product. A u chart which monitors the average number of CAT scans per NYRSB is appropriate.Chapter 6 Exercise Solutions 6-64. the conditions of use and the definition of “failure” may not be consistently applied. The “process characteristic” to be controlled is the rate of CAT scans. Test Failed at points: 15 The rate of monthly CAT scans is out of control. representing an identical “area of opportunity” for each “sample”.00 standard deviations from center line.86 20 15 LCL=15.

A u chart which monitors the average number of office visits per NYRSB is appropriate. The variable NYRSE can be thought of as an “inspection unit”. The “process characteristic” to be controlled is the rate of office visits. representing an identical “area of opportunity” for each “sample”. (a) MTB > Stat > Control Charts > Attributes Charts > U U Chart of Number of Office Visits (Ex6-66aNVIS) Phase 1 Sample Count Per Unit 2500 UCL=2476.0 2300 2200 LCL=2129.5 2400 _ U=2303.Chapter 6 Exercise Solutions 6-66☺.5 2100 JAN94 FEB94 MAR94 APR94 MAY94 Ex6-66aMON JUN94 JUL94 AUG94 Tests performed with unequal sample sizes The chart is in statistical control 6-53 .

0 2200 2100 LCL=2141. 15 The phase 2 data appears to have shifted up from phase 1. 13.0 2400 2300 _ U=2303. One point more than 3.Chapter 6 Exercise Solutions 6-66 continued (b) U Chart of Number of Office Visits (Ex6-66NVIS) Phase 1 Limits 1 2800 Sample Count Per Unit 1 2700 2600 2500 1 1 1 1 1 UCL=2465.0 4 4 4 4 4 4 4 4 5 5 4 5 4 4 4 N9 EB9 A R9 PR9 A Y9 UN9 UL9 UG9 EP9 CT 9 V 9 EC9 A N9 EB9 AR9 A J F M J J F M J S A M A NO D O Ex6-66MON Tests performed with unequal sample sizes Test Results for U Chart of Ex6-66NVIS TEST 1. 6-54 . 10.00 standard deviations from center line. Test Failed at points: 9. 11. 12. 14. The 2nd phase is not in statistical control relative to the 1st phase.

6 2400 9 10 11 12 Sample 13 14 15 Tests performed with unequal sample sizes The Phase 2 data.5 2700 _ U=2623.Chapter 6 Exercise Solutions 6-66 continued (c) U Chart of Number of Office Visits (Ex6-66NVIS) Phase 2 Sample Count Per Unit 2800 UCL=2796. separated from the Phase 1 data. are in statistical control.5 2600 2500 LCL=2450. 6-55 .

010) µˆ − LSL 74.001 − 73. Cˆ pu = 1. µˆ = x = 33.000 ± 0.001. and the new process parameters are used in the process capability analysis.10 Cˆ pu = 3σˆ 3(1. A second exercise number in parentheses indicates that the exercise number has changed.13 Cˆ pu = 3σˆ 3(0.Chapter 7 Exercise Solutions Note: Several exercises in this chapter differ from those in the 4th edition.13 7-2.36 Cˆ pl = 3σˆ 3(1.010) ( ) Cˆ pk = min Cˆ pl . σˆ = R d 2 = 1. New exercises are denoted with an “☺”.17 Cˆ p 6σˆ 6(0.65.010) USL − µˆ 74.65 = = 1.93) USL − µˆ 40 − 33.035 = [73. R = 4.73 Cˆ p 6σˆ 6(1.93 USL = 40. 74. σˆ = R d 2 = 0. LSL = 20 USL − LSL 40 − 20 = = 1.65 − 20 = = 2.965 = = 1. R = 0.326 = 0. In Exercise 5-1.20 Cˆ pl = 3σˆ 3(0.93) ( ) Cˆ pk = min Cˆ pl .010 SL = 74.035 − 73.93) µˆ − LSL 33.965. n = 5.023 2. µˆ = x = 74.965 = = 1.5. samples 12 and 15 are out of control.035 − 74. Cˆ pu = 1.001 = = 1.023. An “*” indicates that the description has changed.10 7-1 .035] USL − LSL 74. 7-1.

with an estimated percent defective much less than 1 ppb.04 USL x = [(350 + 5) − 350] ×10 = 50. this PCR does not tell where the mean is located within the specification band.412 1 + ξˆ 2 7-2 .Chapter 7 Exercise Solutions 7-3. Cˆ ) = 4.48 Cˆ p = 6σˆ x 6(3.25 2.54 2 1+V 1 + (−3. Rx = 6. the mean µ lies within approximately the middle fourth of the specification band.375 − (−50) = = 6.34 Cˆ pu = 3σˆ x 3(3. However.41 σˆ 3. USL x − µˆ 50 − 10.62 Cˆ pl = 3σˆ x 3(3.54 pk = = 0.375 = = −3. Note that the Cpk is less than Cp. T − x 0 − 10. ξˆ = µˆ − T 10. indicating that the process is not centered and is not achieving potential capability. LSL x = [(350 − 5) − 350] × 10 = −50 xi = (obsi − 350) ×10 USL x − LSL x 50 − (−50) = = 5.375 − 0 = = 3.375 = = 4.04) The process produces product that uses approximately 18% of the total specification band.04) µˆ − LSL x 10.25.4128 S 3. µˆ = x = 10.43 1 + 3.04 Cˆ p 5.059 = 3.48 Cˆ pm = = = 1. σˆ x = R d 2 = 6.4128) 2 V= Since Cpm is greater than 4/3.34 pk pu pl This is an extremely capable process.04 Cˆ pkm = Cˆ 1.04) Cˆ = min(Cˆ .375.

00273) Cˆ = min(Cˆ .00635.09 pk pl pu This process is not considered capable.01) Cˆ pl = = = 1. n = 5.35 3σˆ 3(0.33 T − x 0 − 0. the mean µ lies within approximately the middle third of the specification band. USL − µˆ 0.00109 Cˆ pu = = = 1.399 σˆ 0. failing to meet the minimally acceptable definition of capable Cpk ≥ 1. R = 0.00109.00109 = = −0.22 6σˆ 6(0.01 2 2 ˆ 1 0. tolerances: 0 ± 0.01 Cˆ p = = = 1.00109 − 0 = = 0. x = 0.00273) The process produces product that uses approximately 82% of the total specification band.09 3σˆ 3(0.00273) µˆ − LSL 0. ξˆ = µˆ − T 0.01 USL − LSL 0.399) 2 V= Since Cpm is greater than 1.00109 − (−0.399 S 0.00273 Cˆ pkm = Cˆ 1.01 − 0.00273 .Chapter 7 Exercise Solutions 7-4.13 1+V 2 1 + (−0.09 pk = = 1.22 p ˆ C pm = = = 1.00273 Cˆ 1. σˆ x = 0. Cˆ ) = 1.01 + 0.399 + 1+ ξ 7-3 .

000004 85 − 95 ⎫ ⎡ ⎧ ⎧ 105 − 95 ⎫⎤ pˆ Potential = Pr ⎨ z < ⎬ + ⎢1 − Pr ⎨ z ≤ ⎬ 1.49 3σˆ x 3(1.117 ⎭⎥⎦ ⎩ ⎩ = Φ (−8. s = 1.9400 = 1. σˆ x = s c4 = 1.953) + [1 − Φ (8.117 ⎭ ⎣ 1.117 ⎭⎥⎦ ⎩ ⎩ = Φ (−13.49 pk pl pu (c) pˆ Actual = Pr{x < LSL} + Pr{x > USL} = Pr{x < LSL} + [1 − Pr{x ≤ USL}] LSL − µˆ ⎫ ⎡ USL − µˆ ⎫⎤ ⎧ ⎧ = Pr ⎨ z < ⎬ + ⎢1 − Pr ⎨ z ≤ ⎬⎥ σˆ ⎭ ⎣ σˆ ⎩ ⎩ ⎭⎦ 85 − 100 ⎫ ⎡ ⎧ ⎧ 105 − 100 ⎫⎤ = Pr ⎨ z < ⎬ + ⎢1 − Pr ⎨ z ≤ ⎬ 1.476) ] = 0.429) + [1 − Φ (4.48 3σˆ x 3(1.117 (a) USL − LSL (95 + 10) − (95 − 10) = = 2.000000 7-4 .953) ] = 0.Chapter 7 Exercise Solutions 7-5.999996] = 0.000000] = 0.05.117 ⎭ ⎣ 1.05 0. Cˆ ) = 1.000000 + [1 − 1.117) Cˆ = min(Cˆ .0000 + [1 − 0.117) USL x − µˆ (95 + 10) − 100 Actual: Cˆ pu = = = 1.98 Potential: Cˆ p = 6σˆ 6(1. µˆ = x = 100.117) (b) µˆ − LSL x 100 − (95 − 10) Cˆ pl = = = 4.

σˆ x = R d 2 = 3.70 ⎭ ⎩ = 2 × 0.76 3σˆ 3(1.70 ⎭ ⎣ 1. µˆ = x = 199.70) Cˆ = min(Cˆ .70 USL = 200 + 8 = 208. LSL = 200 – 8 = 192 (a) USL − LSL 208 − 192 = = 1.70) µˆ − LSL 199 − 192 = = 1. R = 3.70) The process produces product that uses approximately 64% of the total specification band.0000013 = 0.Chapter 7 Exercise Solutions 7-6☺. (b) USL − µˆ 208 − 199 Cˆ pu = = = 1.059 = 1.0000191 If the process mean could be centered at the specification target.70 ⎭⎥⎦ ⎩ ⎩ = Φ (−4. n = 4.0000191 + [1 − 1] = 0.5.37 Actual: Cˆ pl = 3σˆ 3(1.5 2. the fraction nonconforming would be: ⎧ 192 − 200 ⎫ pˆ Potential = 2 × Pr ⎨ z < ⎬ 1.0000026 7-5 .1176) + [1 − Φ (5.57 Potential: Cˆ p = 6σˆ 6(1.37 pk pl pu (c) The current fraction nonconforming is: pˆ Actual = Pr{x < LSL} + Pr{x > USL} = Pr{x < LSL} + [1 − Pr{x ≤ USL}] LSL − µˆ ⎫ ⎡ USL − µˆ ⎫ ⎤ ⎧ ⎧ = Pr ⎨ z < ⎬ + ⎢1 − Pr ⎨ z ≤ ⎬⎥ σˆ ⎭ ⎣ σˆ ⎩ ⎩ ⎭⎦ 208 − 199 ⎫⎤ ⎧ 192 − 199 ⎫ ⎡ ⎧ = Pr ⎨ z < ⎬ + ⎢1 − Pr ⎨ z ≤ ⎬ 1.2941) ] = 0. Cˆ ) = 1.

7 ⎫⎤ ⎧ ⎧ = Pr ⎨ z < ⎬ + ⎢1 − Pr ⎨ z ≤ ⎬ 2. n = 2. LSL = 40 – 5 = 35 (a) USL − LSL 45 − 35 = = 0.7 Cˆ pu = = = 0.025348 7-6 .39170)] = 0.135) 2 The closeness of estimates for Cp.128 = 2. Cˆ ) = 0.135 s 2.7.5 1.216 USL = 40 + 5 = 45. 2 (d) The current fraction nonconforming is: pˆ Actual = Pr{x < LSL} + Pr{x > USL} = Pr{x < LSL} + [1 − Pr{x ≤ USL}] LSL − µˆ ⎫ ⎡ USL − µˆ ⎫⎤ ⎧ ⎧ = Pr ⎨ z < ⎬ + ⎢1 − Pr ⎨ z ≤ ⎬⎥ σˆ ⎭ ⎣ σˆ ⎩ ⎩ ⎭⎦ 35 − 39.70 1+V 1 + (−0.Chapter 7 Exercise Solutions 7-7☺.7 ⎫ ⎡ 45 − 39.75 Potential: Cˆ p = 6σˆ 6(2.216 Cˆ p 0.216) Cˆ = min(Cˆ .0169634 + [1 − 0.216 ⎭ ⎣ 2. R = 2. σˆ x = R d 2 = 2. µˆ = x = 39. and Cpkm indicate that the process mean is very close to the specification target.75 = = = 0.71 Actual: Cˆ pl = 3σˆ 3(2.216) (b) USL − µˆ 45 − 39.5.216) µˆ − LSL 39.216 ⎭⎥⎦ ⎩ ⎩ = Φ (−2.7 − 40 = = −0.12094) + [1 − Φ (2. Cpm.80 3σˆ 3(2.74 2 1+V 1 + (−0. Cpk.991615] = 0.71 pk pl pu (c) V= Cˆ pm x − T 39.7 − 35 = = 0.71 = 0.135) 2 Cˆ pkm = Cˆ pk = 0.

000086 + 1 − 0.000172 7-7 .756) = 0.13 ⎭ 2.756) + 1 − Φ (3.03 Actual: Cˆ pu = 3σˆ 3(2. σˆ = Sˆ c4 = 2 0.02382 7-8 (7-6).47 pk pl pu (c) Let µˆ = 80 pˆ Potential = Pr{x < LSL} + Pr{x > USL} LSL − µˆ ⎫ USL − µˆ ⎫ ⎧ ⎧ = Pr ⎨ z < ⎬ + 1 − Pr ⎨ z ≤ ⎬ σˆ ⎭ σˆ ⎩ ⎩ ⎭ 72 − 80 ⎫ 88 − 80 ⎫ ⎧ ⎧ = Pr ⎨ z < ⎬ + 1 − Pr ⎨ z ≤ ⎬ 2.216 ⎭ ⎩ = 2 × Pr{z < −2.47 3σˆ 3(2. Cˆ ) = 0. S = 2.13 ⎭ ⎩ ⎩ = Φ (−3. µˆ = 75.13) USL − µˆ 80 + 8 − 75 = = 2.13 (a) USL − LSL 2(8) Potential: Cˆ p = = = 1.9400 = 2.13) (b) µˆ − LSL 75 − (80 − 8) Cˆ pl = = = 0.25 6σˆ 6(2.13) Cˆ = min(Cˆ .26} = 2 × 0.999914 = 0.Chapter 7 Exercise Solutions 7-7 (d) continued If the process mean could be centered at the specification target. the fraction nonconforming would be: 35 − 40 ⎫ ⎧ pˆ Potential = 2 × Pr ⎨ z < ⎬ 2.01191 = 0.

sB = 1. Cˆ ) = 1.064) Cˆ pk = min(Cˆ pl .045 = = = 1.045 2 1+V 1 + (0) 2 pˆ = Pr{x < LSL} + Pr{x > USL} = Pr{x < LSL} + 1 − Pr{x ≤ USL} LSL − µˆ ⎫ USL − µˆ ⎫ ⎧ ⎧ = Pr ⎨ z < ⎬ + 1 − Pr ⎨ z ≤ ⎬ σˆ ⎭ σˆ ⎩ ⎩ ⎭ 90 − 100 ⎫ ⎧ ⎧ 110 − 100 ⎫ = Pr ⎨ z < ⎬ + 1 − Pr ⎨ z ≤ ⎬ 3.133 6(1.045 Cˆ p = 6σˆ 6(3.13) + 1 − Φ (3.191 ⎭ 3.045 pk pl V= Cˆ pm pu x − T 100 − 100 = =0 s 3.00174 Process B µˆ = xB = 105.9400 = 3.699 3σˆ x 3(1.191) Cˆ = min(Cˆ . σˆ A = s A c4 = 3 0.13) = 0.99913 = 0.9400 = 1.566 Cˆ pu = 3σˆ x 3(1.191) µˆ − LSL x 100 − (100 − 10) Cˆ pl = = = 1.045 3σˆ x 3(3. Cˆ pu ) = 1.064 Cˆ p = USL − LSL 6σˆ = (100 + 10) − (100 − 10) = 3.064) µˆ − LSL x 105 − (100 − 10) Cˆ pl = x = = 4. σˆ B = sB c4 = 1 0.064) USL x − µˆ x (100 + 10) − 105 = = 1.191) USL x − µˆ (100 + 10) − 100 Cˆ pu = = = 1.191 Cˆ p 1.191 ⎭ ⎩ ⎩ = Φ (−3. Assume n = 5 Process A µˆ = x A = 100. s A = 3.00087 + 1 − 0.191 USL − LSL (100 + 10) − (100 − 10) = = 1.Chapter 7 Exercise Solutions 7-9 (7-7).045 3σˆ x 3(3.566 7-8 .

For the parts Cˆ pk .000000 + 1 − 0.045 < 1. σˆ B = 20σˆ 2 = 20(1. 7-10 (7-8).064 Cˆ p 3.271 Process B: µˆ B = 20(105) = 2100.064 ⎭ ⎩ ⎩ = Φ (−14.999999 = 0.758 Process B will result in fewer defective assemblies.566 = Cˆ pk .098) + 1 − Φ (4.064 ⎭ 1.064) 2 = 4.652 1+V 2 1 + (−4. A = 1. σˆ A = 20σˆ 2 = 20(3.133 Cˆ pm = = = 0.Chapter 7 Exercise Solutions 7-9 continued x − T 100 − 105 V= = = −4.000001 Prefer to use Process B with estimated process fallout of 0. B indicates that more parts from Process B are within ( ) ( ) specification than from Process A.001726.000001 instead of Process A with estimated fallout 0. Process A: µˆ A = 20(100) = 2000.699) 2 90 − 105 ⎫ ⎧ ⎧ 110 − 105 ⎫ pˆ = Pr ⎨ z < ⎬ + 1 − Pr ⎨ z ≤ ⎬ 1.699 s 1.191) 2 = 14. 7-9 .699) = 0.

LSL = 0. p84 = 1.0200 − 0.950 0.02167 15 0.985 = = 0.0225) LSL − µˆ ⎫ 0.985 − 0.Chapter 7 Exercise Solutions 7-11 (7-9). MTB > Stat > Basic Statistics > Normality Test Probability Plot of 1-kg Containers (Ex7-9Wt) Normal 99 Mean StDev N AD P-Value 95 90 0.0225) = 0.975 1.492 Percent 80 70 60 50 40 30 20 10 5 1 0.9975.025 1.985 kg µˆ − LSL 0.000 Ex7-9Wt 1.289105 0.556) = 0.9975 − 0.9975 = 0.323 0.0225 6σˆ = 6(0.050 A normal probability plot of the 1-kg container weights shows the distribution is close to normal. x ≈ p50 = 0.0200 σˆ = p84 − p50 = 1.9968 0.9975 ⎫ ⎧ ⎧ pˆ = Pr ⎨ z < ⎬ = Pr ⎨ z < ⎬ = Φ (−0.0225 σˆ ⎭ ⎩ ⎩ ⎭ 7-10 .19 C pl = 3σˆ 3(0.1350 7-12☺.

99986 10 5 1 19. mm 20.99986 = 0.00905 − 19.00 0.00919 6σˆ = 6(0.Chapter 7 Exercise Solutions 7-13☺.99 20.99986 p84 = 20.) Probability Plot of Disk Height (Ex7-13Ht) Normal 99 95 90 84 Percent 80 Mean StDev N AD P-Value 20. MTB > Stat > Basic Statistics > Normality Test (Add percentile lines at Y values 50 and 84 to estimate µ and σ. x ≈ p50 = 19.515 0.01 20.05514 7-11 .02 A normal probability plot of computer disk heights shows the distribution is close to normal.00 Disk Height.174 70 60 50 40 30 50 19.00919) = 0.00905 20 20.00905 σˆ = p84 − p50 = 20.009242 25 0.98 19.

401 0.42 7-12 . Days 25 A normal probability plot of reimbursement cycle times shows the distribution is close to normal.2 = 4.07 6σˆ = 6(4.07) = 24.2 p84 = 17.27 σˆ = p84 − p50 = 17.27 − 13.) Probability Plot of Cycle Time (Ex7-14CT) Normal 99 95 90 84 Percent 80 Mean StDev N AD P-Value 13.27 10 10 15 20 Reimbursement Cycle Time. MTB > Stat > Basic Statistics > Normality Test (Add percentile lines at Y values 50 and 84 to estimate µ and σ.340 70 60 50 40 30 50 20 1 13.097 30 0.Chapter 7 Exercise Solutions 7-14☺.2 5 5 17. x ≈ p50 = 13.2 4.

MTB > Stat > Basic Statistics > Normality Test (Add percentile lines at Y values 50 and 84 to estimate µ and σ.2 6σˆ = 6(12.2 (b) USL = 2 hrs = 120 mins USL − µˆ 120 − 98.958983 = 0.) Probability Plot of Response Time (Ex7-15Resp) Normal 99 95 90 84 Percent 80 Mean StDev N AD P-Value 98.78 = 12.463 0.78 12.98 − 98.58 C pu = 3σˆ 3(12.2) = 73.78 ⎫ pˆ = Pr ⎨ z > ⎬ = 1 − Pr ⎨ z < ⎬ = 1 − Pr ⎨ z < ⎬ σˆ σˆ 12.78 = = 0. minutes 120 130 A normal probability plot of response times shows the distribution is close to normal.78 5 70 80 110.98 10 90 100 110 Response Time.27 40 0.041017 7-13 .2 ⎩ ⎭ ⎩ ⎭ ⎩ ⎭ = 1 − Φ (1.78 p84 = 110.Chapter 7 Exercise Solutions 7-15☺.2) USL − µˆ ⎫ USL − µˆ ⎫ ⎧ ⎧ ⎧ 120 − 98.98 σˆ = p84 − p50 = 110.739) = 1 − 0.243 70 60 50 40 30 50 20 1 98. (a) x ≈ p50 = 98.

465 0.69) = 16.96 σˆ = p84 − p50 = 55.) Probability Plot of Hardness Data (Ex5-59Har) Normal 99 95 90 84 Percent 80 Mean StDev N AD P-Value 53.27 5 46 48 50 52 54 Hardness 55.14 7-14 .Chapter 7 Exercise Solutions 7-16 (7-10).712 15 0.27 p84 = 55.69 6σˆ = 6(2.27 2.96 − 53.217 70 60 50 40 30 50 20 1 53.27 = 2.96 10 56 58 60 A normal probability plot of hardness data shows the distribution is close to normal. MTB > Stat > Basic Statistics > Normality Test (Add percentile lines at Y values 50 and 84 to estimate µ and σ. x ≈ p50 = 53.

7-15 .272 0.587 Percent 80 70 60 50 40 30 20 10 5 1 1000 1500 2000 Ex7-17FT 2500 3000 The plot shows that the data is not normally distributed. MTB > Stat > Basic Statistics > Normality Test Probability Plot of Failure Times (Ex7-17FT) Normal 99 Mean StDev N AD P-Value 95 90 1919 507.Chapter 7 Exercise Solutions 7-17 (7-11). so it is not appropriate to estimate capability.1 10 0.

9303 χ12−α .33 at a 95% confidence level.95 2 = 33.40 2 χα2/ 2. 1.36 χ12−α / 2.11 12.42 This confidence interval is wide enough that the process may either be capable (ppm = 27) or far from it (ppm ≈ 16.52 p 1 − α = 0. n −1 0.05 χ2 = χ2 1−α / 2.52 χ12−α .52 ⎝ 1.11 Cˆ p = 6σˆ 6(1.n −1 χα2/ 2.95.Chapter 7 Exercise Solutions 7-18 (7-12). 7-19 (7-13).5 (a) USL − LSL 85 − 75 = = 1.395).40 39.24 = 39. USL = 85. LSL = 75.36 ≤ C p ≤ 1.80 ≤ C p ≤ 1.33 ⎞ χ = 49 ⎜ ⎟ = 37.n −1 = χ 0.n −1 ˆ ˆ Cp ≤ Cp ≤ Cp n −1 n −1 1.12 2 1−α . n = 25. S = 1.975.49 49 = 1. n = 50 Cˆ = 1.33 2 ⎛ 1.49 χ2 Cˆ p 1−α .11 25 − 1 25 − 1 0.n −1 = χ 0.n −1 ≤ C p n −1 1.26 ≤ C p 49 The company cannot demonstrate that the PCR exceeds 1.52 33.025.9303 = 1.49 7-16 .5) (b) α = 0.24 = 12.88 α = 0.52 ⎠ 1 − α = 0.

05 zα / 2 = z0. n = 30.4287 ≤ C pk ≤ 0.6) Cˆ pk = min(Cˆ pl .63) 2(30 − 1) ⎦ ⎣ ⎣ 0.96 ⎥ ≤ C pk ≤ 0. LSL = 90 (a) USL x − µˆ x 100 − 97 Cˆ pu = = = 0.96 ⎥ 2 2 9(30)(0.63 (b) α = 0.6) µˆ − LSL x 97 − 90 Cˆ pl = x = = 1. x = 97.6.Chapter 7 Exercise Solutions 7-20 (7-14). S = 1.46 3σˆ x 3(1.63 ⎢1 + 1. Cˆ pu ) = 0.63 ⎢1 − 1.025 = 1.63 3σˆ x 3(1.63) 2(30 − 1) ⎦ 9(30)(0. USL = 100.8313 7-17 .960 ⎡ 1 1 ⎤ ⎥ ≤ C pk ≤ Cˆ pk + Cˆ pk ⎢1 − zα / 2 2 2(n − 1) ⎥ 9nCˆ pk ⎢⎣ ⎦ ⎡ 1 1 ⎤ ⎢1 + zα / 2 ⎥ + 2 2(n − 1) ⎥ 9nCˆ pk ⎢⎣ ⎦ ⎡ ⎤ ⎡ ⎤ 1 1 1 1 + + 0.

96.Chapter 7 Exercise Solutions 7-21 (7-15). n = 50 (a) USL x − µˆ x 2350 − 2275 = = 0. from Ex. σˆ I = 3.79 The approximation yields a narrower confidence interval.96 ⎥ 2(30 − 1) ⎦ 2(30 − 1) ⎦ ⎣ ⎣ 0. nominal = 2225.5443 7-22 (7-16).05.63 ⎢1+1.63.42 ⎢1 − 1. x = 2275.47 ≤ C pk ≤ 0. LSL = 2100. Cˆ pu ) = 0. s = 60.96 ⎥ 2 2 9(50)(0. σˆ Total = 5 2 2 2 σˆ Total = σˆ Meas + σˆ Process 2 2 σˆ Process = σˆ Total − σˆ Meas = 52 − 32 = 4 7-18 .025 = 1.2957 ≤ C pk ≤ 0.96 + + ⎥ ≤ C pk ≤ 0. 7-23 (7-17).42 Cˆ pu = 3σˆ x 3(60) µˆ − LSL x 2275 − 2100 = = 0.42) 2(50 − 1) ⎦ 9(50)(0.96 ⎥ ≤ C pk ≤ 0.42 ⎢1 + 1. but it is not too far off. n = 30 ⎡ ⎡ 1 ⎤ 1 ⎤ Cˆ pk ⎢1 − zα / 2 ⎥ ≤ C pk ≤ Cˆ pk ⎢1 + zα / 2 ⎥ 2(n − 1) ⎦ 2(n − 1) ⎦ ⎣ ⎣ ⎡ ⎤ ⎡ ⎤ 1 1 0.960 ⎡ 1 1 ⎤ ⎥ ≤ C pk ≤ Cˆ pk Cˆ pk ⎢1 − zα / 2 + 2 2(n − 1) ⎥ 9nCˆ pk ⎢⎣ ⎦ ⎡ 1 1 ⎤ ⎢1 + zα / 2 ⎥ + 2 2(n − 1) ⎥ 9nCˆ pk ⎢⎣ ⎦ ⎡ ⎤ ⎡ ⎤ 1 1 1 1 0. 7-20.42) 2(50 − 1) ⎦ ⎣ ⎣ 0. zα / 2 = z0. zα / 2 = 1.42 (b) α = 0. USL = 2350.97 Cˆ pl = x 3σˆ x 3(60) Cˆ pk = min(Cˆ pl . Cˆ pk = 0. σ OI = 0.63 ⎢1 − 1.

8.8 2 0 LC L=0 2 4 6 8 10 Sample 12 14 16 18 20 Test Results for Xbar Chart of Ex7-24All TEST 1. (b) specs: 25 ± 15 6σˆ Gauge 6(2.482) P = ×100 = ×100 = 49. 20 The R chart is in control. (a) n = 2.15 Sample Range 8 6 4 _ R=2. One point more than 3. x = 21.00 standard deviations from center line. 12.8 20 1 15 2 4 6 8 1 1 10 Sample 12 14 16 18 LC L=16. R = 2.07 25 _ _ X=21. σˆ Gauge = 2.482 MTB > Stat > Control Charts > Variables Charts for Subgroups > X-bar R Xbar-R Chart of Part Measurements (Ex7-24All) 30 1 Sample M ean U C L=27.8. Test Failed at points: 8. The new gauge is more repeatable than the old one. 15.Chapter 7 Exercise Solutions 7-24 (7-18).53 20 U C L=9. and the x chart has a few out-of-control parts.6% T USL − LSL 2(15) 7-19 .

0 1 U C L=100. Test Failed at points: 2.00 standard deviations from center line.553 Sample M ean 100.359 × 100 = 62. σˆ Gauge = R d 2 = 2.5 0.0 LC L=95. 3 The x chart has a couple out-of-control points.0 _ R=2.717 − 1.5% 4.3. One point more than 3. R = 2.2 97. (b) x = 98.272 T USL − LSL 115 − 85 7-20 . LSL = 100 – 15 = 85 6σˆ Gauge P 6(1.Chapter 7 Exercise Solutions 7-25 (7-19).2.359 2 σˆ Total = 4.3 1.717 2 2 2 σˆ Product = σˆ Total − σˆ Gauge = 4.695 (c) σˆ Gauge σˆ Total × 100 = 1.5 99.847 1 1 2 3 4 5 6 7 8 9 10 Sample Sample Range 6.359) = = = 0.3592 = 2.3 1.872 σˆ Product = 1.921 4.693 = 1.0 U C L=5.5 96. MTB > Stat > Control Charts > Variables Charts for Subgroups > X-bar R Xbar-R Chart of Part Mesaurements (Ex7-25All) 102. and the R chart is in control.5 3.0 _ _ X=98. This indicates that the operator is not having difficulty making consistent measurements.717 (d) USL = 100 + 15 = 115.0 LC L=0 1 2 3 4 5 6 7 8 9 10 Sample Test Results for Xbar Chart of Ex7-25All TEST 1.

418 σˆ Measurement Error = 1.70. R2 = 2.30 R = 2.Chapter 7 Exercise Solutions 7-26 (7-20).03.128 σˆ Reproducibility = Rx d 2 = 0.128 = 0.693 σˆ Repeatability = R d 2 = 2.191 (c) specs: 50 ± 10 6σˆ Gauge P 6(1. x2 = 49.17 n = 2 operators d 2 = 1.17 1.7% T USL − LSL 60 − 40 7-21 .151 = 1.693 = 1.151 (b) 2 2 2 ˆ2 ˆ2 σˆ Measurement Error = σ Repeatability + σ Reproducibility = 1. R1 = 1.181 + 0.191) = ×100 = × 100 = 35. (a) Excel : workbook Chap07.87.00 n = 3 repeat measurements d 2 = 1.181 Rx = 0.xls : worksheet Ex7-26 x1 = 50.00 1.

(a) σˆ Gauge = R d 2 = 1.7 20. One point more than 3.00 standard deviations from center line.533 1.0 1 1 2 3 4 5 6 7 8 Sample 9 10 Sample Range 6.Chapter 7 Exercise Solutions 7-27 (7-21).0 11 12 1 1 13 14 15 U C L=5.82 1 15.0 1 2 3 4 5 6 7 8 Sample 9 10 11 12 13 14 15 Test Results for R Chart of Ex7-27All TEST 1.5 _ _ X=20.154 (b) MTB > Stat > Control Charts > Variables Charts for Subgroups > X-bar R Xbar-R Chart of Part Measurements (Ex7-27All) 1 1 Sample M ean 25.0 _ R=1. 7-22 . Test Failed at points: 11.5 LC L=0 0.533 1.010 4.128 = 1. 12 Out-of-control points on R chart indicate operator difficulty with using gage.5 3.359 Gauge capability: 6σˆ = 8.5 LC L=17.0 17.0 U C L=23.58 22.

2. Levels 20 3 3.861 R-Sq(adj) = 90.8832 Error term 3 3 Expected Mean Square for Each Term (using unrestricted model) (3) + 6 (1) (3) + 40 (2) (3) 7-23 . 7.712 2 σˆ Part = = 10. MTB > Stat > ANOVA > Balanced ANOVA In Results.0106 0.9917 Error term 3 3 4 Expected Mean Square for Each Term (using unrestricted model) (4) + 2 (3) + 6 (1) (4) + 2 (3) + 40 (2) (4) + 2 (3) (4) 2 σˆ Repeatability = MS Error = 0. Note the Part × Operator variance component is negative. Ex7-28Op … 1 2 3 Source Ex7-28Part Ex7-28Op Error Variance component 10.712 Error 60 59.992 = = −0.72 P 0. 13.425 62.74% Variance component 10. 20 Ex7-28Op Values 1.0149 -0.10).992 Total 119 1274.2798 = on 3(2) 2 σˆ Part×Operator = The manual calculations match the MINITAB results. 8.1400 ⇒ 0 n 2 MSO − MSP×O 1. 3 Analysis of Variance for Ex7-28Reading Source DF SS MS Ex7-28Part 19 1185.1399 0.308 − 0. R-Sq = 95. 4. 11.84 0.173 0.050 0. 17.Chapter 7 Exercise Solutions 7-28☺.33% Source Ex7-28Part Ex7-28Op Ex7-28Part*Ex7-28Op Error F 87.000 0. 1. 15. 16.712 − 0.391 Ex7-28Op 2 2.992 MSP×O − MSE 0.2798 0. For the reduced model: ANOVA: Ex7-28Reading versus Ex7-28Part. select “Display expected mean squares and variance components” ANOVA: Ex7-28Reading versus Ex7-28Part.65 1. 14.500 0. 12. 5.308 Ex7-28Part*Ex7-28Op 38 27. 2. we can fit a reduced model without that term.995825 1 2 3 4 9. 10.617 1.2513 0.391 − 0.712 2 σˆ Operator = = 0. Since the Part × Operator term is not significant (α = 0. 6.592 S = 0. Ex7-28Op Factor Ex7-28Part Ex7-28Op Type random random Factor Ex7-28Part 18.0149 = pn 20(2) MSP − MSP×O 62. 19.

0106 2 2 σˆ Repeatability = σˆ Error = 0.3083 0.0106 0.33842 20.7178 P 0.59 MS 62.000 0.02 7.32 28.4814 P 0.6470 1.8938 0.50 1274.09 3.6186 Part-To-Part 3.6186 Ex7-28Op 0.1050 /T = P USL-LSL 60 − 6 This gauge is borderline capable since the estimate of P/T ratio just exceeds 0.3083 0.55 1274.43 2.9454 (c) 6 × σˆ Gauge 6 × 0.6447 1.9917 F 87.7118 0.05 59. and total gauge variability may also be found using: MTB > Stat > Quality Tools > Gage Study > Gage R&R Study (Crossed) Gage R&R Study .20176 19.6386 Reproducibility 0. reproducibility.10 0.09 95.93977 5.10310 0.3908 1.10 91.3908 1.8832 (b) 2 2 2 σˆ Gauge = σˆ Reproducibility + σˆ Repeatability = 0.2513 11.6724 Repeatability 0.0106 10.0106 + 0.10.9454 n = = 0.ANOVA Method Two-Way ANOVA Table With Interaction Source Ex7-28Part Ex7-28Op Ex7-28Part * Ex7-28Op Repeatability Total DF 19 2 38 60 119 SS 1185.861 Two-Way ANOVA Table Without Interaction Source Ex7-28Part Ex7-28Op Repeatability Total DF 19 2 98 119 SS 1185.94541 5. repeatability.92 0.15 3.00 Study Var Source StdDev (SD) (6 * SD) Total Gage R&R 0.000 0.43 2.8832 0.Chapter 7 Exercise Solutions (a) 2 2 σˆ Reproducibility = σˆ Operator = 0.8832 = 0.173 0.8938 σˆ Gauge = 0.98 100.59 MS 62.91 100.232 Gage R&R Source Total Gage R&R Repeatability Reproducibility Ex7-28Op Part-To-Part Total Variation VarComp 0.62 86.0305 Number of Distinct Categories = 4 %Study Var (%SV) 28.1451 %Contribution (of VarComp) 8.10310 0.62 27.00 7-24 . Estimates of variance components.8832 F 70.8380 0.2106 Total Variation 3.

757 30 25 2 _ R=1.Chapter 7 Exercise Solutions 7-28 continued Visual representations of variability and stability are also provided: Gage R&R (ANOVA) for Ex7-28Reading Reported by : Tolerance: M isc: G age name: Date of study : Components of Variation Ex7-28Reading by Ex7-28Part 100 % Contribution 30 Percent % Study Var 25 50 20 0 Gage R&R Repeat Reprod 1 Part-to-Part 2 3 4 5 R Chart by Ex7-28Op Sample Range 4 1 2 6 7 9 10 11 12 13 14 15 16 17 18 19 20 Ex7-28Part Ex7-28Reading by Ex7-28Op 3 UCL=3.55 _ _ X=22.15 20 LCL=0 0 1 2 Ex7-28Op Xbar Chart by Ex7-28Op 1 2 3 Ex7-28Op * Ex7-28Part Interaction 3 30 30 25 UCL=24.39 20 LCL=20.23 Average Sample Mean 8 Ex7-28Op 1 2 3 25 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 5 16 17 18 19 20 Ex7-28Part 7-25 .

1451 n= SNR 2 ρˆ P 2(0.79 1 − ρˆ P 1 − 0. 7-30 (7-22). (Also note that the MINITAB Gage R&R output indicates “Number of Distinct Categories = 4”.2513 = = 0.94. this is also the number of distinct categories of parts that the gauge is able to distinguish) DR = 23. exceeding the minimum recommendation of four. µ = µ1 + µ2 + µ3 = 100 + 75 + 75 = 250 σ = σ 12 + σ 22 + σ 33 = 42 + 42 + 22 = 6 Pr{x > 262} = 1 − Pr{x ≤ 262} 262 − µ ⎫ ⎧ = 1 − Pr ⎨ z ≤ ⎬ σ ⎭ ⎩ 262 − 250 ⎫ ⎧ = 1 − Pr ⎨ z ≤ ⎬ 6 ⎩ ⎭ = 1 − Φ (2.1451 ρˆ P = 2 σˆ Part 10.9198 = 23. σˆ Total = 11.000) = 1 − 0. the gauge is capable. By this measure.9198) = = 4.9198 m = 1 + ρˆ P = 1 + 0.2513.9198 2 σˆ Total 11.9772 = 0. This is near the AIAG-recommended value of five levels or more.79 indicates that fewer than five distinct levels can be reliably obtained from the measurements.Chapter 7 Exercise Solutions 7-29☺.0228 7-26 . 2 2 σˆ Part = 10.9198 SNR = 4. but larger than a value of two (or less) that indicates inadequate gauge capability.94 DR 1 − ρˆ P 1 − 0.

01) = 0.6) + 1 − Φ (1.1 or y = x1 − x2 > 0.2 2 )(0.082 ⎡⎣10 2 (0.4 ⎞ ⎛ 0. Volume = L × H × W ≅ µ L µ H µW + ( L − µ L ) µ H µW + ( H − µ H ) µ L µW + (W − µW ) µ L µ H µˆ Volume ≅ µ L µ H µW = 6.00) = 0.32 ) ⎤⎦ = 0.1} + 1 − Pr{ y ≤ 0.22 )(0.9} ⎛ 0. x2 ~ N (19.08(10)(20)(3) = 48 2 σˆ Weight ≅ d 2 ⎡⎣ µˆW2 σˆ L2σˆT2 + µˆ L2σˆW2 σˆT2 + µˆT2σˆW2 σˆ L2 ⎤⎦ = 0.0 2 σ Volume ≅ µ L2σ H2 σ W2 + µ H2 σ L2σ W2 + µW2 σ L2σ H2 = 6.25 σ y = 0.42 = 0.9} = Pr{ y < 0.32 )(0.00181 σˆ Weight ≅ 0.04252 7-27 . Weight = d × W × L × T ≅ d [ µW µ L µT + (W − µW ) µ L µT + ( L − µ L ) µW µT + (T − µT ) µW µ L ] µˆ Weight ≅ d [ µW µ L µT ] = 0.0) = 72.8413 = 0.12 ) + 202 (0.01) + 3.6.1 − 0.50 Pr{Nonconformities} = Pr{ y < LSL} + Pr{ y > USL} = Pr{ y < 0.Chapter 7 Exercise Solutions 7-31 (7-23).02 (0.2743 + 1 − 0.01)(0.0)(4.4330 7-32 (7-24).25 ⎠ = Φ (−0.4 σ y2 = σ 12 + σ 22 = 0.01)(0.01)(0. 0.9 µ y = µ1 − µ2 = 20 − 19.25 ⎠ ⎝ 0.32 ).1} + Pr{ y > 0.6 = 0.4 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 0.02 (0.9 − 0.32 + 0.01) + 4.12 ) + 32 (0.0(3.0061 7-33 (7-25). x1 ~ N (20.42 ) Nonconformities will occur if y = x1 − x2 < 0. 0.02 (0.

05(3.1282 2⎟ 26 ⎜⎝ 3 2 2 ⎣ 26 ⎦ ⎠ 4 4 4 ⎞ 1 ⎛5 2 ⎡1 ⎤ E ( x 2 ) = ∫ x 2 f ( x)dx = ∫ x 2 ⎢ (5 x − 2) ⎥ dx = ⎜ x 4 − x3 ⎟ = 10.1026 26 ⎜⎝ 4 2 3 2 ⎟⎠ 2 ⎣ 26 ⎦ s = (3 + 0. 2 ≤ x ≤ 4 26 4 4 4⎞ 1 ⎛5 3 ⎡1 ⎤ E ( x) = µ x = ∫ xf ( x)dx = ∫ x ⎢ (5 x − 2) ⎥ dx = ⎜ x − x 2 ⎟ = 3.05(3.05 x) 2 and f ( x) = σ x2 = E ( x 2 ) − [ E ( x)] = 10.05 x)2 ⎤ ⎥ =⎢ ∂x ⎢⎣ ⎥⎦ 2 σ x2 µx = 2(3 + 0.1282) 2 = 0.1001 7-35 (7-27).1282) ] (0.3170) = 0.05( µ x )] = [3 + 0.05)σ x2 = 2 [3 + 0.05)(0. I = E ( R1 + R2 ) µI ≅ µE (µR + µR ) 1 σ I2 ≅ σ 2 µE σ 2 + σ R2 2 ( R (µR + µR ) (µR + µR ) 1 2 E + 2 1 1 2 ) 2 7-28 .3170 2 µ s ≅ g ( x) = [3 + 0.9629 2 ⎡ ∂ g ( x) ⎤ σ ≅⎢ ⎣ ∂ x ⎥⎦ 2 2 2 s σ x2 µx ⎡ ∂ (3 + 0. 1 (5 x − 2).1026 − (3.Chapter 7 Exercise Solutions 7-34 (7-26).1282)] = 9.05µ x )(0.

004 = 0.512 0.01 ⎞ 5. 0.346 7-37 (7-29).000005 ⎠ = 1 − Φ (−2.01 γ = 0.Chapter 7 Exercise Solutions 7-36 (7-28).989 1 ⎛ 2 − α ⎞ χ1−γ .006 2 2 σ y2 = σ ID + σ OD = 0.010. x1 ~ N ( µ1 .002236 Pr{positive clearance} = 1 − Pr{interference} = 1 − Pr{ y < 0} ⎛ 0 − 0. ID ~ N (2.0012 = 0.0022 + 0.5(−2. 0.683) = 1 − 0.004.0036 = 0.09] = 1.09 − µ y ⎫⎪ −1 Pr ⎨ z < ⎬ = Φ (0.20.5 Pr{ y < 0.3002 = 0. 0.512) − 0.989 +⎜ = +⎜ = 299 ⎟ ⎟ 2 ⎝ α ⎠ 4 2 ⎝ 0.4 = χ 0.4002 + 0. x2 ~ N ( µ2 .01 ⎠ 4 2 n≅ 7-29 .006) σ y ⎭⎪ ⎩⎪ 0.0012 ) Interference occurs if y = ID – OD < 0 µ y = µID − µOD = 2.006 ⎞ = 1− Φ ⎜ ⎟ ⎝ 0.09} = 0.4002 ).09 − µ y = −2.9964 7-38 (7-30).80 2 χ12−γ .000005 σ y = 0. α = 0.010 − 2.4 = 5.3002 ) µ y = µ1 − µ2 σ y = σ 12 + σ 22 = 0.4 1 ⎛ 2 − 0.5 µ y = −[0.006 ⎧⎪ 0. 0.0022 ) and OD ~ N (2.

α = 0. n = 25.05 ⎠ 4 2 n≅ After the data are collected. one-sided From Appendix VIII: K = 2.08 7-42 (7-34). α = 0.779 +⎜ = +⎜ = 77 ⎟ ⎟ 2 ⎝ α ⎠ 4 2 ⎝ 0.4 1 ⎛ 2 − 0.208 UTL = x + KS = 350 + 2.838 x − KS = 85 − 1.95 .90 2 χ12−γ .95 . x ~ N (350.355 UTL = x + KS = 300 + 2. γ = 0.05 ⎞ 7.10. x ~ N (300. n = 20. one-sided From Appendix VIII: K = 1.102 ). one-sided From Appendix VIII: K = 2. x ~ N (85.12 ). 7-30 .Chapter 7 Exercise Solutions 7-39 (7-31).55 7-40 (7-32). α = 0.779 1 ⎛ 2 − α ⎞ χ1−γ .208(10) = 372.4 = 7. γ = 0.90 . γ = 0.838(1) = 83.10.102 ).4 = χ 0. α = 0.162 7-41 (7-33).355(10) = 323. a natural tolerance interval would be the smallest to largest observations.05. n = 10.05 γ = 0.10.

the interval relates to individual observations (random variables).n −1 = t0.0003) = [0.00032 ( ) (a) α = 0.023 0. γ = 0.n −1 S ( n = 0. while in part (b) the interval refers to a parameter of a distribution (an unknown constant).05) The largest observation would be the nonparametric upper tolerance limit.1271] (b) α = 0.025.1264. part (b) is a confidence interval on mean thickness. 0. 7-31 . 7-44 (7-36).445 TI on x : x ± KS = 0.1257.0003 ) 40 = [0. In part (a).1263.95 log(1 − γ ) log(1 − 0.1265] Part (a) is a tolerance interval on individual thickness observations.05. γ = 0. x ~ N 0. and two-sided From Appendix VII: K = 2.1264 ± 2.39 = 2. 0.Chapter 7 Exercise Solutions 7-43 (7-35).445(0.1264 ± 2. 0.95) n= = = 59 log(1 − α ) log(1 − 0.05.05. α = 0.95. tα / 2.023 CI on x : x ± tα / 2.

µ0 = 1050.Chapter 8 Exercise Solutions Several exercises in this chapter differ from those in the 4th edition. h = 5 1250 Cumulative Sum 1000 750 500 250 UCL=125 0 0 LCL=-125 2 4 6 8 10 12 Sample 14 16 18 20 The process signals out of control at observation 10. 8-1. k = 0. δ = 1σ. (b) σˆ = MR2 d 2 = 38.128 = 34. The assignable cause occurred after observation 10 – 3 = 7. std dev = 25. σ = 25. K = (δ/2)σ = (1/2)25 = 12.5.5. The estimate used for σ is much smaller than that from the data.8421/1. H = 5σ = 5(25) = 125 (a) MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Molecular Weight (Ex8-1mole) target = 1050. The point at which the assignable cause occurred can be determined by counting the number of increasing plot points. New exercises are denoted with an “☺”. 8-1 .4345 No. A number in parentheses gives the exercise number from the 4th edition. An “*” following the exercise number indicates that the description has changed.

MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Standardized Molecular Weight (Ex8-2std) target = 1050. std dev = 25.5. h = 5 50 Cumulative Sum 40 30 20 10 UCL=5 0 0 LCL=-5 2 4 6 8 10 12 Sample 14 16 18 20 The process signals out of control at observation 10. k = 0. 8-2 .Chapter 8 Exercise Solutions 8-2. The assignable cause occurred after observation 10 – 3 = 7.

H/2 = 125/2 = 62.5 FIR = H/2 = 62. the process signals out of control at observation 10.5.Chapter 8 Exercise Solutions 8-3. h = 5.5 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Molecular Weight (Ex8-1mole) FIR=H/2 = 62. k = 0. (a) µ0 = 1050.1045 − (1050 + 12. C1+ = max ⎡⎣0. 8-3 .5. K = 12.5] = 45 Using the tabular CUSUM. σ = 25.5. xi − ( µ0 − K ) + C0+ ⎤⎦ = max [ 0.5) + 62. the same as the CUSUM without a FIR feature.5/25 = 2.5 std dev units) 1250 Cumulative Sum 1000 750 500 250 UCL=125 0 0 LCL=-125 2 4 6 8 10 12 Ex8-1Obs 14 16 18 20 For example. in std dev units = 62.5 (or 2.

5SL=141.5-sigma limits +3. However there does appear to be a trend up from observations 6 through 12—this is the situation detected by the cumulative sum.6 120 80 __ MR=38.5σ limits on the Individuals chart.8 Individual Value 1200 _ X=1116.5SL=995. 8-4 .8 2 4 6 8 10 12 Observation 14 16 18 20 160 Moving Range +3.5SL=0 2 4 6 8 10 12 Observation 14 16 18 20 Using 3. there are no out-of-control signals.8 40 0 -3.Chapter 8 Exercise Solutions 8-3 continued (b) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Molecular Weight (Ex8-1mole) with 3.3 1100 1000 -3.5SL=1236.

5)(5.5) 2 1 1 1 2 = + = = 0.02.5 exp[−2(−0.5)(5.77 0. k = 1/ 2.2 LCL=-0.77.128 = 0.936 δ * = 0.05.2385 (a) MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Can Weight (Ex8-4can) target = 8.0027 + − ARL0 ARL0 ARL0 742. b = h + 1.2385 Cumulative Sum 0. k = 0. h = 4.2 0.936) − 1 = 742.5 = −0.05) = 0.0166 .964 2(−0.77.1 0.0186957 /1.0 0 -0.05 is probably not reasonable. ∆ − = −δ * − k = −0 − 0. h = 4. In Exercise 8-4: µ0 = 8. σ = 0.5.936)] + 2(−0.166 = 4.5.2385 -0. so σ = 0.48 8-5 .77 + 1.05. (b) σˆ = MR2 1. σ = 0. k=1/2. µ0 = 8. and h=4.5 = −0.0027 = 371.3 UCL=0.02.02.128 = 0.3 2 4 6 8 10 12 14 Sample 16 18 20 22 24 There are no out-of-control signals. ∆ + = δ * − k = 0 − 0.Chapter 8 Exercise Solutions 8-4.166 = 5.1 -0. H = hσ = 4.77 (0.964 ARL+0 = ARL−0 ≅ ARL0 = 1/ 0.

2 -0. h = 8. k = 0.1 -0. ∆ − = −δ * − k = −0 − 0.01 + 1.4 -0.166 = 8. h = 8. k = 0.05) = 0. 8-6 .166 = 9.01.4005 -0.25)(9. σ = 0.4005 0.02.176 δ * = 0. H = hσ = 8.84 The theoretical performance of these two CUSUM schemes is the same for Exercises 8-4 and 8-5.01 (0.5 2 4 6 8 10 12 14 Sample 16 18 20 22 24 There are no out-of-control signals.2 0.25)(9. k = 0.05.25. ∆ + = δ * − k = 0 − 0.3 LCL=-0. In Exercise 8-5: µ0 = 8.01.6771 2(−0.0027 = 370.25 = −0.25 exp[−2(−0.4 0.25.1 0 0. h = 8.4005 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Can Weight (Ex8-4can) target = 8.0027 + − ARL0 ARL0 ARL0 741.02. b = h + 1. σ = 0.01 UCL=0.176) − 1 = 741.25) 2 1 1 1 2 = + = = 0.6771 ARL+0 = ARL−0 ≅ ARL0 = 1 0.02.0 -0.3 Cumulative Sum 0.Chapter 8 Exercise Solutions 8-5.25 = −0.05.176)] + 2(−0.25.25. µ0 = 8.

Chapter 8 Exercise Solutions 8-6.3 UCL=0.2385 -0.1 0.385 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Ex8-4can FIR = 2. µ0 = 8. h = 4. FIR in # of standard deviations = h/2 = 4.77/2 = 2. target = 8.3 2 4 6 8 10 12 14 Sample 16 18 20 22 24 The process signals out of control at observation 20.00.0 0 -0.05. σ = 0.77 (0. k = 0.1 -0.2 LCL=-0.77. h = 4.2385 FIR = H/2.05) = 0.2385 Cumulative Sum 0.77 0. H = h σ = 4. k = 1/2.00.5. 8-7 .385 std dev. Process was out of control at process start-up.2 0.

16 (b) µ0 = 950. Test Failed at points: 12. σˆ = 12.128 = 12.7215 /1.8 Cumulative Sum 50 25 0 0 -25 -50 LCL=-60. 8-8 .16. k = 1/ 2. 13 The process signals out of control at observation 12.8 1 8 16 24 32 40 48 Sample 56 64 72 80 Test Results for CUSUM Chart of Ex8-7temp TEST.Chapter 8 Exercise Solutions 8-7. (a) σˆ = MR2 d 2 = 13. h = 5 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Temperature Readings (Ex8-7temp) target = 950.5. h = 5 75 UCL=60. The assignable cause occurred after observation 12 – 10 = 2. One point beyond control limits. k = 0.

std dev = 5. One point beyond control limits. 32 The process signals out of control on the lower side at sample 3 and on the upper side at sample 12.1 0 LCL=-28.35 /1. 18. Assignable causes occurred after startup (sample 3 – 3 = 0) and after sample 8 (12 – 4).629.1 0 -100 -200 3 6 9 12 15 18 Sample 21 24 27 30 Test Results for CUSUM Chart of Ex8-8con TEST. h = 5 400 Cumulative Sum 300 200 100 UCL=28. 28.629.128 = 5.35) (b) µ0 = 175. k = 1/ 2. σˆ = 5. 22. 20. 29. 23.Chapter 8 Exercise Solutions 8-8. k = 1/2.629 (from a Moving Range chart with CL = 6. 8-9 . Test Failed at points: 12. 25. 30. 21. 24. 16. (a) σˆ = MR2 d 2 = 6. h = 5 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Bath Concentrations (Ex8-8con) target = 175. 26. 27. 14. 17. 13. 19. 31. 15.

71/1. Assignable causes occurred after startup (sample 2 – 2) and after sample 9 (16 – 7).25.01 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Viscosity Measurements (Ex8-9vis) target = 3200. h = 8.71) (b) µ0 = 3200.Chapter 8 Exercise Solutions 8-9. σˆ = 5. k = 0. to give longer in-control ARLs with shorter out-of-control ARLs. should be balanced by a larger control limit. (c) Selecting a smaller shift to detect. 8-10 . std dev = 5.01.949. h = 8. (a) σˆ = MR2 d 2 = 6. 18 The process signals out of control on the lower side at sample 2 and on the upper side at sample 16.7 0 0 Cumulative Sum LCL=-47. One point beyond control limits.949.949 (from a Moving Range chart with CL = 6.01 UCL=47. 17. k = 0.25. h = 8. Test Failed at points: 16.25. k = 0.128 = 5.7 -100 -200 -300 -400 -500 4 8 12 16 20 Sample 24 28 32 36 Test Results for CUSUM Chart of Ex8-9vis TEST.

42.2 1.6 0. H = hσ x = 0.14. One point beyond control limits. h = 4.50..2504 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Flow Width Data (Exm5-1x1.14. k = 0.5. k = δ 2 = 0.. n = 5.. h = 4 1.0313. ..250 -0. 41. 8-11 .0626 δ = 1. 45 The CUSUM chart signals out of control at sample 40.250 0.4 4 8 12 16 20 24 Sample 28 32 36 40 44 Test Results for CUSUM Chart of Exm5-1x1. µ0 = 1.Chapter 8 Exercise Solutions 8-10*. The x -R chart shown in Figure 5-4 signals out of control at sample 43.8 0. Exm5-1x5 TEST. .0 -0. 43. σ = 0. K = kσ x = 0.14 5 = 0.0 Cumulative Sum 0.50. Test Failed at points: 40. σ x = σ n = 0. std dev = 0. at sample 40 versus sample 43. This CUSUM detects the shift in process mean earlier.4 UCL=0.2 0 0. and remains above the upper limit.. Exm5-1x5) target = 1..2 LCL=-0. 44.5.

57 1 1.66 9 OOC 7 26.OOC? When? 0 0.049 4 OOC 7 11. Process variability is increasing.56 8 OOC 7 23.49 1008 -1.56 3.61 1146 3. 8-12 .84 one-sided upper cusum Si+ N+ OOC? When? 0 0 0 0 0 0 0 0 0 0.52 -0.36 1050 0 -2. Vi = | yi | − 0.349 ( ) Excel file: workbook Chap08.15 2 0.78 11 OOC 7 33.68 3.38 1146 3.631 1 2.22 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 The process is out of control after observation 10 – 3 = 7.822 0.52 5.40 1128 3.52 4.76 3.04 3.84 3.71 1238 7.498 3 OOC 7 8.2 -1.44 5 OOC 7 14.68 1.8 1.88 13 OOC 7 one-sided lower cusum Si.90 1151 4.848 2 0 0 0.52 3.9 10 OOC 7 30.21 1095 1.86 1 0.36 1087 1.50 1125 3 2.65 4 0 0 0 0 1.94 3 0.84 3.N.29 1064 0.989 1 1.5 5 Obs.55 7 OOC 7 21.48 1. i No FIR xi 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 yi vi 1045 -0.26 1139 3.12 2.07 1055 0.56 -0.07 1037 -0.61 1163 4.2 -1.26 1167 4.738 2 5.35 6 OOC 7 16.54 12 OOC 7 36.Chapter 8 Exercise Solutions 8-11.xls : worksheet Ex8-11 mu0 = sigma = delta = k= h= 1050 25 1 sigma 0.74 1188 5.05 1169 4.13 1125 3 2.

47 13 OOC 0 192 3.9540 1.45 9 OOC 0 195 3.32 12 OOC 0 175 0.27 23.56 10.62 17.36 20.00 19.6646 2.9081 3.04 18 OOC 0 185 1.xls : worksheet Ex8-12 mu0 = sigma = delta = k= h= i No FIR 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 175 5.8 7 OOC 0 162 -2.3 8 OOC 0 186 1.65 23.946 2 150 -4.3093 2. Process variability has been increasing since start-up.31 26.7764 1.349 ( ) Excel file : workbook Chap08.OOC? When? 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1.06 22.74 15 OOC 0 197 3.34 20 OOC 0 one-sided lower cusum Si.4410 3.84 31.68 6 OOC 0 158 -3.46 31 19 OOC 0 182 1.4869 2.822 1 158 -3.6646 2.05 23 10 OOC 0 179 0. 8-13 .62 22.16 30.9081 3.0199 2.129 3 OOC 0 151 -4.2435 0.86 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 … The process was last in control at period 2 – 2 = 0.7106 0.32 28.5528 3.68 7. Vi = | yi | − 0.0199 2.37 17 OOC 0 189 2.18 15.65 20.32 1.7304 3.0000 -2.5987 1.9540 1.62 3.59 14 OOC 0 186 1.N.0199 2.6294 (from Exercise 8-8) 1 sigma 0.5 5 one-sided upper cusum xi yi vi Si+ N+ OOC? When? 0 160 -2.2633 3.822 0.55 16 OOC 0 190 2.19 4 OOC 0 153 -3.56 11 OOC 0 184 1.Chapter 8 Exercise Solutions 8-12.31 13 5 OOC 0 154 -3.

166)] + 2(−0.02 8-14 .3)(9.3) 2 exp[−2(−0.166 exp[−2(−0.166) − 1 ARL−1 = = 381.2 b = h + 1.166 = 8 + 1. 767 2(−0.2 ∆ − = −δ * − k = −0 − 0.2 = −0.2)(9. 767 ARL1 = 1/ 0.Chapter 8 Exercise Solutions 8-13.2 = −0. two-sided cusum with k = 0.23 Out of control ARL Performance: δ * = 0.3 ∆ − = −δ * − k = −0.7)(9.166 = 8 + 1.2 = −0.005 = 215.023 381.040 = 25.556 ARL0 = 1/ 0.5 − 0.023 2(0.556 2(−0. Standardized.3)(9.166) − 1 ARL+0 = ARL−0 ≅ = 430.166)] + 2(0.166) − 1 ARL+1 = = 25.7)(9.5 − 0.7) 2 1 1 1 1 1 = + = + = 0.166 = 9.2 = 0.2) 2 1 1 1 2 = + = = 0.040 + − ARL1 ARL1 ARL1 25.166 exp[−2(0.2 and h = 8 In control ARL performance: δ* = 0 ∆ + = δ * − k = 0 − 0.5 ∆ + = δ * − k = 0.005 + − ARL0 ARL0 ARL0 430.166)] + 2(−0.166 = 9.2)(9.7 b = h + 1.

762 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Viscosity Measurements (Ex8-9vis) target = 3150 1200 Cumulative Sum 1000 800 600 400 200 UCL=30 0 LCL=-30 0 4 8 12 16 20 Sample 24 28 32 36 MINITAB displays both the upper and lower sides of a CUSUM chart on the same graph.95238) = 2. H = hs = 5(5.5.95238. k = 0.976.Chapter 8 Exercise Solutions 8-14. The process signals out of control on the upper side at sample 2.5 (5. The assignable cause occurred at start-up (2 – 2). s = 5.95238) = 29. The upper CUSUM is used to detect upward shifts in the level of the process. h = 5 K = ks = 0. 8-15 . µ0 = 3150. there is no option to display a single-sided chart.

5. starting at about sample 18.5 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Light Velocity (Ex5-60Vel) target = 734.5. from virtually the first sample. of either a trend/drift or a shit in measurements.7) = 543. σˆ = MR2 d 2 = 122. displayed a distinct downward trend in measurements.7 (from a Moving Range chart with CL = 122.7) = 54. The CUSUM chart reflects a consistent run above the target value 734.5. with a centerline at x = 909 .35 H = hσˆ = 5(108.128 = 108. There is a distinct signal on both charts. 8-16 .Chapter 8 Exercise Solutions 8-15☺. k = 0. The outof-control signals should lead us to investigate and determine the assignable cause. h = 5 K = kσˆ = 0.5(108.6) µ0 = 734.5 4000 Cumulative Sum 3000 2000 1000 UCL=544 0 0 LCL=-544 4 8 12 16 20 24 Sample 28 32 36 40 The Individuals I-MR chart.6 /1.

L = 2.8 750 _ _ X=734.1. CL = µ0 = 734. 734.Chapter 8 Exercise Solutions 8-16☺.7.7 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Light Velocity (Ex5-60Vel) lambda = 0.5. σ = 108.7 900 EWMA 850 800 +2.2 4 8 12 16 20 24 Sample 28 32 36 40 The EWMA chart reflects the consistent trend above the target value.7SL=667. L = 2.1.5 700 -2.7SL=801. 8-17 .5. λ = 0. and also indicates the slight downward trend starting at about sample 22.

CL = µ0 = 1050. σ = 25. L = 2.2 (2 − 0.51 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Molecular Weight (Ex8-1mole) lambda = 0.Chapter 8 Exercise Solutions 8-17 (8-15).11.2.4) = [8. the width of the control limits also increases.1. (a) λ = 0. L = 3 limits = µ0 ± Lσ λ (2 − λ ) = 10 ± 3(1) 0.1 (2 − 0. UCL = 1065. the same as the CUSUM chart. 8-18 .4.7SL=1065. L = 3 limits = µ0 ± Lσ λ (2 − λ ) = 10 ± 3(1) 0. λ = 0. L = 3 limits = µ0 ± Lσ λ (2 − λ ) = 10 ± 3(1) 0.1. LCL = 1034.4 _ _ X=1050 1060 1040 -2.6 2 4 6 8 10 12 Sample 14 16 18 20 Process exceeds upper control limit at sample 10.1) = [9. L = 2.11] (c) λ = 0.49.7 1140 1120 EWMA 1100 1080 +2.7.7SL=1034.4 (2 − 0.31.2) = [9.1.10. 8-18 (8-16).5.5] As λ increases.69] (b) λ = 0.

λ = 0. CL = µ0 = 8. L = 3. LCL = 7.05.07. L = 3 8.97 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Can Weight (Ex8-4can) lambda = 0.98 LCL=7.2. Assume σ = 0.0700 8.06 EWMA 8.04 _ _ X=8.Chapter 8 Exercise Solutions 8-19 (8-17).02.02 8.00 7.96 2 4 6 8 10 12 14 Sample 16 18 20 22 24 The process is in control. UCL = 8.2. 8-19 .08 UCL=8.9700 7.02 8.

CL = µ0 = 8.99 -2.02 8.7 +2.7SL=8.98913 2 4 6 8 10 12 14 Sample 16 18 20 22 24 The process is in control. LCL = 7.00 7.Chapter 8 Exercise Solutions 8-20 (8-18).05.02.1.05 8.7SL=7. 8-20 .02 8. UCL = 8. L = 2. L = 2.1.04 EWMA 8. There is not much difference between the control charts. Assume σ = 0.05087 8. λ = 0.99 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Can Weight (Ex8-4can) lambda = 0.7.05.01 8.03 _ _ X=8.

CL = µ0 = 950.7SL=942.1. 8-21 . Test Failed at points: 12. MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Temperature Readings (Ex8-7temp) lambda = 0. 13 Process is out of control at samples 8 (beyond upper limit. but not flagged on chart).7 960 +2. L = 2.7SL=957.7.53 EWMA 955 _ _ X=950 950 945 -2.1. L = 2.47 1 8 16 24 32 40 48 Sample 56 64 72 80 Test Results for EWMA Chart of Ex8-7temp TEST. One point beyond control limits.53. UCL = 957.47.Chapter 8 Exercise Solutions 8-21 (8-19). LCL = 942. λ = 0. 12 and 13.16 . σˆ = 12.

λ = 0.4.24. One point beyond control limits. the process is out of control at observation 70. as compared to the chart in the Exercise 21 (with the smaller λ) which signaled out of control at earlier samples. 8-22 . MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Temperature Readings ( Ex8-7temp) lambda = 0. LCL = 931. CL = µ0 = 950.76.Chapter 8 Exercise Solutions 8-22 (8-20).4.16 . L = 3. UCL = 968.24 EWMA 960 _ _ X=950 950 940 LCL=931.76 930 1 8 16 24 32 40 48 Sample 56 64 72 80 Test Results for EWMA Chart of Ex8-7temp TEST. L = 3 970 UCL=968. Test Failed at points: 70 With the larger λ. σˆ = 12.

σˆ = 5. L = 2.70.Chapter 8 Exercise Solutions 8-23 (8-21). The process average of µˆ = 183.05.6.6SL=177.70 170 3 6 9 12 15 18 Sample 21 24 27 30 Process is out of control. L = 2.05.6SL=172. LCL = 172. MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Bath Concentrations (Ex8-8con) lambda = 0.594 is too far from the process target of µ0 = 175 for the process variability.30 _ _ X=175 175 -2. The data is grouped into three increasing levels.30.6 190 EWMA 185 180 +2. λ = 0.634 . UCL = 177. 8-23 . CL = µ0 = 175.

n.2777th.ecJa F e F e M M a A M Ma Ma J Ju Ju Ju Ju O N N D D S S Ex6-62Day In Exercise 6-62.ov o v . a properly designed EWMA chart is very robust to the assumption of normally distributed data. The EWMA chart also does not signal out of control.1. λ = 0.un n.Oc Oc O c ct.9 26 .0 7.n .5 10.ec.b.a r .0 20 23 25 r .7SL=18. L = 2.0 _ _ X=12. As mentioned in the text (Section 8.b.2 25 28 29 n.7 16 16 22 25 l -6 l.8 l.4 t.Chapter 8 Exercise Solutions 8-24☺.7SL=6.25 12.ep.Ju Ju Ju Ju l.1.7 24 28 .5 10 r .7 20.y .27 17.Sep ep .0 +2.4-3). L = 2.7 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Homicide Data (Ex6-62Bet) lambda = 0.8 19 .5 EWMA 15.5 -2. Individuals control charts of 0.9 22 24 t.n.4 . 8-24 .23 5.1 t.and 0.p ay y .25th-root transformed data showed no out-of-control signals.

68 _ _ X=3200 3200 -2.7 3205 +2. L=2.7SL=3203. λ = 0. L = 2.Chapter 8 Exercise Solutions 8-25 (8-22).1.95 (from Exercise 8-9).7SL=3196.95. sigma=5. lambda=0. µ0 = 3200. σˆ = 5.7 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Viscosity (Ex8-9vis) Target=3200. 8-25 .32 EWMA 3195 3190 3185 3180 3175 4 8 12 16 20 24 Ex5-60Meas 28 32 36 The process is out of control from the first sample.1.

σ = 25. µ0 = 1050. 15. w = 6. 13. LCL = 1019. the same result as for Exercise 8-1. 17. 8-26 . 14.Chapter 8 Exercise Solutions 8-26 (8-23).4 1000 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for Moving Average Chart of Ex8-1mole TEST. target value = 1050. UCL = 1080. One point beyond control limits. CL = 1050. 16. 11. std dev = 25 1200 Moving Average 1150 1100 UCL=1080.6 _ _ X=1050 1050 LCL=1019. 19. 18. Test Failed at points: 10. 12. 20 Process is out of control at observation 10.6.4 MTB > Stat > Control Charts > Time-Weighted Charts > Moving Average Moving Average Chart of Molecular Weight (Ex8-1mole) w = 6.

Chapter 8 Exercise Solutions

8-27 (24).

w = 5, µ0 = 8.02, σ = 0.05, CL = 8.02, UCL = 8.087, LCL = 7.953

MTB > Stat > Control Charts > Time-Weighted Charts > Moving Average

**Moving Average Chart of Can Weight (Ex8-4can)
**

w = 5, process target = 8.02, std dev = 0.05

8.20

Moving Average

8.15

8.10

UCL=8.0871

8.05

_

_

X=8.02

8.00

LCL=7.9529

7.95

7.90

2

4

6

8

10

12 14

Sample

16

18

20

22

24

The process is in control, the same result as for Exercise 8-4.

8-27

Chapter 8 Exercise Solutions

8-28☺.

w=5

MTB > Stat > Control Charts > Time-Weighted Charts > Moving Average

**Moving Average Chart of Homicide Data (Ex6-62Bet)
**

w = 5, target and std dev estimated from data

UCL=25.31

25

Moving Average

20

15

_

_

X=12.25

10

5

0

LCL=-0.81

3

6

9

12

15

Sample

18

21

24

27

**Because these plot points are an average of five observations, the nonnormality of the
**

individual observations should be of less concern. The approximate normality of the

averages is a consequence of the Central Limit Theorem.

8-28

Chapter 8 Exercise Solutions

8-29 (8-25).

Assume that t is so large that the starting value Z 0 = x has no effect.

∞

⎡ ∞

⎤

E ( Z t ) = E[λ xt + (1 − λ )( Z t −1 )] = E ⎢λ ∑ (1 − λ ) j xt − j ⎥ = λ ∑ (1 − λ ) j E ( xt − j )

j =0

⎣ j =0

⎦

∞

Since E ( xt − j ) = µ and λ ∑ (1-λ ) j = 1 , E ( Z t ) = µ

j =0

8-30 (8-26).

⎡ ∞

⎤

var( Z t ) = var ⎢ λ ∑ (1 − λ ) j xt − j ⎥

⎣ j =0

⎦

∞

⎡

⎤

= ⎢λ 2 ∑ (1 − λ ) 2 j ⎥ ⎡⎣ var( xt − j ) ⎤⎦

⎣ j =0

⎦

λ ⎛σ 2 ⎞

=

⎜

⎟

2−λ ⎝ n ⎠

8-31 (8-27).

For the EWMA chart, the steady-state control limits are x ± 3σ

λ

.

(2 − λ )n

⎛ 2 ⎞

⎜

⎟

⎝ w + 1 ⎠ = x ± 3σ 1 = x ± 3σ ,

Substituting λ = 2/(w + 1), x ± 3σ

2 ⎞

wn

⎛

wn

⎜2−

⎟n

w +1 ⎠

⎝

which are the same as the limits for the MA chart.

8-32 (8-28).

1 w−1

w −1

. In the

∑ j=

w j =0

2

EWMA, the weight given to a sample mean j periods ago is λ(1 - λ)j , so the average age

∞

1− λ

. By equating average ages:

is λ ∑ (1 − λ ) j j =

The average age of the data in a w-period moving average is

j =0

λ

1− λ

w −1

λ

2

2

λ=

w +1

=

8-29

Chapter 8 Exercise Solutions

8-33 (8-29).

For n > 1, Control limits = µ0 ±

3 ⎛ σ ⎞

3σ

⎜

⎟ = µ0 ±

w⎝ n⎠

wn

8-34 (8-30).

x chart: CL = 10, UCL = 16, LCL = 4

UCL = CL + kσ x

16 = 10 − kσ x

kσ x = 6

EWMA chart:

UCL = CL + lσ λ [(2 − λ )n]

= CL + l σ

n 0.1 (2 − 0.1) = 10 + 6(0.2294) = 11.3765

LCL = 10 − 6(0.2294) = 8.6236

8-35 (8-31).

λ = 0.4

For EWMA, steady-state limits are ± Lσ λ (2 − λ )

For Shewhart, steady-state limits are ± kσ

kσ = Lσ λ (2 − λ )

k = L 0.4 (2 − 0.4)

k = 0.5 L

8-30

Chapter 8 Exercise Solutions

8-36 (8-32).

The two alternatives to plot a CUSUM chart with transformed data are:

1. Transform the data, target (if given), and standard deviation (if given), then use

these results in the CUSUM Chart dialog box, or

2. Transform the target (if given) and standard deviation (if given), then use the

Box-Cox tab under CUSUM Options to transform the data.

The solution below uses alternative #2.

From Example 6-6, transform time-between-failures (Y) data to approximately normal

distribution with X = Y 0.2777.

TY = 700, TX = 700 0.2777 = 6.167, k = 0.5, h = 5

MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM

**CUSUM Chart of Transformed Failure Data (Ex8-37trans)
**

X = Y^0.277, target - 6.167, k = 0.5, h = 5

UCL=10.46

Cumulative Sum

10

5

0

0

-5

-10

LCL=-10.46

2

4

6

8

10

12

Sample

14

16

18

20

**A one-sided lower CUSUM is needed to detect an increase in failure rate, or equivalently
**

a decrease in the time-between-failures. Evaluate the lower CUSUM on the MINITAB

chart to assess stability.

The process is in control.

8-31

Chapter 8 Exercise Solutions

8-37 (8-33).

**µ0 = 700, h = 5, k = 0.5, estimate σ using the average moving range
**

MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM,

also CUSUM options > Estimate > Average Moving Range

**CUSUM Chart of Valve Failure Data (Ex8-37fail)
**

Target=700, h=5, k=0.5

4000

UCL=3530

3000

Cumulative Sum

2000

1000

0

0

-1000

-2000

-3000

LCL=-3530

-4000

2

4

6

8

10

12

Ex8-37No

14

16

18

20

**A one-sided lower CUSUM is needed to detect an increase in failure rate. Evaluate the
**

lower CUSUM on the MINITAB chart to assess stability.

The process is in control.

Though the data are not normal, the CUSUM works fairly well for monitoring the

process; this chart is very similar to the one constructed with the transformed data.

8-32

Chapter 8 Exercise Solutions

8-38 (8-34).

µ0 = TX = 700 0.2777 = 6.167, λ = 0.1, L = 2.7

MTB > Stat > Control Charts > Time-Weighted Charts > EWMA

**EWMA Chart of Transformed Failure Data (Ex8-37trans)
**

X = Y^0.2777, target = 6.167, lambda = 0.1, L = 2.7

7.5

+2.7SL=7.453

7.0

EWMA

6.5

_

_

X=6.167

6.0

5.5

5.0

-2.7SL=4.881

2

4

6

8

10

12

Sample

14

16

18

20

Valve failure times are in control.

8-39 (8-35).

The standard (two-sided) EWMA can be modified to form a one-sided statistic in much

the same way a CUSUM can be made into a one-sided statistic. The standard (two-sided)

EWMA is

zi = λ xi + (1 − λ ) zi−1

Assume that the center line is at µ0. Then a one-sided upper EWMA is

zi+ = max ⎡⎣ µ0 , λ xi + (1 − λ ) zi −1 ⎤⎦ ,

and the one-sided lower EWMA is

zi− = min ⎡⎣ µ0 , λ xi + (1 − λ ) zi −1 ⎤⎦ .

8-33

**Chapter 9 Exercise Solutions
**

Note: Many of the exercises in this chapter were solved using Microsoft Excel 2002, not

MINITAB. The solutions, with formulas, charts, etc., are in Chap09.xls.

9-1.

σˆ A = 2.530, nA = 15, µˆ A = 101.40

σˆ B = 2.297, nB = 9, µˆ B = 60.444

σˆ C = 1.815, nC = 18, µˆ C = 75.333

**σˆ D = 1.875, nD = 18, µˆ D = 50.111
**

Standard deviations are approximately the same, so the DNOM chart can be used.

R = 3.8, σˆ = 2.245, n = 3

x chart: CL = 0.55, UCL = 4.44, LCL = −3.34

R chart: CL = 3.8, UCL = D4 R = 2.574 (3.8) = 9.78, LCL = 0

Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Chart

**Xbar-R Chart of Measurements (Ex9-1Xi)
**

U C L=4.438

Sample M ean

4

2

_

_

X=0.55

0

-2

LC L=-3.338

-4

2

4

6

8

10

Sample

12

14

16

18

20

Sample Range

10.0

U C L=9.78

7.5

5.0

_

R=3.8

2.5

0.0

LC L=0

2

4

6

8

10

Sample

12

14

16

18

20

Process is in control, with no samples beyond the control limits or unusual plot patterns.

9-1

**Chapter 9 Exercise Solutions
**

9-2.

Since the standard deviations are not the same, use a standardized x and R charts.

Calculations for standardized values are in:

Excel : workbook Chap09.xls : worksheet : Ex9-2.

**n = 4, D3 = 0, D4 = 2.282, A2 = 0.729; RA = 19.3, RB = 44.8, RC = 278.2
**

Graph > Time Series Plot > Simple

Control Chart of Standardized Xbar (Ex9-2Xsi)

1.5

1.0

Ex9-2Xsi

+A2 = 0.729

0.5

0

0.0

-0.5

-A2 = -0.729

-1.0

Ex9-2Samp

Ex9-2Part

2

A

4

A

6

A

8

B

10

B

12

C

14

C

16

C

18

C

20

C

**Control Chart of Standardized R (Ex9-2Rsi)
**

2.5

D4 = 2.282

Ex9-2Rsi

2.0

1.5

1.006

1.0

0.5

0.0

Ex9-2Samp

Ex9-2Part

D3 = 0

2

A

4

A

6

A

8

B

10

B

12

C

14

C

16

C

18

C

20

C

Process is out of control at Sample 16 on the x chart.

9-2

j − µ0. Set up a standardized c chart for defect counts.13 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Standardized Total # of Defects (Ex9-4Zi) 3 UCL=3 Individual Value 2 1 _ X=0 0 -1 -2 -3 LCL=-3 4 8 12 16 20 24 Observation 28 32 36 40 Process is in control. c1130 = 64. Note: In the textbook. j ) / σ j (where j represents the part type) would be used to calculate each plot statistic.Chapter 9 Exercise Solutions 9-3.63 4.13 c1055 = 13. LCL = −3. c4610 = 4.63. with CL = 0. a standardized CUSUM could be used to detect smaller deviations from the target value.67 26.00. The standardized variable ( yi .25. in standard deviation units.00 12.67 50. In a short production run situation. the 4th part on Day 246 should be “1385” not “1395”. c8611 = 50. 9-3 .25 64.67. is the same for each part type. Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Rx9-4Def Rx9-4Def 1055 1130 1261 1385 4610 8611 13.67. c1385 = 26. UCL = +3. 9-4. c1261 = 12. The chart would be designed so that δ. The plot statistic is Z i = ( ci − c ) c.

023 0 2.Chapter 9 Exercise Solutions 9-5.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Sam ple Group Range Control Chart 8 Range 6 Ex9-5Rmax Ex9-5RUCL 4 Ex9-5RLCL 2 0 1 3 5 7 9 11 13 15 17 19 Sam ple There is no situation where one single head gives the maximum or minimum value of x six times in a row.00 Ex9-5Xbmax Xbar 55. not just a specific one.338 4 heads 3 units 1.596 6.000 Group Xbar Control Chart 61. so the process is out-of-control. The assignable cause affects all heads.00 Ex9-5XbLCL 49.00 45.017 0.988 2. Excel : Workbook Chap09.00 47.00 Ex9-5Xbmin 53.00 57. 9-4 .00 59.xls : Worksheet Ex9-5 Grand Avg = Avg R = s= n= A2 = D3 = D4 = Xbar UCL = Xbar LCL = R UCL = R LCL = 52.379 50.574 55.00 Ex9-5XbUCL 51. There are many values of x max and x min that are outside the control limits.

9-5 .Chapter 9 Exercise Solutions 9-6.xls : Worksheet Ex9-6 Group Control Chart for Xbar 65 60 Xbar Ex9-5Xbmax Ex9-5Xbmin 55 Ex9-5XbUCL Ex9-5XbLCL 50 45 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Sample Group Control Chart for Range 7 6 Range 5 Ex9-5Rmax 4 Ex9-5RUCL 3 Ex9-5RLCL 2 1 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Sample The last four samples from Head 4 are the maximum of all heads. Excel : Workbook Chap09. a process change may have caused output of this head to be different from the others.

6 5 Ex9-7aMRmax 4 Ex9-7aMRUCL 3 Ex9-7aMRLCL 2 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Sample See the discussion in Exercise 9-5.727 47.050 0.988 2.128 0 3. Obs. Obs.Chapter 9 Exercise Solutions 9-7.248 7. 70 Individ. 65 Ex9-7aXmax 60 Ex9-7aXmin 55 Ex9-7aXbUCL 50 Ex9-7aXbLCL 45 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Sample Group Control Chart for Moving Range 8 7 Individ.267 58.158 4 heads 2 units 1.000 Group Control Chart for Individual Obs. (a) Excel : Workbook Chap09.xls : Worksheet Ex9-7A Grand Avg = Avg MR = s= n= d2 = D3 = D4 = Xbar UCL = Xbar LCL = R UCL = R LCL = 52. 9-6 .

248 7.267 58.050 0.988 2. Obs. 60 Ex9-7bXmax Ex9-7bXmin 55 Ex9-7bXbUCL Ex9-7bXbLCL 50 45 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Sample Group Control Chart for Moving Range 8 7 MR 6 5 Ex9-7bMRmax 4 Ex9-7bMRUCL 3 Ex9-7bMRLCL 2 1 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Sample The last four samples from Head 4 remain the maximum of all heads.727 47.000 Group Control Chart for Individual Obs.xls : Worksheet Ex9-7b Grand Avg = Avg MR = s= n= d2 = D3 = D4 = Xbar UCL = Xbar LCL = R UCL = R LCL = 52.158 4 heads 2 units 1.Chapter 9 Exercise Solutions 9-7 continued (b) Excel : Workbook Chap09. indicating a potential process change.128 0 3. 65 Individ. 9-7 .

948 2 1 0 LC L=0 2 4 6 8 10 Sample 12 14 16 18 20 Failure to recognize the multiple stream nature of the process had led to control charts that fail to identify the out-of-control conditions in this process.Chapter 9 Exercise Solutions 9-7 continued (c) Stat > Control Charts > Variables Charts for Subgroups > Xbar-S Chart Note: Use “Sbar” as the method for estimating standard deviation.988 52 50 LC L=49. 9-8 . Ex9-7X4) U C L=56. .816 2 4 6 8 10 Sample 12 14 16 18 20 U C L=4. Xbar-S Chart of Head Measurements (Ex9-7X1....159 Sample M ean 56 54 _ _ X=52.415 Sample StDev 4 3 _ S =1.

Xbar-S Chart of Head Measurements (Ex9-7X1.Chapter 9 Exercise Solutions 9-7 continued (d) Stat > Control Charts > Variables Charts for Subgroups > Xbar-S Chart Note: Use “Sbar” as the method for estimating standard deviation.0 LC L=0 3 6 9 12 15 Sample 18 21 24 27 30 Test Results for S Chart of Ex9-7X1.. 9-9 .816 3 6 9 12 15 Sample 18 21 24 Sample StDev 4. Test Failed at points: 27.415 3..6 2.4 _ S =1.988 52 50 LC L=49. 29 Only the S chart gives any indication of out-of-control process.2 0. Ex9-7X4) U C L=56. Ex9-7X4 TEST 1..8 27 1 30 1 U C L=4. One point more than 3.948 1. ..00 standard deviations from center line.159 Sample M ean 56 54 _ _ X=52.. ..

Ex9-8R Variable Ex9-8Xbar Ex9-8R Mean 0. 9-10 .004 0.00227 / 2.003 _ R=0.83 PCR Stat > Control Charts > Variables Charts for Subgroups > R Chart R Chart of Range Values ( Ex9-8R.00227.552 − 0.55025 0. Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Ex9-8Xbar.Chapter 9 Exercise Solutions 9-8. .004800 Sample Range 0.002270 n=5 x = 0.. σˆ = R / d 2 = 0.55025.000 2 4 6 8 10 12 Sample 14 16 18 20 The process variability.00227 0.548 ) [6(0.000976)] = 6.005 UCL=0. Ex9-8Rdum4) 0.326 = 0.. as shown on the R chart is in control..000976 n = ( USL-LSL ) 6σˆ = ( 0. R = 0.001 LCL=0 0.002 0.

Chapter 9 Exercise Solutions

9-8 continued

(a)

3-sigma limits:

δ = 0.01, Zδ = Z 0.01 = 2.33

(

LCL = LSL + ( Z

UCL = USL − Zδ − 3

δ

−3

)

(

n ) σˆ = (0.550 − 0.020) + ( 2.33 − 3

n σˆ = (0.550 + 0.020) − 2.33 − 3

)

20 ) (0.000976) = 0.5316

20 (0.000976) = 0.5684

Graph > Time Series Plot > Simple

**Note: Reference lines have been used set to the control limit values.
**

Control Chart of Xbar Values (Ex9-8Xbar)

0.57

UCL = 0.5684

Ex9-8Xbar

0.56

0.55

0.54

LCL = 0.5316

0.53

2

4

6

8

10

12

Ex9-8Samp

14

16

18

20

**The process mean falls within the limits that define 1% fraction nonconforming.
**

Notice that the control chart does not have a centerline. Since this type of control scheme

allows the process mean to vary over the interval—with the assumption that the overall

process performance is not appreciably affected—a centerline is not needed.

9-11

Chapter 9 Exercise Solutions

9-8 continued

(b)

γ = 0.01, Zγ = Z 0.01 = 2.33

1 − β = 0.90, Z β = z0.10 = 1.28

(

LCL = LSL + ( Z

UCL = USL − Zγ + Z β

γ

+ Zβ

)

(

n ) σˆ = (0.550 − 0.020) + ( 2.33 + 1.28

n σˆ = (0.550 + 0.020) − 2.33 + 1.28

)

20 ) (0.000976) = 0.5326

20 (0.000976) = 0.5674

**Chart control limits for part (b) are slightly narrower than for part (a).
**

Graph > Time Series Plot > Simple

**Note: Reference lines have been used set to the control limit values.
**

Control Chart of Xbar Values (Ex9-8Xbar)

0.57

UCL = 0.5674

Ex9-8Xbar

0.56

0.55

0.54

LCL = 0.5326

0.53

2

4

6

8

10

12

Ex9-8Samp

14

16

18

20

**The process mean falls within the limits defined by 0.90 probability of detecting a 1%
**

fraction nonconforming.

9-12

Chapter 9 Exercise Solutions

9-9.

(a)

3-sigma limits:

n = 5, δ = 0.001, Zδ = Z 0.001 = 3.090

USL = 40 + 8 = 48, LSL = 40 − 8 = 32

(

UCL = USL − Zδ − 3

(

)

(

n σ = 48 − 3.090 − 3

)

(

n σ = 32 + 3.090 − 3

LCL = LSL+ Zδ − 3

)

5 (2.0) = 44.503

)

5 (2.0) = 35.497

Graph > Time Series Plot > Simple

**Note: Reference lines have been used set to the control limit values.
**

Modified Control Chart of Xbar Values (Ex9-9Xbar)

3-sigma Control Limits

45.0

UCL = 44.5

Ex9-9Xbar

42.5

40

40.0

37.5

LCL = 35.5

35.0

2

4

6

8

10

12

Ex9-9Samp

14

16

18

20

Process is out of control at sample #6.

9-13

Chapter 9 Exercise Solutions

9-9 continued

(b)

2-sigma limits:

UCL = USL − Zδ − 2

(

(

LCL = LSL+ Zδ − 2

)

(

n σ = 48 − 3.090 − 2

)

(

n σ = 32 + 3.090 − 2

)

5 (2.0) = 43.609

)

5 (2.0) = 36.391

Graph > Time Series Plot > Simple

**Note: Reference lines have been used set to the control limit values.
**

Modified Control Chart of Xbar Values (Ex9-9Xbar)

2-sigma Control Limits

45

44

UCL = 43.61

43

Ex9-9Xbar

42

41

40

40

39

38

37

LCL = 36.39

36

2

4

6

8

10

12

Ex9-9Samp

14

16

18

20

**With 3-sigma limits, sample #6 exceeds the UCL, while with 2-sigma limits both samples
**

#6 and #10 exceed the UCL.

9-14

Chapter 9 Exercise Solutions

9-9 continued

(c)

γ = 0.05, Zγ = Z 0.05 = 1.645

1 − β = 0.95, Z β = Z 0.05 = 1.645

(

LCL = LSL + ( Z

UCL = USL − Zγ + zβ

γ

+ zβ

5 ) (2.0) = 43.239

)

(

n ) σ = 32 + (1.645 + 1.645 5 ) (2.0) = 36.761

n σ = 48 − 1.645 + 1.645

Graph > Time Series Plot > Simple

**Note: Reference lines have been used set to the control limit values.
**

Acceptance Control Chart of Xbar Values (Ex9-9Xbar)

45

44

UCL = 43.24

43

Ex9-9Xbar

42

41

40

40

39

38

37

LCL = 36.76

36

2

4

6

8

10

12

Ex9-9Samp

14

16

18

20

Sample #18 also signals an out-of-control condition.

9-15

Chapter 9 Exercise Solutions

9-10.

Design an acceptance control chart.

Accept in-control fraction nonconforming = 0.1% → δ = 0.001, Zδ = Z 0.001 = 3.090

with probability 1 − α = 0.95 → α = 0.05, Zα = Z 0.05 = 1.645

Reject at fraction nonconforming = 2% → γ = 0.02, Zγ = Z 0.02 = 2.054

with probability 1 − β = 0.90 → β = 0.10, Z β = Z 0.10 = 1.282

2

⎛ Z + Z β ⎞ ⎛ 1.645 + 1.282 ⎞ 2

n=⎜ α

=

= 7.98 ≈ 8

⎜ Z − Z ⎟⎟ ⎜⎝ 3.090 − 2.054 ⎟⎠

δ

γ

⎝

⎠

(

LCL = LSL + ( Z

UCL = USL − Zγ + Z β

γ

+ Zβ

8 ) σ = USL − 2.507σ

)

(

n ) σ = LSL + ( 2.054 + 1.282 8 ) σ = LSL + 2.507σ

n σ = USL − 2.054 + 1.282

9-16

Chapter 9 Exercise Solutions

9-11.

µ = 0, σ = 1.0, n = 5, δ = 0.00135, Zδ = Z0.00135 = 3.00

For 3-sigma limits, Zα = 3

(

UCL = USL − zδ − zα

)

(

n σ = USL − 3.000 − 3

)

5 (1.0) = USL − 1.658

⎛ USL − 1.658 − µ0 ⎞

⎛ UCL − µ0 ⎞

Pr{Accept} = Pr{x < UCL} = Φ ⎜

=

Φ

⎜

⎟ = Φ (∆ − 1.658) 5

⎟

⎜

⎟

σ

n

1.0

5

⎝

⎠

⎝

⎠

where ∆ = USL − µ0

(

(

For 2-sigma limits, Zα = 2 ⇒ Pr{Accept} = Φ (∆ − 2.106) 5

)

)

⎛ USL − µ0 ⎞

p = Pr{x > USL} = 1 − Pr{x ≤ USL} = 1 − Φ ⎜

⎟ = 1 − Φ (∆ )

σ

⎝

⎠

Excel : Workbook Chap09.xls : Worksheet Ex9-11

DELTA=USL-mu0

3.50

3.25

3.00

2.50

2.25

2.00

1.75

1.50

1.00

0.50

0.25

0.00

CumNorm(DELTA)

0.9998

0.9994

0.9987

0.9938

0.9878

0.9772

0.9599

0.9332

0.8413

0.6915

0.5987

0.5000

p

0.0002

0.0006

0.0013

0.0062

0.0122

0.0228

0.0401

0.0668

0.1587

0.3085

0.4013

0.5000

Pr(Accept@3)

1.0000

0.9998

0.9987

0.9701

0.9072

0.7778

0.5815

0.3619

0.0706

0.0048

0.0008

0.0001

Pr(Accept@2)

0.9991

0.9947

0.9772

0.8108

0.6263

0.4063

0.2130

0.0877

0.0067

0.0002

0.0000

0.0000

Operating Curves

Pr{Acceptance}

1.0000

0.8000

0.6000

0.4000

0.2000

0.0000

0.0000

0.1000

0.2000

0.3000

0.4000

0.5000

Fraction Defective, p

Pr(Accept@3)

Pr(Accept@2)

9-17

Chapter 9 Exercise Solutions

9-12.

Design a modified control chart.

n = 8, USL = 8.01, LSL = 7.99, S = 0.001

δ = 0.00135, Zδ = Z0.00135 = 3.000

For 3-sigma control limits, Zα = 3

(

UCL = USL − Zδ − Zα

(

LCL = LSL+ Zδ − Zα

)

(

n σ = 8.01 − 3.000 − 3

)

(

n σ = 7.99 + 3.000 − 3

)

8 (0.001) = 8.008

)

8 (0.001) = 7.992

9-13.

Design a modified control chart.

n = 4, USL = 70, LSL = 30, S = 4

δ = 0.01, Zδ = 2.326

1 − α = 0.995, α = 0.005, Zα = 2.576

(

LCL = LSL + ( Z

UCL = USL − Zδ − Zα

δ

− Zα

4 ) (4) = 65.848

)

(

n ) σ = (50 − 20) + ( 2.326 − 2.576 4 ) (4) = 34.152

n σ = (50 + 20) − 2.326 − 2.576

9-14.

Design a modified control chart.

n = 4, USL = 820, LSL = 780, S = 4

δ = 0.01, Zδ = 2.326

1 − α = 0.90, α = 0.10, Zα = 1.282

(

LCL = LSL + ( Z

UCL = USL − Zδ − Zα

δ

− Zα

4 ) (4) = 813.26

)

(

n ) σ = (800 − 20) + ( 2.326 − 1.282 4 ) (4) = 786.74

n σ = (800 + 20) − 2.326 − 1.282

9-18

Chapter 9 Exercise Solutions

9-15.

n = 4, R = 8.236, x = 620.00

(a)

σˆ x = R d 2 = 8.236 2.059 = 4.000

(b)

pˆ = Pr{x < LSL} + Pr{x > USL}

= Pr{x < 595} + [1 − Pr{x ≤ 625}]

⎛ 595 − 620 ⎞ ⎡

⎛ 625 − 620 ⎞ ⎤

= Φ⎜

⎟ + ⎢1 − Φ ⎜

⎟⎥

⎝ 4.000 ⎠ ⎣

⎝ 4.000 ⎠ ⎦

= 0.0000 + [1 − 0.8944]

= 0.1056

(c)

δ = 0.005, Zδ = Z 0.005 = 2.576

α = 0.01, Zα = Z 0.01 = 2.326

(

LCL = LSL + ( Z

UCL = USL − Zδ − Zα

δ

− Zα

4 ) 4 = 619.35

)

(

n ) σ = 595 + ( 2.576 − 2.326 4 ) 4 = 600.65

n σ = 625 − 2.576 − 2.326

9-19

4 0.072562 -0.2 -0.6 0.0 2 4 6 8 10 Lag 12 14 16 18 Autocorrelation Function: Ex9-16mole Lag 1 2 3 4 5 ACF 0.0 -0.0 0.2 0.4 -0.0 Partial Autocorrelation 0.30 0.96 -0. Note: In the textbook.110802 -0.16 49.4 0. AR(1).84 48.29 -0.92 0.23 LBQ 33.6 0.055640 T 5.2 0.70 -0. 9-20 .0 2 4 6 8 10 Lag 12 14 16 18 Partial Autocorrelation Function: Ex9-16mole Lag 1 2 3 4 5 PACF 0.6 -0.0 -0.29 … Stat > Time Series > Partial Autocorrelation rtial Autocorrelation Function for Molecular Weight Measurements (Ex9-16mo (with 5% significance limits for the partial autocorrelations) 1.37 1.8 0.42 -0. the 5th row should be “2000” not “2006”.8 Autocorrelation 0.6 -0.658253 0.658253 -0.70 2.4 -0.039599 T 5.48 … The decaying sine wave of the ACFs combined with a spike at lag 1 for the PACFs suggests an autoregressive process of order 1.Chapter 9 Exercise Solutions 9-16.81 44.033132 -0.220536 0.2 -0.105969 0.8 -1. the 5th column. (a) Stat > Time Series > Autocorrelation Autocorrelation Function for Molecular Weight Measurements (Ex9-16mole) (with 5% significance limits for the autocorrelations) 1.373245 0.74 49.8 -1.

violating many runs tests. 13. 9. 13. 9-21 . 70 6. 13. 69 2. 14. 10. LCL = 1953 σˆ = MR d 2 = 17. Failed at points: 12. 12.Chapter 9 Exercise Solutions 9-16 continued (b) x chart: CL = 2001. 6 points in a row all increasing or all decreasing. 7.97 1. 8 points in a row more than 1 standard deviation from center line (above and below CL).00 standard deviations from center line. with big swings and very few observations actually near the mean.1 1 1 14 21 28 35 42 49 Observation 56 63 70 Test Results for I Chart of Ex9-16mole TEST Test TEST Test TEST Test TEST Test TEST Test TEST Test 1. 32. Failed at points: 12. 12. 34. 35. 14. 11. 37 The process is out of control on the x chart. 32. One point more than 3. 15. Failed at points: 8. 40. Failed at points: 7. 16. 15. 35. 37 8. 13. Failed at points: 7. Failed at points: 6. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 36. 15 3. 12. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 36. 8. 11. 8. 14. 31. 14. UCL = 2049.9 2000 6 1975 6 2 1950 2 5 2 1 1 1 1 7 1 1 LCL=1953.7 8 6 6 6 6 6 2025 Individual Value 3 _ X=2000. 53 5. 9 points in a row on same side of center line.93 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Molecular Weight Measurements (Ex9-16mole) 1 1 2050 UCL=2048.128 = 15. 33.

0 50 Individual Value 25 _ X=-0.4 -75 1 7 14 21 28 35 42 49 Observation 56 63 70 Test Results for I Chart of Ex9-16res TEST 1.1 0.4 0.0 0.942 1 41717.9 0.000 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Residuals from Molecular Weight Model (Ex9-16res) 1 UCL=58.843 2 35687.693 4 30929.196 2.21 7. One point more than 3. 9-22 .756 3 32083.000 0.0852 8.693 613.6979 0.00 standard deviations from center line.1 0.400 1200.0010 Final Estimates of Parameters Type Coef SE Coef T AR 1 0.956 7 30897.364 256.697 606.675 650.02 Mean 2003.3 0.6 0.250 1500.698 605.Chapter 9 Exercise Solutions 9-16 continued (c) Stat > Time Series > ARIMA ARIMA Model: Ex9-16mole Estimates at each iteration Iteration SSE Parameters 0 50173.7 0.550 900.1 0.7 0 -25 -50 LCL=-59.19 Constant 605. There are no other violations of special cause tests.100 1800.998 6 30897.698 605.494 8 30897. Test Failed at points: 16 Observation 16 signals out of control above the upper limit.196 Relative change in each estimate less than 0.82… P 0.197 5 30898.

5. δ = 1 sigma.9 -100 1 7 14 21 28 35 42 Sample 49 56 63 70 No observations exceed the control limit. h = 5. Let µ0 = 0.9 Cumulative Sum 50 0 0 -50 LCL=-97. h = 5 100 UCL=97. k = 0.5. 9-23 . The residuals are in control.Chapter 9 Exercise Solutions 9-17. k = 0. Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Residuals from Molecular Weight Model (Ex9-16res) mu0 = 0.

1. Let λ = 0.Chapter 9 Exercise Solutions 9-18.83 -15 1 7 14 21 28 35 42 Sample 49 56 63 70 Process is in control.42 10 EWMA 5 _ _ X=-0.7 (approximately the same as a CUSUM with k = 0.7SL=11. Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Residuals from Molecular Weight Model (Ex9-16res) lambda = 0. 9-24 .71 0 -5 -10 -2.5 and h = 5). L = 2.7 +2.1 and L = 2.7SL=-12.

Stat > Time Series > ARIMA ARIMA Model: Ex9-16mole … Final Estimates of Parameters Type Coef SE Coef T MA 1 0.057 1982.1)).128 1940.09 … P 0.608 2016.212 No 1890.930 λ = 1 – MA1 = 1 – 0.772 No 1907.842 2048.217 2074.947 2044.780 No 1899.282 1964.xls : Worksheet Ex9-19 t xt 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 2048 2025 2017 1995 1983 1943 1940 1947 1972 1983 1935 1948 1966 1954 1970 2039 zt 2000.000 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 Obs.255 No 1934.1) (= EWMA = IMA(1.128 1940.731 1984.211 2.467 1946.046 1946.644 LCL OOC? 1953.574 1954.393 -0.057 1982.269 1994.479 2017.613 No 1978.524 2044.Chapter 9 Exercise Solutions 9-19.046 1946.040 above UCL … Xt.000 2000.000 1750.65 Constant -0.816 1986.0762 0.282 1964. CL UCL LCL xt Observation 6 exceeds the upper control limit compared to one out-of-control signal at observation 16 on the Individuals control chart.93 Excel : Workbook Chap09.9238 σˆ = MR d 2 = 17.145 No 1996.521 0.304 2017.281 2065.244 No 1898.000 1850.502 1970.326 No 1892.000 1950.929 No 1936.665 No 1898.1.859 2029.677 No 1969.749 2092.842 2033.806 1968.0762 = 0.731 1984.700 No 1922.000 1800.97 1.848 1993. Molecular Weight EWMA Moving Center-Line Control Chart for Molecular Weight 2150.479 2017.480 No 1916.014 1938.574 1954.000 2050.947 2044.128 = 15.004 No 1921.000 1900.533 2031.920 No 1948. 9-25 .806 1968.415 2026.582 1947.000 2100.014 1938. fit an ARIMA (0.415 2026.722 1996.722 1996.376 2002. To find the optimal λ.1181 0.467 1946.084 2012.502 1970.930 1988.384 1995. No.582 1947.654 CL UCL 2000.

12 … Stat > Time Series > Partial Autocorrelation Partial Autocorrelation Function for Concentration Readings (Ex9-20conc) (with 5% significance limits for the partial autocorrelations) 1.4 -0.0 -0.8 Autocorrelation 0.36 99.05 2.Chapter 9 Exercise Solutions 9-20 (a) Stat > Time Series > Autocorrelation Autocorrelation Function for Concentration Readings (Ex9-20conc) (with 5% significance limits for the autocorrelations) 1.6 -0.6 -0.0 -0.23 LBQ 57.238198 T 7.746174 0.8 0.8 -1.0 2 4 6 8 10 12 14 16 18 20 22 24 Lag Autocorrelation Function: Ex9-20conc Lag 1 2 3 4 5 ACF 0.6 0.520417 0.390108 0.58 … The decaying sine wave of the ACFs combined with a spike at lag 1 for the PACFs suggests an autoregressive process of order 1.77 -0.635375 0.4 0. AR(1).95 -1.46 4.2 0.37 3.4 0.38 127.8 -1.0 2 4 6 8 10 12 14 16 18 20 22 24 Lag Partial Autocorrelation Function: Ex9-20conc Lag 1 2 3 4 5 PACF 0.46 1.177336 -0.03 150. 9-26 .04 -0.10 1.2 -0.746174 0.0 0.004498 -0.6 0.095134 -0.158358 T 7.4 -0.2 0.86 144.2 -0.0 Partial Autocorrelation 0.

34 1 20 _ X=200. 37. 29. 89. 44. 39. 41. 10. 86. 97. 71. 34. 12. 18.68 6 200 2 195 6 62 5 190 1 1 1 10 6 2 22 2 2 2 2 2 2 2 55 2 6 66 LCL=190. 95 TEST 6. 13. 34. 22. with big swings and very few observations actually near the mean. 41. 93. 42. 43. 44 The process is out of control on the x chart. 9-27 . 17. 40. 16. 66. 87. 88. Test Failed at points: 8. 95. 40. 96. 39. 21. 86. 88.01 11 1 11 30 40 50 60 Observation 70 80 90 100 Test Results for I Chart of Ex9-20conc TEST 1. 21. 93. 40. Test Failed at points: 11. 42. 36. 29. 22. 39. 9 points in a row on same side of center line. 41. 69. 44. 23. 41. One point more than 3. 66. 30. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 37. 12. Test Failed at points: 15. 21. 15. 38. 94. 43. 8 points in a row more than 1 standard deviation from center line (above and below CL). 88. 95 TEST 2. 68. 69.128 = 3. 38.00 standard deviations from center line. 37. 36. violating many runs tests. 100 TEST 5. 72. 38. 14. 42. Test Failed at points: 15. 89. 73. 89. 28.227 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Concentration Readings (Ex9-20conc) 215 1 1 1 1 210 5 Individual Value 5 6 205 1 1 22 52 2 11 UCL=209. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 20. 42. 43. 65.64 1. 94. Test Failed at points: 10.Chapter 9 Exercise Solutions 9-20 continued (b) σˆ = MR d 2 = 3. 99 TEST 8. 68. 99. 19. 98. 94. 36. 43.

Test Failed at points: 29 Observation 29 signals out of control for test 4.76 Mean 200.122 1.4155 120. however this is not unlikely for a dataset of 100 observations.Chapter 9 Exercise Solutions 9-20 continued (c) Stat > Time Series > ARIMA ARIMA Model: Ex9-20conc … Final Estimates of Parameters Type Coef SE Coef T AR 1 0.7493 0.05 0 -5 -10 LCL=-13.62 Individual Value 10 5 4 _ X=-0. Consider the process to be in control. 9-28 .000 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Residuals from Concentration Model (Ex9-20res) 15 UCL=13.000 0.73 -15 1 10 20 30 40 50 60 Observation 70 80 90 100 Test Results for I Chart of Ex9-20res TEST 4.657 … P 0. 14 points in a row alternating up and down.20 Constant 50.1734 0.0669 11.

0 -0.2 0.0 -0.8 Autocorrelation 0.6 0.4 0.2 -0.6 0.0 0.8 -1.0 Partial Autocorrelation 0.4 -0.0 2 4 6 8 10 12 14 Lag 16 18 20 22 24 Stat > Time Series > Partial Autocorrelation tial Autocorrelation Function for Residuals from Concentration Model (Ex9-20r (with 5% significance limits for the partial autocorrelations) 1.8 -1.0 2 4 6 8 10 12 14 Lag 16 18 20 22 24 9-29 .6 -0.4 -0.2 -0.8 0.4 0.Chapter 9 Exercise Solutions 9-20 continued (d) Stat > Time Series > Autocorrelation Autocorrelation Function for Residuals from Concentration Model (Ex9-20res) (with 5% significance limits for the autocorrelations) 1.6 -0.2 0.

9 Mean StDev N AD P-Value 99 95 Percent 90 -0.407 0.Chapter 9 Exercise Solutions 9-20 (d) continued Stat > Basic Statistics > Normality Test Probability Plot of Residuals from Concentration Model (Ex9-20res) Normal 99.05075 4.133 100 0.1 -15 -10 -5 0 Ex9-20res 5 10 Visual examination of the ACF. PACF and normal probability plot indicates that the residuals are normal and uncorrelated.343 80 70 60 50 40 30 20 10 5 1 0. 9-30 .

k = 0.79 1 10 20 30 40 50 60 Sample 70 80 90 100 No observations exceed the control limit. The residuals are in control.Chapter 9 Exercise Solutions 9-21.5. δ = 1 sigma. k = 0. 9-31 . h = 5. Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Residuals from Concentration Model (Ex9-20res) mu0 = 0. and the AR(1) model for concentration should be a good fit.5.79 Cumulative Sum 20 10 0 0 -10 -20 LCL=-22. Let µ0 = 0. h = 5 UCL=22.

Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Residuals from Concentration Model (Ex9-20res) lambda = 0.7 (approximately the same as a CUSUM with k = 0. The residuals are in control. 9-32 . Let λ = 0.5 and h = 5).7SL=2.7 3 +2.7SL=-2.051 0 -1 -2 -2.1 and L = 2. L = 2.Chapter 9 Exercise Solutions 9-22.874 -3 1 10 20 30 40 50 60 Sample 70 80 90 100 No observations exceed the control limit.773 2 EWMA 1 _ _ X=-0.1.

366 201.003 0.660 191.144 192.15 … P 0.239 198.64 1.128 = 3. xt CL UCL = LCL = The control chart of concentration data signals out of control at three observations (8.558 188.243 201.23 CL 200.813 198.7055 σˆ = MR d 2 = 3.0975 3.920 208.924 211.182 0 0 0 0 0 0 0 0 0 … Xt.139 193.706 sigma^ = t xt 0 1 2 3 4 5 6 7 8 9 10 204 202 201 202 197 201 198 188 195 189 zt 200.xls : Worksheet Ex9-23 lamda = 0.825 202.737 190.691 212.341 200.243 201. Concentration EWMA Moving Center-Line Chart for Concentration 230 220 210 200 190 180 170 160 150 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 Obs.418 200.132 188.1.418 200.863 UCL = LCL = OOC? 209.660 191.010 202.1)). To find the optimal λ.329 193. fit an ARIMA (0. 90).979 below LCL 181. 56.139 193.239 198.432 3.2945 0. Stat > Time Series > ARIMA ARIMA Model: Ex9-20conc … Final Estimates of Parameters Type Coef SE Coef T MA 1 0.506 211.Chapter 9 Exercise Solutions 9-23.010 202.2945 = 0.3034 -0.099 209.02 Constant -0.820 203.047 211.562 191. No.0452 0.863 190.882 λ = 1 – MA1 = 1 – 0. 9-33 .825 202.458 184.544 190.1) (= EWMA = IMA(1.494 208.685 192.366 201.227 Excel : Workbook Chap09.813 198.

66 4.6 0.865899 -0.78 -1. (a) Stat > Time Series > Autocorrelation Autocorrelation Function for Temperature Measurements (Ex9-24temp) (with 5% significance limits for the autocorrelations) 1.047106 -0.489422 0.4 0.13… Slow decay of ACFs with sinusoidal wave indicates autoregressive process.4 -0.4 0.25 133.31 211.67 3.592580 0.8 -1.2 0.31… Stat > Time Series > Partial Autocorrelation Partial Autocorrelation Function for Temperature Measurements (Ex9-24temp) (with 5% significance limits for the partial autocorrelations) 1.0 2 4 6 8 10 12 14 16 18 20 22 24 Lag Autocorrelation Function: Ex9-24temp Lag 1 2 3 4 5 ACF 0.71 LBQ 77.6 -0.8 -1.078040 -0.94 170.0 0.737994 0. 9-34 .86 196.36 1.373763 T 8.13 2.0 -0.0 -0.112785 T 8.2 -0.66 -0.47 -1.6 0.Chapter 9 Exercise Solutions 9-24. PACF graph suggest order 1.8 Autocorrelation 0.43 0.4 -0.143236 0.2 0.2 -0.0 2 4 6 8 10 12 14 16 18 20 22 24 Lag Partial Autocorrelation Function: Ex9-24temp Lag 1 2 3 4 5 PACF 0.6 -0.865899 0.0 Partial Autocorrelation 0.8 0.

25. The temperature measurements appear to wander over time. 36. 38. 21. 20. 22. 26. 25. 19. 4. 22. 27. 37. 27. Test Failed at points: 2.23 5 11 1 1 1 1 1 1 1 470 1 1 10 20 30 40 50 60 Observation 70 80 90 1 100 Test Results for I Chart of Ex9-24temp TEST 1. … Process is out of control. 3. 22. 16. 3.00 standard deviations from center line. 8 points in a row more than 1 standard deviation from center line (above and below CL). 22. 23. 5. 71 TEST 5. 9-35 . 21. 6. One point more than 3. 21. … TEST 3. 23. 23. 19.52 6 6 5 6 66 5 52 2 2 2 1 1 1 11 1 5 11 1 LCL=491. 23. 18. 39. 21. 18. Test Failed at points: 65. 18. 18. 9 points in a row on same side of center line. 21. violating many of the tests for special causes. 24. 23. Test Failed at points: 4. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 17. Test Failed at points: 1.Chapter 9 Exercise Solutions 9-24 continued (b) Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Temperaure Measurements (Ex9-24temp) 540 1 1 11 530 11 1 1 Individual Value 520 52 2 510 480 2 2 2 8 1 1 11 1 1 1 6 11 1 UCL=521. 17. 28. 20. Test Failed at points: 17. … TEST 2. 24. 24. 26.81 22 22 2 2 500 490 5 8 2 1 _ X=506. … TEST 6. 20. … TEST 8. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 19. 24. 19. 16. 6 points in a row all increasing or all decreasing. Test Failed at points: 20. 2. 24. 34. 33. 32. 22.

000 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Residuals from Temperature Model (Ex9-24res) UCL=22.7263 72.67 Constant 52. 9-36 . The residuals do not exhibit cycles in the original temperature readings. and points are distributed between the control limits.727 6. Test Failed at points: 94 TEST 5. The chemical process is in control.3794 0.23 20 5 Individual Value 10 _ X=0.000 0.80 1 1 10 20 30 40 50 60 Observation 70 80 90 100 Test Results for I Chart of Ex9-24res TEST 1. Test Failed at points: 71 Observation 94 signals out of control above the upper limit.985 … P 0.0480 18. One point more than 3. and observation 71 fails Test 5.Chapter 9 Exercise Solutions 9-24 continued (c) Stat > Time Series > ARIMA ARIMA Model: Ex9-24temp … Final Estimates of Parameters Type Coef SE Coef T AR 1 0.8960 0.22 0 -10 -20 LCL=-21. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).12 Mean 503.00 standard deviations from center line.

69 30 Cumulative Sum 20 10 0 0 -10 -20 -30 LCL=-36.5. h = 5 40 UCL=36. 9-37 .Chapter 9 Exercise Solutions 9-25.69 -40 1 10 20 30 40 50 60 Sample 70 80 90 100 No observations exceed the control limits. The residuals are in control. MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Residuals from Temperature Model (Ex9-24res) k = 0. This is the same conclusion as applying an Individuals control chart to the model residuals. indicating the process is in control.

indicating the process is in control.0 1 10 20 30 40 50 60 Sample 70 80 90 100 No observations exceed the control limits.Chapter 9 Exercise Solutions 9-26.33 -5. This is the same conclusion as applying the Individuals and CUSUM control charts to the model residuals.1.0 +2.7SL=4.76 EWMA 2.22 0. L = 2. 9-38 .5 _ _ X=0. The residuals are in control.7 5.5 -2.7SL=-4.0 -2. MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Residuals from Temperature Model (Ex9-24res) lambda = 0.

098 UCL 506.429 494.465 509. 94).75 1.070 487.0794 0.438 0.558 498.679 491.921 sigma^ = zt 506.xls : Worksheet Ex9-27 lambda = t xt 0 1 2 3 4 5 6 7 8 9 10 491 482 490 495 499 499 507 503 510 509 0.813 507.758 OOC? 491.105 504.265 521.812 489. Temperature 570 550 530 510 490 470 450 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 Sample No.9206 σˆ = MR d 2 = 5.647 498. 58.10 … P 0. 69) and the lower limit (1.812 489.465 LCL 521.917 λ = 1 – MA1 = 1 – 0.0794 = 0.558 498.974 494.173 0 0 0 0 0 0 0 0 0 EWMA Moving Center-Line Chart for Temperature Xt.265 483.940 514.647 498.232 482.232 482.267 509. fit an ARIMA (0.Chapter 9 Exercise Solutions 9-27. 9-39 .128 = 5.137 479.972 506.355 483. 94).363 503. xt CL UCL LCL A few observations exceed the upper limit (46.429 494.1)).850 513.525 498.78 Constant -0.267 509.1. similar to the two out-of-control signals on the Individuals control chart (71.972 506.227 below LCL 476. To find the optimal λ.722 509.1019 0.75) Excel : Workbook Chap09. Stat > Time Series > ARIMA ARIMA Model: Ex9-24temp … Final Estimates of Parameters Type Coef SE Coef T MA 1 0.1) (= EWMA = IMA(1.363 503.655 518.6784 -0.560 524.520 474.940 467.520 492.0975 (from a Moving Range chart with CL = 5.0711 0.520 492.037 CL 5.

(c) If assignable causes are present. adjacent observations will tend to be similar. it is generally a bad practice to use S2 to estimate σ2.Chapter 9 Exercise Solutions 9-28. This would tend to produce an estimate of the process standard deviation that is too small. therefore making the moving ranges smaller. (b) S2 is still an unbiased estimator of σ2 when the data are positively autocorrelated. it is not good practice to estimate σ2 from S2. Since it is difficult to determine whether a process generating autocorrelated data – or really any process – is in control. There is nothing in the derivation of the expected value of S2 = σ2 that depends on an assumption of independence. 9-40 . (a) When the data are positively autocorrelated.

54 LBQ 25.0 -0.41 -2.41 32. (a) Stat > Time Series > Autocorrelation Autocorrelation Function for Viscosity Readings (Ex9-29Vis) (with 5% significance limits for the autocorrelations) 1.494137 -0.16 25.49.85 r1 = 0. There is a serious problem with autocorrelation in viscosity readings.29 41.Chapter 9 Exercise Solutions 9-29. 9-41 .049610 -0.264612 -0.071963 T 4.0 0.94 -0.283150 -0.6 0.4 -0.8 Autocorrelation 0.4 0.2 -0.2 0.22 -0.6 -0. indicating a strong positive correlation at lag 1.8 -1.78 41.17 -2.0 2 4 6 8 10 12 14 Lag 16 18 20 22 24 Autocorrelation Function: Ex9-29Vis Lag 1 2 3 4 5 … ACF 0.

Test Failed at points: 40. 86. The viscosity measurements appear to wander over time. Test Failed at points: 64 Process is out of control. 15 points within 1 standard deviation of center line (above and below CL).11 5 Individual Value 35 6 6 7 30 _ X=28.00 standard deviations from center line. Test Failed at points: 22 TEST 8. Test Failed at points: 38. 75 TEST 7. 58.Chapter 9 Exercise Solutions 9-29 continued (b) Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Viscosity (Ex9-29Vis) 40 UCL=37. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 64. 92 TEST 5. 38. 86 TEST 6. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 9-42 . 59.57 25 6 20 6 55 LCL=20. Test Failed at points: 2. 8 points in a row more than 1 standard deviation from center line (above and below CL).03 1 1 1 1 10 20 30 40 50 60 Observation 1 70 80 90 100 Test Results for I Chart of Ex9-29Vis TEST 1. One point more than 3. 63. 60. violating many of the tests for special causes.

569.569 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Viscosity (Ex9-29Vis) target = 28. k = 0. 9-43 . h = 5 20 Cumulative Sum UCL=14.24 10 0 0 -10 LCL=-14.Chapter 9 Exercise Solutions 9-29 continued (c) Let target = µ0 = 28.5.24 -20 1 10 20 30 40 50 60 Sample 70 80 90 100 Several observations are out of control on both the lower and upper sides.

569 29 28 27 -2.7 32 31 +2.7SL=26.Chapter 9 Exercise Solutions 9-29 continued (d) MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM EWMA Chart of Ex9-29Vis lambda = 0. L = 2.380 26 25 1 10 20 30 40 50 60 Sample 70 80 90 100 The process is not in control. 9-44 . There are wide swings in the plot points and several are beyond the control limits.7SL=30.15.759 EWMA 30 _ _ X=28.

007 27.128 = 2.1.243 0 0 0 0 0 0 0 0 0 EWMA Moving Center-Line Chart for Viscosity Xt. No. Viscosity 50.909 29.75) Excel : Workbook Chap09.756 24.0231 0.1579 0.898 33.025 35. 9-45 .479 29.479 29.464 18.482 26.1)).962 λ = 1 – MA1 = 1 – (– 0.144 1.464 18.527 26.367 20.560 27.05 P 0.21 1.760 29.841 41. Xi CL UCL LCL A few observations exceed the upper limit (87) and the lower limit (2.xls : Worksheet Ex9-29 lambda = l Xi 0 1 2 3 4 5 6 7 8 9 10 … 29.57 Constant 0.750 30.1579) = 1. 55.330 31. fit an ARIMA (0.1007 -1.909 29.122 17.940 18.158 sigma^ = CL UCL 28. 37.4839 0.527 26.665 26.356 25.665 26.441 42.000 31.873 39.000 10.550 Zi 28.000 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 Obs.482 26.922 below LCL 9.785 37.8457 (from a Moving Range chart with CL = 5. Stat > Time Series > ARIMA ARIMA Model: Ex9-29Vis … Final Estimates of Parameters Type Coef SE Coef T MA 1 -0.980 25.1) (= EWMA = IMA(1.000 40.85 LCL OOC? 19.937 20.452 37.000 30.785 31.030 32.1579 σˆ = MR d 2 = 3.984 18.788 22.298 32.120 0.680 33. 85).207 35.069 35.330 19.000 20.000 0.Chapter 9 Exercise Solutions 9-29 continued (e) To find the optimal λ.022 38.298 32.330 31.328 2.898 33.500 26.

Test Failed at points: 18. 21.4581 … P 0.7193 0.000 0.4349 0.5017 0. 9-46 .Chapter 9 Exercise Solutions 9-29 continued (f) Stat > Time Series > ARIMA ARIMA Model: Ex9-29Vis … Final Estimates of Parameters Type Coef SE Coef T AR 1 0. 19.54 Mean 28. with a good distribution of points between the control limits and no patterns.000 0.000 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Residuals from Viscosity AR(2) Model (Ex9-29res) 10 UCL=9. 20.79 AR 2 -0.6514 0.3278 62.68 -10 1 10 20 30 40 50 60 Observation 70 80 90 100 Test Results for I Chart of Ex9-29res TEST 7.04 0 7 7 -5 LCL=-9. 22 The model residuals signal a potential issue with viscosity around observation 20.0923 7.72 Constant 20. 15 points within 1 standard deviation of center line (above and below CL).60 Individual Value 5 7 7 7 _ X=-0.0922 -4. Otherwise the process appears to be in control.

05hr/sample. α = 0.27 −λh λh 1− e ( ) E(L) = $3. λ = 0.0 a1 = $0.4992 τ≅ − 2 12 α e−λh α ≅ = 0.10/unit.xls : worksheet Ex9-30a n = 5.0705 h λ h2 = 0.6098/hr 9-47 .0705 − 0.895 E(L) = $3. δ = 2.472) = 0. D = 2hr (a) Excel : workbook Chap09.Chapter 9 Exercise Solutions 9-30. a'3 = $5. k = 3. 1 − β = 0.00. a4 = $100/hr g = 0.01/hr or 1/λ = 100hr.210.472) − Φ (−7. a3 = $2. hopt = 1.0027 ⎛ ( µ + k σ n ) − ( µ0 + 2σ ) ⎞ ⎛ ( µ0 − k σ n ) − ( µ0 + 2σ ) ⎞ β = Φ ⎜⎜ 0 ⎟⎟ − Φ ⎜⎜ ⎟⎟ σ n σ n ⎝ ⎠ ⎝ ⎠ ( ) ( = Φ 3 − 2 5 − Φ −3 − 2 5 ) = Φ (−1.0000 = 0.027. h = 1. a2 = $0. α = 0.50/sample.231.79/hr (b) n = 3. kopt = 2.50.

12/hr (b) n = 5.0027.01/hr or 1/λ = 100hr. δ = 2.472) − Φ (−7. a3 = $25. α = 0. k = 3.01392/hr 9-48 . β = 0. λ = 0.4992 τ≅ − 2 12 2 12 α e−λh α 0.01(0.0705 h λ h 2 0. α = 0.0705 − 0.918 E(L) = $4.xls : worksheet Ex9-31 n = 5. D = 2hr (a) Excel : workbook Chap09.2498 τ≅ − 2 12 2 12 0.0 a1 = $0.27 −λh λ h 0.Chapter 9 Exercise Solutions 9-31.01(1) 1− e ( ) E(L) = $4. α = 0.52 ) = − = 0.05hr/sample. hopt = 1.0705 h λ h 2 1 0.54 −λh λ h 0.0027 ≅ = = 0. a'3 = $50.00207. kopt = 3.080.10/unit.01(0. h = 1.0027 α e−λh α ≅ = = 0.5 0. a4 = $100/hr g = 0.50/sample.0000 = 0.472) = 0.5) 1− e ( ) E(L) = $4. a2 = $0.98/hr (c) n = 5. h = 0. 1 − β = 0.0027 ⎛ ( µ + k σ n ) − ( µ + δσ ) ⎞ ⎛ ( µ − k σ n ) − ( µ + δσ ) ⎞ 0 0 ⎟−Φ⎜ 0 ⎟ β = Φ⎜ 0 ⎜ ⎟ ⎜ ⎟ σ n σ n ⎝ ⎠ ⎝ ⎠ ( ) ( ) = Φ ( 3 − 2 5 ) − Φ ( −3 − 2 5 ) = Φ k − δ n − Φ −k − δ n = Φ (−1.5.01(12 ) = − = 0. k = 3.368.

k = 3. h = 1.0 a1 = $2/sample a2 = $0.17/hr (b) n = 10. β = 0.472) = 0. hopt = 2.472) − Φ (−7. kopt = 2.8218083 E(L) = $10.5 0.01(0.0027 β = Φ k − δ n − Φ −k − δ n ( ) ( ) = Φ ( 3 − 1 5 ) − Φ ( −3 − 1 5 ) = Φ (−1.05 hr/sample D = 1 hr (a) n = 5.54 −λh λ h 0.0027 α e−λh α ≅ = = 0.xls : worksheet Ex9-33 λ = 0.Chapter 9 Exercise Solutions 9-32.0705 E(L) = $13.025091. α = 0.5.489018. k = 3.240.0027. h = 0.01(0.775 h λ h 2 0.xls : worksheet Ex9-32 D0 = 2hr. Excel : workbook Chap09. Excel : workbook Chap09.50/unit a'3 = $75 a3 = $50 a4 = $200/hr g = 0.52 ) = − = 0.775 − 0. ∆ = $25 n = 5. D1 = 2hr V0 = $500. α = 0.0000 = 0.2498 τ≅ − 2 12 2 12 0. α = 0.16/hr 9-33.39762/hr 9-49 . 1 − β = 0.5) 1− e ( ) E(L) = $16.01/hr or 1/λ = 100hr δ = 2.

00064.0030 1.00064 2. Graph > Time Series Plot > With Groups Time Series Plot of Ex9-34Xb USL = 1.00266 + 3 0.00304 LCL = CL − 3σ x ( ) = 1.00028 CL = R = 0.00234 + 3 0. UCL = D4 R = 2.00028 5 ) = 1. σˆ = R d 2 = 0.00266 − 3 ( 0. R = 0.00234 − 3 ( 0.0025 1. The plot below shows that the tool has been reset to approximately the same level as initially and the rate of tool wear is approximately the same after reset.0020 LSL = 1.00028) = 1.00064.00266 (maximum permissible average) UCL = CL + 3σ x = 1.326 = 0.0015 + 3(0.00064) = 0.0035 Ex9-34Reset A fter Before Ex9-34Xb 1.0035 − 3(0.00228 9-50 .0015 1.00028 5 = 1.0015 1 2 3 4 5 6 7 8 9 10 11 12 Ex9-34Sample n = 5.00196 x chart at tool reset: CL = USL − 3σ = 1.00234 UCL = CL + 3σ x = 1.00135.00272 LCL = CL − 3σ x ( ) = 1.114(0.00028) = 1.Chapter 9 Exercise Solutions 9-34. LCL = 0 x chart initial settings: CL = LSL + 3σ = 1.00028 5 = 1. It is good practice visually examine data in order to understand the type of tool wear occurring.00028 5 ) = 1.0035 1.

1850 0 14. 10-1.4254 10.186 LCL = 0 Excel : workbook Chap10.1850 0 14.00 T^2 40.00 30.9127 22.00 10.1850 0 14.Chapter 10 Exercise Solutions Note: MINITAB’s Tsquared functionality does not use summary statistics. xbar1 xbar2 1 58 32 2 60 33 3 50 27 4 54 31 5 63 38 6 53 30 7 42 20 8 55 31 9 46 25 3 2 5 3 -5 -3 -1 1 8 8 -2 0 -13 -10 0 1 -9 -5 0.0423 0. p = 2 characteristics. n = 25 sample size.xls : worksheet Ex10-1 Sample No.5408 13.00 50.001.948) = 14.1268 3.0451 1.7042 0.1690 0. Let α = 0.0817 2.001.1850 0 14.1850 0 diff1 diff2 matrix calc t2 = n * calc UCL = LCL = OOC? In control In control In control In control In control In control Above UCL In control In control … T^2 Control Chart for Quality Characteristics 60.8169 0.0676 1.5211 0. 10-1 .1268 0.2.6901 0.1850 0 14. p (m + 1)(n − 1) UCL = Fα .1850 0 14. Phase 2 T 2 control charts with m = 50 preliminary samples.1199 50(25) − 50 − 2 + 1 = ( 2448 1199 ) (6.0282 0.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sample Number Process is out of control at samples 7 and 14.1850 0 14.1850 0 14.1268 3. so many of these exercises have been solved in Excel.6338 14.00 0.mn − m − p +1 mn − m − p + 1 2(50 + 1)(25 − 1) = F0.00 20.1690 0. p .

mn − m − p +1 mn − m − p + 1 3(30 + 1)(10 − 1) = F0.2 3.2 14 3. p = 3 characteristics.1 0.4249 17.1 -0.1399 1.5 -0.2 0.7 3 2 3.00 T^2 15.1 0 0.1 4.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sample Number Process is out of control at sample 13.3 2.4249 17.4249 17.6 -0.5279 0.0593 10.8 6 4 4. p .4249 17.4249 17.5 7 3. Let α = 0.2 1.2 3 8 3 3.7 4 3 13 4.2 -0.001.3 2.0397 10. xbar1 xbar2 xbar3 1 3.00 10.6 3.0808 0.4 0.3122 3.2 0.2 0 -0.1 3 2.4 4 2.6 2.4249 17.8 11 3.00 20.4249 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 In control In control In control In control In control In control In control In control In control In control In control In control Above UCL In control In control T^2 Control Chart for Quality Characteristics 25.8 0.8 4.1189 1.4 -0.2 0.7927 1.1 0.3.7 0.4249 17.8525 matrix calc t2 = n * calc UCL = LCL = OOC? 17.2 0.5 5 3 3. p (m + 1)(n − 1) UCL = Fα .9 15 3.6 3 2.4 0.1 1.8084 1.9 3.579) ⎝ 268 ⎠ = 17.00 5.5741 2.1271 11.4 0.4 -0.8800 0.4249 17.6 1.8 diff1 diff2 diff3 0.8 4 2.Chapter 10 Exercise Solutions 10-2.2 0.8 0.5 -0.4249 17.5932 0.9 -1 0.6922 0.2 3.4249 17.4249 17.1887 0.1 3.2706 0.2 -0.1880 1.3 0 -0.0852 0.8692 8.2 0.3966 1.0528 0.5 0.3 -0.9 3 12 3.3 0.7 9 2.5 0.3990 0.6844 0.1216 0.5 -0.2372 2.7 3.6 -1 -0.2 10 2 2.1 0.4249 17. n = 10 sample size.4249 17.268 30(10) − 30 − 3 + 1 ⎛ 837 ⎞ =⎜ ⎟ (5.6574 6.xls : worksheet Ex10-2 Sample No.2 0 1 1. 10-2 .8 3 2.4 3 2.4249 17.3 3.3719 0.425 LCL = 0 Excel : workbook Chap10. Phase 2 T 2 control limits with m = 30 preliminary samples.7 0.001.7 0.00 0.0684 0.0793 20.

10-3 .0000 30.0817 2.2676 6.6901 0.8150 0 13.8150 0 13.0000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sample Number Process is out of control at samples 7 and 14. p = χ 0.8150 0 13. the chi-square formula will be used for the upper 2 = 13.1268 3. Let α = 0.1690 0. Same results as for parameters estimated from samples.2535 6.6338 0.8150 0 13.0676 1.816 control limit: UCL = χα2 .1127 52.001.5408 13.0000 T^2 40.0000 20.8150 0 13. xbar1 xbar2 1 2 UCL = LCL = OOC? 4 5 6 7 8 9 10 11 12 13 14 15 60 33 50 27 54 31 63 38 53 30 42 20 55 31 46 25 50 29 49 27 57 30 58 33 75 45 55 27 3 2 5 3 -5 -3 -1 1 8 8 -2 0 -13 -10 0 1 -9 -5 -5 -1 -6 -3 2 0 3 3 20 15 0 -3 0.8150 0 13.0704 0.3380 13.2028 5.9014 2. Phase 2 T 2 control limits with p = 2 characteristics.8150 0 13.8169 0.0761 1. Since population parameters are known.Chapter 10 Exercise Solutions 10-3.1268 0.1268 3.2 Excel : workbook Chap10.7042 0.001.xls : worksheet Ex10-3 Sample No.0000 0.1690 0.0676 1.8150 0 diff1 diff2 matrix calc t2 = n * calc 3 58 32 In control In control In control In control In control In control Above UCL In control In control In control In control In control In control Above UCL In control Phase II T^2 Control Chart 60.0451 1.0423 0.5211 0.8150 0 13.8150 0 13.8150 0 13.0000 10.0000 50.8150 0 13.4254 10.8150 0 13.6901 0.9127 22.0282 0.6901 0.8150 0 13.8169 0.8150 0 13.

0808 0.2706 0.8 6 4 4.2660 16.8 diff1 diff2 diff3 0.2660 16.2 0.4 -0.2 3 8 3 3.5 0.2 0.001.1 4.7 0.2 0.266 control limit: UCL = χα2 .5932 0.7927 1.5741 2. 10-4 .1399 1.6 3 2.4 0.2 0. Phase 2 T 2 control limits with p = 3 characteristics.8 0.1 0.00 5.8 3 2.1887 0.2 1.8 4.6 1.7 3 2 3.3 -0.00 10. the chi-square formula will be used for the upper 2 = 16.1880 1. Same as results for parameters estimated from samples.2660 16.7 0.2 3.6 -0.2660 16.4 -0. xbar1 xbar2 xbar3 1 3.3 3.9 -1 0.2 0 1 1.2 14 3.5 -0.8084 1.2372 2.7 9 2.7 3.2660 16.2 10 2 2.3966 1.1 0 0.5 -0.4 0.2660 16.2660 16. p = χ 0.2 3.1 0.4 0.Chapter 10 Exercise Solutions 10-4.xls : worksheet Ex10-4 Sample No.6 -1 -0.6 2.3719 0.3 0 -0.00 0.2660 16.1 -0.9 3 12 3. Let α = 0.7 0.5 7 3.1 1.2660 16.6922 0. Since population parameters are known.001.6844 0.00 T^2 15.1216 0.1 3 2.3 0.4 4 2.2660 16.5 0.0593 10.9 3.1 3.00 20.8 11 3.3 2.2 0.0793 20.7 4 3 13 4.8525 matrix calc t2 = n * calc UCL = LCL = OOC? 16.1 0.8692 8.8 4 2.2660 16.8 0.3 2.5 -0.0397 10.6 3.4 3 2.2 0.1189 1.2660 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 In control In control In control In control In control In control In control In control In control In control In control In control Above UCL In control In control Phase II T^2 Control Chart 25.2660 16.8800 0.2660 16.2 -0.2 -0.2660 16.9 15 3.0684 0.0528 0.5279 0.1271 11.2 0 -0.3 Excel : workbook Chap10.5 5 3 3.3990 0.6574 6.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sample Number Process is out of control at sample 13.0852 0.3122 3.

n = 3 sample size.7324 … 1429 1431 1433 1435 1437 1439 720 preliminary samples must be taken to ensure that the exact Phase II limit is within 1% of the chi-square limit.7331 18. (c) Quality characteristics. Find "m" such that exact Phase II limit is within 1% of chi-square limit.9447 717 718 719 720 721 722 8616 8628 8640 8652 8664 8676 3.107 18.55 30(3) − 30 − 6 + 1 ⎛ 372 ⎞ =⎜ ⎟ (3.005 (a) Phase II limits: p (m + 1)(n − 1) UCL = Fα .548 The Phase II UCL is almost 30% larger than the chi-square limit.6 = 18.531 3. p = 6 characteristics.1095 19.8820 22.6.548) = 18. m = 30 preliminary samples. Samples size.5042 20.263 3. 1.338 3.7328 18. 10-5 .107 3. p .107 3. p = χ 0.005.240 3. α = 0.005.xls : worksheet Ex10-5 m 30 40 50 60 70 80 90 100 num denom 372 55 492 75 612 95 732 115 852 135 972 155 1092 175 1212 195 F 3.882 LCL = 0 (b) 2 chi-square limit: UCL = χα2 . α = 0. Excel : workbook Chap10.Chapter 10 Exercise Solutions 10-5.407 3.9650 20.107 3.209 UCL 23.733.7337 18.01(18.3527 21.3184 20.294 3.mn − m − p +1 mn − m − p + 1 6(30 + 1)(3 − 1) = F0.531) ⎝ 55 ⎠ = 23. p = 6.005.223 3. n = 3.7325 18.7332 18.5920 20.107 3.107 3.

Samples size.294) ⎝ 115 ⎠ = 21.7424 18.733. Excel : workbook Chap10.005 (a) Phase II UCL: p (m + 1)(n − 1) UCL = Fα .Chapter 10 Exercise Solutions 10-6.8641 19. m = 30 preliminary samples.105 3.mn − m − p +1 mn − m − p + 1 6(30 + 1)(5 − 1) = F0.105 3.xls : worksheet Ex10-6 m 30 40 50 60 70 80 90 100 num 744 984 1224 1464 1704 1944 2184 2424 denom 115 155 195 235 275 315 355 395 F 3.164 3. p . 10-6 . n = 5 sample size.105 18.209 3. 1.174 3.155 3.5237 19.005. p = χ 0.1422 19. n = 5.5692 20.6. α = 0.548 The Phase II UCL is almost 15% larger than the chi-square limit.005. p = 6.4119 19.240 3.106 3.01(18.3087 20. (c) Quality characteristics. α = 0.105 3.189 3.6685 19.7318 … 411 preliminary samples must be taken to ensure that the exact Phase II limit is within 1% of the chi-square limit.005.548) = 18.115 30(5) − 30 − 6 + 1 ⎛ 744 ⎞ =⎜ ⎟ (3. Find "m" such that exact Phase II limit is within 1% of chi-square limit. p = 6 characteristics.309 (b) 2 chi-square UCL: UCL = χα2 .3232 390 400 410 411 412 9384 9624 9864 9888 9912 1555 1595 1635 1639 1643 3.149 UCL 21.7376 18.294 3.7330 18.7324 18.6 = 18.

188 The Phase II UCL is more than 55% larger than the chi-square limit.4401 25.Chapter 10 Exercise Solutions 10-7.188) = 25.7953 30.mn − m − p +1 mn − m − p + 1 10(25 + 1)(3 − 1) = F0.8549 28.641 2.1991 31.530 2.10.0154 27. 10-7 .3808 28.005 (a) Phase II UCL: p (m + 1)(n − 1) UCL = Fα . m = 25 preliminary samples. Samples size.101 2.41 25(3) − 25 − 10 + 1 ⎛ 520 ⎞ =⎜ ⎟ (3.530 2. 1. n = 3.326 (b) 2 chi-square UCL: UCL = χα2 . Excel : workbook Chap10.3259 34.7246 986 987 988 989 990 19740 19760 19780 19800 19820 1963 1965 1967 1969 1971 2. α = 0.530 2.629 UCL 39. (c) Quality characteristics.440.4399 25. n = 3 sample size.530 2.4024 29. p .101) ⎝ 41 ⎠ = 39.704 2.10 = 25.897 2. Find "m" such that exact Phase II limit is within 1% of chi-square limit.xls : worksheet Ex10-7 m 25 35 45 55 65 75 85 95 105 num 520 720 920 1120 1320 1520 1720 1920 2120 denom 41 61 81 101 121 141 161 181 201 F 3.4405 25. α = 0.01(25. p = χ 0.005. p = 10 characteristics.677 2.4398 25.4940 28.530 25.005.742 2.799 2. p = 10.4394 … 988 preliminary samples must be taken to ensure that the exact Phase II limit is within 1% of the chi-square limit.657 2.005.

p = 10. α = 0.188 The Phase II UCL is more than 25% larger than the chi-square limit.4408 25.005.578 26.4405 25.4413 25.623 27. p = 10 characteristics.529 2.005 (a) Phase II UCL: p (m + 1)(n − 1) UCL = Fα .188) = 25.6812 4240 411 2.01(25. p = χ 0.529 2.Chapter 10 Exercise Solutions 10-8. n = 5.529 2.8495 2640 251 2. 1.8651 3840 371 2.10.4399 25.6251 1440 131 2. p . 10-8 . m = 25 preliminary samples.mn − m − p +1 mn − m − p + 1 10(25 + 1)(5 − 1) = F0.5595 1840 171 2.4141 3040 291 2. (c) Quality characteristics. n = 5 sample size.005.767 31.4394 544 preliminary samples must be taken to ensure that the exact Phase II limit is within 1% of the chi-square limit.10 = 25.5335 540 541 542 543 544 545 21640 21680 21720 21760 21800 21840 2151 2155 2159 2163 2167 2171 2.529 25.529 2.625 (b) 2 chi-square UCL: UCL = χα2 .606 27.91 25(5) − 25 − 10 + 1 ⎛ 1040 ⎞ =⎜ ⎟ (2.689 29.585 26.594 27.572 26.005. Excel : workbook Chap10.767) ⎝ 91 ⎠ = 31.648 28.xls : worksheet Ex10-8 m 25 35 45 55 65 75 85 95 105 num denom F UCL 1040 91 2. Samples size. Find "m" such that exact Phase II limit is within 1% of chi-square limit.529 2.440.4419 25. α = 0.4967 2240 211 2.1011 3440 331 2.

p .7 0.Chapter 10 Exercise Solutions 10-9. p = 10 quality characteristics.01.7 1 0.10.7 ⎤ ⎢0.7 0. n = 3 sample size.mn − m − p +1 mn − m − p + 1 10(25 + 1)(3 − 1) = F0.7 ⎥ ⎢ ⎥ ⎣0.638 Phase II UCL: p (m + 1)(n − 1) UCL = Fα . Excel : workbook Chap10.788) ⎝ 41 ⎠ = 35. Assume α = 0.360 10-10. p . p = χ 0.01.277 10-9 .788) ⎝ 41 ⎠ = 32.10.7 ⎥ ⎥ Σ=⎢ ⎢0.41 25(3) − 25 − 10 + 1 ⎛ 520 ⎞ =⎜ ⎟ (2.7 0.7 0.mn − m − p +1 mn − m − p + 1 10(25 − 1)(3 − 1) = F0.4 = 13.01.01.7 1 ⎦ (b) 2 UCL = χα2 .41 25(3) − 25 − 10 + 1 ⎛ 480 ⎞ =⎜ ⎟ (2. m = 25 preliminary samples. Phase I UCL: p (m − 1)(n − 1) UCL = Fα .7 1 0.7 0.xls : worksheet Ex10-10 (a) ⎡ 1 0.7 0.

(e) Since (T 2 = 28.7 ⎤ −1 ⎛ ⎡3.7 1 0.479.5⎤ ⎡ 0⎤ ⎞ ⎜ ⎟ ⎜ ⎟ = 1⎜ ⎢⎢3.7 0.Chapter 10 Exercise Solutions 10-10 continued (c) T 2 = n ( y .7 1 0.800 2 T(4) = 25. 10-10 .5 0 0.694.µ (1) )′ Σ (1) ( y (1) .01.806 − 15.806 Yes.7 ⎥⎥ ⎜ ⎢⎢3.635 2 No. since all di are smaller than χ 0. d 2 = 6.590.313 di = T 2 − T(2i ) d1 = d 2 = d3 = d 4 = 15.01.277 ) . First.1 = 6.689 Investigate variables 1 and 3.493 2 χ 0.5⎥⎥ ⎢⎢ 0⎥⎥ ⎟ − 0.5⎦ ⎣ 0⎦ ⎟⎠ ) > ( UCL = 13. no variable is identified as a relatively large contributor.7 0.5⎤ ⎡ 0 ⎤ ⎞′ ⎡ 1 ⎜⎢ ⎥ ⎢ ⎥⎟ ⎢ 3. d3 = 13. d1 = 12. Second.1 = 6.585 2 T(2) = 21.5⎤ ⎡0 ⎤ ⎞′ ⎡ 1 0.7 ⎥ ⎜ ⎢3.7 = 1⎜ ⎢ ⎥ − ⎢ ⎥ ⎟ ⎢ ⎜ ⎢3.5⎤ ⎡ 0⎤ ⎞ ⎜ ⎟ 1 0.7 0.01.280) > (UCL = 13. (d) −1 T(1)2 = n ( y (1) .806 ( 0.313 = 0.300 T(3)2 = 14.979.635 T(1)2 = 15.5⎦ ⎣ 0 ⎦ ⎠ ⎣ 0. an out-of-control signal is generated.1 .µ )′ Σ −1 ( y .7 1 ⎥ ⎜ ⎢3. this information does not assist in identifying which a process variable shifted. d 4 = 2.7 ⎜⎜ ⎢ ⎥ ⎢ ⎥ ⎟⎟ ⎢ ⎝ ⎣3.5⎥⎥ − ⎢⎢0 ⎥⎥ ⎟ ⎢⎢ 0. an out-of-control signal is generated. since the standardized observations are equal (that is.277) .5⎥ ⎢ 0⎥ ⎟ ⎥ ⎜⎢ ⎥ ⎢ ⎥⎟ 0.5⎥ ⎢0 ⎥ ⎟ ⎢ 0.µ (1) ) ⎛ ⎡3.7 0.7 ⎥⎥ ⎜ ⎢⎢3.5⎥ ⎢ 0 ⎥ ⎟ ⎢ 0.7 0. all variables had the same shift).313 2 2 2 2 T(1) = T(2) = T(3) = T(4) = 15.5⎥⎥ − ⎢⎢ 0⎥⎥ ⎟ ⎜ ⎢3.7 ⎤ −1 ⎛ ⎡3.µ ) ⎛ ⎡3.7 1 ⎦ ⎜⎝ ⎣3. Since T 2 = 15.7 0. (f) 2 χ 0.7 = 15.5⎥ ⎢ 0⎥ ⎟ ⎦ ⎝⎣ ⎦ ⎣ ⎦⎠ ⎝⎣ ⎦ ⎣ ⎦⎠ ⎣ = 15.

d3 = 2.538) > (UCL = 7.043 2 T(2) = 2. d 2 = 1. p = χ 0.8 1 0.000.8⎥⎥ ⎢⎣ 0. Excel : workbook Chap10. (f) 2 χ 0.000.111.8 0.154 > ( UCL = 7.1 = 3.154 Yes.3 = 7.094 Since an out-of-control signal was not generated in (e). d3 = 6.000.538 T(3)2 = 4. Since T 2 = 11. ( ) (d) 2 χ 0.xls : worksheet Ex10-11 (a) ⎡ 1 0. it is not necessary to calculate the diagnostic quantities.841 T(1)2 = 11. an out-of-control signal is generated. (e) Since (T 2 = 6.8 0. d1 = 0.538 2 T(2) = 5.815 (c) T 2 = 11.05. 10-11 . an out-of-control signal is not generated.8 1 ⎥⎦ (b) 2 UCL = χα2 .815 ) .05.815) .05.841 T(1)2 = 5.444.376 T(3)2 = 5. This is confirmed since none of the di’s exceeds the UCL.154 Variables 2 and 3 should be investigated.8⎤ Σ = ⎢⎢ 0.778. d 2 = 8.Chapter 10 Exercise Solutions 10-11.1 = 3. d1 = 1.

607 21.950 ⎥ .553 −0.001 0.392 1.561⎤ ⎡121.012 ⎥⎥ ⎢⎣ −0.599 ⎥⎦ ⎡ 1.003 0.740 80.101 −0.014 ⎥⎥ ⎢⎣ 5.xls : worksheet Ex10-13 m = 40 ⎡ 4.003⎤ .104 88.561 0.780 80.526 ⎥⎦ ⎢⎣ 43. S1 = ⎢⎢ −0.001 ⎦ 10-13.553 −0.305 3. S 2 = ⎢⎢ −0. S1 = ⎢ ⎥ ⎣ −0.207 S1 3.014 27.158 10-12 .101 −0.012 7.016 0.305 5. Excel : workbook Chap10.256 43.339 0.950 587. Excel : workbook Chap10.720 ⎤ ⎢ ⎥ V′V = ⎢ −0.150 S2 2.282 3. S2 = ⎢ V′V = ⎢ ⎥ ⎥ ⎣ −0.003 0.641 V'V 133.256 0.Chapter 10 Exercise Solutions 10-12.016 0.740 67.001 ⎦ ⎡121.720 0.016 ⎤ x′ = [15.071 0.395 ⎤ x′ = [15.125] .256 0.104] .440 −0.003 −0.xls : worksheet Ex10-12 m = 40 ⎡ 4. Excel : workbook Chap10.016 5.395 −0.339 0.440 −0.256 ⎤ ⎡ 1.000 ⎥⎦ 10-14.392 1.001 −0.xls : worksheet Ex10-14 xbar xbar1 xbar2 10.307 1.071 ⎦ ⎣ −0.

1 with an UCL = H = 12.9231 3.75 1 0.Chapter 10 Exercise Solutions 10-15.17 6.9231 -0.2 lambda = 12.9231 y= -0.73.231 delta = 1.9231 -0.9231 -0.75 0. This gives an ARL1 between 7.75 0.0769 -0.9231 -0.9231 3.75 0.5 delta = 0.9231 3.9231 -0. 10-13 .73 13.1 0.9231 -0.22 and 12.308 0.87 UCL = H = 12.75 0.75 1 1 1 1 1 y' Sigma-1 = 0.9231 -0.75 0.17.0769 -0. Excel : workbook Chap10.308 0.75 0.0769 1 1 1 1 From Table 10-3.75 1 0.75 0.xls : worksheet Ex10-15 p= mu' = Sigma = y' = 4 0 0 0 0 1 0.0769 -0.53 ARL1 = Select λ = 0.308 y' Sigma-1 y = 1.75 0.75 0.109 ARL0 = 200 Sigma-1 = 3. select (lambda.9231 -0. H) pair that closely minimizes ARL1 1 1.308 0.

9 0.105 with an UCL = H = 15.9 0.432 -2.9 1 0.432 -2.432 -2.568 -2.270 0.432 7.9 0.081 delta = 1.9 0.60.432 -2.60 7.432 -2.9 0.03 ARLmin = 14.105 0.432 -2.432 7.568 1 1 1 1 From Table 10-4.270 0.432 -2.9 0.9 0.568 -2.26. H) pair delta = 1 1.9 0.270 y' Sigma-1 y = 1.432 -2.270 0.9 0. Excel : workbook Chap10.432 -2.26 16.9 1 1 1 1 1 y' Sigma-1 = 0.Chapter 10 Exercise Solutions 10-16. select (lambda. 10-14 .5 lambda = 0.9 1 0. This gives an ARLmin near 14.xls : worksheet Ex10-16 p= mu' = Sigma = y' = 4 0 0 0 0 1 0.432 7.65 Select λ = 0.040 ARL0 = 500 Sigma-1 = y= 7.18 UCL = H = 15.568 -2.

054 ARL0 = 200 Sigma-1 = 2.49 5.xls : worksheet Ex10-17 p= mu' = Sigma = y' = 2 0 0 1 0.65.3 lambda = 8. 10-15 .556 y' Sigma-1 y = 1. This gives an ARL1 between 5.15 10. Excel : workbook Chap10. H) pair that closely minimizes ARL1 1 1 1.5 delta = 0.2 0.20.8 1 1 1 y' Sigma-1 = 0.08 UCL = H = 10. select (lambda.49 and 10.7778 y= 1 1 From Table 10-3.7778 -2.65 9.64 9.20 5.1 0.8 0.Chapter 10 Exercise Solutions 10-17.65 10.2222 -2.111 delta = 1.556 0.5 1.48 ARL1 = Select λ = 0.2 with an UCL = H = 9.2 0.2222 2.

Chapter 10 Exercise Solutions 10-18. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). One point more than 3. Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of y2 Output Variable (Tab10-5y2) 110 1 Individual Value 105 1 5 100 UCL=100. 37. (a) Note: In the textbook Table 10-5.12 5 6 95 _ X=91. Test Failed at points: 40 10-16 . 11. 40 TEST 6. 9. Test Failed at points: 8. and 107. and 10 should be 100. 12. 39. 103. 36. Test Failed at points: 9.25 90 6 85 6 6 5 5 5 5 LCL=82. 11. 10. 10. 35.38 80 4 8 12 16 20 24 Observation 28 32 36 40 Test Results for I Chart of Tab10-5y2 TEST 1. the y2 values for Observations 8. 8 points in a row more than 1 standard deviation from center line (above and below CL).00 standard deviations from center line. 38. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). Test Failed at points: 5. 39. 37. 40 TEST 8. 9. 10 TEST 5.

215 Tab10-5x9 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Regression Model Residuals (Ex10-18Res) 7. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL).5 _ X=-0. Test Failed at points: 19.5 -5. 25 TEST 6.192 Tab10-5x4 .9 Tab10-5x8 .73 Tab10-5x6 + 6. 18 TEST 5.Chapter 10 Exercise Solutions 10-18 continued (b) Stat > Regression > Regression Regression Analysis: Tab10-5y2 versus Tab10-5x1. .6 Tab10-5x2 + 0.3.00 0.435 Tab10-5x3 + 0. Test Failed at points: 7.. The regression equation is Tab10-5y2 = 215 ..57 4 8 12 16 20 24 Observation 28 32 36 40 Test Results for I Chart of Ex10-18Res TEST 1. 10-17 .2 Tab10-5x5 + 0. 21.0 -2.666 Tab10-5x1 . One point more than 3.1 Tab10-5x7 + 10.0.00 standard deviations from center line. Test Failed at points: 21 Plot points on the residuals control chart are spread between the control limits and do not exhibit the downward trend of the response y2 control chart.0 5 LCL=-6.11.57 5 5 Individual Value 5.5 1 1 UCL=6. Tab10-5x2. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).0 2.

while the ACF for the residuals suggests a random process.Chapter 10 Exercise Solutions 10-18 continued (c) Stat > Time Series > Autocorrelation Autocorrelation Function for y2 Output Variable (Tab10-5y2) (with 5% significance limits for the autocorrelations) 1.0 1 2 3 4 5 6 7 8 9 10 Lag Autocorrelation Function for Regression Model Residuals (Ex10-18Res) (with 5% significance limits for the autocorrelations) 1.4 -0.8 -1.2 -0.0 -0.6 -0.4 -0.6 0.8 Autocorrelation 0.6 0.8 Autocorrelation 0. 10-18 .2 0.4 0.2 0.8 -1.0 -0.4 0.6 -0.2 -0.0 0.0 1 2 3 4 5 6 7 8 9 10 Lag The decaying sine wave of ACFs for Response y2 suggests an autoregressive process.0 0.

5684 X X X X X 7 67.6 0.0 0.7 8.9 60.1 34.4 60.8 34.3 3.0 52.6387 X X X 5 67.8 62.1 3.5 5 5 5 5 5 5 5 Mallows x x x x x x x x Vars R-Sq R-Sq(adj) C-p S 1 2 3 4 5 6 7 8 1 36.4 3.9 76.1665 X X 3 61.0 0. Response is Tab10-5y1 T T T T T T T T a a a a a a a a b b b b b b b b 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 .8 69.0760 X 2 55.6816 X 1 35.4 4.3 1.8 6.1 4..1 68.0273 X X 4 72.6 0.4378 X 2 62.8 76.5 64.5 5 5 5 5 5 5 5 Mallows x x x x x x x x Vars R-Sq R-Sq(adj) C-p S 1 2 3 4 5 6 7 8 1 43.83914 X X X X X X 7 80.9 3..6 58.7 48.9 4.1 52.9 2. Stat > Regression > Best Subsets Best Subsets Regression: Tab10-5y1 versus Tab10-5x1..9 2.8 64.8 7.83550 X X X X X 6 79.2 5.5 4..9751 X 2 50.0 4..Chapter 10 Exercise Solutions 10-19. x4..3 69.3 0.3087 1 31.9 13.9 60.7 3.8160 X X 4 64. Response is Tab10-5y2 T T T T T T T T a a a a a a a a b b b b b b b b 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 .2 60. Different approaches can be used to identify insignificant variables and reduce the number of variables in a regression model.8 56.5 76.. x8.5 0.83693 X X X X X 7 80.5 24. Tab10-5x2. .8 60..1 75.1 24...9 6.6149 X X X X X X 7 67.9 1.0 3.9 1. Tab10-5x2..0171 X X 3 66.6 60.9 1.8 0.1 12.5 1..7 3.0 19.1 41.1 13.7288 X X X 3 59.84292 X X X X X X … Best Subsets Regression: Tab10-5y2 versus Tab10-5x1.5 36.0 6.1 5..6921 2 55.4 1.7 62.6526 X X X X 6 67.5 71.1 0. a regression model of input variables x1.0 4.83020 X X X X 5 73.2 3..9 13.2 4.6 52. and x9 maximize adjusted R2 (minimize S) and minimize Mallow’s C-p.93966 X X X X 6 79.95522 X X X 5 79.95201 X X X 4 72.5660 X X X X X 6 67.7 18.. 10-19 .6147 X X X 4 64.6200 X X X X X X … T a b 1 0 5 x 9 X X X X X X X X X X X X X ****** T a b 1 0 5 x 9 X X X X X X X X X X X ****** For output variables y1 and y2.4 24.1799 X 3 67. x3.2 3.4 5.3 29..2 4.1 3.3 76.4 3.5208 X X X X 5 65. This solution uses MINITAB’s “Best Subsets” functionality to identify the best-fitting model with as few variables as possible. .8 61.3 3.

000 10-20 .52081 SE Coef 123.5% Analysis of Variance Source DF SS Regression 5 90.000 R-Sq(adj) = 76.23 P 0.10 5.03530 0.6329 0.001 0.80 0.43080 -0.09146 2.35 P 0.689 F 26.6367 114.236 Tab10-5x9 Predictor Constant Tab10-5x1 Tab10-5x3 Tab10-5x4 Tab10-5x8 Tab10-5x9 Coef 244.124 Tab10-5x3 .Chapter 10 Exercise Solutions 10-19 continued Stat > Regression > Regression Regression Analysis: Tab10-5y1 versus Tab10-5x1.000 0.225 100.98 -1.075 0.005 0.7% Analysis of Variance Source DF SS Regression 5 882.3 R-Sq = 67.47 -2. The regression equation is Tab10-5y2 = 244 .454 Tab10-5x3 + 0.633 Tab10-5x1 + 0.84 3.001 0.0.51 -3.12396 -0.. Tab10-5x3.85 P 0.000 Regression Analysis: Tab10-5y2 versus Tab10-5x1.5% MS 18.990 Residual Error 34 23. The regression equation is Tab10-5y1 = 819 + 0.02438 0.41 12.198 0.1034 3.08113 0.47 4. .424 T 28.4540 0.1497 0.9% MS 176.434 Total 39 114.7 S = 3.7604 23.098 0.830201 SE Coef 29.056 0. . Tab10-5x3.64 Tab10-5x8 + 115 Tab10-5x9 Predictor Constant Tab10-5x1 Tab10-5x3 Tab10-5x4 Tab10-5x8 Tab10-5x9 Coef 818.70 3.3441 0.0915 Tab10-5x4 + 2..001 0.175 -235.431 Tab10-5x1 .2 Tab10-5x8 .50 T 1..03 1.4 -0.81 S = 0.001 0.000 0.47 Total 39 1303.1758 11.03 Residual Error 34 421.65 R-Sq = 79.14 0..0.6 0.31 -3.025 R-Sq(adj) = 62.0.40 F 14.40 P 0.176 Tab10-5x4 + 11.75 3.

105 4 8 12 16 20 24 O bser vation 28 32 36 40 1 M oving Range 3 U C L=2.586 2 __ M R=0. One point more than 3. Failed at points: 11 10-21 .00 standard deviations from center line. 9 points in a row on same side of center line.00 standard deviations from center line. 9 points in a row on same side of center line. One point more than 3. 11 Test Results for MR Chart of Ex10-19Res1 TEST Test TEST Test 1. Failed at points: 10.105 1 _ X=-0. Failed at points: 26 2.000 0 2 -1 2 -2 LC L=-2.Chapter 10 Exercise Solutions 10-19 continued Stat > Control Charts > Variables Charts for Individuals > Individuals I-MR Chart of y1 Regression Model Residuals (Ex10-19Res1) 1 Individual V alue 2 U C L=2. Failed at points: 25 2.791 1 0 LC L=0 2 4 8 12 16 20 24 O bser vation 28 32 36 40 Test Results for I Chart of Ex10-19Res1 TEST Test TEST Test 1.

52 5 4 _ X=-0. 21.Chapter 10 Exercise Solutions 10-19 continued I-MR Chart of y2 Regression Model Residuals (Ex10-19Res2) Individual Value 8 1 1 5 U C L=6. no out-of-control observations) and the subset regression model (observation 25 is OOC). Test Failed at points: 7. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL).52 -8 4 8 12 16 20 24 O bser vation 28 32 36 40 1 M oving Range 8 U C L=8.45 2 0 LC L=0 4 8 12 16 20 24 O bser vation 28 32 36 40 Test Results for I Chart of Ex10-19Res2 TEST 1. 21 Test Results for MR Chart of Ex10-19Res2 TEST 1. observations 7 and 18 are OOC) and the subset regression model (observations 7 and 18 are OOC). One point more than 3. One point more than 3. 25 TEST 6. there is not a significant difference between control charts for residuals from either the full regression model (Exercise 10-18. Test Failed at points: 19. there is not a significant difference between control charts for residuals from either the full regression model (Figure 10-10. 10-22 . 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).00 0 -4 5 LC L=-6. Test Failed at points: 26 For response y1.00 standard deviations from center line.02 6 4 __ M R=2. Test Failed at points: 7. 18 TEST 5.00 standard deviations from center line. For response y2.

Chapter 10 Exercise Solutions 10-20.2 0.5 +2.3 -0.7.0 1. A potential advantage to using the EWMA control chart for residuals from a regression model is the quicker detection of small shifts in the process.7SL=-1.1 -0. However the chart for y2 residuals still indicates a problem beginning near observation 20.1 _ _ X=-0.0 -2.4 0.435 4 8 12 16 20 24 Sample 28 32 36 40 EWMA Chart of y2 Regression Model Residuals (Ex10-19Res2) 2.0 -0.7SL=0.7SL=-0.5 -1.5 4 8 12 16 20 24 Sample 28 32 36 40 Test Results for EWMA Chart of Ex10-19Res2 TEST.000 0.5 _ _ X=-0.435 0.1 and L = 2.000 0. Test Failed at points: 21.3 EWMA 0. Use λ = 0.347 EWMA 1.0 0.5 +2. 10-23 .347 -1.7SL=1.4 -2.0 -0. 22 The EWMA control chart for residuals from the response y1 subset model has no out-ofcontrol signals. Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of y1 Regression Model Residuals (Ex10-19Res1) 0.2 -0. One point beyond control limits.

580 0. Principal Component Analysis: Ex10-21X1.64094 1.94763 0.20327 -0.49153 0.985 1.334 -0.27825 -2.09129 -0.17976 -1.30494 -0.12378 -1.70241 -0.86838 10-24 . Ex10-21X4 Eigenanalysis of the Correlation Matrix Eigenvalue 2.444 PC2 -0.19592 0. select Correlation Matrix.60867 -0.10423 -0. Ex10-21X2.832 0.286 0.21893 -1.253 0.0118 0.83902 3.33164 0.152 0.29428 0.41131 -2. Note: To obtain principal component scores.02496 1. (a) Stat > Multivariate > Principal Components Note: To work in standardized variables in MINITAB.20787 0.580 0.3181 1.56081 -0.29168 0.534 -0.000 Variable Ex10-21X1 Ex10-21X2 Ex10-21X3 Ex10-21X4 PC1 0.03929 0. select Storage and enter columns for Scores.87917 -2.60340 0.17849 Ex10-21z3 0.08767 1.0613 Proportion 0.801 PC4 0.594 0.94470 -1.36089 0.13529 -1.46958 0.19166 -0.50471 -0.104 Principal Component Scores Ex10-21z1 0.36528 0.23100 0.387 PC3 0.47259 -0.330 0.99498 0.82157 0.14662 Ex10-21z2 -0.061 0.65608 2.Chapter 10 Exercise Solutions 10-21.14246 -0.718 -0.23823 -0.32286 0.01299 -1.64480 -0.62545 -0.685 -0.11243 0.794 0.015 Cumulative 0.20488 0.58815 1.257 0.29514 1.31445 -0. Ex10-21X3.31493 0.11787 -1.12420 0.21950 2.67046 -0.6088 0.99211 -1.607 0.19734 0.42176 -0.083 0.

5 2.0 -0.5 -2 0 2 -0.5 Ex1 0 -2 1 z2 all -3. See Excel : workbook Chap10.9 0.Chapter 10 Exercise Solutions 10-21 continued (b) Graph > Matrix Plot > Simple Matrix of Plots Matrix Plot of Ex10-21z1. Ex10-21z3 Principal Component Scores -2 0 2 2 0 Ex10-21z1 -2 2 0 Ex10-21z2 -2 1. Ex10-21z2.5 0. Ex10-21z3all Principal Component Scores -3. Graph > Matrix Plot > Matrix of Plots with Groups Matrix Plot of Ex10-21z1all. 10-25 . Ex10-21z2all.7 1.0 Ex10-21Obs New Original 4 Ex1 0 -2 1 z1 all 0 -4 2.9 (c) Note: Principal component scores for new observations were calculated in Excel.xls : worksheet Ex10-21.0 -0. the majority of new observations are clearly different from the original observations.7 Ex10-21z3 -0.0 2 Ex1 0 -2 1 z3 all 0 -2 -4 0 4 -2 0 2 Although a few new points are within area defined by the original points.

115 PC4 -0.513 -0.089 0.7% of the variability is explained by the first 3 principal components.232 0. Ex10-22x9 Eigenanalysis of the Correlation Matrix Eigenvalue 3.Chapter 10 Exercise Solutions 10-22.089 0.1407 2.123 0.003 1. (a) Stat > Multivariate > Principal Components Note: To work in standardized variables in MINITAB.156 -0.021 0.074 -0.465 0.247 0.200 0.368 0.727 0.204 -0.162 0.997 PC8 0.035 0.323 0.105 0. Ex10-22z2.261 -0.372 0.508 0. Ex10-22z3 Principal Component Scores -3 0 3 3 0 Ex10-22z1 -3 3 Ex10-22z2 0 -3 2 0 Ex10-22z3 -2 -3 0 3 -2 0 2 10-26 .406 0.014 0.022 0.035 0.467 -0.277 0.471 0.575 0.068 0.048 0.349 0.202 -0. select Correlation Matrix. Note: To obtain principal component scores.651 -0.256 0.632 0.1973 0.163 -0.579 0.127 -0.124 -0.419 0.028 0.0520 Proportion 0.912 PC5 0.148 0.135 0.152 -0.240 0.349 0.175 0.322 0.425 -0.244 -0.357 0.0287 0.117 Cumulative 0.006 0.022 0.3292 1.032 -0.145 -0.407 0.018 0.525 -0.3121 0.380 0.0730 1.050 0.199 0.000 PC9 0. ….099 -0.6129 0. Principal Component Analysis: Ex10-22x1.529 0.068 0.022 -0.436 -0.262 0.297 0.448 PC2 0.723 0.267 0. select Storage and enter columns for Scores.193 -0.120 0.409 -0.230 0.262 0.166 -0.108 0. Ex10-22x3.947 PC6 -0.332 -0. (c) Graph > Matrix Plot > Simple Matrix of Plots Matrix Plot of Ex10-22z1.844 Variable Ex10-22x1 Ex10-22x2 Ex10-22x3 Ex10-22x4 Ex10-22x5 Ex10-22x6 Ex10-22x7 Ex10-22x8 Ex10-22x9 PC1 -0.874 0.662 -0.602 -0. Ex10-22x2.579 -0.000 0.090 -0.188 0.2542 0.975 PC7 0.133 (b) 72.431 -0.238 PC3 -0.

xls : worksheet Ex10-22. Ex10-22z3all All Principal Component Scores -4 0 4 3 Ex10-22Obs First Last 0 Ex1 0 -2 2 z1 all -3 4 0 Ex1 0 -2 2 z2 all -4 2 0 Ex1 0 -2 2 z3 all -2 -3 0 3 -2 0 2 Several points lie outside the area defined by the first 30 observations. Ex10-22z2all. 10-27 . Graph > Matrix Plot > Matrix of Plots with Groups Matrix Plot of Ex10-22z1all. See Excel : workbook Chap10.Chapter 10 Exercise Solutions 10-22 continued (d) Note: Principal component scores for new observations were calculated in Excel. indicating that the process is not in control.

Chapter 11 Exercise Solutions 11-1. yt : observation zt : EWMA (a) zt = λ yt + (1 − λ ) zt −1 zt = λ yt + zt −1 − λ zt −1 zt − zt −1 = λ yt + zt −1 − zt −1 − λ zt −1 zt − zt −1 = λ yt − λ zt −1 zt − zt −1 = λ ( yt − zt −1 ) (b) zt −1 − zt − 2 = λ et −1 (as a result of part (a)) zt −1 − zt − 2 + (et − et −1 ) = λ et −1 + (et − et −1 ) zt −1 + et − zt − 2 − et −1 = et − (1 − λ )et −1 yt − yt −1 = et − (1 − λ )et −1 11-11 .

975 -0.2 gives SS = 9780 and average deviation from target = 1.000 0.375 0.0 0 2.0 -20 6.984 0.975 -13.766 9.0 -10.0 L = +10 10 4.0 -30 8.690 Bounded Adjustment Chart for Ex 11-2 50 -12.975 -1.975 -13.800 10.0 40 -8.0 -18.000 -3.625 10.xls : worksheet Ex 11-2 0 0.715 -5.375 -18.854 0.700 1.375 -18.3 exhibits less variability and is closer to target on average.0 -9.0 -9.375 -18.24 6526.375 -18. 11-12 .625 19.975 -10.0 0. The chart with λ = 0.975 -13.375 -22.000 25.359 45 46 47 48 49 50 22 -9 3 12 3 12 9 -31 12 9 -9 9 8.000 0.8 T= lambda = L= g= Obs Orig_out Orig_Nt Adj_out_t EWMA_t |EWMA_t|>L? Adj_Obs_t+1 Cum_Adj 1 2 3 4 5 6 7 8 9 10 0 16 24 29 34 24 31 26 38 29 0 16 8 5 5 -10 7 -5 12 -9 16 24 20.975 -1.000 0.923 10.0 12.975 -13.000 0.969 5.0 -50 1 6 11 16 21 26 31 36 41 46 Obs Orig_out Adj_out_t EWMA_t Adj_Obs_t+1 Chart with λ = 0.025 -22.975 -13.0 -9.000 -13.76.000 0.625 4.0 -9.0 0.982 -8.688 4.134 no yes no yes no no no no yes 0.975 … SS = Average = 21468 17.Chapter 11 Exercise Solutions 11-2.0 30 -6.3 10 0.000 5.000 0.975 -10.318 -7.625 7.625 12.560 6.000 0.127 -6.0 L = -10 -10 Adjustment Scale -4.993 no no no no no no 0.0 -40 10.000 0.000 11.179 -6. Excel : workbook Chap11.0 20 -2.

800 12.440 3.5 -8.288 no no no no no Adj_Obs_t+1 Cum_Adj 0 -12 0 -11 0 0 0 0 0 0 -12 -12 -23 -23 -23 -23 -23 -23 0 0 0 0 0 -11.440 6. but is further from target on average than for the chart with λ = 0.061 -7.5 -6.037 -4.264 7.880 0.Chapter 11 Exercise Solutions 11-3.4 10 0.25 0.958 7.5 -11.4 exhibits less variability.5 0.5 -11.022 -5. Excel : workbook Chap11.5 -11.5 0.24 5610.400 3.5 -11.175 no yes no yes no no no no no 46 47 48 49 50 -9 3 12 3 12 -31 12 9 -9 9 -20. 11-13 .5 -8.3.5 … SS = Average = 21468 17.xls : worksheet Ex 11-3 Target yt = lambda = L= g= 0 0.91 Bounded Adjustment Chart for Ex 11-3 80 -15 70 60 -10 50 40 30 L = +10 10 0 0 L = -10 -10 -20 Adjustment Scale -5 20 5 -30 -40 -50 10 -60 -70 15 -80 1 6 11 16 21 26 31 36 41 46 Obs Orig_out Adj_out_t EWMA_t Adj_Obs_t+1 The chart with λ = 0.400 13.813 -3.8 Obs Orig_out Orig_Nt Adj_out_t EWMA_t |EWMA_t|>L? 1 2 3 4 5 6 7 8 9 10 0 16 24 29 34 24 31 26 38 29 0 16 8 5 5 -10 7 -5 12 -9 16 24 17 22 1 8 3 15 6 6.

3 -27.3 -7.7 -6.3 -22.3 -2.6 1.0 5.3 1.0 -5.3 -4.0 -5.8 -3.3 -13.0 20. 11-14 .0 -1.9 5.2 -1.Chapter 11 Exercise Solutions 11-4.9 0.6 -22.6 -0.2 -1.0 -9.9 -23.7 -27.5 -7.3 -2.24 5495.0 10.0 -19.0 -14.9 -6.0 20.8 3.7 -4.2 0.7 5.4 1.4.7 Integral Control for Ex 11-4 50 -12 -10 40 -8 30 -6 20 -2 0 0 2 -10 Adjustment Scale -4 10 4 -20 6 -30 8 -40 10 -50 12 1 6 11 16 Orig_out 21 26 Adj_out_t 31 36 41 46 Adj_Obs_t+1 The chart with process adjustment after every observation exhibits approximately the same variability and deviation from target as the chart with λ = 0.0 -20.7 -0.7 45 46 47 48 49 50 22 -9 3 12 3 12 9 -31 12 9 -9 9 11.4 -4.9 5.0 -5.xls : worksheet Ex 11-4 T= lambda = g= 0 0.5 -4.0 … SS = Average = 21468 17.1 14.0 20. Excel : workbook Chap11.7 -8.8 Obs Orig_out Adj_out_t Orig_Nt Adj_Obs_t+1 Cum_Adj 1 2 3 4 5 6 7 8 9 10 0 16 24 29 34 24 31 26 38 29 0 16 8 5 5 -10 7 -5 12 -9 16.

39 162.53 138.xls : worksheet Ex 11-5 t Yt 1 2 3 4 5 6 7 8 9 10 … / m => 0 16 24 29 34 24 31 26 38 29 Var_m = Var_m/Var_1 = 1 2 3 4 5 6 7 8 9 16 8 5 5 -10 7 -5 12 -9 24 13 10 -5 -3 2 7 3 29 18 0 2 -8 14 -2 34 8 7 -3 4 5 24 15 2 9 -5 31 10 14 0 26 22 5 38 13 29 147.000 m 1.02 136.72 147.943 Variogram for Ex 11-5 2.029 1.500 1.929 1.11 175.000 0 5 10 15 20 25 Vm/V1 11-15 .000 1.47 179.500 0.217 0.002 1.500 2.000 0.Chapter 11 Exercise Solutions 11-5.70 1. Excel : workbook Chap11.104 1.195 1.370 0.60 151.43 201.

68 1.03 55.14 32.80 41.55 48.164698 0.0 -0.86 63. 11-16 .316478 0.31 16.28 0.01 2.8 Autocorrelation 0.299822 0.0 0.71 62.Chapter 11 Exercise Solutions 11-5 continued MTB : Chap11.33 0.149321 0.2 0.85 1.87 67.325056 0.59 54.345327 0.6 0.6 -0. The slow decline in the sample ACF also indicates the data are correlated and potentially nonstationary.05 0.39 27.4 0.228540 0.066173 T 3.0 1 2 3 4 5 6 7 Lag 8 9 10 11 12 13 Autocorrelation Function: Yt Lag 1 2 3 4 5 6 7 8 9 10 11 12 13 ACF 0.54 1.75 Variogram appears to be increasing.90 0.440855 0. so the observations are correlated and there may be some mild indication of nonstationary behavior.12 2.mtw : Yt Stat > Time Series > Autocorrelation Function Autocorrelation Function for Data in Table 11-1 (Yt) (with 5% significance limits for the autocorrelations) 1.59 0.012158 0.334961 0.4 -0.26 LBQ 10.8 -1.389094 0.45 1.2 -0.41 63.44 67.58 1.440819 0.

3 2.8 0.0 195.0 140.2 1.3 -6.5 49 50 SS = Average = Variance = 11.4 Adjusted 1.0 193.7 187.4 1.3 2.5 192.6 -13.3 467.0 10.5 24.4 4.7 3.0 -5.0 30.5 208.0 -0.0 0.2 T= lambda = g= Obs.4 62.9 180.5 3.7 4.0 195.9 12.8 161.5 -7.Chapter 11 Exercise Solutions 11-6.0 15.4 0.8 -15.7 180.4 170.871.1 216.9 -39.0 260.9 320.0 5.2 1.510.0 -20.8 -1.5 Unadjusted 1.1 13.3 192. t 1 2 3 4 5 6 7 8 9 10 Orig_out Orig_Nt Adj_out_t Adj_Obs_t+1 Cum_Adj 215.0 290.0 63.0 185.0 12.0 1 6 11 16 21 26 31 36 41 46 Obs.0 170.818. 11-17 .0 -15.0 230.0 Adjustment Scale Integral Control for Ex 11-6(a) 20.5 176.7 220.323.xls : worksheet Ex 11-6a 200 0.3 -4.0 110.1 -3.1 177.4 191.8 162.2 7.9 200.8 … 145.1 4. (a) and (b) Excel : workbook Chap11.0 200.8 176.0 -10.8 -20.0 129.0 80.0 1.2 0.7 191.2 160.7 0.0 187.8 0. t Orig_out Adj_out_t Adj_Obs_t+1 Significant reduction in variability with use of integral control scheme.

5 5.7 0.0 198.6 8. 11-18 .6 -13.4 176.323.6 -2.995.0 3.0 290.0 80.0 30.0 Orig_out Orig_Nt Adj_out_t Adj_Obs_t+1 Cum_Adj 215.5 24.0 260.5 216.0 -10.3 -4.8 -0.0 30.5 -7.0 170.1 21.0 0.0 -0.0 200.0 8.5 -7.4 Adjusted 1.4).5 13.6 7.0 195.871.9 -39.4 1.0 4.7 … 129.0 10.0 Adjustment Scale Integral Control for Ex 11-6(c) 40.0 6.0 -30.1 187.9 2.8 185.0 7.0 195.3 467.8 161.0 2.5 7.8 0.0 189.5 192.xls : worksheet Ex 11-6c 200 0.0 9.Chapter 11 Exercise Solutions 11-6 continued (c) Excel : workbook Chap11. t Orig_out Adj_out_t Adj_Obs_t+1 Variances are similar for both integral adjustment control schemes (λ = 0. t 1.2 and λ = 0.0 20.0 -40.5 176.0 140.8 193.888.0 -20.1 176.4 3.2 T= lambda = g= Obs.0 5.0 230.0 320.0 1 6 11 16 21 26 31 36 41 46 Obs.3 -6.8 1.0 10.2 180.8 -20.1 3.4 1.4 18.0 162.9 164.9 183.0 66.7 223.9 18.7 2.5 Unadjusted 1.9 208.5 50.4 191.0 110.8 7.0 195.4 175.0 SS = Average = Variance = -15.

633 187.253 180.000 0.516 48.5 191.304 467.717 0.000 0.177 176.000 6.0 1 6 11 16 21 26 31 36 41 46 Obs.0 0.0 260.9 -39.000 0.000 0.0 0.0 -15.840 191.7 320.5 -7 186.xls : worksheet Ex 11-7 200 0.7 216.733 -3.0 Adjustment Scale Bounded Adjustment Chart for Ex 11-7 20.870 216.000 0.051 172.516 48.000 0.000 6.000 0.733 -1.000 0.233 0.3 -6 185.000 48.364 1.3 -4.8 -15.016 -12.699 174.3 7 11.5 1. Excel : workbook Chap11. t Orig_out Adj_out_t EWMA_t Adj_Obs_t+1 Behavior of the bounded adjustment control scheme is similar to both integral control schemes (λ = 0.969 -5.000 -0.0 170.000 0.0 80.Chapter 11 Exercise Solutions 11-7.516 48.000 0.696 176.8 -20 196 -0.0 200.300 -4.4).323.0 -20.0 -5.2 and λ = 0.516 48.0 15.900 -5.941 162.265 182.0 140.0 -0.057 -7.2 145.5 46 47 48 49 50 -7 3.000 -8.2 133.632.412 185.2 Target yt = lambda = L= g= Obs.233 6.6 -13.9 126.0 129.0 0.233 yes no no no no 5.0 290.0 -10.000 0.5 24.300 -2.0 5.716 181.8 SS = Average = Variance = 165.233 6.716 193.000 0.1 176.4 162.0 30.0 0.516 … 122. t 1 2 3 4 5 6 7 8 9 10 Orig_out Orig_Nt Adj_out_t EWMA_t 215.459 -10.872 161.702 -7.0 230.9 193.516 178.0 10.2 12 1.656 |EWMA_t|>L? Adj_Obs_t+1 Cum_Adj no no no no no no yes no no 0.0 110.600 -14.8 0 195. 11-19 .9 176.

0 -8.000 0.9 -39.758 58.326 -4.0 0.0 170.000 0.6 -13.758 58.3 -6 185.000 0.0 1 6 11 16 21 26 31 36 41 46 Obs.0 16.9 126.0 40.9 199.8 0 195.775 187.958 191.467 0.000 0.5 46 47 48 49 50 SS = Average = Variance = -7 3.658 184.467 no no no no no 0.693 -7.486 176.967 -0.256 176.000 0.958 203.758 … 122.872 161.5 191.000 0.000 0. 11-110 .4 162.3 7 11.0 0.81 -9.0 24.0 48.488 185.0 129.0 -32.600 -23.7 216.720 -11.900 -8.323.0 32.9 176.2 145.573 216.5 -7 192.758 58.0 290.000 12.0 0.467 12.0 80.8 -15.xls : worksheet Ex 11-8 200 0.304 467.000 0.000 0.2 133. t Orig_out Adj_out_t EWMA_t Adj_Obs_t+1 Behavior of both bounded adjustment control schemes are similar to each other and simlar to the integral control schemes.849 -10.0 -40.692 162.758 58.000 -14.2 T= lambda = L= g= Obs.0 30.0 0.0 230.784 170. Excel : workbook Chap11.467 12.0 Adjustment Scale -16.0 320.258 1.000 0.000 0.Chapter 11 Exercise Solutions 11-8.1 176.083 189.000 1.0 200.000 12.680 191.013 180.0 140.513 1.4 15 1. t 1 2 3 4 5 6 7 8 9 10 Orig_out Orig_Nt Adj_out_t EWMA_t 215.967 -2.86 Bounded Adjustment Chart for Ex 11-8 350.5 181.0 110.0 260.758 188.821 |EWMA_t|>L? Adj_Obs_t+1 Cum_Adj no no no no no no yes no no 0.0 50.000 58.000 0.000 0.773.0 8.5 24.8 -20 196 -1.0 -0.0 -48.3 -4.300 -4.0 -24.300 -8.

2 1.6 0.2 51.7 -5.262 78.0 -1.6 55.8 -3.7 0.3 -1.6 -1.0 34 2.0 45.8 -0.0 -9.9 0.xls : worksheet Ex 11-9a T= lambda = g= 50 0.9 … SS = Average = Variance = Adjusted 108.0 66 -2.4 65.0 58 -1.0 58.4 -0.0 -11.0 10.2 -1.0 53.0 -4.Chapter 11 Exercise Solutions 11-9.6 Obs Orig_out Adj_out_t Orig_Nt Adj_Obs_t+1 Cum_Adj 1 2 3 4 5 6 7 8 9 10 50 58 54 45 56 56 66 55 69 56 8.8 -5.0 42 1.0 -13. t Orig_out Adj_out_t Adj_Obs_t+1 11-111 .2 22.0 Adjustment Scale Integral Control for Ex 11-9 (a) 5.0 0.0 11.7 22.0 -1.32 Unadjusted 109.51 Significant reduction in variability with use of an integral control scheme.4 54.4 223. (a) and (b) Excel : workbook Chap11.1 48.0 -0.0 82 -4.2 50.0 43.6 -1.0 74 -3.4 0.629 46.7 -0.0 50 0. 90 -5.0 3.0 14.8 -0.7 49 50 23 26 3.0 10 1 6 11 16 21 26 31 36 41 46 Obs.0 18 4.0 26 3.520 44.3 -1.6 -3.7 64.

1 -3.2.0 50.4 -1.8 27.833 56.5 114.0 -0.4 … 49 50 SS = Average = Variance = 23 26 109.2 55 -11.0 56 0.0 48.9 69 14. as compared to λ = 0.0 63.8 -8.5 1.5 -2.0 42.40 There is a slight reduction in variability with use of λ = 0.0 52.0 -1.0 66 10.3 -5.2 56 -13.3 -2.51 3.3 -9. with a process average slightly closer to the target of 50.0 58.4 26.8 0.6 56 11.5 53.Chapter 11 Exercise Solutions 11-9 continued (c) Excel : workbook Chap11.3 -6.520 44.0 -2.0 -2.0 -3.4 223.0 54.1 -0.8 0.6 Obs 1 2 3 4 5 6 7 8 9 10 Orig_out Orig_Nt Adj_out_t Adj_Obs_t+1 Cum_Adj 50 58 8.4.4 1.4 -0.0 63.0 55.0 -3. 11-112 .0 54 -4.9 -0.5 45 -9.0 46.xls : worksheet Ex 11-9c T= lambda = g= 50 0.819 47.0 3.

0 1 6 11 16 21 26 31 36 41 46 Obs.0 50 0.0 66 -4.0 58 -2.0 74 -6.0 34 4.0 82 -8.0 18 8.Chapter 11 Exercise Solutions Integral Control for Ex 11-9(c) 90 -10.0 10 10.t Orig_out Adj_out_t Adj_Obs_t+1 11-113 .0 26 6.0 42 2.

664 56 11 56. t Orig_out Adj_out_t EWMA_t Adj_Obs_t+1 Nearly the same performance as the integral control scheme.00 5.00 0.00 0.0 58 -1.904 |EWMA_t|>L? Adj_Obs_t+1 Cum_Adj no no no no no yes no no no 0.48 31.00 0.00 -2. 11-114 .505 -5.515 -1.2 4 1.0 10 5.842 -4.00 -2.620 121.00 2.00 0.00 1.00 0.880 56 -13 54.600 69 14 67.00 0.00 -2.0 18 4.709 -2.00 3.00 -2.0 66 -2.79 38.0 0.79 15.0 0.00 0.731 56 0 56.0 2 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 Obs.79 43.0 42 1.822 45.Chapter 11 Exercise Solutions 11-10.19 17.080 45 -9 45.00 13.00 0.362 107.48 15.00 0.00 no yes no yes no 0.6 Target yt = lambda = L= g= Obs.0 90 -5.00 0.268 55 -11 53.00 0.40 0.00 3.00 0.19 109.600 54 -4 54.51 SS = Average = Variance = -2.4 223.00 -2. Excel : workbook Chap11.00 1.0 50 0.0 26 3. t 1 2 3 4 5 6 7 8 9 10 Orig_out Orig_Nt Adj_out_t EWMA_t 50 0 58 8 58 1. with similar means and sums of squares.72 98 -6.520 44.585 66 10 66.48 35.19 … 24 18 20 23 26 46 47 48 49 50 8 -6 2 3 3 37.xls : worksheet Ex 11-10 50 0.0 34 2.00 2.0 Adjustment Scale Bounded Adjustment Chart for Ex 11-10 6.32 0.0 82 -4.0 74 -3.79 17.00 2. but different variances (bounded adjustment variance is larger).

Enter the data into the MINITAB worksheet using the first three columns: one column for glass type. therefore some information required by MINITAB is not included. one column for phosphor type. Since the experiment layout was not created in MINITAB. The remaining exercises are worked in a similar manner. Detailed MINITAB instructions are provided for Exercises 12-1 and 12-2 to define and create designs. The dialog box should look: 12-1 . This experiment is three replicates of a factorial design in two factors—two levels of glass type and three levels of phosphor type—to investigate brightness. 12-1. This is how the Excel file is structured (Chap12. the initial experimental layout must be created in MINITAB or defined by the user. the design must be defined before the results can be analyzed. and one column for brightness. and only the solutions are provided. check “General full factorial”.Chapter 12 Exercise Solutions Note: To analyze an experiment in MINITAB. Select the two factors (Glass Type and Phosphor Type). select Stat > DOE > Factorial > Define Custom Factorial Design. then for this exercise.xls). After entering the data in MINITAB. The Excel data sets contain only the data given in the textbook.

point type. The design and data are in the MINITAB worksheet Ex12-1. DO NOT insert or delete columns between columns 1 through 7. so leave the selections as below.Chapter 12 Exercise Solutions 12-1 continued Next. or blocks. no information is provided on standard order. run order. 12-2 . MINITAB recognizes these contiguous seven columns as a designed experiment.MTW. and click “OK” twice. select “Designs”. inserting or deleting columns will cause the design layout to become corrupt. Note that MINITAB added four new columns (4 through 7) to the worksheet. For this exercise.

since the factor levels in this experiment are categorical. 2. General Linear Model: Ex12-1Bright versus Ex12-1Glass.3 52. click “OK”. Glass type (A) and phosphor type (B) significantly affect television tube brightness (P-values are less than 0. verify that the selected terms are Glass Type.26483 R-Sq = 96. and click “OK” twice. Phosphor Type.7 8.10). do not select this option. Residual Plots for Ex12-1Bright Normal Probability Plot of the Residuals Residuals Versus the Fitted Values 99 10 Residual Percent 90 50 0 10 -10 1 -10 0 Residual 10 220 Histogram of the Residuals 240 260 280 Fitted Value 300 Residuals Versus the Order of the Data 10 3 Residual Frequency 4 2 0 1 0 -10 -10 -5 0 5 Residual 10 15 2 4 6 8 10 12 14 Observation Order 16 18 12-3 .7 1. 3 Analysis of Variance for Ex12-1Bright.000 Ex12-1Phosphor 2 933.3 466. Click on “Graphs”.26 0. Click on “Storage”. The option to plot residuals versus variables is for continuous factor levels.0 14450.44% No indication of significant interaction (P-value is greater than 0.3 133.08% R-Sq(adj) = 94.10).Chapter 12 Exercise Solutions 12-1 continued Select Stat > DOE > Factorial > Analyze Factorial Design. 2 1.0 14450. select “Residuals Plots : Four in one”. Ex12-1Phosphor Factor Ex12-1Glass Ex12-1Phosphor Type fixed fixed Levels 2 3 Values 1.3 66.8 Total 17 16150.79 0. and their interaction.318 Error 12 633.0 273. Click “OK”.84 0.0 S = 7.3 633. select “Fits” and “Residuals”. Select the response (Brightness).3 933. using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Ex12-1Glass 1 14450.004 Ex12-1Glass*Ex12-1Phosphor 2 133. then click on “Terms”.

If the model were re-fit with only Glass Type and Phosphor Type. and versus fitted values reveals no problems. Individual Value Plot of RESI1 vs Ex12-1Glass 15 RESI1 10 5 0 0 -5 -10 1 2 Ex12-1Glass Individual Value Plot of RESI1 vs Ex12-1Phosphor 15 RESI1 10 5 0 0 -5 -10 1 2 Ex12-1Phosphor 3 12-4 .Chapter 12 Exercise Solutions 12-1 continued Visual examination of residuals on the normal probability plot. Click on “Scale”. then click “OK” twice. select the “Reference Lines” tab. Select the column with stored residuals (RESI1) as the Graph variable and select one of the factors (Glass Type or Phosphor Type) as the Categorical variable for grouping. To plot residuals versus the two factors. histogram. select Graph > Individual Value Plot > One Y with Groups. The plot of residuals versus observation order is not meaningful since no order was provided with the data. and enter “0” for the Y axis. the residuals should be re-examined.

Chapter 12 Exercise Solutions 12-1 continued Note that the plot points are “jittered” about the factor levels. To remove the jitter. however. select the graph to make it active then: Editor > Select Item > Individual Symbols and then Editor > Edit Individual Symbols > Jitter and de-select Add jitter to direction. 12-5 . there appears to be more variability in results with phosphor type 2. Individual Value Plot of RESI1 vs Ex12-1Glass 15 RESI1 10 5 0 0 -5 -10 1 2 Ex12-1Glass Individual Value Plot of RESI1 vs Ex12-1Phosphor 15 RESI1 10 5 0 0 -5 -10 1 2 Ex12-1Phosphor 3 Variability appears to be the same for both glass types.

Interaction Plot (data means) for Ex12-1Bright 310 Ex12-1Glass 1 2 300 290 Mean 280 270 260 250 240 230 220 1 2 Ex12-1Phosphor 3 The absence of a significant interaction is evident in the parallelism of the two lines. Final selected combination of glass type and phosphor type depends on the desired brightness level. Select “Interaction Plot” and click on “Setup”. 12-6 . and click “OK” twice. select the response (Brightness) and both factors (Glass Type and Phosphor Type).Chapter 12 Exercise Solutions 12-1 continued Select Stat > DOE > Factorial > Factorial Plots.

79 0.318 Error 12 633.08% R-Sq(adj) = 94.3 66.7 8. and complete the dialog box as below.5.26 0.667 235.004 Interaction 2 133.000 Ex12-1Phosphor 1 2 3 Individual 95% CIs For Mean Based on Pooled StDev -----+---------+---------+---------+---(--*-) (--*-) -----+---------+---------+---------+---240 260 280 300 Mean 260. Two-way ANOVA: Ex12-1Bright versus Ex12-1Glass.Chapter 12 Exercise Solutions 12-1 continued Alternate Solution: This exercise may also be solved using MINITAB’s ANOVA functionality instead of its DOE functionality. select Stat > ANOVA > Two-Way.7 1. The DOE functionality was selected to illustrate the approach that will be used for most of the remaining exercises. To obtain results which match the output in the textbook’s Table 12.0 272.0 14450.0 12-7 .0 264.0 273.265 R-Sq = 96.000 Ex12-1Phosphor 2 933. Ex12-1Phosphor Source DF SS MS F P Ex12-1Glass 1 14450.44% Ex12-1Glass 1 2 Mean 291.333 256.8 Total 17 16150.0 S = 7.3 52.84 0.667 Individual 95% CIs For Mean Based on Pooled StDev -------+---------+---------+---------+-(-------*-------) (-------*-------) (-------*-------) -------+---------+---------+---------+-256.3 466.0 280.000 273.

009 ** 0. ABC) are included. change number of replicates to “2”. This solution uses the first approach. Another approach would be to create a worksheet containing the data. Cutting Speed will also be retained to maintain a hierarchical model. then define a customer factorial design. and click “OK” twice.25 -12.25 42. Cutting Angle. Metal Hardness.014 0.199 R-Sq(adj) = 72. Leave the design type as a 2-level factorial with default generators. a full factorial model is not necessary.199 Based on ANOVA results. select Stat > DOE > Factorial > Analyze Factorial Design.62 12.357 0.41 -0.10.81 8.25 -59. Both approaches would achieve the same result. to change to standard order (and ease data entry) select Stat > DOE > Display Design and choose standard order. (a) To analyze the experiment. B.25 -5.09 1.483 0.75 35.89 Cutting Speed*Metal Hardness -11.020 ** 0. 12-8 .6236 R-Sq = 85.41 2.Chapter 12 Exercise Solutions 12-2.13 12.41 0. one approach to solving this exercise is to create a 23 factorial design in MINITAB. Select “Factors”. a reduced model in Metal Hardness.96 P 0.41 -0.40 Cutting Angle S = 49.008 0. and change the Number of factors to “3”. AB. AC.662 0.75 -17. Select Stat > DOE > Factorial > Create Factorial Design. C.40 Cutting Angle 71.81 Metal Hardness*Cutting Angle -24. and click “OK”.41 33.001 ** 0.30 Cutting Speed 18. enter the factor names.37 12. The design and data are in the MINITAB worksheet Ex12-2.56% Analysis of Variance for Life (coded units) Source DF Seq SS Adj SS Adj MS Main Effects 3 50317 50317 16772 2-Way Interactions 3 59741 59741 19914 3-Way Interactions 1 4830 4830 4830 Residual Error 8 19700 19700 2462 Pure Error 8 19700 19700 2463 Total 15 134588 … F 6.88 12. BC. and Cutting Speed*Cutting Angle is more appropriate. The worksheet is in run order.45 Cutting Speed*Cutting Angle -119.13 12. Select “Designs”.MTW. Cutting Angle Estimated Effects and Coefficients for Life (coded units) Term Effect Coef SE Coef T Constant 413. Factorial Fit: Life versus Cutting Speed. Based on P-values less than 0.12 12. Select “Terms” and verify that all terms (A.36% P 0. then enter the data.000 0.74 Metal Hardness 84. leave factor types as “Numeric” and factor levels as -1 and +1.12 12. Since the standard order (Run) is provided.62 12.41 -4.41 -1.41 3.25 9. highlight full factorial.98 Cutting Speed*Metal Hardness* -34.

Chapter 12 Exercise Solutions 12-2(a) continued Factorial Fit: Life versus Cutting Speed.47 0.78 S = 49.001 1.25 9.0 512. Cutting Angle Estimated Effects and Coefficients for Life (coded units) Term Effect Coef SE Coef T Constant 413.12 12.47 -4.47 3.13 12.38 Cutting Angle 71.5 1 266.006 0. Cube Plot (data means) for Life Exercise 12-2(b) 552.85 P 0.04 0.88 12.001 R-Sq(adj) = 72.5 391.65% P 0. 12-9 .73 Metal Hardness 84.25 42.5 405.0 380.015 0.74 22.5. Choose and set-up a “Cube Plot”. Select Stat > DOE > Factorial > Factorial Plots. we see that the low level of cutting speed and the high level of cutting angle gives good results regardless of metal hardness.425 (b) The combination that maximizes tool life is easily seen from a cube plot. B+ and C+.480 0.008 0.0 1 Metal Hardness 446.47 33.12 Cutting Speed 18.25 -59.5 351. for an average predicted life of 552.13 12. Metal Hardness.88 Cutting Speed*Cutting Angle -119.47 2.0 -1 1 -1 Cutting Speed Cutting Angle -1 Longest tool life is at A-.8988 R-Sq = 79. (c) From examination of the cube plot.25% Analysis of Variance for Life (coded units) Source DF Seq SS Adj SS Adj MS Main Effects 3 50317 50317 16772 2-Way Interactions 1 56882 56882 56882 Residual Error 11 27389 27389 2490 Lack of Fit 3 7689 7689 2563 Pure Error 8 19700 19700 2463 Total 15 134588 F 6.75 35.62 12.000 0.

select “Storage” and choose “Residuals”. AC) are included. To save residuals to the worksheet. Select “Graphs”. Select “Terms” and verify that all terms for the reduced model (A. C. select Stat > DOE > Factorial > Analyze Factorial Design. 12-10 . Residuals versus fitted values plot shows that the equal variance assumption across the prediction range is reasonable. To find the residuals.Chapter 12 Exercise Solutions 12-3. B. Normal Probability Plot of the Residuals (response is Life) 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 -100 -50 0 Residual 50 100 Residuals Versus the Fitted Values (response is Life) 50 Residual 25 0 -25 -50 -75 250 300 350 400 Fitted Value 450 500 550 Normal probability plot of residuals indicates that the normality assumption is reasonable. and for residuals plots choose “Normal plot” and “Residuals versus fits”.

To preserve model hierarchy. carbonation.500 28. along with interactions involving both sweetener and the ratio of syrup to water.188 -0.938 -0.9504 -0.91 Total 31 1967.9504 -1.923 0.188 -1.500 0.9504 -1. syrup to water*carbonation.563 -0.594 0.37 2-Way Interactions 6 199.9504 -0.63 852.69 Carbonation -2.28 Carbonation*Temperature Analysis of Variance for Total Score (coded units) Source DF Seq SS Adj SS Adj MS F Main Effects 4 852.50 462.687 0.10 Carbonation*Temperature 1. 12-11 . and then enter the data.9504 -2.15 Syrup to Water*Carbonation -0.688 33.Chapter 12 Exercise Solutions 12-4.49 Temperature Sweetener*Syrup to Water* 2.469 0. sweetener*syrup to water*temperature).938 -0.031 0.77 Syrup to Water -1. Create a 24 factorial design in MINITAB.629 * * * * * 0.094 0.094 0. Syrup to Water.28 1.14 Sweetener*Carbonation 0.688 -1.055 0.9504 -0. sweetener*syrup to water*carbonation.531 47.031 0.062 2.9504 2.47 P 0.781 0.218 P 0. Use an α = 0.07 Sweetener*Syrup to Water 4. Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef SE Coef T Constant 182.73 Sweetener*Syrup to Water*Temperature 4.15 3-Way Interactions 4 405.771 0.47 Sweetener*Carbonation*Temperature -0.9504 -4.625 213.000 0.688 2.281 0. the reduced model will contain the significant terms (sweetener. as well as lower-order terms included in the significant terms (main effects: syrup to water.16 7.50 462. syrup to water*temperature).344 0.438 1. sweetener*syrup to water.500 28.31 Sweetener -9. select “Terms” and verify that all terms are selected.9504 0.9504 192.10.10 and select terms with P-value less than 0.531 0.36 Sweetener*Temperature -2.53 1. Since there are two replicates of the experiment.125 101.28 3.344 0.13 405.062 -4.49 Syrup to Water*Carbonation* -0.69 199.000 0.015 0. sweetener*temperature.722 0. Select Stat > DOE > Factorial > Analyze Factorial Design.9504 -0. .025 0.64 Residual Error 16 462.218 From magnitude of effects.219 0.9504 2.91 Pure Error 16 462.688 0.313 -0. temperature.629 0.9504 2..187 -2.344 0..41 Temperature 3. two-factor interactions: sweetener*carbonation.844 0.53 47.938 1. type of sweetener is dominant.001 0. Factorial Fit: Total Score versus Sweetener.469 0.89 Sweetener*Syrup to Water*Carbonation -5.048 0.177 0.50 4-Way Interactions 1 47.MTW.388 0.969 0.30 Syrup to Water*Temperature -0.9504 1.9504 -0.379 0.267 0.656 0. The design and data are in the MINITAB worksheet Ex12-4.9504 0.

63 852.969 0.45 Temperature 3.486 0.9244 -1.9244 -4.688 0.688 -1. .29 3-Way Interactions 2 391.9244 0.37 Sweetener*Temperature -2.9244 -2.54 Analysis of Variance for Total Score (coded units) Source DF Seq SS Adj SS Adj MS F Main Effects 4 852.094 0.88 546.15 Residual Error 20 546.000 0.. Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef SE Coef T Constant 182.344 0.714 0.9244 -0.020 P 0.313 -0.09 0.764 0.80 2-Way Interactions 5 176.Chapter 12 Exercise Solutions 12-4 continued Factorial Fit: Total Score versus Sweetener.188 -0.71 Carbonation -2.9244 2.50 462.81 Sweetener*Syrup to Water*Temperature 4.162 0.344 0.046 0.001 0.188 -1.13 Sweetener*Syrup to Water 4.38 1.34 Lack of Fit 4 84.30 Syrup to Water*Temperature -0.9244 -0.585 12-12 .88 27.306 0.06 391.594 0.062 2.005 0.63 213.920 0.011 0.781 0.344 0.18 Syrup to Water*Carbonation -0.06 195.531 0.73 Sweetener -9.563 -0.10 Sweetener*Syrup to Water*Carbonation -5.53 7.062 -4.094 0.91 Total 31 1967.000 0.9244 2.9244 -1.688 2.50 28.040 0.9244 -0.188 -2.73 Pure Error 16 462.281 0.91 176.20 Sweetener*Carbonation 0.9244 197.9244 2.38 84.. Syrup to Water.251 0.031 0.16 7.91 35.38 21.90 Syrup to Water -1.47 P 0.656 0.938 1.

0 Sweetener 0.0 Carbonation 0.5 1.0 -1. Normal Probability Plot of the Residuals (response is Total Score) 99 Percent 90 50 10 1 -10 -5 0 Residual 5 10 Residuals Versus Sweetener Residuals Versus Syrup to Water (response is Total Score) (response is Total Score) 8 Residual Residual 8 0 -8 -8 -1.0 -0.0 8 Residual 8 Residual 0 0 -8 0 -8 -1.5 0.0 -1. To find the residuals.0 Syrup to Water Residuals Versus Carbonation Residuals Versus Temperature (response is Total Score) (response is Total Score) 0. Select “Terms” and verify that all terms for the reduced model are included.0 Temperature There appears to be a slight indication of inequality of variance for sweetener and syrup ratio. select Stat > DOE > Factorial > Analyze Factorial Design.0 -0.Chapter 12 Exercise Solutions 12-5. This is not serious enough to warrant concern.0 0. and then select the variables.0 -0.5 1.5 0.5 0. choose “Normal plot” of residuals and “Residuals versus variables”.5 1.5 0.0 -0. 12-13 . Select “Graphs”.5 1. as well as a slight indication of an outlier.

9244 2.18 Syrup to Water*Carbonation -0.688 2. Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef SE Coef T Constant 182. . Select “Terms” and verify that all terms for the reduced model are selected.594 0. Select Stat > DOE > Factorial > Analyze Factorial Design.10 Sweetener*Syrup to Water*Carbonation -5.781 0.37 Sweetener*Temperature -2. Factorial Fit: Total Score versus Sweetener. and ABD appear to be significant. factors A.344 0.30 Syrup to Water*Temperature -0.9244 -2.000 0.281 0.9244 -1.45 Temperature 3.9244 -4.251 0.920 0..046 0. Also..9244 -0.040 0.938 1. From examination of the above table.764 0.81 Sweetener*Syrup to Water*Temperature 4.344 0. Syrup to Water.20 Sweetener*Carbonation 0.062 2.031 0.9244 0.313 -0. ABC. D.000 0.9244 -0.344 0.90 Syrup to Water -1.9244 -0.011 0.71 Carbonation -2.020 The ratio of the coefficient estimate to the standard error is distributed as t statistic.13 Sweetener*Syrup to Water 4.563 -0.73 Sweetener -9.656 0.188 -0. the factor is not significant.688 -1.188 -2.54 P 0.688 0.094 0. AB.9244 2.531 0.486 0.188 -1.714 0. if the confidence interval includes zero.9244 197. 12-14 .9244 2.162 0.Chapter 12 Exercise Solutions 12-6.9244 -1.062 -4. and a value greater than approximately |2| would be considered significant.094 0.969 0.

MTW. and set alpha to 0. .750 0.500 Sweetener*Syrup to Water*Carbonation -7.000 -1.125 Sweetener*Carbonation*Temperature 0. and then enter the data.125 Syrup to Water*Carbonation* -2.5 12-15 ..10 Factorial Fit: Total Score versus Sweetener. select “Terms” and verify that all terms are selected.875 Carbonation*Temperature … Normal Probability Plot of the Effects (response is Total Score.000 0.125 Syrup to Water*Carbonation -1. Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef Constant 183. Syrup to Water.500 Carbonation*Temperature 1. Create a 24 factorial design in MINITAB.000 2.750 -0. choose the normal effects plot.500 2.000 0.625 Sweetener -10. The design and data are in the MINITAB worksheet Ex12-7.250 0.10) 99 Effect Type Not Significant Significant 95 90 F actor A B C D Percent 80 70 60 50 40 N ame S w eetener S y rup to Water C arbonation Temperature 30 20 10 5 1 A -10 -5 0 Effect 5 10 Lenth's PSE = 4.Chapter 12 Exercise Solutions 12-7..750 Sweetener*Syrup to Water*Temperature 4. Then select “Graphs”.250 Syrup to Water -0.875 Syrup to Water*Temperature -3.250 2. Since there is only one replicate of the experiment.500 -5. Alpha = .000 Sweetener*Carbonation 1.250 -0.750 1.375 Temperature 5.750 Sweetener*Syrup to Water 4.500 Sweetener*Temperature -6.500 -1.125 Carbonation 0.250 Temperature Sweetener*Syrup to Water* 3.500 -3. Select Stat > DOE > Factorial > Analyze Factorial Design.250 -3.

Re-fit and analyze the reduced model.15% R-Sq(adj) = 31.014 S = 7.750 55.5 190.5 0. This is not serious enough to warrant concern.63 Total 15 1219.00 441.000 Sweetener -10.0 182.00 7.45822 R-Sq = 36.5 185.750 55.250 1.63 Pure Error 14 778.865 98.5 1.75 778.Chapter 12 Exercise Solutions 12-7 continued From visual examination of the normal probability plot of effects.014 Normal Probability Plot of the Residuals (responseisTotal Score) 99 Percent 90 50 10 1 -20 -10 0 Residual 10 20 Residuals Versus the FittedValues (responseisTotal Score) Residual 10 0 -10 180.500 -5.0 Sweetener 0.48 0.865 -2.75 778. as well as in the predicted values.75 P 0.0 187. 12-16 . Factorial Fit: Total Score versus Sweetener Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef SE Coef T P Constant 183.93 Residual Error 14 778.0 There appears to be a slight indication of inequality of variance for sweetener.625 1.59% Analysis of Variance for Total Score (coded units) Source DF Seq SS Adj SS Adj MS F Main Effects 1 441.0 -0.0 FittedValue Residuals Versus Sweetener (responseisTotal Score) Residual 10 0 -10 -1.000 441. only factor A (sweetener) is significant.82 0.

or days. 12-9. The ABCD interaction is confounded with blocks. Day 1 a b c abc Day 2 d abd acd bcd (1) ab ac ad bc bd cd abcd Treatment combinations within a day should be run in random order.Chapter 12 Exercise Solutions 12-8. A 25 design in two blocks will lose the ABCDE interaction to blocks. Block 1 (1) ae ab be ac ce bc abce ad de bd abde cd acde abcd bcde Block 2 a e b abe c ace abc bce d ade abd bde acd cde bcd abcde 12-17 .

4275 0.7700 Solv/React 1. and then enter the data.1200 0.2125 0. The design and data are in the MINITAB worksheet Ex12-10. Alpha = .1825 React Purity*React pH 0.1625 0.1462 Cat/React*React Purity 0. select “Terms” and verify that all main effects and interaction effects are selected.MTW. and set alpha to 0. Since there is only one replicate of the experiment.4650 -0.5450 2.3650 -0.0812 Temp*React Purity -0.2725 -0.8475 12-18 .Chapter 12 Exercise Solutions 12-10.0600 Cat/React*React pH 0.1500 0.1062 Normal Probability Plot of the Effects (response is Color.5750 Solv/React*Temp -0.7025 -0.. Factorial Fit: Color versus Solv/React.10.8375 -0.7175 Cat/React -1.4187 Temp*React pH -0.9125 -0. Then select “Graphs”. Select Stat > DOE > Factorial > Analyze Factorial Design.2925 0. .. choose the normal effects plot.4350 0.2300 -0.4562 Solv/React*React Purity -1. Estimated Effects and Coefficients for Color (coded units) Term Effect Coef Constant 2.7325 Temp -0.6150 Solv/React*React pH 0. (a) Create a 25-1 factorial design in MINITAB.2725 React pH -0.2138 Cat/React*Temp 0.1363 React Purity 4. Cat/React.10) 99 Effect Ty pe Not Significant Significant D 95 90 F actor A B C D E Percent 80 70 60 50 40 N ame S olv /React C at/React Temp React P urity React pH 30 20 10 5 1 -2 -1 0 1 2 Effect 3 4 5 Lenth's PSE = 0.3513 Solv/React*Cat/React 1.

5 0.4147 6.03 Residual Error 14 38.68 0.751 Pure Error 14 38.272 0.52 2.000 S = 1.93% Analysis of Variance for Color (coded units) Source DF Seq SS Adj SS Adj MS F Main Effects 1 82.0 Residual plots indicate that there may be problems with both the normality and constant variance assumptions.000 (b) Nor mal P r obability P lot of the Residuals (response is Color) 99 Percent 90 50 10 1 -4 -3 -2 -1 0 Residual 1 2 3 4 Residuals V er sus the Fitted V alues (response is Color) Residual 2 0 -2 0 1 2 3 4 5 Fitted Value Residuals Ver sus React P ur ity (response is Color) Residual 2 0 -2 -1.15 P 0.770 0.751 Total 15 121. 12-19 . Factorial Fit: Color versus React Purity Estimated Effects and Coefficients for Color (coded units) Term Effect Coef SE Coef T P Constant 2.63 82.20% R-Sq(adj) = 65. only factor D (reactant purity) is significant.48 0.000 React Purity 4.628 30.52 2.52 38.0 -0.0 React Purity 0.65876 R-Sq = 68.545 2.5 1.4147 5.Chapter 12 Exercise Solutions 12-10 (a) continued From visual examination of the normal probability plot of effects.52 38.63 82. Re-fit and analyze the reduced model.

425 1 0. A cube plot in these factors shows how the design can be collapsed into a replicated 23 design.385 1. the 2nd and 3rd largest effects in absolute magnitude are A (solvent/reactant) and B (catalyst/reactant). D (reactant purity).865 -2.875 4.Chapter 12 Exercise Solutions 12-10 continued (c) There is only one significant factor.005 5. The highest color scores are at high reactant purity. so this design collapses to a onefactor experiment. Looking at the original normal probability plot of effects and effect estimates. or simply a 2-sample t-test.715 1. the lowest at low reactant purity.795 1 Cat/React 6.865 -1 1 -1 React Purity -1 Solv/React 12-20 . Cube Plot (data means) for Color 3.

675 -0. Factorial Fit: yield versus A:Temp.138 D:Time -0. the complete defining relation is: I = ACE = BDE = ABCD. Since there is only one replicate of the experiment.MTW.275 -0.Chapter 12 Exercise Solutions 12-11.175 -2.275 1.525 -0.913 A:Temp*D:Time -1. then define the experiment using Stat > DOE > Factorial > Define Custom Factorial Design. The aliases are: A*I = A*ACE = A*BDE = A*ABCD ⇒ A = CE = ABDE = BCD B*I = B*ACE = B*BDE = B*ABCD ⇒ B = ABCE = DE = ACD C*I = C*ACE = C*BDE = C*ABCD ⇒ C = AE = BCDE = ABD … AB*I = AB*ACE = AB*BDE = AB*ABCD ⇒ AB = BCE = ADE = CD The remaining aliases are calculated in a similar fashion. Enter the factor levels and yield data into a MINITAB worksheet.762 B:Matl1 -5.587 C:Vol 2.638 … Alias Structure I + A:Temp*C:Vol*E:Matl2 + B:Matl1*D:Time*E:Matl2 + A:Temp*B:Matl1*C:Vol*D:Time A:Temp + C:Vol*E:Matl2 + B:Matl1*C:Vol*D:Time + A:Temp*B:Matl1*D:Time*E:Matl2 B:Matl1 + D:Time*E:Matl2 + A:Temp*C:Vol*D:Time + A:Temp*B:Matl1*C:Vol*E:Matl2 C:Vol + A:Temp*E:Matl2 + A:Temp*B:Matl1*D:Time + B:Matl1*C:Vol*D:Time*E:Matl2 D:Time + B:Matl1*E:Matl2 + A:Temp*B:Matl1*C:Vol + A:Temp*C:Vol*D:Time*E:Matl2 E:Matl2 + A:Temp*C:Vol + B:Matl1*D:Time + A:Temp*B:Matl1*C:Vol*D:Time*E:Matl2 A:Temp*B:Matl1 + C:Vol*D:Time + A:Temp*D:Time*E:Matl2 + B:Matl1*C:Vol*E:Matl2 A:Temp*D:Time + B:Matl1*C:Vol + A:Temp*B:Matl1*E:Matl2 + C:Vol*D:Time*E:Matl2 From the Alias Structure shown in the Session Window.275 1. (a) and (b) Select Stat > DOE > Factorial > Analyze Factorial Design. select “Terms” and verify that all main effects and twofactor interaction effects are selected. D:Time.825 0. C:Vol.238 A:Temp -1. The design and data are in the MINITAB worksheet Ex12-11.337 E:Matl2 2. 12-21 .138 A:Temp*B:Matl1 1. B:Matl1. E:Matl2 Estimated Effects and Coefficients for yield (coded units) Term Effect Coef Constant 19.

826 79.9 -16.5 – 16.8 23.Chapter 12 Exercise Solutions 12-11 continued (c) A B -1 1 1 -1 -1 1 -1 1 C -1 1 -1 1 -1 -1 1 1 D -1 -1 -1 1 1 1 -1 1 E -1 -1 1 -1 1 -1 1 1 1 -1 -1 -1 -1 1 1 1 yield 23.8 18. Factorial Fit: yield versus A:Temp.965 2-Way Interactions 2 9. D:Time.10. Since there is only one replicate of the experiment.2 + 15.2 15.9 16.8 + 18.3) = 1. select “Terms” and verify that all main effects and twofactor interaction effects are selected.1 [A] = A + CE + BCD + ABDE = ¼ (–23.2 +15.2 – 23.2 +23.1) = ¼ (7.739 … F * * P * * 12-22 .913 4.913 9.5 + 16.956 Residual Error 0 * * * Total 7 89. Then select “Graphs”.4 16.4 – 16. The other main effects and interaction effects are calculated in the same way. C:Vol.826 15. (d) Select Stat > DOE > Factorial > Analyze Factorial Design.5 16.525 [AB] = AB + BCE + ADE + CD = ¼ (+23.4 – 16.1) = ¼ (–6.825 This are the same effect estimates provided in the MINITAB output above. E:Matl2 … … Analysis of Variance for yield (coded units) Source DF Seq SS Adj SS Adj MS Main Effects 5 79.9 – 16.8 – 23. and set alpha to 0.1) = –1. B:Matl1.2 23.8 + 23. choose the normal effects plot.8 + 18.

Factorial Fit: yield versus B:Matl1 Estimated Effects and Coefficients for yield (coded units) Term Effect Coef SE Coef T P Constant 19. and select the “Normal plot” and “Residuals versus fits” residual plots. Select Stat > DOE > Factorial > Analyze Factorial Design.0 -2.8682 -2. Select “Terms” and select “B”.56 53. main effect B (amount of material 1) is more than twice as large as the 2nd largest effect (absolute values) and falls far from a line passing through the remaining points.030 Total 7 89.5 0.88 0.56 53.18 6.18 6.5 5.238 0.18 36. Alpha = .000 B:Matl1 -5.0 Lenth's PSE = 2.16 0.10.10) 99 Effect Ty pe Not Significant Significant 95 90 F actor A B C D E Percent 80 70 60 50 40 N ame A :Temp B:M atl1 C :V ol D :Time E :M atl2 30 20 B 10 5 1 -7.025 Residual Error 6 36.030 Pure Error 6 36. Re-fit a reduced model containing only the B main effect.561 8. Then select “Graphs”.7375 Although none of the effects is significant at 0.Chapter 12 Exercise Solutions 12-11 (d) continued Normal Probability Plot of the Effects (response is yield.8682 22.18 36.025 … Analysis of Variance for yield (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 1 53.0 Effect 2.98 0.587 0.74 … 12-23 . and pool the remaining terms to estimate error.5 -5.175 -2.

5 0. The run should be investigated for any issues which occurred while running the experiment.0 Residual plots indicate a potential outlier. If no issues can be identified.0 -2.0 Residual 2.Chapter 12 Exercise Solutions 12-11 continued (e) Residuals Versus the Fitted Values Exercise 12-11 (e) (response is yield) 2 1 Residual 0 -1 -2 -3 -4 -5 16 17 18 19 Fitted Value 20 21 22 Normal Probability Plot of the Residuals Exercise 12-11(e) (response is yield) 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 -5. it may be necessary to make additional experimental runs 12-24 .5 5.

00 -15.79 0.26 0.036 * A*C -2.MTW.75 3.50 -1. Since this is a single replicate of the experiment.50 1.151 Residual Error 5 1250 1250 250.555 B*C -20. The design and data are in the MINITAB worksheet Ex12-12.953 -2.953 -4.50 3. 12-25 .7 2.50 3. Then select “Graphs”.10) 99 95 AB 90 Percent 80 70 60 50 40 30 F actor A B C D N ame A B C D BC 20 C 10 A 5 1 Effect Ty pe Not Significant Significant -5 -4 -3 -2 -1 0 Standardized Effect 1 2 3 The main effects A and C and two two-factor interactions with B (AB. AB.25 3.0 Total 15 15100 Normal Probability Plot of the Standardized Effects (response is Mole Wt.50 -3.95 0.953 1.953 -3.50 11.953 0. choose the normal effects plot. Create a 24 factorial design in MINITAB.005 * B 10. C.953 2. D Estimated Effects and Coefficients for Mole Wt (coded units) Term Effect Coef SE Coef T P Constant 837. C. and then enter the data.765 A*D 5.00 2. select “Terms” and verify that all main effects and twofactor interaction effects are selected.95 0. Re-fit and analyze a reduced model containing A.765 C*D 7.262 C -30.5 9.50 -18.386 A*B 22.00 -10.10.014 2-Way Interactions 6 4000 4000 666. B.74 0.75 3.75 3. Alpha = .Chapter 12 Exercise Solutions 12-12.32 0. B.953 211.953 -0.32 0.25 3.953 0.953 0. The main effect B must be kept in the model to maintain hierarchy.53 0. Factorial Fit: Mole Wt versus A.386 … Analysis of Variance for Mole Wt (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 4 9850 9850 2462. (a) Select Stat > DOE > Factorial > Analyze Factorial Design.85 0.00 3.50 3.013 * D -7.953 -0.67 0.63 0.053 * B*D 2.00 3. and set alpha to 0.25 3.85 0.00 3.00 5.000 A -37. and BC. BC) are significant.87 0.

0 200. 12-26 . Factorial Fit: Mole Wt versus A.0 … The same terms remain significant. AB. C.00 -15.0 125.004 Residual Error 10 1850.000 * B 10.51 0. B.0 9625.015 * … Analysis of Variance for Mole Wt (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 3 9625. B.400 -2.00 3.008 * B*C -20.000 A -37.0 1600.0 3208.5 9.31 0. C Estimated Effects and Coefficients for Mole Wt (coded units) Term Effect Coef SE Coef T P Constant 837.34 0.0 0.80 0.400 1.0 3625. A.3 17.0 1850.25 3. and BC.50 -18.41 0. and select the “Normal plot” and “Residuals versus fits” residual plots.0 185.00 5.94 0.47 0.400 -4.001 * A*B 22.0 Lack of Fit 2 250. C.400 -5.400 246.559 Pure Error 8 1600.00 3. Then select “Graphs”.0 1812.172 C -30.00 3. AB.Chapter 12 Exercise Solutions 12-12 continued (b) Select Stat > DOE > Factorial > Analyze Factorial Design. BC”.30 0.75 3.50 3.50 11. Select “Terms” and select “A.0 250.63 0.400 3.000 2-Way Interactions 2 3625.0 Total 15 15100.00 -10.

12-27 . however neither plot reveals a major problem with the normality and constant variance assumptions.Chapter 12 Exercise Solutions 12-12 continued (c) Normal Probability Plot of the Residuals (response is Mole Wt) 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 -30 -20 -10 0 Residual 10 20 30 Residuals Versus the Fitted Values (response is Mole Wt) 20 Residual 10 0 -10 -20 -30 790 800 810 820 830 840 Fitted Value 850 860 870 880 A “modest” outlier appears on both plots.

all three (curvature. To summarize MINITAB’s functionality. Also. Select “Terms” and verify that all main effects and two-factor interactions are selected. regardless of whether center points are included in the model. the main and interaction effects are tested for significance against the correct residual error (lack of fit + curvature + pure error). Assuming that lack of fit and curvature tests are not significant. This will ensure that if both lack of fit and curvature are not significant.Chapter 12 Exercise Solutions 12-13. (a) Select Stat > DOE > Factorial > Analyze Factorial Design.MTW. See the dialog box below. lack of fit. Create a 24 factorial design with four center points in MINITAB. 12-28 . curvature is always tested against pure error and lack of fit (if available). DO NOT include the center points in the model (uncheck the default selection). and then enter the data. The inclusion/exclusion of center points in the model affects the total residual error used to test significance of effects. The design and data are in the MINITAB worksheet Ex12-13. and pure error) should be included in the residual mean square.

50 -18.57 0.822 Residual Error 9 13070 13070 1452.898 A*D 5. the first test to consider is the “lack of fit” test.60 0.25 9.898 C*D 7.Chapter 12 Exercise Solutions 12-13 (a) continued When looking at results in the ANOVA table.52 0. and lack of fit (if available) and use this as the basis for testing for significant effects.527 -1.5 1. B.13 0.50 9.25 9.004).70 0. D Estimated Effects and Coefficients for Mole Wt (coded units) Term Effect Coef SE Coef T P Constant 848.26 0.13 0. pure error.00 9.46 0.18 0.97 0.0 Total 19 26920 … (b) The test for curvature is significant (P-value = 0.799 B*C -20.081 B 10.00 8. 12-29 . C.527 0. the three-factor and four-factor interactions).527 0.00 9.000 A -37. which is a test of significance for the pure quadratic terms.521 99.50 -1.527 0. If tests for both lack of fit and curvature are not significant.321 B*D 2.2 Curvature 1 8820 8820 8820.234 2-Way Interactions 6 4000 4000 666.50 -3. (In MINITAB.527 0. no further statistical analysis should be performed because the model is inadequate.75 9. If this test is significant.50 3.703 A*B 22.50 1. this is accomplished by not including center points in the model. and some terms need to be added to the model.00 -10. If lack of fit is not significant.39 0.268 A*C -2.00 2.) Factorial Fit: Mole Wt versus A.915 Pure Error 3 3000 3000 1000.75 9.00 5.52 0.0 0. If lack of fit is significant. then it is reasonable to pool the curvature.527 1.703 … Analysis of Variance for Mole Wt (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 4 9850 9850 2462. This approach to sequential experimentation is presented in Chapter 13.25 9.0 16.75 9.25 0.150 D -7.05 0.00 -15.527 -0. a better strategy is to add runs that would enable estimation of the quadratic effects.00 9. Although one could pick a “winning combination” from the experimental runs.527 -1. the model is not correctly specified.004 * Lack of Fit 5 1250 1250 250.527 -1.7 0.612 C -30.50 11.39 0.527 -0. which is a test of significance for terms not included in the model (in this exercise. the next test to consider is the “curvature” test.

Chapter 12 Exercise Solutions 12-14. From Table 12-23 in the textbook. and at least some 2-factor interactions are aliased with each other. A lower resolution design would have some 2-factor interactions and main effects aliased together. this problem is of concern. so this is a resolution IV design. Main effects are clear of 2-factor interactions. a 28IV− 4 design has a complete defining relation of: I = BCDE = ACDF = ABCG = ABDH = ABEF = ADEG = ACEH = BDFG = BCFH = CDGH = CEFG = DEFH = AFGH = ABCDEFGH The runs would be: Run A 1 – 2 + 3 – 4 + 5 – 6 + 7 – 8 + 9 – 10 + 11 – 12 + 13 – 14 + 15 – 16 + B – – + + – – + + – – + + – – + + C – – – – + + + + – – – – + + + + D – – – – – – – – + + + + + + + + E=BCD – – + + + + – – + + – – – – + + F=ACD – + – + + – + – + – + – – + – + G=ABC – + + – + – – + – + + – + – – + H=ABD – + + – – + + – + – – + + – – + A=BCDE=CDF=BCG=BDH=BEF=DEG=CEH=ABDFG=ACDGH=ABCFH=ACEFG=ADEFH=FGH=BCDEFGH B=CDE=ACDF=ACG=ADH=AEF=ABDEG=ABCEH=DFG=CFH=BCDGH=BCEFG=BDEFH=ABFGH=ACDEFGH C=BDE=ADF=ABG=ABDH=ABCEF=ACDEF=AEH=BCDFG=BFH=DGH=EFG=CDEFH=ACFGH=ABDEFGH D=BCE=ACF=ABCG=ABH=ABDEF=AEG=ACDEH=BFG=BCDFH=CGH=CDEFG=EFH=ADFGH=ABCEFGH E=BCD=ACDEF=ABCEG=ABDEH=ABF=ADG=ACH=BDEFG=BCEFH=CDEGH=CFG=DFH=AEFGH=ABCDFGH F=BCDEF=ACD=ABCFG=ABDFH=ABE=ADEFG=ACEFH=BDE=BCH=CDFGH=CEG=DEH=AGH=ABCDEGH G=BCDEG=ACDFG=ABC=ABDGH=ABEFG=ADE=ACEGH=BDF=BFGH=CDH=CEF=DEFGH=AFH=ABCDEFH H=BCDEH=ACDFH=ABCGH=ABD=ABEFH=ADEGH= ACE=BDFGH=BCF=CDG=CEFGH=DEF=AFG=ABCDEFG AB=ACDE=BCDF=CG=DH=EF=BDEG=BCEH=ADFG=ACFH=ABCDFH=ABCEFG=ABDEFH=BFGH=CDEFGH AC=ABDE=DF=BG=BCDH=BCEF=CDEG=EH=ABCDFG=ABFH=ADGH=AEFG=ACDEFH=CFGH=BDEFGH etc. Since significant 2-factor interactions often occur in practice. The source of interest for any combined main and 2-factor interaction effect would be in question. 12-30 .

000 -1.000 0.31 0.100 0.MTW.260 … Analysis of Variance for Resist (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 4 17555.7621 0.33 Total 11 28184.510 A*C 72.10) 99 Effect Ty pe Not Significant Significant 95 C 90 Percent 80 F actor A B C D AC 70 A 60 50 40 N ame A B C D 30 20 10 5 1 0 10 20 30 40 Standardized Effect 50 60 12-31 .65 Curvature 1 5.190 A*B 1. Enter the factor levels and resist data into a MINITAB worksheet. (a) Select Stat > DOE > Factorial > Analyze Factorial Design.500 -0.300 0.200 0. The design and data are in the MINITAB worksheet Ex12-15.700 23.6 5. 0 = center point).0 Normal Probability Plot of the Standardized Effects (response is Resist.250 0.1 10610.57 0.51 0.7621 31. choose the normal effects plot.3 4388.000 Residual Error 4 18.550 0.7621 -1.3 17555.433 0. Then define the experiment using Stat > DOE > Factorial > Define Custom Factorial Design. including a column indicating whether a run is a center point run (1 = not center point.7621 -1.600 40.1 3536.Chapter 12 Exercise Solutions 12-15. Alpha = .000 2-Way Interactions 3 10610.13 0. C.7621 52.7621 47. and set alpha to 0.6 5.000 * D -2.30 0.7621 -0.0 4.850 0.12 0.400 -1.000 * B -0. Also.800 36.83 944.6 18. DO NOT include the center points in the model (uncheck the default selection).000 A 47.88 0.338 Pure Error 3 13.759 C 80. B.10. Select “Terms” and verify that all main effects and two-factor interactions are selected. D Estimated Effects and Coefficients for Resist (coded units) Term Effect Coef SE Coef T P Constant 60. Then select “Graphs”.61 1.70 761.76 0. Factorial Fit: Resist versus A.000 * A*D -2.33 0.6223 97.29 0.72 0.0 13.6 4.400 0.

0 41.6 1710.7 2066.1 Curvature 1 5.6 5.1 Total 11 28184.6 5.70 23. Select Stat > DOE > Factorial > Analyze Factorial Design.43 0.11 0.30 0.000 A 47.43 0. Re-fit and analyze a reduced model containing A. and AC.60 40.8007 45.44 0.327).89 0.85 0.000 2-Way Interactions 1 10599.7 10599. Then select “Graphs”. C.0 Curvature is not significant (P-value = 0.8007 50.80 36.000 * A*C 72.327 Pure Error 7 35. the main effects A and C and their twofactor interaction (AC) are significant.40 0. C Estimated Effects and Coefficients for Resist (coded units) Term Effect Coef SE Coef T P Constant 60. AC”.Chapter 12 Exercise Solutions 12-15 continued Examining the normal probability plot of effects.8007 29. Select “Terms” and select “A.7 10599. so continue with analysis.000 * C 80.4 35.79 0. C. (c) Residuals Versus the Fitted Values (response is Resist) 3 2 Residual 1 0 -1 -2 -3 0 20 40 60 80 100 Fitted Value 120 140 160 180 A funnel pattern at the low value and an overall lack of consistent width suggest a problem with equal variance across the prediction range.000 Residual Error 8 41.3 17543.4 5. and select the “Normal plot” and “Residuals versus fits” residual plots.000 * … Analysis of Variance for Resist (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 2 17543.0 5.3 8771.6 1.33 0.6537 92. 12-32 .46 0. (b) Factorial Fit: Resist versus A.

12-33 .0 -2.Chapter 12 Exercise Solutions 12-15 continued (d) Normal Probability Plot of the Residuals (response is Resist) 99 95 90 Percent 80 70 60 50 40 30 20 10 5 1 -5. The concern with variance in the predicted resistivity indicates that a data transformation may be needed.5 0.5 5.0 The normal probability plot of residuals is satisfactory.0 Residual 2.

70 70 . 13-1.1 ≤ x2 ≤ 1 x2 6 = = 0.85 > 85 Ex13-1x2 0. (a) Graph > Contour Plot Contour Plot of Ex13-1y vs Ex13-1x2. the initial experimental layout must be created in MINITAB or defined by the user.5 0.5 -1.80 80 .0 − 1 ≤ x1 ≤ 1.75 75 .0 Ex13-1y < 60 60 . Ex13-1x1 y = 75 + 10x1 + 6x2 1. therefore some information required by MINITAB is not included.6 13-1 .5 0.0 (b) yˆ = 75 + 10 x1 + 6 x2 -0. The Excel data sets contain only the data given in the textbook.6 x1 10 ∆x1 = 1 ∆x2 = 0. The MINITAB instructions provided for the factorial designs in Chapter 12 are similar to those for response surface designs in this Chapter.65 65 .Chapter 13 Exercise Solutions Note: To analyze an experiment in MINITAB.5 1.0 -1.0 Ex13-1x1 0.0 -0.

and set ∆x2 = 1.0 13-3.5.13 βˆ ∆x −15 1. 13-2 .0 2 βˆ1 ∆x1 = = = −0.3 select x2 with largest absolute coefficient.Chapter 13 Exercise Solutions 13-2. (a) This design is a CCD with k = 2 and α = 1. The design is not rotatable.0 2 ∆x3 = 2 βˆ 3 βˆ2 ∆x2 = 3 = −0. βˆ2 = −15. i = 1.20 −15 1. 2. yˆ = 50 + 2 x1 − 15 x2 + 3 x3 − 1 ≤ xi ≤ +1.

8682 x1 -58.5 83.000 Square 2 1647.85 0.314 0. The design and data are in the MINITAB worksheet Ex13-3.017 x2*x2 15.5 15.1 173.2 28934.3 0.461 2.0 1647.471 x1*x1 -24.769 0.285 -0. two-factor interactions. Select “Terms” and verify that all main effects.5 5.3 2. including a column indicating whether a run is a center point run (1 = not center point.164 0.088 0.704 0.991 Pure Error 3 486.8546 x2*x2 6.688 10.7 73.0 486.09 0.0 Total 11 31084.555 35.Chapter 13 Exercise Solutions 13-3 continued (b) Enter the factor levels and response data into a MINITAB worksheet.9231 x1*x2 -0.4 30583.875 Residual Error 6 501.082 x1*x2 -1. 0 = center point).4118 x1*x1 -10.461 -3.3 2.9 … Estimated Regression Coefficients for y using data in uncoded units Term Coef Constant 160.618 4. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 160.273 0.441 4.5 9.6 Lack-of-Fit 3 15.000 Linear 2 28934.590 0.2941 x2 2.000 x1 -87.875 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS F P Regression 5 30583.423 7.868 4.2 0.000 x2 3.704 -18. Response Surface Regression: y versus x1. Select Stat > DOE > Response Surface > Analyze Response Surface Design.18 0.MTW.7500 13-3 .4 6116.5 501.03 0. and quadratic terms are selected.03 0.013 Interaction 1 2.577 7. x2 The analysis was done using coded units. Then define the experiment using Stat > DOE > Response Surface > Define Custom Response Surface Design.0 162.2 14467.0 823.

5.0 -0.Chapter 13 Exercise Solutions 13-3 continued (c) Stat > DOE > Response Surface > Contour/Surface Plots Contour Plot of y vs x2.6101 0. such as cost. A flag is planted at one option on the contour plot above.22) + 30 = 26.50) + 750 = 825 Time = 15 x2 + 30 = 15(−0. x1 Surface Plot of y vs x2.5 x1 = 1. x1 1.5 40 50 75 100 125 150 175 200 y < > 40 50 75 100 125 150 175 200 225 225 -1.217615 y = 49.5 -1 0 x1 1 x1 0 x2 -1 1 From visual examination of the contour and surface plots.7 13-4 .49384 x2 = -0. (d) Temp = 50 x1 + 750 = 50(+1.5 and letting x2 (temperature) range from −1.471). it appears that minimum purity can be achieved by setting x1 (time) = +1.0 240 180 y 120 60 1 0 -1 -1.5 to + 1. The level for temperature could be established based on other considerations.5 1.0 x2 0. The range for x2 agrees with the ANOVA results indicating that it is statistically insignificant (P-value = 0.

0 + 1.569 13-5 .58 58 .60 60 .70 > 70 Ex13-4x2 1 0 -1 -2 -2 -1 0 Ex13-4x1 1 2 yˆ max = 70.885)] (−2.885 x2 = [−1.1 − 0.64 64 .1.1x2 − 1x12 − 1.3(0.4 x2 + 0.68 68 .3x1x2 2 Ex13-4y < 56 56 .6x1 + 1.6 (b) ∂ yˆ ∂ (69.4) = 0.62 62 .3x1 x2 ) = ⇒0 ∂ x1 ∂ x1 = 1.0 + 1. Graph > Contour Plot Contour Plot of Ex13-4y vs Ex13-4x2.Chapter 13 Exercise Solutions 13-4.012 at x1 ≈ +0.3 x1 = 0 ∂ x2 x1 = −13.2x2^2 + 0.6 x1 + 1.2 x22 + 0.1x2 . Ex13-4x1 y = 69.1x1^2 .66 66 .3 x2 = 0 ∂ yˆ = 1. x2 ≈ +0.9.7) = 0.9 (−15.1 − 2.6 − 2 x1 + 0.

The worksheet is in run order.54 0. Select “Designs”. (a) The design is a CCD with k = 2 and α = 1.291 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS F P Regression 5 2.00781 108.01210 0.4071 0. The design is rotatable.1249 0.132 0.00807 Total 12 2. To analyze the experiment. then enter the data.889 0.Chapter 13 Exercise Solutions 13-5. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 13. central composite design. Response Surface Regression: y versus x1.142 0.000 Linear 2 2.06499 0.2980 0. x2 The analysis was done using coded units. The design and data are in the MINITAB worksheet Ex13-5.01210 1. Select Stat > DOE > Response Surface > Create Response Surface Design.000 x1*x1 -0.0790 x1*x2 0.19 0.703 0. (b) Since the standard order is provided. Select “Terms” and verify that a full quadratic model (A.01090 1.377 Pure Error 4 0.000 Square 2 0.03228 0.16128 0.22628 … Estimated Regression Coefficients for y using data in uncoded units Term Coef Constant 13. B. AB) is selected.0550 13-6 .43226 46.06499 0.06678 7. B2.01563 1.7273 0.012 x2*x2 -0.291 Residual Error 7 0.35 0.01563 2.03228 0.03271 0.03706 -2.4. Leave the design type as a 2-factor. A2.020 Interaction 1 0.0550 0.56 0.MTW.070 x1*x2 0. and enter a custom alpha value of exactly 1.000 x1 0. highlight the design with five center points (13 runs).03271 0.30 0.04309 318.03424 8.00928 Lack-of-Fit 3 0.03706 -3.13355 0.04818 1. to change to standard order (and ease data entry) select Stat > DOE > Display Design and choose standard order.000 x2 -0. select Stat > DOE > Response Surface > Analyze Response Surface Design.4 (the rotatable design is α = 1.580 0.16128 2.13355 0. one approach to solving this exercise is to create a two-factor response surface design in MINITAB.03424 -11.0790 0.41421).7273 x1 0.371 0.1249 x2*x2 -0.01210 0.4071 x1*x1 -0.2980 x2 -0.

00 12.Chapter 13 Exercise Solutions 13-5 (b) continued Values of x1 and x2 maximizing the Mooney viscosity can be found from visual examination of the contour and surface plots.00 13.4 will maximize viscosity.5 y 13.75 14.5 -1.4 and setting x2 between -1 and -1.75 14.25 14.00 14.50 12.25 13.0 x1 0 x2 -1 1 Stat > DOE > Response Surface > Response Optimizer In Setup.0 x1 0.5 1.50 13. setting x1 in a range from 0 to +1.75 13. or using MINITAB’s Response Optimizer.0 13. x1 0.25 12.5 x2 Surface Plot of y vs x2.0 -0. x1 1.25 14.50 13. and Weight = 7.00 12.5 0.5 1 0 -1 -1. From the plots and the optimizer.00 13. let Goal = maximize. Stat > DOE > Response Surface > Contour/Surface Plots Contour Plot of y vs x2.0 12.0 y < > 12. 13-7 . Lower = 10. Target = 20.25 13.50 12.0 12.25 12.00 0.75 13.0 -0.

0283 x1*x3 75. and select the number of factors as “3”.76 8.030 0.03 21.4656 x1*x1 32.74 0.000 Linear 3 1090558 1090558 363519 62.5833 13-8 .064 … Analysis of Variance for y1 Source DF Seq SS Adj SS Adj MS F P Regression 9 1248237 1248237 138693 23.58 21.47 17.0011 x2 109.317 x2*x2 -22.94 9.099 0.MTW.00 17.0056 x2*x2 -22.453 0.06 31.000 x3 131. x2.000 Square 3 14219 14219 4740 0.000 x2 109. Select “Terms” and verify that a full quadratic model is selected.008 x1*x3 75.43 17. Select “Designs” to establish three levels for each factor. (a) To analyze the experiment. Change the design type to a general full factorial design.4256 x3 131.005 0. Since the runs are listed in a patterned (but not standard) order.97 3. The design is a full factorial of three factors at three levels.38 31.6237 x1 177. The design and data are in the MINITAB worksheet Ex13-6.62 38.01 31.720 0.001 Residual Error 17 98498 98498 5794 Total 26 1346735 … Estimated Regression Coefficients for y1 using data in uncoded units Term Coef Constant 327. one approach to solving this exercise is to create a general full factorial design in MINITAB. Response Surface Regression: y1 versus x1. then select “Factors” to specify the actual level values.328 0. it must also be defined using Stat > DOE > Response Surface > Define Custom Response Surface Design.000 x1*x1 32. x3 The analysis was done using coded units.0578 x1*x2 66. Estimated Regression Coefficients for y1 Term Coef SE Coef T P Constant 327.Chapter 13 Exercise Solutions 13-6.4708 x2*x3 43.82 0.3844 x3*x3 -29.94 0.435 0.000 x1 177.47 21.25 0. In order to analyze this experiment using the Response Surface functionality.502 Interaction 3 143461 143461 47820 8.94 6. and then enter the data.97 3.983 0.94 7.481 x3*x3 -29.935 0.363 x1*x2 66.08 -0. Select Stat > DOE > Factoriall > Create Factorial Design.97 1. select Stat > DOE > Response Surface > Analyze Response Surface Design.003 x2*x3 43.08 -0.866 0.08 1.

012 x1*x1 4.3 21957.89 0.890 22. x2.192 10.528 10.361 x1*x2 7.89 -0. select Stat > DOE > Response Surface > Analyze Response Surface Design.1978 x2*x2 -1.3 7319.198 17.2 32650.136 x1 11.Chapter 13 Exercise Solutions 13-6 continued (b) To analyze the experiment.0 1135.7 3018.404 0.5278 x2 15.7 27170.33 1.59 P 0.319 17.116 0.65 0. Estimated Regression Coefficients for y2 Term Coef SE Coef T P Constant 34.8896 x1 11.09 Square 3 1805.281 … Analysis of Variance for y2 Source DF Seq SS Adj SS Adj MS Regression 9 27170.082 12.323 10.65 1.610 0.1083 x2*x3 14. x3 The analysis was done using coded units.1917 x1*x1 4.89 0.719 12.280 x2 15.156 x3 29.550 x1*x3 5.5 601.0 3408.815 0.942 x3*x3 16.235 0.7192 x1*x3 5.691 x2*x3 14.99 Residual Error 17 32650.938 0. Select “Terms” and verify that a full quadratic model is selected.0825 F 1.33 1.629 data in uncoded units 13-9 .5 1805.108 12.113 0.33 2.31 0.97 Linear 3 21957.826 0.9 … Estimated Regression Coefficients for y2 using Term Coef Constant 34. Response Surface Regression: y2 versus x1.483 0.564 0.202 0.31 1.2 1920.3189 x3*x3 16.779 17.81 0.57 3.7794 x1*x2 7.65 0.82 Interaction 3 3408.60 Total 26 59820.030 0.817 x2*x2 -1.074 0.3233 x3 29.

0 x1 0.0 x1 0. y2 1. select the first two factors x1 and x2.5 0. Stat > DOE > Response Surface > Overlaid Contour Plot After selecting the responses.Chapter 13 Exercise Solutions 13-6 continued (c) Both overlaid contour plots and the response optimizer can be used to identify settings to achieve both objectives.0 13-10 . y2 1.5 x2 Hold Values x3 -1 0.0 -0.0 -1.0 -0. set low = 400 and high = 600. Since the goal is to hold y1 (resistivity) at 500.5 -1. Overlaid Contour Plot of y1.5 1. The goal is to minimize y2 (standard deviation) set low = 0 (the minimum of the observed results) and high = 80 (the 3rd quartile of the observed results).0 -0.5 0.5 1.0 Overlaid Contour Plot of y1.5 -1.0 y1 400 600 y2 0 80 0.5 x2 Hold Values x3 0 0.0 -1.0 -0.0 y1 400 600 y2 0 80 0. Select “Contours” to establish the low and high contours for both y1 and y2.

0 -0. For y2.5 -1.Chapter 13 Exercise Solutions 13-6 (c) continued Overlaid Contour Plot of y1.0. x2 = 0. 13-11 . set Goal = Minimize. for y1 set Goal = Target. At x1 = 1.0 -0.0 -1.75.5 1.5 0.0 Stat > DOE > Response Surface > Response Optimizer In Setup. y2 1.3 and x3 = -0. Target = 0.4.0 y1 400 600 y2 0 80 0. Leave all Weight and Importance values at 1. the predicted resistivity mean is 495.5 x2 Hold Values x3 1 0.16 and standard deviation is 44. The graph below represents one possible solution. Upper = 600. Lower = 400. Target = 500. and Upper = 80.0 x1 0.

D. E … Alias Structure I + A*B*C*D A + B*C*D B + A*C*D C + A*B*D D + A*B*C E + A*B*C*D*E A*B + C*D A*C + B*D A*D + B*C A*E + B*C*D*E B*E + A*C*D*E C*E + A*B*D*E D*E + A*B*C*E 13-12 .MTW. B. A+BCD B+ACD C+ABD D+ABC E AB+CD AC+BD AD+BC AE+BCDE BE+ACDE CE+ABDE DE+ABCE ABE+CDE ACE+BDE ADE+BCE This is a resolution IV design. All main effects are clear of 2-factor interactions. Enter the factor levels and response data into a MINITAB worksheet. Stat > DOE > Factorial > Analyze Factorial Design Factorial Fit: Mean versus A. but some 2-factor interactions are aliased with each other. (a) The defining relation for this half-fraction design is I = ABCD (from examination of the plus and minus signs).Chapter 13 Exercise Solutions 13-7. and then define the experiment using Stat > DOE > Factorial > Define Custom Factorial Design. The design and data are in the MINITAB worksheet Ex13-7. C.

62707 0. E Estimated Effects and Coefficients for Height (coded units) Term Effect Coef SE Coef T P Constant 7.0912 0.04726 0. C.02021 -5.62707 0. B.07 0.000 B*E 0.2806 0.05 0.2806 0.0010 0.0012 0.0765 0.000 2-Way Interactions 7 0.0148 0.03 0.05400 2.0456 0.01960 Pure Error 32 0.2421 0.89078 The reduced model for mean: Factorial Fit: Height versus A.0229 -0.000 2-Way Interactions 1 0.0196 0.Chapter 13 Exercise Solutions 13-7 continued (b) The full model for mean: Stat > DOE > Factorial > Analyze Factorial Design Factorial Fit: Height versus A.469 A*C 0.0396 0.150 … Analysis of Variance for Height (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 5 1. D.28060 14.6256 0.000 A*B -0.000 … Analysis of Variance for Height (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 4 1.02021 0.02021 0.1742 0.001 C*E -0.02021 -0.0819 0.1742 0.0021 0.58 0.41 0.229 D 0.78 0.421 D*E 0.2387 -0.57 0.0248 0.0456 0.02021 5.48 0.71 0.0296 -0.0596 -0.04726 0.0819 0.6256 0.000 Residual Error 42 0.0496 -0.02021 -4.02021 -1.0913 0.02021 377.62707 0.83846 0.01994 -5.2421 0.89 0.23 0.29 0.124 B*E 0.0319 0.01994 382.0098 0.05 0.0165 0.84 0.71 0.11 0.501 Residual Error 32 0.37800 0.1638 -0.8908 13-13 .000 B -0.01994 2.01960 Total 47 2.1638 -0.554 Pure Error 32 0.2387 -0.02021 0.000 A 0.8090 0.02021 -0.01994 6.81 0.023 3-Way Interactions 3 0.1529 0.000 C -0.0298 0.01742 0.1194 0.6271 0.031 E -0.01908 Lack of Fit 10 0.80 0.959 A*C*E 0.37800 0.01994 -4.976 A*D -0.36769 18.62707 0. D.02021 -0.01960 Total 47 2.1210 0.02021 0.000 B -0.76 0.83846 1.8090 1.45224 23.1194 0.0006 0.1529 0.01994 3.6271 0.91 0.0198 0.335 A*B*E 0.47 0.027 E -0.73 0.76 0.02021 3.26 0.0115 0.575 A*E 0.02021 -1. E Estimated Effects and Coefficients for Height (coded units) Term Effect Coef SE Coef T P Constant 7.631 A*D*E -0.01575 0.0637 0.99 0.0329 -0.02021 1.8012 0.02021 2.98 0.0765 0. B.000 A 0.51 0.000 D 0.8012 0.99 0.1210 0.

07701 0.01688 60 50 0.01625 13.077006 18.03380 0.13875 0.01625 0.096 E -0.01812 40 -0.01313 A DE 95 0.23 0.02125 -0.00687 80 0. Alpha = .683 C*E -0.074256 17.00687 0.19 0.01625 1.04375 0.003 … Analysis of Variance for Range (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 5 0.04875 0. C.05688 0.11375 0.443 D 0.03625 0.01063 0.06312 0.003 3-Way Interactions 1 0.31909 … 13-14 .003 A*D*E 0.06937 0.06125 0.21937 0.03375 -0.13875 0.03380 0.01625 -3.00188 30 20 0.13625 -0.07426 0.58 0. D.000 A 0.01562 1 0. C.00812 B 10 -0.004225 Total 15 0.05688 (response is Range.42 0.01625 -0.21937 Normal Probability Plot of the Effects 0.15 -0. E Effects and Coefficients for Range (coded units) Effect Coef 0.03062 0.12625 -0.34 0.00 0.81 0.07701 0. D.06937 Percent Estimated Term Constant A B C D E A*B A*C A*D A*E B*E C*E D*E A*B*E A*C*E A*D*E Effect Ty pe Not Significant Significant F actor A B C D E N ame A B C D E Lenth's PSE = 0.03062 90 A -0.01313 0. E Estimated Effects and Coefficients for Range (coded units) Term Effect Coef SE Coef T P Constant 0.01625 3.011 2-Way Interactions 1 0.01375 -0.008 B -0. B.15 Effect 0.20) -0.00375 -0.12625 -0.05 0.01625 0.88 0.026806 6.050625 The reduced model for range: Factorial Fit: Range versus A.10 -0.05 0.50 0.02625 0.003 Residual Error 8 0.06125 0.01625 -4.13403 0.50 0.13625 -0.88 0.06812 0.Chapter 13 Exercise Solutions 13-7 continued (c) The full model for range: Factorial Fit: Range versus A.02625 0.01625 4.06312 99 0. B.02437 -0.07426 0.27 0.10 0.11375 0.13403 0.06812 5 CE -0.02188 70 -0.01375 -0.005 C 0.03125 0.

along with factors B (heating time) and A (furnace temperature) are significant.03536 0.007559 -4.023369 0.20) 0.06 0.07149 -0.00684 -0.11744 (response is StdDev.54 0.00528 0.36 3-Way Interactions 1 0.04 0.03574 0. B.73 0.Chapter 13 Exercise Solutions 13-7 (c) continued The full model for standard deviation: Factorial Fit: StdDev versus A.00770 70 60 -0.06259 0.06259 0. interactions CE (transfer time × quench oil temperature) and ADE=BCE.001 A*D*E 0.00877 0.04 -0.047 E -0.03536 0.07643 0.73 0.00342 D 80 0.001 … Analysis of Variance for StdDev (coded units) Source DF Seq SS Adj SS Adj MS F Main Effects 5 0.041748 0.01906 0.34 0.07148 -0. B.504 D 0.003 B -0. E Effects and Coefficients for StdDev (coded units) Effect Coef Normal Probability Plot of the Effects 0.01093 50 40 0.00528 A DE 95 0.02 0.03822 0.01540 0.00953 30 -0.03129 0.70 0.00165 20 0. C.06 0.06 -0.007559 2.01057 0.004 0.001 0.007559 5.56 Residual Error 8 0. Factors C and E are included to keep the models hierarchical.08 0.0009142 Total 15 0.007559 -4. D. Alpha = .001 For both models of variability. 13-15 .00234 0.00438 CE 10 -0. D.03574 0.00684 -0.45 0.00342 0.03822 Percent Estimated Term Constant A B C D E A*B A*C A*D A*E B*E C*E D*E A*B*E A*C*E A*D*E Effect Ty pe Not Significant Significant F actor A B C D E N ame A B C D E Lenth's PSE = 0.0232179 The reduced model for standard deviation: Factorial Fit: StdDev versus A. E Estimated Effects and Coefficients for StdDev (coded units) Term Effect Coef SE Coef T P Constant 0.02 0.00468 -0.020438 0.023369 0.007559 4.001 C 0.041748 0.007559 0.03574 5 B -0.0233690 25.007314 0.00778 1 -0.13 2-Way Interactions 1 0.07148 -0.01997 0.07149 -0.14 0.00329 -0.01556 0.01768 0.07643 0.11744 0.663 C*E -0.020438 0. C.02185 -0.03129 99 -0.03574 0.007314 0.00 0.08 -0.00999 Effect 0.01768 90 A -0.007559 15.000 A 0.007559 -0.092869 P 0.01057 0.0083496 9.0204385 22.

5 1.0 -0.0 0.4 Residuals Versus the Order of the Data 1 0 -1 1 2 3 4 5 6 7 8 (response is Range) -1.0 -0.5 1.0 0.0 -1.5 1.0 0.5 Residuals Versus C (response is Height) 2 -1.5 1.0 -0.5 2 0 -2 1.0 -1.0 -0.0 C Residuals Versus E (response is Height) (response is Height) Standardized Residual Standardized Residual 1.0 2 -1.0 A Residuals Versus C (response is Range) (response is Range) 0.0 A 0.5 Standardized Residual Residuals Versus A Residuals Versus the Fitted Values (response is Height) 2 0 -2 7.5 1.5 0.5 -1.5 1.0 1.0 (response is Height) 0 -0.0 1.5 1.0 0 Residuals Versus D 0 -2 -0.5 0.5 0.0 C Residuals Versus E (response is Range) (response is Range) Standardized Residual Residuals Versus D 1 0 -1 -0.5 -1.0 E Standardized Residual Standardized Residual For range: Residual Plots for Range Histogram of the Residuals 3.5 1 Observation Order 1 -1.0 -0.0 B 0.0 0 9 10 11 12 13 14 15 16 0 -0.5 0.5 0.0 D 0.5 Residuals Versus B Standardized Residual Standardized Residual -2 0.5 0.0 D 0.5 0.0 0.0 -0.0 -1.00 Residuals Versus the Order of the Data 2 0 Standardized Residual Percent Normal Probability Plot of the Residuals 99 90 50 10 1 2 0 -2 -2 1 5 -1.0 -2 -1.0 1.0 10 15 20 25 30 35 40 45 Observation Order -0.0 1.0 -0.2 Fitted Value 0.0 1 0 -1 -1.0 -1.5 2 1.5 0.5 -1 Residuals Versus B -1 -1.0 0.5 0.0 B 0.5 0.75 Fitted Value 8.5 Standardized Residual 0.50 7.5 0.0 0.Chapter 13 Exercise Solutions (d) Standardized Residual Standardized Residual For mean height: Residual Plots for Height -2 0 Standardized Residual 2 Frequency Histogram of the Residuals 10 5 0 -2.5 1 0 -1 1.0 0.0 Standardized Residual 0 2 Standardized Residual 1 0 -1 Standardized Residual -2 Residuals Versus A Residuals Versus the Fitted Values Standardized Residual Standardized Residual Frequency Percent Normal Probability Plot of the Residuals 99 90 50 10 1 0.0 E 13-16 .

0 -0.0 D 1.0 1.5 0. (e) This is not the best 16-run design for five factors.0 -0. Plots of residuals versus each factor shows that variance is less at low level of factor E.0 2 0 -2 -1. Range Plot of residuals versus predicted shows that variance is approximately constant over range of predicted values. A resolution V design can be generated with E = ± ABCD.5 0.5 1.5 -1.5 0. Normal probability plot of residuals support normality assumption.0 Residuals Versus A Residuals Versus the Fitted Values (response is StdDev) 2 0 -2 0.0 Residuals Versus E (response is StdDev) Standardized Residual Standardized Residual 0.5 -2 (response is StdDev) 0 -2 0.5 Standardized Residual 2.2 Fitted Value Residuals Versus the Order of the Data 2 0 -2 1 2 3 4 5 6 7 8 Standardized Residual Frequency Percent Normal Probability Plot of the Residuals 99 90 50 10 1 2 0 -2 -1.1 0. Plots of residuals versus each factor indicate that the variance may be different at different levels of factor D. Residuals normal probability plot indicate normality assumption is reasonable Plots of residuals versus each factor indicate that the variance may be different at different levels of factor D.0 B 0.5 -1.0 (response is StdDev) 0 -0.0 2 -1.0 E Mean Height Plot of residuals versus predicted indicates constant variance assumption is reasonable.0 0. Standard Deviation Residuals versus predicted plot and residuals normal probability plot support constant variance and normality assumptions.5 0.0 1.5 Observation Order Residuals Versus B Standardized Residual Standardized Residual -2 0.0 9 10 11 12 1 3 14 15 16 -0. then none of the 2-factor interactions will be aliased with each other. 13-17 .0 -0.0 0.0 C 1.5 0.0 0.5 2 1.0 0 Residuals Versus D -0.5 Residuals Versus C (response is StdDev) 2 -1.0 A 0.5 1.Chapter 13 Exercise Solutions 13-7 (d) continued Standardized Residual Standardized Residual For standard deviation: Residual Plots for StdDev -2 0 Standardized Residual 2 Histogram of the Residuals 4 2 0 -1.

50 0 .5 0 -1.0 0 D: h o ld ti m e = 0 .50 1.02 For the current factor levels.056 at A = –0.63 + 0. and optimal settings could be identified from visual examination of both plots.12 + 0.081B + 0.Chapter 13 Exercise Solutions 13-8.99. The overlaid contour plot below (constructed in Design-Expert) shows one solution with mean Free Height ≅ 7.0 0 -1 .00 A: fu rn te m p 13-18 .15 PO E(f re e 0. so Variance of Free Height = (–0.44 and B = 0. and then contour plots in factors A. 55 0. h 14 51 04 X -0 . 49 59 4 POE( free height): 0.49 and minimum standard deviation of 0.02 . and D could be constructed using the Graph > Contour Plot functionality.4 1 Y 0.12 + 0.00 -0.00 O ve rl a y Pl o t X = A: fu rn te m p Y = B: h e a t tim e f ree he igh7. This approach is fully described in the solution to Exercise 13-12. Factor E is hard to control (a “noise” variable). Overlay Plot DES IG N-E XP ERT P lo t 1.50 f r ee h eight: 7. B.077B)2 + 0.0 0 B: h ea t tim e 0. Using equations (13-6) and (13-7) the mean and variance models are: Mean Free Height = 7. Free Height Variance could be calculated in the MINITAB worksheet.0 0 E : o il te m p = 0 .00 0. 99 A ctu a l Fa cto rs C: tra n s tim e = 0 .077B)2 + σ2 Assume (following text) that σ2E = 1 and σˆ 2 = MSE = 0.00 -0 .046D Variance of Free Height = σ2E (–0.12A – 0. These contour plots could be compared with a contour plot of Mean Free Height.

00 -0.12A – 0. Overlay Plot DES IG N-E XP ERT P lo t 1.0 0 D: h o ld ti m e = 0 .99.046)2 + σ2E (–0. Factors D and E are noise variables. Using equations (13-6) and (13-7).00 A: fu rn te m p 13-19 .50 0 .0 0 E : o il te m p = 0 . 55 0. These contour plots could be compared with a contour plot of Mean Free Height.16 Y 0 .63 + 0. the mean and variance are: Mean Free Height = 7.50 and minimum standard deviation of Free Height to be: A = – 0.00 0. and optimal settings could be identified from visual examination of both plots.12 + 0.00 O ve rl a y Pl o t X = A: fu rn te m p Y = B: h e a t tim e f ree he igh7 .077B)2 + 0.12 + 0.046)2 + (–0. and D could be constructed using the Graph > Contour Plot functionality.5 0 -1.1 52 23 3 X -0.4 95 01 PO E(f re e 0h. The overlaid contour plot below (constructed in Design-Expert) shows one solution with mean Free Height ≅ 7. Free Height Variance could be calculated in the MINITAB worksheet.0 0 -1 .02 For the current factor levels.9 9 A ctu a l Fa cto rs C: tra n s tim e = 0 . and then contour plots in factors A.00 -0 .077B)2 + σ2 Using σˆ 2 = MSE = 0. Assume σ D2 = σ E2 = 1 . B.081B Variance of Free Height = σ2D (+0.42 and B = 0. This approach is fully described in the solution to Exercise 13-12.50 1.0 0 B: h ea t tim e 0.Chapter 13 Exercise Solutions 13-9.50 f r ee h eight: 7. 42 POE( free height): 0.02 : Variance of Free Height = (0.

699 20. 100.801 1. 80.000 x1*x1 -14.8750 x1*x3 -2.17 989.95 Square 3 1217.8862 x3*x3 0.205 x3 -10.466 2. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 87.446 x1*x2 8.764 -8.710 -2. 63.053 x1*x3 -7. x2.000 x2 2.073 x2*x3 -13.710 -3.66 137.17 329.8279 x2 1.067 Total 19 2636. 100.805 0.000 0. 88.305 2.730 0.766 Lack-of-Fit 5 92. The correct values are: 66.764 -5.29 277.195 2. 60.005 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS F Regression 9 2499.289 1.689 -6. 78.38 97.081 3.192 0.33 92.001 0.458 7.Chapter 13 Exercise Solutions 13-10. 83.33 18.689 5.176 1. 82.95 … Estimated Regression Coefficients for y using data in Term Coef Constant 87. 91.6250 P 0.794 0.069 0. x3 The analysis was done using coded units. 70.526 0. one approach to solving this exercise is to create a general full factorial design in MINITAB. 65.29 2499. 80. 85.0509 x1*x1 -5.66 13.305 2. and then enter the data.0578 x2*x2 -7.33 9.175 0.425 3.38 292. These values are used in the Excel and MINITAB data files. Since the runs are listed in a patterned (but not standard) order.49 Interaction 3 292.356 0. 87.359 1.000 0.MTW. 70. Note: Several y values are incorrectly listed in the textbook.513 57.3613 x3 -6. 75.7759 x1*x2 2.000 0.000 x1 9. 90. 68.764 0.723 23.6250 x2*x3 -4.027 0.227 uncoded units 13-20 .33 45.710 2. The design and data are in the MINITAB worksheet Ex13-10.17 Linear 3 989.74 405.914 29.04 Pure Error 5 45.3589 x1 5.000 x3*x3 2.000 x2*x2 -22. Stat > DOE > Response Surface > Analyze Response Surface Design Response Surface Regression: y versus x1.689 1.74 1217.008 0.132 3.08 Residual Error 10 137.

587 0.0 2.000 0.534 Interaction 3 292.34 13.0 -5 0 -4 -2 0 2 Residual 4 2 4 -2 6 8 10 12 14 16 18 20 Observation Order -1 Residuals Versus x2 1 2 1 2 Residuals Versus x3 (response is y) (response is y) 5.5 0 -5.131 0.33 P 0.5 2.132 3.86 0.647 2.704 -5.660 -6.685 0.257 Residual Plots for y Normal Probability Plot of the Residuals (response is y) 5 90 Residual 50 10 5.17 989.327 Linear 3 989. x3 The analysis was done using coded units.78 45.229 0.458 Residual Error 11 146. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 87.006 1.081 3.723 Square 2 1209.5 Residual Residual 0 x1 0.523 2.07 1209.289 1.00 101.004 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS Regression 8 2490.834 Pure Error 5 45.61 2490.34 146.425 3.048 x1*x3 -7.176 1.801 1.5 -5 1 -5 0 Residual 5 60 Histogram of the Residuals 80 Fitted Value 100 Residuals Versus the Order of the Data Residual Percent Residuals Versus x1 Residuals Versus the Fitted Values 99 Residual Frequency 5 4 2 0.33 9.195 x3 -10.660 1.994 1.647 -2.40 24.905 0.Chapter 13 Exercise Solutions 13-10 continued Reduced model: Response Surface Regression: y versus x1.263 69.00 16.329 0.38 292.067 Total 19 2636.0 -5.17 329.0 5.33 45.0 -2.647 -3.44 7.303 Lack-of-Fit 6 101.5 0.000 0.000 0.0 -2.704 -8.036 0.0 -2 -1 0 x2 1 2 -2 -1 0 x3 13-21 .0 0 2.660 5.61 311.38 97.5 -5.07 604.379 0.371 0.000 x2*x2 -22.000 x1 9.000 x1*x2 8.000 x2 2.523 2.000 x1*x1 -14.0 -2. x2.067 x2*x3 -13.95 … F 23.

x1 Surface Plot of y vs x3. Upper = 120.0 > 1 Hold Values x2 0 70.807. Weight = 1. Importance = 1 One solution maximizing growth is x1 = 1.0 60.0 90.0 60.Chapter 13 Exercise Solutions 13-10 continued Stat > DOE > Response Surface > Contour/Surface Plots Contour Plot of y vs x2.0 100.0 Hold Values x3 0 0 80 y 60 2 40 -1 0 -2 0 x1 -1 0 x1 -2 1 Contour Plot of y vs x3. and x3 = −1.0 50.0 1 x2 Surface Plot of y vs x2. x1 40. x2 = 0.0 80.682.0 80.0 90. 13-22 . Predicted yield is approximately 108 grams.0 70. Lower = 60.292.0 80.0 50.0 Hold Values x2 0 x3 2 x2 105 0 90 y 75 2 60 -1 0 -2 0 x1 -1 0 x1 2 x3 -2 1 Stat > DOE > Response Surface > Response Optimizer Goal = Maximize.0 70. x1 y < 70.0 100.0 80. x1 y < > Hold Values x3 0 40.

011 Square 2 123.000 Residual Error 7 154.Chapter 13 Exercise Solutions 13-11.186 0.017 13-23 .58 61.534 Pure Error 4 14.80 3. Stat > DOE > Response Surface > Analyze Response Surface Design Response Surface Regression: y versus x1. Since the runs are listed in a patterned (but not standard) order.000 x1 -1.388 0.58 123.60 315.660 -1.80 14.00 144.60 139.02 24.00 144.712 1.058 Lack-of-Fit 3 139.40 22.274 x2 1.076 x2*x2 2.102 0.60 46.09 2.119 Linear 2 48. one approach to solving this exercise is to create a general full factorial design in MINITAB.038 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS Regression 5 315.200 2.700 Total 12 470. and then enter the data.00 F 2.02 48.60 63.409 x1*x1 3.555 0.788 Interaction 1 144.457 1. The design and data are in the MINITAB worksheet Ex13-11.MTW.878 0.660 0.781 2. x2 The analysis was done using coded units.80 6.40 154.53 P 0.209 x1*x2 6.616 0.463 1.100 19.383 0.86 1.038 12.128 0. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 41.085 0.970 1.781 1.348 2.000 2.58 0.

0 1 0 -1 x1 -1. x1 y < > 40 45 50 55 60 60 70 0.Chapter 13 Exercise Solutions 13-11 continued Contour Plot of y vs x2. Target = 0.4142. Importance = 1 Recommended operating conditions are temperature = +1.4109 and pressure = -1. Weight = 1.0.5 50 40 -1.5 x2 Surface Plot of y vs x2. Weight = 10. Upper = 50. to achieve predicted filtration time of 36. to achieve predicted filtration time of 46. 13-24 . (b) Goal = Target.0785.5 0 x2 -1 1 1. x1 1.0 60 y -0.7.5 0.0 40 45 50 55 0. Target = 46.0 x1 0. Upper = 55. Importance = 1 Recommended operating conditions are temperature = +1.0 -0.3415 and pressure = -0.0 (a) Goal = Minimize. Lower = 42.

764 18.94 226.232 0.8279 x2 1.6250 P 0.7783 2.003 x2*x2 -21.945 x1*x2 8.50 … Estimated Regression Coefficients for y using data in Term Coef Constant 87.001 0.MTW Stat > DOE > Response Surface > Analyze Response Surface Design Response Surface Regression: y versus x1.3333 1.804 0.56 135.084 x1*z -4.873 1.33 9.109 x2*z -7.22 30.968 0.53 Square 3 953.1250 x1*x1 -4.75 Interaction 3 292.38 292.001 0.7192 z*z 0.Chapter 13 Exercise Solutions 13-12.681 51.29 953.22 90.067 Total 17 2170.8333 3.10).1250 1.013 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS F Regression 9 2034.001 x2 2.6250 x2*z -4.75 Residual Error 8 135.32 Pure Error 5 45. so there is a robust design problem.8750 x1*z -2.116 0.873 5.455 -4.28 789.072 0.38 97.1517 2.074 3.3613 z -6.021 0.1517 x1*x2 2.8013 1.94 2034. The design and data are in the MINITAB worksheet Ex13-12.000 x1 9.003 x1*x1 -13.105 13.209 0.34 Linear 3 789.3333 x1 5. x2.496 0.222 0. 13-25 .945 Lack-of-Fit 3 90.975 0.000 z*z 0.56 16.2894 1.1317 4.458 5.448 -3.361 -4. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 87.092 15.361 -6.8333 3.116 1.448 -1.8908 x2*x2 -7.178 0.4147 2.115 uncoded units The coefficients for x1z and x2z (the two interactions involving the noise variable) are significant (P-values ≤ 0.28 263. z The analysis was done using coded units.29 317.001 0.33 45.256 z -6.116 0.

0 Frequency (response is y) 5 90 1.5 0 -5.092 Square 2 953.5 -5.308 -3.778 2.767 1.0 -5.019 -4.0 -5 -4 -2 0 2 Residual 4 2 4 6 8 10 12 14 16 Observation Order -2 18 -1 Residuals Versus x2 1 2 Residuals Versus z (response is y) (response is y) 5.0 5 Residual 3.5 1.308 -1.760 3.088 x2*z -7.172 Residual Plots for y Normal Probability Plot of the Residuals Residual 10 5.86 254.000 x1*x2 8.675 0.20 476.64 135.0 2.072 Lack-of-Fit 4 90.760 3.361 1.31 90.0 Residuals Versus x1 Residuals Versus the Fitted Values 99 -2.0 -2.0 z 0.49 0.548 0.0 -0.38 97.000 0.370 0.5 -5 -5 0 Residual 5 60 Histogram of the Residuals 80 Fitted Value 100 Residuals Versus the Order of the Data Residual Percent 50 1 0. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 87.125 1.912 0.801 1.000 0.28 263.000 0.000 x1 9. x2.47 P 0.31 22.013 2.5 2.88 17.882 2.462 0.Chapter 13 Exercise Solutions 13-12 continued Reduced model: Response Surface Regression: y versus x1.62 6.0 -2 -1 0 x2 1 2 -1.067 Total 17 2170.28 789.095 0.50 … F 16.289 1.227 z -6.0 0 2.132 3.0 5.541 56.002 x1*x1 -13.20 953.558 0. z The analysis was done using coded units.208 0.0 -2.767 5.0 13-26 .33 9.357 Linear 3 789.38 292.46 31.000 x2 2.296 0.5 0.066 x1*z -4.578 Pure Error 5 45.5 0.008 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS Regression 8 2034.5 0.373 -4.019 -7.64 15.86 2034.415 2.5 Residual Residual 0 x1 0.001 x2*x2 -21.33 45.458 Residual Error 9 135.602 Interaction 3 292.

36 + 5.10 10 .86x12 – 7. 13-27 .109756 x2 = -0.0 70.0 Hold Values z 1 0 -1 -1 0 x1 1 -1 0 x1 1 Examination of contour plots for Free Height show that heights greater than 90 are achieved with z = –1.0 70.Chapter 13 Exercise Solutions 13-12 continued yPred = 87.0 1 x2 x2 -1 0 x1 Contour Plot of y vs x2. x1 y < > -1 -1 Contour Plot of sqrt{Vz(y(x.0 70.36x2 – 4.308367 y = 90.13 – 2.0 80.63x2)2 + σˆ 2 = (–6.36x2 – 4.63x2)2 + 15.0 60. x1 sqrt{Vz(y(x.31 (mean yield of about 90 with a standard deviation between 6 and 8).0 70.69x22 For the variance model.) Contour Plot of y vs x2.63x2)z For the mean yield model.86x12 – 7.0 60.293782 y = 90.708407 x2 = -0.12 > 12 0 40. x1 x2 x1 = -0.2054 0 x1 40.0 1 x2 50. Comparison with the contour plot for variability shows that growth greater than 90 with minimum variability is achieved at approximately x1 = – 0.36 + 5.0 90.0 80.83x1 + 1.393500 x2 = 0.11 and x2 = – 0. (Refer to MINITAB worksheet Ex13-12.0 50.0 80. There are other combinations that would work.MTW.z)]} < 6 6 8 8 .0 80.0 50.63x1 – 4.0 80.0 60.0 60.0 Hold Values z 0 0 1 1 y < > y < > 30.072 This equation can be added to the worksheet and used in a contour plot with x1 and x2.0 80.0 60.13 – 2.555312 y = 90.0 70. assume σz2 = 1: Variance of Yield = σz2 (–6.0 40.13 – 2.69x22 + (–6.0 60.1692 50.0 50.0 70.0 Hold Values z -1 0 x1 = -0.0 1 -1 Contour Plot of y vs x2.63x1 – 4. set z = 0: Mean Yield = 87.z)]} vs x2.0 40.0 90.83x1 + 1.63x1 – 4.0198 -1 x1 = 0. x1 1 30.0 50.

z ) r = ∑ γ + ∑ ∑ δ ui xu + ∑ ∑ λij ( zi + z j ) . r k k ∂h(x. 13-14. r k r r r i< j =2 i =1 If h(x. z ) ] = σ ∑ γ i + ∑ δ ui xu 2 z r r i =1 k u =1 k ) +σ 2 r 2 r If h(x. z ) = γ i + ∑ δ ui xu . and i =1 i =1 i i =1 u =1 i< j =2 ∂zi r then ∑ r ⎡ k r ⎤ V [ y (x. then i =1 i =1 j =1 r ∂h(x. z ) = ∑ γ i zi + ∑ ∑ δ ij xi z j + ∑ ∑ λij zi z j + ∑ θi zi2 . and u =1 ∂zi r If h(x. z ) = ∑ γ i + ∑ ∑ δ ui xu + ∑ ∑ λij ( zi + z j ) + 2∑ θ i zi . z ) = ∑ γ i zi + ∑ ∑ δ ij xi z j . i =1 i =1 j =1 i< j =2 r k r ∂h(x. z ) ] = V ∑ ⎢γ i + ∑ δ ui xu + ∑ λij ( zi + z j ) + 2θi zi2 ⎥zi + σ 2 i =1 ⎣ u =1 j >i ⎦ There will be additional terms in the variance expression arising from the last two terms inside the square brackets. and i =1 i =1 i =1 u =1 i< j =2 i =1 ∂zi r ∑ r r k r r ⎡ k r ⎤ V [ y (x.Chapter 13 Exercise Solutions 13-13. 13-28 . then i =1 i =1 j =1 ( V [ y (x. z ) = ∑ γ i zi + ∑ ∑ δ ij xi z j + ∑ ∑ λij zi z j . z ) ] = V ∑ ⎢γ i + ∑ δ ui xu + ∑ λij ( zi + z j ) ⎥zi + σ 2 i =1 ⎣ u =1 j >i ⎦ There will be additional terms in the variance expression arising from the third term inside the square brackets.

98998 0.20 1.70382 0.30556 0.92528 0.030 0.24807 0.60501 0.14467 0.009 0.12649 0.40048 0.00 Pr{acceptance} 0.27060 0.93910 0.20 0.02656 0.060 0.02863 Pr{d<=c} 0.36417 0.37160 0.55528 0.040 0.07694 0.100 f(d=0) 0.003 0.63633 0.000 0.008 0.060 0.96353 0.040 0.007 0.09994 0.95190 0.00 0. c=1 1.08271 0.90475 0.26986 0.070 0.05324 0.40 0.73577 0.03379 Type-B OC Curve for n=50.16434 0.00896 0.77831 0.Chapter 14 Exercise Solutions Note: Many of the exercises in this chapter are easily solved with spreadsheet application software.080 0.04761 0.004 0.002 0.98274 0. The BINOMDIST.020 0.66924 0.27943 0.06725 0.19000 0.100 0.21807 0.20249 0.74015 0.006 0. and graphing functions in Microsoft® Excel were used for these solutions.81840 0.95121 0.97387 0.12989 0.28895 0.04533 0.005 0.140 p 14-1 .80 0.001 0.09066 0.080 0.33721 0.99881 0.12947 0.010 0. Solutions are in the Excel workbook Chap14.91056 0.050 0.020 0.01547 0.86051 0.04428 0.090 0. 14-1.xls.19556 0.00515 f(d=1) 0.99540 0. p 0.22339 0. HYPGEOMDIST.120 0.60 0.

01312 0.60 0.00000 Pr{d<=c} 0.100 0.96641 0.22515 0.18486 0.92063 0.090 0.99649 0.90479 0.060 0.74048 0.01127 0.Chapter 14 Exercise Solutions 14-2.00476 0.020 0.54782 0.60577 0.008 0.003 0.00449 0.36120 0.03319 0.00531 0.00000 f(d=2) 0.40 0.004 0.050 0.03116 0. c=2 1.80 0.09057 0.95327 0.00 Pr{acceptance} 0.100 0.00003 0.00208 0.11826 0.080 0.20 1.00030 0.44789 0.13262 0.09880 0.02579 0.04144 0.01978 0.14419 0.36973 0.00592 0.99888 0.04755 0.12185 0.00079 0.120 0.00 0.66978 0.080 0.00162 0.97730 0.49536 0.000 0.22281 0.060 0.98590 0.01627 0.00205 0.36773 0.16404 0.00388 0.070 0.00071 0.08118 0.99985 0.140 p 14-2 .040 0.93796 0.00024 0.009 0.01687 0.23214 0.010 0.27341 0.00194 0.20 0.002 0.005 0.14707 0.00000 Type-B OC Curve for n=100.001 0.05661 0.006 0.14498 0.30441 0.07572 0.36603 0.030 0.07029 0.41978 0.00895 0.16531 0.33068 0.00008 0.67669 0.81857 0.040 0.26899 0.40492 0.200 f(d=0) 0.020 0.99225 0.05347 0.27065 0.007 0. p 0.34920 0.00000 f(d=1) 0.

or α ≅ 0. (a) Type-A OC Curve for N=5000.080 0.000 0.020 0.40 0.60 0.00 0. n=50.10 (b) Type-B OC Curve for N=5000.800 0.10133. n=50.000 0.020 0.200 0.600 0.9521.10 (c) Based on values for α and β. or α ≅ 0.060 0.120 0. the difference between the two curves is small.05 Pa (p = 0.140 p Pa (p = 0.060 0.80 0.10133. or β ≅ 0. 14-3 .9521.080 0.100 0.007) = 0.075) = 0. either is appropriate. c=1 1.400 0. or β ≅ 0.000 Pr{acceptance} 0.000 0.05 Pa (d = 375) = 0.140 p Pa (d = 35) = 0.100 0.00 Pr{acceptance} 0. c=1 1.120 0.040 0.040 0.20 0.Chapter 14 Exercise Solutions 14-3.

1 − α = 1 − 0.06.95. select n = 35 and c = 1.05 = 0. p1 = 0. 14-5. p2 = 0.01. the sampling plan is n = 80 and c = 7.04786 and β = 0.99. β = 0. 14-4 .15.95. β = 0.10 From the binomial nomograph. resulting in actual α = 0. p1 = 0.12238. 1 − α = 1 − 0. p2 = 0.01 = 0. p2 = 0.05.02.Chapter 14 Exercise Solutions 14-4. p1 = 0. select a sampling plan of n = 300 and c = 12.05 = 0.10 From the binomial nomograph. 14-6.10. β = 0.10 From the binomial nomograph. 1 − α = 1 − 0.

00027 -0. but pays for the high probability of rejecting acceptable lots like those with p = 0.01175 0.00000 5000 500 0.0001 0.0600 0.00263 -0.99359 0. 14-5 .58304 0.00095 -0.0015 0.00001 -0.0004 0.4409 0.0200 10 N2 = n1 = pmax = cmax = binomial Pr{reject} Pr{d<=20} 0.0250 0.99959 0.294.0200 20 Pr{reject} 0.99850 0.0132 0.05 by Pa(N = 5. Also.0100 0.00000 0.0064 0.00001 0.00938 0.0010 0.0050 0.0000 0.00046 and Pa(N = 10.025) = 0.Chapter 14 Exercise Solutions 14-7.0000 1.8178 0.00046 0.0060 0.00000 1.05 p 0.99972 0.0000 0.99991 0.00006 -0.000) = 0. LTPD = 0.0000 1.9667 0.18221 0.9997 1.0300 0.00046 0.98676 0.0000 0.000.0001 1.00000 0. the consumer is protected from a LTPD = 0.0020 0.9903 0. Pa(p = 0.0000 1.08151 0.29404 0.99994 0.0000 1.0000 0.0000 difference 0.55910 0.03328 0.0040 0.99999 0.0090 0.0200 0.0000 1.00030 0.0000 0.0000 0.02395 0.7060 0.00967 0.0000 0. For N = 5.0400 0.00000 0. while for N = 10.00000.0000 0.00000 10000 1000 0.025. Pa(p = 0.0027 0.0003 1.9995 0.00000 -0.0000 0.00000 1.11183 0.0070 0.00000 0.0500 0.000.00000 Different sample sizes offer different levels of protection.4170 0.99903 0.11479 0.00000 1.182.0010 0.0080 0.0700 N1 = n1 = pmax = cmax = binomial Pr{d<=10} 1.00000 0.00000 1.00001 0.8852 0.00600 -0.00000 0.99729 0.0000 1.00000 0.00000 0.000) = 0.0030 0.025) = 0.99999 0.00000 0.

0090 0.0184 0.95886 0.9868 0.0300 0.4949 for N = 5000.0060 0. For example.2204 0.00000 1.0200 0.0700 0.0139 0.1168 0.1897 0.11858 0.2959 0.10543 0.1000 0.99382 0.24312 0.73230 0.0900 0.75558 0. Pa(p = 0.1957 0.3884 0.8814 0.0800 0.93236 0.00063 0.0020 0.38898 0.01 1 Pr{reject} 0.5052 0.14688 0.9618 0.2677 0.9938 0.0100 0.0000 1.7961 0.00000 5000 71 0.1439 0.53457 0.00622 0.9971 0.86924 0.05374 0.0008 0.0000 0. p 0.98610 0.81033 0.00000 1.97550 0.53107 0.0676 0.9724 0.0276 0.98157 0.2446 0. and Pa(p = 0.0411 0.0006 0.0000 This plan offers vastly different protections at various levels of defectives.9426 0.0002 0.7323 for N = 1000.80432 0.94552 0. 14-6 .00002 0.01.98767 0.75536 0.56992 0.01315 0.9994 1.0245 0.70407 0.0545 0.0050 0.0400 0. depending on the lot size.61157 0.00099 0.9990 0.01 0 N2 = n1 = pmax = cmax = binomial Pr{reject} Pr{d<=1} 0.3000 0.93982 0.7569 0.4689 0.0602 0.05741 0.9246 0.01) = 0.3438 0.1308 0.07541 0.0600 0.4654 0.00136 0.3500 N1 = n1 = pmax = cmax = binomial Pr{d<=0} 0.1702 0.4301 0.7179 0.2000 0.20391 0.96946 0.0889 0.8531 0.0123 0.00000 1000 32 0.49484 0.0500 0.03815 0.0305 0.0030 0.0080 0.0040 0.0000 1. at p = 0.82981 0.0000 0.00292 0.85608 0.77958 0.2444 0.9463 0.0010 0.65622 0.0004 0.6110 0.28210 0.9986 0.02758 0.88316 0.0070 0.97238 0.Chapter 14 Exercise Solutions 14-8.0062 0.01) = 0.8946 0.91107 0.

000 0. N = 2.000 ATI = n + (1 − Pa )( N − n) = 35 + (1 − Pa )(2000 − 35) = 2000 − 1965 Pa Pa p( N − n) N = (1965 2000 ) Pa p AOQ = AOQL = 0.025 0.0234 ATI Curve for n=35.10 0.00 0. c = 1.20 0.020 AOQ 0.25 0. c=1 2500 2000 ATI 1500 1000 500 0 0.05 0.015 0.Chapter 14 Exercise Solutions 14-9.00 0.15 0.10 0.005 0.20 0.30 p 14-7 .15 0. c=1 0. n = 35.010 0.25 p AOQ Curve for n=35.05 0.

009 0.98927 0.025 0.020 0.004 0.000 0.0003 2985 0.0022 2834 0.003 0.00523 0.010 0.00 0.00 Pr{accept} 0.00002 AOQ ATI 0.0009 151 0.Chapter 14 Exercise Solutions 14-10.100 Pa=Pr{d<=2} 0.040 0.0034 2712 0.0065 2221 0.88019 0.0072 588 0.045 0.006 0.05840 0.060 0.42093 0.99951 0.0087 AOQL 1265 0.050 0.0019 160 0.070 0.0053 328 0. c = 2 p 0.60 0.015 0. c=2 1.97716 0.0000 3000 0.0009 2948 0.050 0.91092 0.080 p 14-8 .030 0.00009 0.030 0.80948 0.95991 0.0000 2999 0.002 0.27341 0.0061 404 0.84615 0.16932 0.60884 0.01815 0.20 0.80 0.040 0.0048 2517 0.40 0.060 0.0067 491 0. n = 150.008 0.03292 0. N = 3000.0077 693 0.0001 2996 0.93769 0.0014 2906 0.035 0.00142 0.00036 0.10098 0.0000 3000 (a) OC Curve for n=150.020 0.0046 264 0.070 0.090 0.007 0.001 0.0080 1800 0.0028 181 0.080 0.99646 0.010 0.005 0.0037 215 0.

020 0.030 0.0010 0.0050 0.040 0.0090 0.070 0.0087 0.020 0.Chapter 14 Exercise Solutions 14-10 continued (b) AOQ Curve for n=150.0070 AOQ 0.060 0.050 0.070 0. c=2 AOQL ≅ 0.030 0.080 p 14-9 .0040 0. c=2 3500 3000 2500 ATI 2000 1500 1000 500 0 0.000 0.0030 0.0000 0.010 0.080 p (c) ATI Curve for n=150.0080 0.0100 0.000 0.050 0.010 0.0020 0.0060 0.040 0.060 0.

1200 0.0800 0.90015 0.0000 0.2000 0.77403 0.99871 0.00758 0.22597 0.31079 0.16054 0.01043 0.1050 0.99891 0.00048 0.88827 0.00 Probabilty of Acceptance.0200 0.1000 0.80 0.00 0.00002 0.58375 0. p 14-10 .99657 0.67671 0.92157 0.0020 0.10764 0.0090 0.0500 0.00015 0.20 0.0100 0.Chapter 14 Exercise Solutions 14-11.83946 0.0400 0.99242 0.89632 0. c=2 1.1020 0.40 0.0200 0.0060 0.1600 0.1010 0.00206 0.00109 0. (a) N = 5000.1800 0.11173 0.89236 0.99985 0.1000 0.09985 0.0030 0.45947 0.00129 0.81080 0.00343 0.0800 0.99952 0.0700 0.90386 0. n = 50.0600 0.0010 0.10368 0.00000 OC Curve for n=50.0050 0.32329 0.98957 0.0080 0.1400 0.3000 0.0040 0.00000 1.0900 0.09255 0.0300 0.54053 0.68921 0.1030 0.0400 0.00526 0.98618 0.0600 0.07843 0.0070 0.60 0.2000 Fraction defective.1040 0.41625 0.90745 0.01382 0.99998 0.09614 0.18920 0.99474 0. Pa 0.99794 0. c = 2 p Pa=Pr{d<=1} Pr{reject} 0.

66761 0.0200 0.0500 0.87842 0.09539 0.05832 0.0400 0.85160 0.01153 0.70989 0.Chapter 14 Exercise Solutions 14-11 continued (b) p = 0. yielding Pa = 1 – 0.54379 0.45621 0. will be extremely hard on the vendor because the Pa is low even if the lot fraction defective is low.0050 0.1000 0.0900 0.90.15164 0.86893 0.0070 0.00080 0.13107 0.90461 0.16541 0.98847 0.11372 = 0.0020 0.2000 0.18869 0.55800 0.94168 0.92297 0.5000 Pa=Pr{d<=0} 0.99920 0.0010 0.29011 0.99996 1.11340 0.12158 0. The OC curve for this zero-defects plan is much steeper.35849 0. quality improvement begins with the manufacturing process control.81131 0.14840 0.0090 0. Generally.0300 0.4000 0. not the sampling plan.96075 0. p 0.0040 0.23424 0.18209 0.81791 0.64151 0.88638 ≈ 0.0700 0.1030 will be rejected about 90% of the time.07703 0. select n = 20.00000 Pr{reject} 0.98019 0.44200 0.01981 0.0030 0.0080 0.00000 14-11 .3000 0.0100 0.03925 0.0600 0.33239 0. with acceptance number c = 0.0060 0.0800 0.76576 0.88660 0. (d) From the nomograph.83459 0.84836 0.00004 0. (c) A zero-defects sampling plan.

00 Probability of acceptance.0500 0. c=0 1. c = 2} = 0.20 0.00206)(5000 − 50) = 60 The c = 2 plan is preferred because the c = 0 plan will reject good lots 10% of the time.09539 Pr{reject | p = 0.80 0.09539)(5000 − 20) = 495 ATIc = 2 = 50 + (0.60 0.Chapter 14 Exercise Solutions 14-11 (d) continued OC Curve for n=20. c = 0} = 0. 14-12 .005.1000 0.0000 0.1500 0.2000 0.00206 ATIc =0 = n + (1 − Pa )( N − n) = 20 + (0. p (e) Pr{reject | p = 0.005.00 0.40 0. Pa 0.2500 Fraction defective.

2260 0.3274 0.0073 0.0820 0.9862 0.0002 0.0011 0. n1 = 50.0000 0.0001 0.0791 0.100 0. c1 = 2.0844 0.0023 0.0015 0.075 0.005 0.3848 0.060 0.0004 0.0111 0.0514 0.0142 Primary and Supplementary OC Curves for n1=50.0238 0.5974 0.0444 0.0000 0.0098 0. c2 = 6 P d1 = PaI PrI 0. proportion defective Pr{accept on 1st sample} Pr{reject on 1st sample} Pr{accept lot} 14-13 .6078 0.4494 0.Chapter 14 Exercise Solutions 14-12.0001 0.80 Pr 0.0001 0.2548 0.0000 0.0019 0.0000 0.1635 0.d2<=3} Pr{d1=4.0568 0.d3<=2} Pr{d1=5.0029 0.025 0.9661 0.055 0.140 0.40 0.9999 0.0001 0.0000 0. n2=100.0002 0.0000 0.4595 0.0627 0.6390 0.8928 0.0011 0.0000 0.0793 0.050 0.0000 0.9487 0.20 0.0000 0.110 0.8296 0.115 0.8706 0.9216 0.0001 0.030 0.0011 0.120 0.7452 0.60 0.035 0.0000 0.0051 0.0002 0.9373 0.0330 0.0004 0.0001 0.0000 0.1117 0.045 0.9979 0.0002 0.0006 0.010 0.2658 0.0142 0.0009 0.0002 0.0000 0.0000 0. n2 = 100.0025 0.020 0.0001 0.0165 0.0028 0.6767 0.0000 0.4763 0.0035 0.0129 0.1892 0.2333 0.0000 0.8883 0.0193 0.2770 0.0020 0.0000 0.7740 0.0339 0.0629 0.0001 0.070 0.3610 0.0152 0.0098 0.0001 0.0120 0.0567 0.0030 0.100 0.0004 0.7560 0.0000 PaII Pa 0.0000 0. c1=2.6770 0.0013 0.0331 0.0708 0.065 0.0767 0.0166 0.0120 0.0690 0.080 0.5405 0.020 0.0004 0.0024 0.9237 0.00 0.160 p.0000 0.0692 0.0073 0.0000 0.0019 0.8395 0.5207 0.8108 0.3108 0.0842 0.120 0.0209 0.d2=0} 0.0029 0.0513 0.9858 3 4 5 6 Pr{d1=3.0008 0.5237 0. c2=6 1.0818 0.6892 0.0000 0.0763 0.0000 0.0000 0.0000 0.9982 0.0029 0.4162 0.0000 0.0339 0.0112 0.0001 0.0442 0.140 0.0221 0.0002 0.0021 0.0001 0.150 0.060 0.0238 0.1294 0.090 0.0221 0.040 0.0784 0.9779 0.0212 0.040 0.0001 0.00 0.0072 0.7342 0.d2<=1} Pr{d1=6.9737 0.0000 0.0019 0.0001 0.9414 0.3233 0.0011 0.3922 0.1605 0.1128 0.0001 0.0000 0.0138 0.0000 0.0002 0.0000 0.0521 0.0190 0.0161 0.0522 0.0707 0.130 0.080 0.5838 0.000 0.

859 -0.10. (b) Three points on the OC curve are: p1 = 0. Acc = 1.064 -0. Pa = h2 1.2054 p2 = 0.850 0.285 1.9389.10 k = 1.034 3.0414.01. h2 = 1.000 2.2054.10 14-14 .1 − α = 1 − 0.199 … 2. p2 = 0.015 0.0397 n.159 2.890 0.95.994 3.05 = 0.0397 X A = −0.0397n n 1 2 3 4 5 … 20 21 22 23 24 25 … 45 46 47 48 49 50 XA -0.820 -0.95 p = s = 0. Rej = 4.404 … 2.929 0.074 3. s = 0. β = 0.9389 + 0.040 2.5621 h1 + h2 0.969 1.780 -0.049 XR 1.245 1.2054 = = 0.055 … 0.Chapter 14 Exercise Solutions 14-13.113 3.325 1.364 1.01.153 3.144 -0. Pa = β = 0. Pa = 1 − α = 0. h1 = 0. X R = 1.193 Acc n/a n/a n/a n/a n/a … n/a n/a n/a n/a 0 0 … 0 0 0 0 1 1 Rej 2 2 2 2 2 … 2 3 3 3 3 3 … 3 4 4 4 4 4 The sampling plan is n = 49.104 -0.740 … -0.025 0.899 -0.080 2.009 1.10.9389 + 1. (a) p1 = 0.120 2.2054 + 0.0397.

670 … 2.0436 + 0.0436 + 1.309 4.472 1. Pa = β = 0.057 2.406 1.912 -0.3399 + 0. p2 = 0.375 4.659 2.857 2.923 2.02.5622 h1 + h2 1.15.780 -0.0660n n 1 2 3 4 5 … 20 21 22 23 24 25 … 45 46 47 48 49 50 XA -0.3399 = = 0.0660. h2 = 1.978 -0.Chapter 14 Exercise Solutions 14-14.791 2.846 -0. h1 = 1. Pa = α = 0.638 Acc n/a n/a n/a n/a n/a … n/a n/a n/a n/a 0 0 … 0 0 0 0 1 1 Rej 2 2 2 2 2 … 2 3 3 3 3 3 … 3 4 4 4 4 4 The sampling plan is n = 49.0660n.342 0.95 p = s = 0.95.0660 X A = −1.1 − α = 1 − 0.02.276 0. Acc = 1 and Rej = 4.725 2.441 4.15.989 … 4.572 4.123 2.991 2.3399.0436. (b) p1 = 0.05 = 0. X R = 1.255 XR 1.408 0.604 1.540 0.714 … 0. s = 0. β = 0.925 1. Pa = h2 1.10 14-15 .474 0.9369.538 1.3399 p2 = 0.606 … 1. (a) p1 = 0.189 2.507 4.10 k = 0.

n = 20. Re = 1 Reduced: up to letter K. N = 10. n = 125. Ac = 1. N = 3000. Ac = 1. Ac = 3. General inspection level II. Re = 3 Reduced sampling plan: n = 50. Re = 2 Tightened sampling plan: Sample size code letter = J. Re = 1 Tightened: n = 200. Ac = 0. AQL = 0. AQL = 1% General level I Normal sampling plan: Sample size code letter = H. AOQ = [ Pa × p × ( N − n)] [ N − Pa × (np) − (1 − Pa ) × ( Np) ] 14-16. Ac = 2.10%. Re = 4 14-17. n = 50. Ac = 1. Ac = 0. Re = 4 Tightened sampling plan: n = 125. Sample size code letter = L Normal: up to letter K. Ac = 0. Re = 2 Reduced sampling plan: Sample size code letter = H.000. Re = 1 14-16 . N = 3000. n = 80. Ac = 0. AQL = 1% General level II Sample size code letter = K Normal sampling plan: n = 125. n = 50.Chapter 14 Exercise Solutions 14-15. Re = 2 14-18.

0200 0.1654 0.0593 0.0003 0.0900 0.1472 0.9529 0.0125 0. Re = 4 Tightened sampling plan: n = 200.0001 0.9202 0.8916 0.9911 0.0090 0.5230 0.0018 0.0000 0.9992 0.0020 0.0070 0.9469 0.0861 0. proportion defective Normal Tightened Reduced 14-17 .9756 0.8092 0.9588 0.0400 0.8340 0.2351 0. AQL = 0.9220 0.00 Pa.9989 0.0030 0.0001 0.0000 0.0022 OC Curves for N=2000.0101 0.1000 p.0023 0.6767 0.9771 0.9389 0.40 0.9886 0.0800 0. Ac = 2.0000 0.8377 0.0090 0. AQL=0.0047 0.00 0.0000 tightened 200 2 Pa=Pr{d<=2} 0.9667 0.8800 0.80 0.0200 0. Pr{acceptance} 0.65% 1. Sample size code letter = L Normal sampling plan: n = 200.0700 0.0080 0.0004 0.9813 0.65%.9922 0.20 0.0211 0.60 0.0395 0.0400 0.0000 reduced 80 1 Pa=Pr{d<=1} 0.0300 0.0100 0.0433 0.0600 0.7838 0.0010 0.9999 0. General level II.0500 0.8580 0. (a) N = 5000.9163 0.9967 0.0060 0.0050 0.8922 0.9970 0. Re = 3 Reduced sampling plan: n = 80.7309 0. Re = 4 (b) N = 5000 n= c= p 0.3038 0.0600 0. II.0000 0.Chapter 14 Exercise Solutions 14-19. Ac = 3.8653 0.4315 0.1000 normal 200 3 Pa=Pr{d<=3} 0.0800 0. Ac = 1.0040 0.

p = 0.020 D = N*p 2 4 6 8 10 12 14 15 16 18 20 22 24 26 28 30 32 34 36 38 40 Pa 0.5564 0.0001 0.003 0.011 0. AOQL = 0. N = 2000.0006 0.2% p 0.0031 ATI 511 605 767 962 1160 1341 1497 1564 1624 1724 1801 1858 1900 1930 1952 1967 1977 1985 1990 1993 1995 AOQ 0.00 2000 0.018 0.21%.0018 0.004 0. proportion defective 0.015 0.0002 0.21% ATI Curve for N=2000.0001 0.010 0.010 0.004 0.0012 0.8165 0.002 0.016 0.018 0.0464 0. Pr{acceptance} 2500 ATI 1500 1000 0.006 0.0942 0.0015 0. LTPD = 1%.0021 AOQL 0.0150 0.0068 0.008 0.6875 0.0010 0.c=2. 14-18 .016 0.012 0.0003 0.2886 0.0016 0.014 0.20 500 0 0. n=490.25% n = 490.002 0.005 0. without replacement.0008 0.3330 0.006 0.0014 0.012 p.017 0.0007 0.1827 0.010 0.016 0.0000 OC Curve for N=2000. A more precise solution would use the hypergeometric distribution to represent this sampling plan of n = 490 from N = 2000.2489 0.014 0.006 0.000 0.00 0.9864 0.020 p.001 0.0102 0.4361 0.018 0.1320 0. Note that this solution uses the cumulative binomial distribution in a spreadsheet formulation.008 0.019 0.000 0. c=2 1.020 0.0046 0.002 0.0321 0. AOQL=0.0001 0.0005 0.0664 0.80 Pa.008 0. proportion defective The AOQL is 0.007 0.40 0.012 0. c = 2.013 0.9235 0.014 0.0018 0.Chapter 14 Exercise Solutions 14-20.004 0.60 0.0002 0.008 0.0021 0.0220 0.n=490.0020 0.009 0.

000 10.250 p let N = 50. Dodge-Romig single sampling.001 ≤ N ≤ 100.150 0.200 0 0.005) = 0.80000 40.001.000 0.20000 0. n=65.40000 20. AOQL = 3%. 001 − 65) = 82 On average.99967 ATI = n + (1 − Pa )( N − n) = 65 + (1 − 0.100 0.050 0.120 0. p 0.000 0. the average inspection required will be 82 units.000. Pa OC Curve for Dodge-Romig.00000 0.160 Fraction defective. 001 Pa = Binom(3. c=3 60. c = 3. (b) ATI Curve for N=50.99967)(50. average process fallout = p = 0.060 0.080 0.000 0.180 0.100 0. n = 65.000 ATI Probability of acceptance. if the vendor’s process operates close to process average.040 0.140 0.000 1.020 0. c=3 1. 0.50% defective (a) Minimum sampling plan that meets the quality requirements is 50. (c) LTPD = 10.20000 0.Chapter 14 Exercise Solutions 14-21. 65. n=65.000 0.000 30.200 0.3% 14-19 .00000 50.60000 0.

6% c 2 d =0 d =0 Pa = ∑ binomial(n. c = 3. LTPD = 10. p) = ∑ b(65. LTPD = 11. 0.9998)(8000 − 46) ≈ 48 14-20 . 0.9958)(8000 − 65) ≈ 98 (c) N = 8000. AOQL = 3%.25% n = 46.Chapter 14 Exercise Solutions 14-22. p ≤ 0.9998 ATI = n + (1 − Pa )( N − n) = 46 + (1 − 0. p ≤ 1% n = 65.9958 ATI = n + (1 − Pa )( N − n) = 65 + (1 − 0.3% (b) c 3 d =0 d =0 Pa = ∑ binomial(n. c = 2.01) = 0.0025) = 0. AOQL = 3%. (a) N = 8000. p ) = ∑ b(46.

200 0.016 0.050 0.560 0.7 ⎣ ⎦ Accept the lot.005 0.400 0.38 (from nomograph) 15-1 .200 p Pa{p = 0.020 0.02.180 0.Chapter 15 Exercise Solutions 15-1.988 0.73 − 0.70 g/cm3.060 0.730 0.000 0.010 0. Accept the lot if ⎡⎣ Z LSL = ( x − LSL ) S ⎤⎦ ≥ 1.05 (a) From the variables nomograph.150 0.7 . p1 = 0.900 1-alpha 0.190 0.030 0.140 0.7.020 0.90.120 0.7: p2 OC Curve for n=25.xls : worksheet Ex15-1 From the variables nomograph at n = 35 and k = 1.001 0.70 ) 1.025 0. S = 1.040 0. β = 0.800 0.100 0. k = 1.7 1.190 0.10 = 0.160 0. p2 = 0.070 0.600 Pr{accept} p1 p Pr{accept} 0. (b) x = 0.05 × 10−2 ( ) ⎡ Z LSL = ( 0.945 0. k=1.080 0.05 × 10−2 = 2.8571⎤ ≥ 1.050 beta 0. (c) Excel workbook Chap15.820 0.400 0. the sampling plan is n = 35.000 0.10.05} ≈ 0.040 0.73. Calculate x and S.000 0. 1 – α = 1 – 0.100 0. LSL = 0.

Accept the lot if ⎡⎣ Z LSL = ( x − 150 ) S ⎤⎦ ≥ 2. Now.93 A reduced sampling (nreduced = 20.10 From variables nomograph. knormal = 1. Then a k can be found to satisfy the AOQL equation.80. .Chapter 15 Exercise Solution 15-2. Repeat until the n and k that minimize ATI are found.005.95. and special Romig tables are usually employed. The table required to do this is available on the Montgomery SQC website: www.wiley.Form 1. and generally involves some trial-and-error search. p2 = 0. AQL = 1. we know n and k are related. LSL = 150.05 = 0.51) can be obtained from the full set of tables in MIL-STD-414 using Table B-3. for a given AOQL = Pa pm (N – n) / N (where pm is the value of p that maximizes AOQ). standard deviation unknown Assume single specification limit . kreduced = 1.5%.02. The design of a variables plan in rectifying inspection is somewhat different from the attribute plan design.3 15-3. N = 7000.com/college/montgomery 15-5. Suppose n is arbitrarily specified. Calculate x and S. The equations do not change: AOQ = Pa p (N – n) / N and ATI = n + (1 – Pa) (N – n). ATI is found from the above equation. No convenient mathematical method exists to do this. σ = 5 p1 = 0. Inspection level IV From Table 15-1 (A-2): Sample size code letter = M From Table 15-2 (B-1): n = 50. Finally. n and k will define Pa. n = 120 and k = 2. For example. β = 0. for a specified process average. because both Pa and pm are functions of n and k. Under MIL STD 105E. Inspection level II. Sample size code letter = L: n Ac Re Normal 200 7 8 Tightened 200 5 6 Reduced 80 3 6 The MIL STD 414 sample sizes are considerably smaller than those for MIL STD 105E.3. 15-4.1 − α = 1 − 0. ktightened = 1.

k = 2. k = 2.645 0. N = 100. AQL = 1%.00 Tightened sampling: n = 100. x = 255. LSL = 225psi. S = 10 ⎡⎣ Z LSL = ( x − LSL ) S = ( 255 − 225 ) 10 = 3. N = 500.33 15-7.005 g/cm3 x1 = 0. sample size code letter = O Normal sampling: n = 100.1 − α = 1 − 0. k = 1.1527 .282 0. β = 0. so accept the lot.15 = +1.15 Tightened sampling: n = 7. 15-8.05 x A − x1 = Φ (1 − α ) σ n x A − 0.95 = 0.Chapter 15 Exercise Solution 15-6.15.145 = −1.145.10 x A − x2 = Φ(β ) σ n x A − 0.005 n x2 = 0.005 n n ≈ 9 and the target x A = 0.000 Assume inspection level IV. σ = 0. AQL = 4% Sample size code letter = E Assume single specification limit Normal sampling: n = 7. inspection level II.000 ⎤⎦ > 2.14 Assume normal sampling is in effect.00. k = 1.

. N = 5000.95. N = 5. Ac = 3. M = 4.01.08) Reduced: n = 20.000 Assume inspection level II: sample size code letter = L Normal: n = 200.10. the sampling plan is n = 30 and k = 1. c = 2 The sample size is slightly larger than required for the variables plan (a). Re = 6 Tightened: n = 200.95. p1 = 0.08 (a) 1 – α = 0. Re = 4 Reduced: n = 80. Variables sampling would be more efficient if σ were known. p1 = 1% = 0.wiley.93) Tightened: n = 50.10ppm. (c) 1 – α = 0. β = 0. Re = 5 The sample sizes required are much larger than for the other plans. σ unknown Double specification limit. p2 = 8% = 0.8.09 (k = 1. β = 1 – 0. Ac = 2. M = 1.69) σ known allows smaller sample sizes than σ unknown.00 (k = 1. Ac = 5.01. target = 3ppm.71 (k = 2.com/college/montgomery AQL = 1%.90 = 0. p2 = 0.08 From nomograph (for attributes): n = 60.10 From the nomograph. σ = 0.Chapter 15 Exercise Solution 15-9. assume inspection level IV From Table A-2: sample size code letter = M From Table A-3: Normal: n = 50. (d) AQL = 1%. (b) Note: The tables from MIL-STD-414 required to complete this part of the exercise are available on the Montgomery SQC website: www. M = 1.

00 0 0.20 0. c=0 1.02 0.40 0. Excel workbook Chap15.14 0.20 1.04 0.1 0. .60 0.08 0. chain sampling plans with c = 0 have slightly less steep OC curves.2 p single I=1 I=2 I=5 I=7 Compared to single sampling with c = 0.xls : worksheet Ex15-10 OC Curves for Various Plans with n=25.12 0.80 0.00 Pa 0.06 0.Chapter 15 Exercise Solution 15-10.18 0.16 0.

0100 0. c=0 1.8248 0.0900 0.1752 0.8619 0. c = 0 Excel workbook Chap15.00 0.9657 0. Pa 0.10% = 0.02 0.0300 0.xls : worksheet Ex15-11 (a) p 0.0080 0.0010 0.0694 0.0500 0.0200 0.18 0.04 0.7733 0.0008 0.1000 0.3773 0.1937 0.0 0.0020 0.7250 0.0000 OC Chart for n=32.0 Probability of Acceptance.7488 0.5239 0.9083 0.2267 0.Chapter 15 Exercise Solution 15-11.2512 0.12 Fraction Defective.0600 0. p 0.9306 0.8518 0.7292 0.9992 0.20 .8796 0.0981 0.1381 0.4 0.06 0.9511 0.8063 0. n = 32.10 0.0060 0.2 0.1482 0.16 0.0090 0.2000 0.000.6 0.9379 0.6227 0.0400 0.0800 0.0040 0.0917 0.8 0.0489 0.7987 0.14 0.0030 0.0621 0.0050 0.0000 1.2013 0.2708 0.1204 0.3000 Pa Pr{reject} 0.0700 0.4761 0.0070 0. average process fallout = 0.2750 0.0343 0.9019 0.0315 0. N = 30.9685 0.001.08 0.

9685 + (0. n) = P(1. n)]i P(0. there is little change in performance by increasing i. (d) Pa = 0.9685)3 = 0.0310 Pa = 0. c = 0.Chapter 15 Exercise Solution 15-11 continued (b) ATI = n + (1 − Pa )( N − n) = 32 + (1 − 0.9685 P(1. n) + P(1.9967)(30000 − 32) = 131 Compared to conventional sampling.9958)(30000 − 32) = 158 .9685)(30000 − 32) = 976 (c) Chain-sampling: n = 32.32) = 0.9958.0310)(0. p = 0. n) = P (0.9967 ATI = 32 + (1 − 0. i = 3. the Pa for chain sampling is slightly larger. but the average number inspected is much smaller.32) = 0. n)[ P(0. ATI = 32 + (1 − 0.001 Pa = P(0.

0040 0.00 .0600 0.6561 0.70 0.2713 0.60 0.2492 0.4377 0.0000 OC Curve for ChSP-1 n=4.60 0.4096 0.0388 0.1095 0.7000 0.3456 0.0626 0.4096 0.90 1.0081 0.4) 0.3000 0.9072 0.90 0.0256 0.8423 0.0100 0.50 p 0.0300 0.8756 0.9613 0.40 0.0756 0.1715 0.8000 0.40 0.9960 0.9815 0.30 0.4) 0.1296 0.0256 0.4116 0.xls : worksheet Ex15-1 p 0.6000 0. i = 3 Excel workbook Chap15.00 0.0001 0.9500 P(0.9950 0.8853 0.50 0.20 0.1000 0. c = 0.2252 0.0256 0.1993 0.30 0.7732 0.5000 0.7385 0.00 0.0800 0.0010 0.20 0.0700 0.00 0.1304 0.2401 0.7481 0.0625 0.80 0.Chapter 15 Exercise Solution 15-12.0081 0.0753 0.2000 0.8145 0.2458 0.2916 0.8080 0.0016 0.0900 0.10 0.6857 0.9000 0.10 0.9224 0.80 0.70 Pa 0.9606 0.4000 0.0500 0. n = 4.0005 Pa 0.c=0 1.7164 0.7807 0.0200 0.0016 0.2500 0.0001 0.9999 0.0036 0.0000 P(1.1536 0.

0000 1.8863 1.0450 0. n = 6 If c = 0.2749 6.2222 0.6266E+04 40.2151 7.000.88584 + 0.0000 3.2024.0000 f = 1/100 and i = 1302 u v Pa 2.8266 1.0000 1.0000 0.02} Pa = P(0.0346E+02 400.5714 0. v = 1333.0000 .5035E+02 2000.9985 1.0001 4.9429 6.2804E+15 4000.4085E+11 250.8863 3.3521E+06 12500.8608 1.0100 0.3333 0.0040 0.0300 0.7927 1.9739 4.0800 0.0000 0.1111 0.8284E+02 222.5471E+07 20.0000 6.Chapter 15 Exercise Solution 15-13.6628 2.0000 6.0000 3.1429 0.0001 1.1429 0. f = 1/10 and i = 550 3.0000 0.95264 15-14.0000 0.2867E+08 333.5056E+04 1000.0250 0.0000 0.6667 0. Average process fallout.9562E+02 444.1116 1.7157E+05 666.0000 0.5984E+02 250.0350 0.0000 0.7.9301 1.3333 0.0010 0.8192E+07 10000.0162E+03 2500.1088 1.0020 0.4383E+07 11111.4121E+27 2222.0150 0.5536 2.3395E+06 14285.0000 0.8106E+02 571.0000 2.4676E+26 100.7594 1.7.2222 0.5602E+03 50.3333 0.8816E+02 500.0050 0.4774 2.7998 1.5592E+02 1333.2131E+05 16666.0500 0.3333 0.0000 0.0000 0.0561 2. v = 66.2222 0.2451E+09 285.3373E+02 10000.8648E+02 133.0384 2.0000 0. f = 1/10 and i = 550: u = 855.0000 0.3333 0.8885 1. Three different CSP-1 plans with AOQL = 0.6667 0.0318E+21 125.5714 0. AFI = 0.3638E+05 20000.0080 0.0000 0.0035E+16 166.0025 0.0024 3.7143 0.4329E+03 2222.4400 3.0010E+04 40000.0401E+03 66666.6788E+02 800.6619E+03 1428.0000 0.6667 0.3439E+10 6666.3526 7.2222 0. 6)]4 = 0.6667 0. f = ½ and i = 140 2.0000 0.1279 4.6667 0.0106 4.6331E+04 33333.0000 0.0000 0.0035 0.0700 0.7431E+02 666.5714 0.6667 0.7161E+04 28571.0032 2.3.0000 2.4286 0.5553E+02 6666.4066E+03 3333.0000 0.0000 0.3333 0.3984E+04 44.0000 0.0000 0.0381 1.5127 4.0000 8. Pa = 0.6667 0.4444 0.0000 0.0015 9.0015 0.0238E+04 1250.0000 2.3972E+03 1666.0000 2.0000 0.6195E+36 1666.0030 0.9590E+02 100.7266 1.4286 0.0000 0. AFI = 0.0045 0.9502E+03 2000.7692E+60 1000.0000 0.0070 0.8953 1.0090 0.6175E+02 1000.0000 1. 6) + P(1.0600 0.6355E+04 33.0135 7.0000 3.0000 0.5492E+42 1428.7410E+23 111.8328 1. Pa = 0.198% would be: 1.6944 2.0652 2.0000 1.0037E+03 5000.666.0000 0.1111 0.0277 4.6712 2.2222 0.5912E+04 25000.3652E+54 1111.0179E+30 2000.3467E+06 500.0666.0000 0.9316 8.6667 0.0251E+06 22.0000 2.9936E+21 2857. AFI = 0.3909E+02 285.3371E+03 66.6667 0.0900 0.0000 0.7586E+48 1250. p = 0.3333 0. Pa = 0.4904 2.915.6790E+03 100000.0000 0.2128E+12 222.1000 f = 1/2 and i = 140 u v Pa 1.3449E+03 80. If c = 1. 6)[ P(0.1111 0. N = 500. accept if i = 4.1429 0.4619E+07 400. Need to find Pa{p = 0.6020 2.2222 0.9429 p 0.88584) 4 = 0.0000 0.5523.5930E+04 1111.7032 1.1848E+03 4000.8675E+04 22222.0000 9.5729E+18 3333.0000 1.0000 1.15% = 0.5731E+13 200.4444 0.0002 2.6667 0.3333 0. f = ½ and i = 140: u = 155.0060 0.0000 0.5444 2.10847(0.0000 0.6795E+03 2857.1766 1.8953 2. f = 1/100 and i = 1302 15-15.3262E+13 5000.0000 5.0000 f = 1/10 and i = 550 u v Pa 7.0003 3.2765E+03 50000.0000 3.3934 4.4040 4.0852E+18 142.8571 0. v = 6666. f = 1/100 and i = 1302: u = 4040.0000 0.6921E+05 28.0000 1.0000 0.0000 3.0000 0.0200 0.1604E+03 57.0000 0.2037E+02 333.0255E+24 2500.7143 0.0000 3. accept.4676E+06 25.7143 0.7715 1.0015 and q = 1 – p = 0.0000 1.0066 1.0000 0.530.0000 0.1111 0.6298 2.0400 0.2202 1.0029 1.0839E+02 200.0008 4.

Plan A: AFI = 0.90% Plan A: f = 1/5 and i = 38 Plan B: f = 1/25 and i = 86 15-17. CSP-1 with AOQL = 1. .0375} = 0.5272 and Pa{p = 0.6043 Plan B: AFI = 0.Chapter 15 Exercise Solution 15-16.4925 Prefer Plan B over Plan A since it has a lower Pa at the unacceptable level of p.5165 and Pa{p = 0.0375} = 0.

Sign up to vote on this title

UsefulNot useful- Facilities Planning,Tompkins ,A. White,A. Bozer, Tanchoco,4ed Solution Manualby mustang_mujtaba
- 77500896 Statistical Quality Controlby ENRIQUE_POMALES683
- Douglas C. Montgomery-Design and Analysis of Experiments, Seventh Edition-Wiley(2008)by Thais Beraldo Amaral
- Applied+Multivariate+Statistical+Analysis+6E【课后习题答案】by xyunchao

- Introduction to Statistical Quality Control, 5th Edition
- Douglas C Montgomery - Introduction to Statistical Quality Control
- Introduction to Statistical Quality Control, 5th Edition
- Quality 2
- SQC Chapter 6-11
- QC Solution Manual Chapter 6, 6th Edition
- facility manual
- Introduction to Statistic Quality Control
- Facilities Planning,Tompkins ,A. White,A. Bozer, Tanchoco,4ed Solution Manual
- 77500896 Statistical Quality Control
- Douglas C. Montgomery-Design and Analysis of Experiments, Seventh Edition-Wiley(2008)
- Applied+Multivariate+Statistical+Analysis+6E【课后习题答案】
- Ch02 Solution
- charts_QP-104
- Exercises Solutions
- Design And Analysis Of Experiments, 5Th Edition (Douglas C Montgomery)
- Production and Operation Analysis
- Methods, Standards and Work Design, (11th Ed) Solution Manual
- 1 Introduction to Statistical Quality Control
- SQC CH6
- Facilities planning Manual
- Facilities Planning - James Tompkins
- Assignment #3
- Introduction to Statistical Quality Control
- Engineering Economy 7th Edition Solution Manual [Blank & Tarquin]
- tugas5ppmNadPrils
- Introduction to Statistical Quality Control
- Statistical Quality Control_final
- Solution Manual -Quality Control 5th Edition Montgomery