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

with no bends or curves.36 80 1. However.Chapter 2 Exercise Solutions 2-13 continued Probability Plot of Viscosity Data (Ex2-8) Weibull Percent 99.010 10 5 3 2 1 0. the plot points on the Weibull probability plot are not straight—particularly in the tails— indicating it is not a reasonable model.10 15.9 99 Shape Scale N AD P-Value 90 80 70 60 50 40 30 20 16.1 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. 2-10 . the plot points tend to fall along a straight line.

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 0.8537 20 0. MTB > Graph > Probability Plot > Single (In the dialog box.Chapter 2 Exercise Solutions 2-14 ☺.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 8.163 Percent 80 70 60 50 40 30 20 10 5 1 1000 10000 Cycles to Failure 100000 2-11 .549 0.

though the Weibull distribution appears to best fit the data in the tails.464 9624 20 0.336 >0.250 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.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. 2-12 .

Chapter 2 Exercise Solutions 2-15 ☺. ppm 100.426 <0.56 40 8.9347 1.0 2-13 .1 1.005 Loc Scale N AD P-Value 0.651 40 0.0 Concentration.0 10. 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. MTB > Graph > Probability Plot > Single (In the dialog box.873 Percent 80 70 60 50 40 30 20 10 5 1 -50 0 50 Concentration.470 22. select Distribution to choose the distributions) Probability Plot of Concentration (Ex2-15) Normal 99 95 90 Mean StDev N AD P-Value 9.201 0.

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 0.000 10.001 0.6132 5. 2-14 .091 30 20 10 5 3 2 1 0. Plotted points on the normal and Weibull probability plots tend to fall off a straight line.010 0.000 Concentration.782 40 0.000 The lognormal distribution appears to be a reasonable model for the concentration data.637 0. ppm 100.100 1.

as evidenced by potentially cyclic behavior. shifts or obvious patterns in the 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 . indicating that time is not an important source of variability.Chapter 2 Exercise Solutions 2-16* (2-9). 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. 2-17* (2-10). there are no trends.

as well as occasional “spikes” around 20.438 StDev 4.476 Maximum 98.000 SE Mean 0.100 Median 89. MTB > Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Ex2-7 Variable Ex2-7 Variable Ex2-7 N N* Mean 90 0 89.250 Q3 93. 2-19 (2-11). 35).Chapter 2 Exercise Solutions 2-18 ☺. The order in which the data were collected may be an important source of variability.600 Q1 86.125 2-16 .158 Minimum 82. 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. there are two unusually large readings (9.

04 16.02 16. and 7. The stem-and-leaf plot suggests that certain values may occur more frequently than others. 6. The data appear to be fairly well scattered.06 Fluid Ounces 16. those ending in 1. for example.00 2-17 . MTB > Graph > Boxplot > Simple Boxplot of Detergent Data (Ex2-1) 16. 2-21 (2-13).07 16.Chapter 2 Exercise Solutions 2-20 (2-12). 4. MTB > Graph > Stem-and-Leaf Stem-and-Leaf Display: Ex2-7 Stem-and-leaf of Ex2-7 Leaf Unit = 0.05 16.10 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.01 16.03 16.

6. 4.9975 49. ( ) ( ( ) ) ( ) ( ) ⎧1/ 36. otherwise 2-24 (2-16).. 7. 2} + Pr{3. 2} + Pr{2. x = 11 1/ 36.0050 mm 50. 10. 9. x = 2 2 / 36. x = 9 3 / 36.3} + Pr{2.1} = 1 × 1 + 1 × 1 = 2 6 6 6 6 36 Pr{x = 4} = Pr{1.92 − 7 2 11 = 0.1} = 1 × 1 + 1 × 1 + 1 × 1 = 3 6 6 6 6 6 6 36 .Chapter 2 Exercise Solutions 2-22 (2-14). 3. x = 8 4 / 36. x = 12 0. x = 7 p ( x) = ⎨ ⎩5 / 36. x: {the sum of two up dice faces} sample space: {2. MTB > Graph > Boxplot > Simple Boxplot of Bearing Bore Diameters (Ex2-2) 50.0000 49. 5. x = 5 5 / 36. 11.9950 2-23 (2-15). 12} Pr{x = 2} = 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. 8. x = 6 6 / 36.. x = 10 2 / 36.0075 50. x = 4 4 / 36.1} = 1 × 1 = 1 6 6 36 Pr{x = 3} = Pr{1.0025 50. x = 3 3 / 36.38 10 2-18 .

Chapter 2 Exercise Solutions 2-25 (2-17). µ = 1/λ = 1 (Eqn. 2-33) 2-27 (2-19). This is a Poisson distribution with parameter λ = 0. x = 1 (1 + 2k ) / 3. e −0.0100.01)0 = 1 − 0.02). x = 3 0. ⎧(1 + 3k ) / 3. 2-32) σ2 = 1/λ2 = 1 (Eqn. from 0.5 + 5k ) / 3.01).0198 0! (c) This is a Poisson distribution with parameter λ = 0.0100 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. +∞ For f(x) to be a probability distribution.01 (0. ∫ f ( x)dx must equal unity. x ~ POI(0. x ~ POI(0.02 (0.5 + 5k ) =1 3 10k = 0.01.05 2-19 .0198 to 0.02)0 = 1 − 0. 2-26 (2-18).02. x = 2 p ( x) = ⎨ ⎩(0. otherwise (a) ∞ To solve for k. −∞ ∞ −x −x ∞ ∫ ke dx = [−ke ]0 = − k[0 − 1] = k ⇒ 1 0 This is an exponential distribution with parameter λ=1.0196 1! (b) Pr{x ≥ 1} = 1 − Pr{x = 0} = 1 − p(0) = 1 − e −0. (a) Pr{x = 1} = p(1) = e −0.02 (0. use F ( x) = ∑ p( xi ) = 1 i =1 (1 + 3k ) + (1 + 2k ) + (0.02)1 = 0.9802 = 0.5 k = 0.9900 = 0.

750.5 + 5(0. (a) This is an exponential distribution with parameter λ = 0.15 ⎧ = 0.75 = 1.05) ⎤ ⎡1 + 2(0. x = 3 ⎪ 3 ⎩ 2-28 (2-20).000. 0 < r < 1. x = 0.15 + 1.383) + 22 (0.615 i =1 (c) 1.1 + 0.90/calculator.367) + 32 (0. x = 2 3 ⎪ ⎪1. p ( x) = kr x . cost = $50/calculator Sale profit = $25/calculator Net profit = $[-50(1 + 0.1.1 ⎪ F ( x) = ⎨ = 0.125(1) = 0. 2.05) ⎤ ⎡ 0.15 + 1. x = 1 ⎪ 3 ⎪ 1. 2-20 .867 2 = 0.8% will fail during the first year.10/calculator. The effect of warranty replacements is to decrease profit by $5.118 Approximately 11. (b) Mfg.867 3 3 3 ⎣ ⎦ ⎣ ⎦ ⎣ 3 µ = ∑ xi p( xi ) = 1× ⎢ i =1 3 σ 2 = ∑ xi2 p( xi ) − µ 2 = 12 (0.Chapter 2 Exercise Solutions 2-27 continued (b) ⎡1 + 3(0.05) ⎤ + 2× ⎢ + 3× ⎢ ⎥ ⎥ ⎥⎦ = 1.383.118) + 75]/calculator = $19.… ∞ F ( x) = ∑ kr x = 1 by definition i =0 k ⎡⎣1 (1 − r ) ⎤⎦ = 1 k = 1− r 2-29 (2-21).250) − 1.125: Pr{x ≤ 1} = F (1) = 1 − e −0.

Chapter 2 Exercise Solutions 2-30 (2-22).02) x (1 − 0. 0. and the process will be stopped to look for a cause.875 − 11. This exercise may also be solved using Excel or MINITAB: (1) Excel Function BINOMDIST(x. x ~ BIN(25. 4 ⎛ 50 ⎞ Pr{ pˆ ≤ 0.777821 2-32* (2-24).02)(50− x ) x =0 ⎝ x ⎠ ⎛ 50 ⎞ ⎛ 50 ⎞ = ⎜ ⎟ (0.04} = Pr{x ≤ 2} = ∑ ⎜ ⎟(0.75) dx = 11.02)1 (1 − 0. ⎛ 25 ⎞ Pr{x ≥ 1} = 1 − Pr{x < 1} = 1 − Pr{x = 0} = 1 − ⎜ ⎟ (0.01)0 (1 − 0.02) 4 (1 − 0.78 = 0.125 12 11. The process is stopped if x ≥ 1.75 4 x2 2 12 − 47 x 11. p.75 2-31* (2-23).64 ⎝0⎠ 2-33* (2-25).01) 25 = 1 − 0.02)0 (1 − 0.02 and n = 50. TRUE) (2) MTB > Calc > Probability Distributions > Binomial Cumulative Distribution Function Binomial with n = 25 and p = 0. This is a binomial distribution with parameter p = 0.02)50 + ⎜ ⎟ (0. This is a somewhat difficult operating situation. ⎛ 25 ⎞ Pr{x ≥ 1} = 1 − Pr{x < 1} = 1 − Pr{x = 0} = 1 − ⎜ ⎟ (0. 12 Pr{x < 12} = F (12) = ∫ −∞ 12 f ( x) dx = ∫ 4( x − 11.75 = 0.01 x P( X <= x ) 0 0.22 ⎝0⎠ This decision rule means that 22% of the samples will have one or more nonconforming units.04) Stop process if x ≥ 1.75 = 11.02) 49 + ⎝0⎠ ⎝1⎠ ⎛ 50 ⎞ + ⎜ ⎟ (0.04)0 (1 − 0. This is a binomial distribution with parameter p = 0.04) 25 = 1 − 0.36 = 0.02) 46 = 0.01 and n = 25. n.921 ⎝4⎠ 2-21 .

This is a binomial distribution with parameter p = 0.0100) + 0.079 k=2 1 − Pr{x ≤ n(kσ + p )} = 1 − Pr{x ≤ 100(2(0.01) (0.99) + ⎜ ⎟ (0.99)100− x = 1 − ⎢0.Chapter 2 Exercise Solutions 2-34* (2-26).992] = 0.0100) + 0.01) (0.99)97 ⎥ = 1− ∑ ⎜ ⎟ ⎟ x =0 ⎝ x ⎠ ⎝ 3 ⎠ ⎣ ⎦ = 1 − [0.01) (0.01 and n = 100.982 + ⎜ (0.99) ⎥ ⎝ 1 ⎠ ⎝ 2 ⎠ ⎣⎝ 0 ⎠ ⎦ = 1 − [0.01) x (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(0.01) x (0.01) (1 − 0.982] = 0.0100) + 0.921 + ⎜ (0.99)100− x = 1 − ⎢0.01) x =0 ⎝ x ⎠ ⎡⎛100 ⎞ ⎤ ⎛ 100 ⎞ ⎛100 ⎞ 0 100 1 99 2 98 = 1 − ⎢⎜ ⎟ (0.01)} = 1 − Pr{x ≤ 2} 2 ⎛ 100 ⎞ 100 − x x = 1− ∑ ⎜ ⎟(0.01) 100 = 0.921] = 0.003 2-22 .01)} = 1 − Pr{x ≤ 4} 4 ⎛ 100 ⎞ ⎡ ⎤ ⎛100 ⎞ (0.99) + ⎜ ⎟ (0. σ = 0.01(1 − 0.01) 4 (0.01)3 (0.01)} = 1 − Pr{x ≤ 3} 3 ⎛ 100 ⎞ ⎡ ⎤ ⎛100 ⎞ (0.99)96 ⎥ = 1− ∑ ⎜ ⎟ ⎟ x =0 ⎝ x ⎠ ⎝ 4 ⎠ ⎣ ⎦ = 1 − [0.018 k=3 1 − Pr{x ≤ n(kσ + p )} = 1 − Pr{x ≤ 100(3(0.

08 and n = 5.1.633333 (b) For the binomial approximation to the hypergeometric. p = D/N = 2/25 = 0. 649) Pr{Acceptance} = p (0) = ⎝ ⎠ ⎝ = = 0. This is a hypergeometric distribution with N = 25 and n = 5.20 be less than the suggested 0. n. The binomial approximation is not satisfactory in this case.130) ⎛ 25 ⎞ ⎜ ⎟ ⎝5⎠ This exercise may also be solved using Excel or MINITAB: (1) Excel Function HYPGEOMDIST(x. n/N = 5/150 = 0. (c) For N = 150. N) (2) MTB > Calc > Probability Distributions > Hypergeometric Cumulative Distribution Function Hypergeometric with N = 25. violates the rule-ofthumb that n/N = 5/25 = 0. though close to the exact solution for x = 0. 2-23 . M = 2.1. and n = 5 x P( X <= x ) 0 0. ⎛5⎞ Pr{acceptance} = p (0) = ⎜ ⎟ (0.08)0 (1 − 0.Chapter 2 Exercise Solutions 2-35* (2-27).633 (53. without replacement. so the binomial approximation would be a satisfactory approximation the hypergeometric in this case.659 ⎝0⎠ This approximation. D.08)5 = 0. (a) Given D = 2 and x = 0: ⎛ 2 ⎞ ⎛ 25 − 2 ⎞ ⎜ ⎟⎜ ⎟ 0 5 − 0 ⎠ (1)(33.033 ≤ 0.

457. ⎛ 5 ⎞ ⎛ 25 − 5 ⎞ ⎛ 5 ⎞ ⎛ 20 ⎞ ⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟ 0⎠⎝ n − 0 ⎠ ⎝ 0⎠⎝ n ⎠ ⎝ = p(0) = ⎛ 25 ⎞ ⎛ 25 ⎞ ⎜ ⎟ ⎜ ⎟ ⎝n⎠ ⎝n⎠ try n = 10 ⎛ 5 ⎞ ⎛ 20 ⎞ ⎜ ⎟⎜ ⎟ 0 10 (1)(184.369 (142. 760) ⎛ 25 ⎞ ⎜ ⎟ ⎝ 10 ⎠ try n = 11 ⎛ 5 ⎞ ⎛ 20 ⎞ ⎜ ⎟⎜ ⎟ 0 11 (1)(167. and D = 3.057 p(0) = ⎝ ⎠ ⎝ ⎠ = (3. 756) = 0. This is a hypergeometric distribution with N = 30. ⎛ 3 ⎞⎛ 30 − 3 ⎞ ⎜ ⎟⎜ ⎟ 1 ⎠⎝ 5 − 1 ⎠ (3)(17.05. n = 5. or equivalently Pr{x = 0 | D = 5} < 0. 2-36 (2-28).506) ⎛ 30 ⎞ ⎜ ⎟ ⎝5⎠ ⎛ 3 ⎞ ⎛ 27 ⎞ ⎜ ⎟⎜ ⎟ 0 5 Pr{x ≥ 1} = 1 − Pr{x = 0} = 1 − p(0) = 1 − ⎝ ⎠ ⎝ ⎠ = 1 − 0.960) = 0.95.038 p(0) = ⎝ ⎠ ⎝ ⎠ = (4.550) ⎝ Pr{x = 1} = p (1) = = = 0.Chapter 2 Exercise Solutions 2-35 continued (d) Find n to satisfy Pr{x ≥ 1 | D ≥ 5} ≥ 0. 400) ⎛ 25 ⎞ ⎜ ⎟ ⎝ 11 ⎠ Let sample size n = 11. 268.433 ⎛ 30 ⎞ ⎜ ⎟ ⎝5⎠ 2-24 .567 = 0.

364 ⎝0⎠ Pr{x ≥ 2} = 1 − Pr{x ≤ 1} = 1 − [Pr{x = 0} + Pr{x = 1}] = 1 − p(0) − p(1) ⎛ 50 ⎞ = 1 − 0.095 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.264 ⎝1⎠ 2-38 (2-30). x.Chapter 2 Exercise Solutions 2-37 (2-29).020)1 (1 − 0. the binomial distribution can be used to approximate the hypergeometric.01)1 = 0. This is a Poisson distribution with λ = 0.1 x P( X <= x ) 0 0.99999 = 0.364 − ⎜ ⎟ (0.364) = 0.01 errors/bill.020)0 (1 − 0.1 (0.364 − 0.020)50−0 = (1)(1)(0.00001 0! 2-40 (2-32).1 defects/unit. Pr{x ≥ 1} = 1 − Pr{x = 0} = 1 − e −0.00001 (0.1)0 = 1 − 0. Checking n/N = 50/500 = 0. TRUE) (2) MTB > Calc > Probability Distributions > Poisson Cumulative Distribution Function Poisson with mean = 0. This is a hypergeometric distribution with N = 500 pages. This is a Poisson distribution with λ = 0.020)50−1 = 1 − 0. n = 50 pages. with p = D/N = 10/500 = 0.0099 Pr{x = 1} = p (1) = 1 2-25 .372 = 0. e −0.904837 2-39 (2-31). and D = 10 errors.905 = 0.00001)0 = 1 − 0. This is a Poisson distribution with λ = 0. e −0.020.00001 stones/bottle.01 (0. ⎛ 50 ⎞ Pr{x = 0} = p(0) = ⎜ ⎟ (0.1.1 ≤ 0.

and x = 1 + (5000/100) = 51. Pr(t ) = p(1 − p)t −1 .6} 5 ⎭ ⎩ = 1 − Φ (1.010 0.159) = 7950.3. is (50.617290) = 0.? 35 − 40 ⎫ ⎧ Pr{x ≤ 35} = Pr ⎨ z ≤ ⎬ = Pr{z ≤ −1} = Φ (−1) = 0.000001)(0. σ. r = 3 welds.01.? 48 − 40 ⎫ ⎧ Pr{x > 48} = 1 − Pr{x ≤ 48} = 1 − Pr ⎨ z ≤ ⎬ = 1 − Pr{z ≤ 1. LSL = 35 lb. 2. This is a Pascal distribution with Pr{defective weld} = p = 0. This exercise may also be solved using Excel or MINITAB: (1) Excel Function NORMDIST(X.01) 3 1 − ⎝ ⎠ Pr{x > 51} = Pr{r = 0} + Pr{r = 1} + Pr{r = 2} ⎛ 50 ⎞ ⎛ 50 ⎞ ⎛ 50 ⎞ = ⎜ ⎟ 0.0110. t = 1.159 5 ⎭ ⎩ So.Chapter 2 Exercise Solutions 2-41 (2-33).055 So.945 = 0. the number that fail the minimum specification are (50. TRUE) (2) MTB > Calc > Probability Distributions > Normal Cumulative Distribution Function Normal with mean = 40 and standard deviation = 5 x P( X <= x ) 35 0.055) = 2750.9862 ⎝0⎠ ⎝1⎠ ⎝2⎠ 2-43* (2-35). 52). x ~ N (40. n = 50.000) × (0.9948 = 0. µ. ⎛ 51 − 1⎞ 3 51−3 Pr{x = 51} = p(51) = ⎜ = (1225)(0.000) × (0.… ∞ d ∞ 1 µ = ∑ t ⎡⎣ p(1 − p)t −1 ⎤⎦ = p ⎡ ∑ q t ⎤ = ⎢ ⎥ t =1 dq ⎣ t =1 ⎦ p 2-42 (2-34).9949 ⎜ ⎟ 0.9950 + ⎜ ⎟ 0. the number that exceed 48 lb.000 How many fail the minimum specification.0008 ⎟ (0. 2-26 .6) = 1 − 0.01) (1 − 0.012 0.158655 How many exceed 48 lb.

015 z −3.95 − 5 ⎞ = Φ⎜ ⎟−Φ⎜ ⎟ = Φ (2.5) = 0. Φ −1 (0. Shifting the process mean in either direction would increase the number of nonconformities produced.5) − Φ (−2. 42 ).02 ⎠ ⎝ 0. is currently centered between the specification limits (target = 5 V).9991) + 32 = 40.8571) = 1 − 0. 2-46 (2-38). USL = 5. 0. Assume that the process remains centered between the specification limits at 5 V.00621 = 0.9991 4 µ = −4(−1.99379 − 0.0 2-47 (2-39).0005) = −3. x ~ N(5.0005 z = Φ −1 (0. Desire Pr{Conformance} = 1 / 1000 = 0. x ~ N ( µ . x ~ N(900.0228.9979 = 0.05 − 5 ⎞ ⎛ 4.98758 ⎝ 0.0228) = −1.022). with mean 5 V.001 / 2 = 0.95 V.02 ⎠ 2-45* (2-37). so σ = LSL − µ 4.0021 2-27 .05 V Pr{Conformance} = Pr{LSL ≤ x ≤ USL} = Pr{x ≤ USL} − Pr{x ≤ LSL} ⎛ 5. 352) Pr{x > 1000} = 1 − Pr{x ≤ 1000} 1000 − 900 ⎫ ⎧ = 1 − Pr ⎨ x ≤ ⎬ 35 ⎩ ⎭ = 1 − Φ (2. Need Pr{x ≤ LSL} = 0.0005.29 z= LSL − µ σ .95 − 5 = = 0.9991 32 − µ = −1. Find µ such that Pr{x < 32} = 0. The process. LSL = 4.001.0152 to have at least 999 of 1000 conform to specification.Chapter 2 Exercise Solutions 2-44* (2-36). Φ ( z ) = 0.29 Process variance must be reduced to 0.

000 1.500 ⎫ ⎧ ⎧ = 1 − Pr ⎨ z2 ≤ ⎬ + Pr ⎨ z2 ≤ ⎬ 500 500 ⎩ ⎭ ⎩ ⎭ = 1 − Φ (5) + Φ (−5) = 1 − 1.0000 + 0.500 ⎫ 5. 000 − 7. USL = 10. x1 ~ N(7500. then choose process 1 2-28 .0000 profit for process 2 = 10 (1 – 0.5758 LSL − 5000 = −2.500 ⎫ 5.0062 = 0.005 Φ −1 (0.9380 – c1 For Process 2 proportion defective = p2 = 1 − Pr{LSL ≤ x2 ≤ USL} = 1 − Pr{x2 ≤ USL} + Pr{x2 ≤ LSL} 10.000 h.5758) + 5000 = 4871 2-49 (2-41). x2 ~ N(7500.0000 = 0.0124 profit for process 1 = 10 (1 – 0.0000) – c2 = 10 – c2 If c2 > c1 + 0. 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. 000 − 7. 000 − 7. σ22 = 5002).9938 + 0.0124) + 5 (0.500 ⎫ ⎧ ⎧ = 1 − Pr ⎨ z1 ≤ ⎬ + Pr ⎨ z1 ≤ ⎬ 1. x ~ N(5000.0000) + 5 (0.005) = −2. defect = $5/unit.5) = 1 − 0. 000 ⎩ ⎭ ⎩ ⎭ = 1 − Φ (2.5758 50 LSL = 50(−2.5) + Φ (−2.0620. σ12 = 10002).000 h sales = $10/unit.0124) – c1 = 9. 502).Chapter 2 Exercise Solutions 2-48 (2-40). 000 − 7. Find LSL such that Pr{x < LSL} = 0. LSL = 5.

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.

x = 0.. For the Poisson distribution.. x = 0. ⎛n⎞ For the binomial distribution.… 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 .. p( x) = ∞ ∞ ⎛ e− λ λ x ⎝ x! µ = E[ x] = ∑ xi p( xi ) = ∑ x ⎜ i =1 x =0 e− λ λ x .Chapter 2 Exercise Solutions 2-51 (2-43).1.1.. 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). p ( x) = ⎜ ⎟ p x (1 − p) n − x .

For the exponential distribution. f ( x ) = λ e − λ x .Chapter 2 Exercise Solutions 2-53 (2-45). x ≥ 0 For the mean: +∞ +∞ 0 0 µ = ∫ xf ( x)dx = ∫ x ( λ e − λ x )dx Integrate by parts. 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. 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 .

σ = 0. H1: µ ≠ 8. and conclude that the mean bearing ID is not equal to 8. σ = 2 psi (a) µ0 = 125.002 ) ( 15 ≤ µ ≤ 8.78)] = 2[1 − 1.025 = 1. x = 127 psi.002 15 Zα/2 = Z0.002 N Mean SE Mean 95% CI 15 8.96 0.25 vs.00000] = 0 (c) ⎞ ≤ µ ≤ x + Z ⎛σ ⎞ x − Zα / 2 ⎛⎜ σ ⎟ ⎟ α /2 ⎜ n⎠ n⎠ ⎝ ⎝ ( 8.05/2 = Z0.78 σ n 0. (b) P-value = 2[1 − Φ(Z0)] = 2[1 − Φ(6. Reject H0 if |Z0| > Zα/2.05 Test H0: µ = 125 vs.05 Test H0: µ = 8.000 3-2.25101) Z -6.249 ≤ µ ≤ 8.251 MTB > Stat > Basic Statistics > 1-Sample Z > Summarized data One-Sample Z Test of mu = 8. n = 8. 8.25.25 cm.002 cm (a) µ0 = 8.2535 The assumed standard deviation = 0.24899.645 Reject H0: µ = 125.25. Reject H0 if Z0 > Zα.Chapter 3 Exercise Solutions 3-1. and conclude that the mean tensile strength exceeds 125 psi. 3-1 . H1: µ > 125.05 = 1. α = 0.25 − 1.2535 vs not = 8.2535 − 8. x − µ0 127 − 125 Z0 = = = 2.002 15 ) 8.96 0. x = 8. α = 0. x − µ0 8.25 Z0 = = = 6.828 σ n 2 8 Zα = Z0.2535 cm.25.25000 0.78 P 0.96 Reject H0: µ = 8.00052 (8. n = 15.25 + 1.

α = 0. 10−1 = 1. A one-sided hypothesis test lets us do this. µ0 = 25.000 0. H1: µ > 25.0. and conclude that the mean life exceeds 25 h.05 Test H0: µ = 25 vs.042 3-2 .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. x ~ N(µ.83 P 0. we usually are interested in whether some minimum requirement is met.00234 (c) In strength tests.0 − 25 t0 = = = 1. Reject H0 if t0 > tα. x − µ0 26.833 Reject H0: µ = 25.Chapter 3 Exercise Solutions 3-2 continued (b) P-value = 1 − Φ(Z0) = 1 − Φ(2.62 10 tα.5138 95% Lower Bound 25.828) = 1 − 0. σ).99766 = 0.0581 T 1. n−1 = t0.62.837 2.0000 StDev 1. s = 1.707 125. n = 10 (a) x = 26.05.95 P 0.002 3-3. (d) ( n) ≤ µ 127 − 1. not simply that the mean does not equal the hypothesized value.645 ( 2 8 ) ≤ µ x − Zα σ 125.952 S n 1. 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 26.6248 SE Mean 0.

95 P 0.n −1 S ( 26. 3-3 .n −1 S ) n ( 10 ≤ µ ≤ 26.0 − 1.Chapter 3 Exercise Solutions 3-3 continued (b) α = 0.5138 90% CI (25.114 0.0000 1.833 1.0 + 1.95% CI 99 Mean StDev N AD P-Value 95 90 26 1.62 10 ) 25.62 n ≤ µ ≤ x + tα / 2. so the assumption that battery life is normally distributed is appropriate.986 Percent 80 70 60 50 40 30 20 10 5 1 20 22 24 26 28 Lifetime.9419) T 1.10 x − tα / 2.6248 SE Mean 0. 26. Hours 30 32 The plotted points fall approximately along a straight line.0581.94 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 26.083 (c) MTB > Graph > Probability Plot > Single Probability Plot of Battery Service Life (Ex3-3) Normal .833 1.06 ≤ µ ≤ 26.625 10 0.

n −1 S ( 26. (a) x ~ N(µ. σ).3922.089 S n 0.0012 99% CI (13.0012 (b) α = 0.40020 MTB > Stat > Basic Statistics > 1-Sample t > Samples in columns One-Sample T: Ex3-5 Test of mu = 13.00391 P 0.39618 × 1000 Å. α = 0.2498 ( 0.3990) T -3.39618 − 3.833 1.3962 0.0 − 1.39618 + 3.013 n ≤ µ ≤ x + tα / 2.39216 ≤ µ ≤ 13. α = 0. 13.4.833 ( ) 10 ) ≤ µ n ≤µ x − tα .9 = 1.4 t0 = = = −3.09 P 0. tα. σ).4 Variable N Mean StDev SE Mean Ex3-5 10 13.3934.06 ≤ µ The manufacturer might be interested in a lower confidence interval on mean battery life when establishing a warranty policy. MTB > Stat > Basic Statistics > 1-Sample t > Samples in columns One-Sample T: Ex3-5 Test of mu = 13.n −1 S ( 13.3962 0.4 × 1000 Å. x − µ0 13.Chapter 3 Exercise Solutions 3-4.39618 − 13. x ~ N(µ. x = 26.262 Reject H0: µ = 13.4002) T -3.09 ) ( n) 10 ) ≤ µ ≤ 13. and conclude that the mean thickness differs from 13.013 3-4 .0039 0.00391 µ0 = 13.05 Test H0: µ = 13. s = 1.62 h.4.4 Variable N Mean StDev SE Mean Ex3-5 10 13.2498 0. 3-5.4 vs not = 13. n−1 = t0. n = 10.05.01 ( x − tα / 2. H1: µ ≠ 13.n −1 S 10 ) 13.00391 10 tα/2.0039 0.4 vs not = 13. s = 0. Reject H0 if |t0| > tα/2.62 25. 9 = 2.4 × 1000 Å.05. n−1 = t0.4 vs.025. x = 13. 13. n = 10.00391 95% CI (13.0 h.

030 Test H0: µ = 12 vs.40 0. x = 12.711 Percent 80 70 60 50 40 30 20 10 5 1 13. µ0 = 12.57 P 0. x − µ0 12.015. so the assumption that battery life is normally distributed is not appropriate. s = 0. Reject H0 if t0 > tα. (a) x ~ N(µ. instead of a straight line.0303 10 tα/2.9880 T 1. x1000 Angstroms 13.0303 SE Mean 0.5655 S n 0.01 n = 10.0096 99% Lower Bound 11.410 The plotted points form a reverse-“S” shape.237 0.395 13. α = 0.015 − 12 t0 = = = 1.390 13. n−1 = t0.076 3-5 . 3-6.95% CI 99 Mean StDev N AD P-Value 95 90 13. σ). 9 = 3. H1: µ > 12.385 13.405 Thickness.005. 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 12.003909 10 0. and conclude that there is not enough evidence that the mean fill volume exceeds 12 oz.380 13.400 13.0150 StDev 0.Chapter 3 Exercise Solutions 3-5 continued (c) MTB > Graph > Probability Plot > Single Probability Plot of Photoresist Thickness (Ex3-5) Normal .250 Do not reject H0: µ = 12.

90 11.025.03028 10 0. 12.9600.05.274 0.9600 4 ⎣ ) ) n ⎤ = total width ⎦ n ⎤ =1 ⎦ n = 246 3-6 .152 (c) MTB > Graph > Probability Plot > Single Probability Plot of Fill Volume (Ex3-6) Normal . ounces 12.95% CI 99 Mean StDev N AD P-Value 95 90 12. so the assumption that fill volume is normally distributed is appropriate.015 − 2. α = 0.025 = 1.262 ( S 10 ) ≤ µ ≤ 12.05 tα/2.582 Percent 80 70 60 50 40 30 20 10 5 1 11.00 12.57 P 0.0150 0. Zα/2 = Z0.993 ≤ µ ≤ 12.015 + 2.95 12.02 0.262 x − tα / 2.15 The plotted points fall approximately along a straight line.0096 95% CI (11. 3-7.n −1 S n ≤ µ ≤ x + tα / 2. total confidence interval width = 1 lb.10 12.037 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 12.05 Fill Volume.62 ( S 10 ) 11.9933. σ = 4 lb.0367) T 1. n−1 = t0.Chapter 3 Exercise Solutions 3-6 continued (b) α = 0.0303 SE Mean 0. find n ( ( 2 ⎡ Zα / 2 σ ⎣ 2 ⎡1. 9 = 2.n −1 S ( ) ( n) 12.

0.5046 − 0.0001 ) ( n) 25 ) ≤ µ ≤ 0.50456 ≤ µ ≤ 0. x = 10. 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 10.5025 vs not = 0.5025.5025 The assumed standard deviation = 0.5025.5025 vs.504600 0. µ0 = 0. Reject H0 if |Z0| > Zα/2.000020 (0. (a) x ~ N(µ. H1: µ ≠ 12.0001 N Mean SE Mean 95% CI 25 0.5025.0001 in Test H0: µ = 0.999 V (a) µ0 = 12.5025 = 105 Z0 = σ n 0. 15 = 2.259 − 12 t0 = = = −6. MTB > Stat > Basic Statistics > 1-Sample Z > Summarized data One-Sample Z Test of mu = 0.504639) Z 105.2498 95% CI (9.131 Reject H0: µ = 12. s = 0.000 3-7 .96 Reject H0: µ = 0.025.5046 in. α = 0. σ).5025. and conclude that the mean rod diameter differs from 0. α = 0.000 (b) P-value = 2[1 − Φ(Z0)] = 2[1 − Φ(105)] = 2[1 − 1] = 0 (c) ( x − Zα / 2 σ ( 0.Chapter 3 Exercise Solutions 3-8.7270.9990 SE Mean 0.00 P 0.7917) T -6. and conclude that the mean output voltage differs from 12V.05/2 = Z0.50464 3-9. x ~ N(µ.2594 0. x − µ0 10.05 Test H0: µ = 12 vs.960 ( 0. σ). n = 16.97 P 0.960 0.259 V. 10.5046 + 1.0001 n ≤ µ ≤ x + Zα / 2 σ 25 ) 0.504561. H1: µ ≠ 0. x −µ 0 = 0. x = 0. n−1 = t0.999 16 tα/2.0001 25 Zα/2 = Z0. σ = 0. Reject H0 if |t0| > tα/2.05 n = 25.025 = 1.971 S n 0.5046 − 1.

α = 0.9992 = 14.n −1 S 16 ) 9. n-1 or χ20 < χ 21-α/2.999 n ≤ µ ≤ x + tα / 2. n−1 = χ20. Reject H0 if χ 20 > χ 2α/2.488 6.262 Do not reject H0: σ 2 = 1.545 ≤ σ ≤ 2. σ 02 = 1.05 Test H0: σ 2 = 1 vs. the null hypothesis.792 (c) σ 02 = 1.999 (16 − 1)0.259 − 2.131( 0.9992 2 ≤σ ≤ 27.999 ) ( n) 16 ) ≤ µ ≤ 10.970 1 σ 02 χ2α/2. cannot be rejected. H1: σ 2 ≠ 1.391 2 0.025.259 + 2.975.131 0. n−1 = χ20.488 χ21−α/2. H0: σ 2 = 1.16−1 = 27.n −1 ≤σ2 ≤ (n − 1) S 2 χ2 1−α / 2.262 2 0.738 ≤ σ ≤ 1.16−1 = 6.727 ≤ µ ≤ 10. 3-8 . and conclude that there is insufficient evidence that the variance differs from 1.n −1 S ( 10. χ 02 = (n − 1) S 2 = (d) (n − 1) S 2 χα2 / 2. n −1 (16 − 1)0.Chapter 3 Exercise Solutions 3-9 continued (b) ( x − tα / 2. n-1. (16 − 1)0.546 Since the 95% confidence interval on σ contains the hypothesized value.

23 0.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 0.05.767 M ean S tD ev V ariance S kew ness Kurtosis N 10.738 1.00 (e) 2 = 7.998 0.75 11.999 0.116487 -0.150 12.492793 16 M inimum 1st Q uartile M edian 3rd Q uartile M aximum 12 8.n −1 = χ 0.2609 α = 0.50 9.75 10.2609 2 σ ≤ 2.546 Mean Median 9.00 10.436 2 3-9 .140 11.430 10.945 95% C onfidence Interv al for S tD ev 95% Confidence Intervals 0.50 10.370 9.000 95% C onfidence Interv al for M ean 9.062 σ ≤ 1.25 10.792 95% C onfidence Interv al for M edian 9.533 10.95.727 10.9992 σ ≤ 7.n −1 (16 − 1)0.259 0. χ12−α .15 σ2 ≤ (n − 1) S 2 χ12−α .

05/2 = Z0.96 −Zα/2 = −1. the assumption of a normal distribution for output voltage seems appropriate.682 2 2 0. (a) α = 0. n1 = 25. x2 = 2.07) − 0 Z0 = = = −7.04 − 2.025 = 1.0152 20 σ 1 n1 + σ 2 n2 Zα/2 = Z0.015 l.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. σ1 = 0.05. and conclude that there is a difference in mean net contents between machine 1 and machine 2.26 0. ( x1 − x2 ) − ∆ 0 (2. n2 = 20.07 l. 3-10.04 l. x1 = 2.96 Reject H0: µ1 – µ2 = 0.010 l.00000] = 0 3-10 .Chapter 3 Exercise Solutions 3-9 continued (f) MTB > Graph > Probability Plot > Single Probability Plot of Output Voltage (Ex3-9) Normal .230 0. (b) P-value = 2[1 − Φ(Z0)] = 2[1 − Φ(−7.9990 16 0.0102 25 + 0.95% CI 99 Mean StDev N AD P-Value 95 90 10. Reject H0 if Z0 > Zα/2 or Z0 < –Zα/2. ∆ 0 = 0 Test H0: µ1 – µ2 = 0 versus H0: µ1 – µ2 ≠ 0.682)] = 2[1 − 1. σ2 = 0.

376 0.120 n1 + n2 − 2 7+8−2 ( x1 − x2 ) − tα / 2. n2 = 8.11 Both use Pooled StDev = 0.115. S1 = 0.Chapter 3 Exercise Solutions 3-10 continued (c) ( x1 − x2 ) − Zα / 2 2 (2.006607 95% CI for difference: (-0.015 2 n2 σ 12 ≤ ( µ1 − µ 2 ) ≤ ( x1 − x2 ) + Zα / 2 n1 + σ 22 n2 2 ≤ ( µ1 − µ 2 ) ≤ (2. x2 = 1. and conclude that there is not sufficient evidence of a difference between measurements obtained by the two technicians. 3-11 . x1 = 1. 0.1152 + (8 − 1)0. We can conclude that the machines do not fill to the same volume.1204 P-Value = 0.125 0.383 − 1.n + n 1 2 −2 S p 1 n1 + 1 n2 (1.043 Ex3-11T2 8 1.1604(0. 13 = 2.07) − 1. (b) The practical implication of this test is that it does not matter which technician measures parts. the readings will be the same.n + n 1 2 −2 S p 1 n1 + 1 n2 ≤ ( µ1 − µ 2 ) ≤ ( x1 − x2 ) + tα / 2.141183) T-Test of difference = 0 (vs not =): T-Value = 0. (c) n1 = 7.917 DF = 13 Do not reject H0: µ1 – µ2 = 0. and an investigation should be undertaken to understand why.010 20 −0. 3-11. we would have been concerned that the technicians obtained different measurements.04 − 2.038 ≤ ( µ1 − µ 2 ) ≤ −0.376) + 2.376) − 2.022 25 + 0.376.383.025.1604(0.9600 0.383 0. Ex3-11T2 Two-sample T for Ex3-11T1 vs Ex3-11T2 N Mean StDev SE Mean Ex3-11T1 7 1.07) + 1. We can conclude that there is no difference in measurements obtained by the two technicians. n1+n2−2 = t0. (a) MTB > Stat > Basic Statistics > 2-Sample t > Samples in different columns Two-Sample T-Test and CI: Ex3-11T1.015 2 20 The confidence interval for the difference does not contain zero.125 α = 0.05.044 Difference = mu (Ex3-11T1) .127969.383 − 1. If the null hypothesis had been rejected.127 ≤ ( µ1 − µ 2 ) ≤ 0.04 − 2.120) 1 7 + 1 8 −0.010 σ 12 25 n1 + σ 22 + 0.mu (Ex3-11T2) Estimate for difference: 0.9600 0.1604 Sp = (n1 − 1) S12 + (n2 − 1) S22 (7 − 1)0. S2 = 0.141 The confidence interval for the difference contains zero.120) 1 7 + 1 8 ≤ ( µ1 − µ 2 ) ≤ (1.115 0.1252 = = 0. tα/2.

25 95% Bonferroni Confidence Intervals for StDevs 0.854 Lev ene's Test Test Statistic P-Value Ex3-11T2 0.05/ 2.05/ 2.920 0.15 0.8464 Fα / 2.01 0.n −1.975. Ex3-11T2 F-Test Test Statistic P-Value Ex3-11T1 0.119 2 1 1 2 2 2 2 F1−α / 2.8−1 = F0.05 Test H 0 : σ 12 = σ 22 versus H1 : σ 12 ≠ σ 22 .6. 3-12 .n 1 2 −1 2 or F0 < F1−α / 2. we would have been concerned about the difference in measurement variability between the technicians.7 = 0.7 −1.10 0.7 −1. If the null hypothesis is rejected.6.2 1. and conclude that there is no difference in variability of measurements obtained by the two technicians. 1 2 F0 = S S = 0.n −1.025.7 = 5.6 Do not reject H0.5 1.n −1.20 0.n −1.n 1 2 −1 = F1−0.8−1 = F0.30 Ex3-11T1 Ex3-11T2 1.176 MTB > Stat > Basic Statistics > 2 Variances > Summarized data Test for Equal Variances for Ex3-11T1.n −1 = F0.4 Data 1.85 0.3 1.125 = 0.115 0.Chapter 3 Exercise Solutions 3-11 continued (d) α = 0. and an investigation should be undertaken to understand why. Reject H 0 if F0 > Fα / 2.n −1.

n −1. n1 −1 = F0.n −1 ≤σ2 ≤ (n − 1) S 2 χ12−α / 2.125 2 2 = 16. S2 = 0.025.6 = 5.n 2 −1. x2 = 1.8 The normality assumption seems reasonable for these readings.6955) ≤ ≤ 0.693 70 60 50 40 30 20 10 5 1 1.1954) (5.7 2 (n − 1) S 2 χα2 / 2.1252 2 ≤σ ≤ 16.1954. n1 −1 2 1 S2 S σ σ 12 0.n −1 ≤ ≤ 2 1−α / 2.1252 σ 22 0.235 0.n −1 = χ 0.0128 1.0 1. n1 −1 2 2 1 1 2 2 2 2 = F0. χ12−α / 2. 3-13 .1249 8 0.n 2 −1.6 1.0128.943 0.1252 0.007 ≤ σ ≤ 0.2 1.065 (g) MTB > Graph > Probability Plot > Multiple Probability Plot of Surface Finish by Technician (Ex3-11T1.142 0.n −1 (8 − 1)0.383 1.95% CI 99 Variable Ex3-11T1 Ex3-11T2 95 90 Mean 1. χα2 / 2.05 F1−α / 2.6899 2 0.05.7.165 ≤ σ 12 ≤ 4.975. Ex3-11T2) Normal . n2 −1.1152 0.6 = 0.376.6955 S12 σ S F Fα / 2.6899 α = 0.376 Percent 80 StDev N AD P 0.7 = 1.975.n2 −1 = χ 0.1252 (8 − 1)0. Fα / 2.821 σ 22 (f) n2 = 8.7.Chapter 3 Exercise Solutions 3-11 continued (e) α = 0.025.1152 (0.4 Data 1.1148 7 0.

18 = 2. S2 = 5.22) 1 10 + 1 10 ≤ ( µ1 − µ 2 ) ≤ (147. From Eqn. Saltwater quench: n1 = 10.mu (Ex3-13OQ) Estimate for difference: -1.97 2 + (10 − 1)5.ν S12 n1 + S 22 n2 3-13.n + n 1 2 −2 S p 1 n1 + 1 n2 (147. 3.97 Oil quench: n2 = 10. x2 = 149. for σ 12 ≠ σ 22 and both unknown. tα/2.97 1.2217 P-Value = 0.40 5.1009(5. S1 = 4.6 Ex3-13OQ 10 149.4.025.6 − 149.6.451 DF = 18 Do not reject H0.7 Difference = mu (Ex3-13SQ) .4) − 2.05. x1 = 147. and conclude that there is no difference between the quenching processes.1009 Sp = (n1 − 1) S12 + (n2 − 1) S 22 (10 − 1)4.22 10 + 10 − 2 n1 + n2 − 2 ( x1 − x2 ) − tα / 2. (b) α = 0.1 3-14 .70615.462 = = 5.46 (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. 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.ν S12 n1 + S 22 n2 ≤ ( µ1 − µ2 ) ≤ ( x1 − x2 ) + tα / 2.1009(5.n + n 1 2 −2 S p 1 n1 + 1 n2 ≤ ( µ1 − µ 2 ) ≤ ( x1 − x2 ) + tα / 2.4) + 2.80000 95% CI for difference: (-6.6 − 149.7 ≤ ( µ1 − µ 2 ) ≤ 3.60 4.46 1.Chapter 3 Exercise Solutions 3-12. n1+n2−2 = t0.10615) T-Test of difference = 0 (vs not =): T-Value = -0.22) 1 10 + 1 10 −6. Ex3-13OQ Two-sample T for Ex3-13SQ vs Ex3-13OQ N Mean StDev SE Mean Ex3-13SQ 10 147.77 Both use Pooled StDev = 5. 3-54 and 3-55.

2484) (4.9.906 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.779 149.461 10 0.025.4 5.95% CI 99 Variable Ex3-13SQ Ex3-13OQ 95 AD P Mean StDev N 147.169 0.9 = 4. 1 1 σ 22 S S22 σ 12 4. Ex3-13OQ) Normal . 3-15 .6 4. n1 −1 2 1 2 α / 2. 2 1.218 0.462 0.n =F = 0. the assumption of equal variances seems reasonable.971 10 0.2484.97 2 ≤ ≤ (0.0260) σ 22 5. Fα / 2. n1 −1 = F0. n1 −1 2 2 −1.462 5.05 F1−α / 2.975.n 0.21 ≤ σ 12 ≤ 3.9 2 −1.34 σ 22 Since the confidence interval includes the ratio of 1.0260 S12 σ 12 S ≤ ≤ F F − α n − n − 1 / 2.9.97 2 4. (d) MTB > Graph > Probability Plot > Multiple Probability Plot of Quench Hardness (Ex3-13SQ. n2 −1.Chapter 3 Exercise Solutions 3-13 continued (c) α = 0.

09(1 − 0.057 ≤ p ≤ 0.645 0.025 = 1.09) 200 ≤ p ≤ 0.10) Zα/2 = Z0. use the normal approximation to the binomial for x < np0.3536|] = 2[1 − 0. np0 = 200(0.09(1 − 0.5) − 20 = = −0. Z0 = ( x + 0.09 + 1.09 − 1. (b) α = 0.123 3-16 .05.10.05/2 = Z0.090000 (0.129662) Z-Value -0.Chapter 3 Exercise Solutions 3-14.7236 MTB > Stat > Basic Statistics > 1 Proportion > Summarized data Test and CI for One Proportion Test of p = 0.10 versus H1: p ≠ 0.6382] = 0.050338.1 Sample X N Sample p 95% CI 1 18 200 0. pˆ = x/n = 18/200 = 0. and conclude that the sample process fraction nonconforming does not differ from 0.09) 200 0.47 P-Value 0. Test H0: p = 0. 0. n = 200.645 0.10.10) = 20 Since (x = 18) < (np0 = 20).645 pˆ − Zα / 2 pˆ (1 − pˆ ) n ≤ p ≤ pˆ + Zα / 2 pˆ (1 − pˆ ) n 0. P-value = 2[1 − Φ|Z0|] = 2[1 − Φ|−0. x = 18. Zα/2 = Z0.96 Do not reject H0. α = 0. Reject H0 if |Z0| > Zα/2.3536 np0 (1 − p0 ) 20(1 − 0.10.5) − np0 (18 + 0.1 vs p not = 0.10/2 = Z0.09 (a) p0 = 0. not an approximation.10.637 Note that MINITAB uses an exact method.05 = 1.

x = 65. MTB > Stat > Basic Statistics > 1 Proportion > Summarized data Test and CI for One Proportion Test of p = 0.08.5) − np0 (65 − 0. Test H0: p = 0.130000 (0. use the normal approximation to the binomial for x > np0. (b) P-value = 2[1 − Φ|Z0|] = 2[1 − Φ|4.08 Sample X N Sample p 95% CI 1 65 500 0.13(1 − 0. and conclude the sample process fraction nonconforming differs from 0.159478) Z-Value 4.000 Note that MINITAB uses an exact method.96 Reject H0. pˆ = x/n = 65/500 = 0.08) Zα/2 = Z0.05 = 1.08) = 40 Since (x = 65) > (np0 = 40). n = 500.05. ( x − 0.13 + 1.645 p ≤ pˆ + Zα pˆ (1 − pˆ ) n p ≤ 0.0387|] = 2[1 − 0.08 vs p not = 0.12 P-Value 0.130 (a) p0 = 0.100522.5) − 40 Z0 = = = 4. not an approximation.05/2 = Z0.05.08.0387 np0 (1 − p0 ) 40(1 − 0. Reject H0 if |Z0| > Zα/2.025 = 1. α = 0.13) 500 p ≤ 0.08.99997] = 0. Zα = Z0. np0 = 500(0.645 0.155 3-17 . 0.08 versus H1: p ≠ 0.00006 (c) α = 0.Chapter 3 Exercise Solutions 3-15.

0580079.0246745) Test for difference = 0 (vs not = 0): Z = -0.067) 300 pˆ 1 (1 − pˆ 1 ) + pˆ 2 (1 − pˆ 2 ) n1 n2 ≤ ( p1 − p2 ) ≤ (0.645 0.77 P-Value = 0.05(1 − 0. Test H0: p1 = p2 versus H1: p1 ≠ p2.050 − 0.052 ≤ ( p1 − p2 ) ≤ 0.Chapter 3 Exercise Solutions 3-16.066667 Difference = p (1) .96 Do not reject H0.7842 −Zα/2 = −1.067) − 1.06) (1 200 + 1 300 ) = −0.06(1 − 0.067 (b) Use α = 0.067) + 1.05. x1 = 10. pˆ1 = x1/n1 = 10/200 = 0.05/2 = Z0. pˆ 2 = x2/n2 = 20/300 = 0.05) 200 + 0.645 0.05) 200 + 0.018 3-18 . Reject H0 if Z0 > Zα/2 or Z0 < –Zα/2 pˆ = x1 + x2 10 + 20 = = 0.067) 300 −0.05 n2 = 300.0166667 95% CI for difference: (-0.05 − 0.06 n1 + n2 200 + 300 Z0 = pˆ1 − pˆ 2 pˆ (1 − pˆ ) (1 n1 + 1 n2 ) 0. MTB > Stat > Basic Statistics > 2 Proportions > Summarized data Test and CI for Two Proportions Sample X N Sample p 1 10 200 0. (a) n1 = 200.05(1 − 0.96 Zα/2 = Z0.067(1 − 0.05 − 0. x2 = 20. Conclude there is no strong evidence to indicate a difference between the fraction nonconforming for the two processes.050000 2 20 300 0.p (2) Estimate for difference: -0.025 = 1.067 = 0.442 (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(1 − 0. 0.

70449 2.025.2383.10.85.n1+n2−2 = t0.746 SP = t0 = SP ( n1 − 1) S12 + ( n2 − 1) S22 n1 + n2 − 2 = (10 − 1) 6.69405 F-Test (normal distribution) Test statistic = 1.60576 5. 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.n2 − 2 or F0 < F1−α / 2.975. F1−α / 2.n1 −1.77000 95% lower bound for difference: -0.554 10 + 8 − 2 x1 − x2 9.83 2 8 8.9.08.9.48596 5.85 2.922 The impurity variances before and after installation are the same.08 2.88 Difference = mu (1) .85 − 8.* before: n1 = 10. x2 = 8.55525 2.61 0.8232 and > 0.5582 P-Value = 0.n1 −1.0987 < 4.Chapter 3 Exercise Solutions 3-17.34856 T-Test of difference = 0 (vs >): T-Value = 1.18 = 1.79 + (8 − 1) 6.24710 2 8 1. α = 0. 3-19 . 10+8−2 = 1.7 = 0.n1 −1.7 = 4. at α = 0.2383 F0 = S12 S 22 = 6. x1 = 9.49 0.n2 − 2 = F0.554 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 9.8232. Reject H0 if t0 > tα.79 6.n2 −1 Fα / 2. tα.05 Reject H 0 if F0 > Fα / 2.n1+n2−2.461 1 n1 + 1 n2 2. p-value = 0.05.18 = 2. S12 = 6.n2 −1 = F0.05.082 DF = 16 The mean impurity after installation of the new purification unit is not less than before.79 after: n2 = 8.mu (2) Estimate for difference: 1.08 = = 1.46 Both use Pooled StDev = 2.18 (a) Test H 0 : σ 12 = σ 22 versus H1 : σ 12 ≠ σ 22 .n1 −1. S 22 = 6. (b) Test H0: µ1 = µ2 versus H1: µ1 > µ2.0987 F0 = 1.

8 psi.645 x x − ( 1 2 ) − ∆ 0 = (175. 3-20 .0000 P-Value = 0. So test a difference ∆0 = −5.3) − (−5) = −0.47 Both use Pooled StDev = 3. The mean strength of Design 2 does not exceed Design 1 by 5 psi. Reject H0 if Z0 < − Zα .Chapter 3 Exercise Solutions 3-18.00 0.69978 T-Test of difference = -5 (vs <): T-Value = -0. σ1 = σ2 = 3.8 − 181.0 psi Want to demonstrate that µ2 is greater than µ1 by at least 5 psi.320 DF = 30 Note: For equal variances and sample sizes.4714) = 0.75 Difference = mu (1) .mu (2) Estimate for difference: -5. so H1: µ1 + 5 < µ2. so do not reject H0. ∆0 = −5 −Zα = −Z0.80 3. test H0: µ1 − µ2 = − 5 versus H1: µ1 − µ2 < − 5.30 3.645.05 = −1.4714) > −1.3 psi.75 2 16 181.50000 95% upper bound for difference: -3. the Z-value is the same as the t-value. P-value = Φ(Z0) = Φ(−0. The P-values are close due to the sample sizes. n2 = 16.3187 MTB > Stat > Basic Statistics > 2-Sample t > Summarized data Two-Sample T-Test and CI Sample N Mean StDev SE Mean 1 16 175.4714 Z0 = 32 16 + 32 16 σ 12 n1 + σ 22 n2 (Z0 = −0. x1 = 175. x2 = 181. n1 = 16.00 0.

001250.000468 Difference 12 -0.8188.0013112 ( ) n = −0.150 − 0. j − xVernier.000417 0.000241 Ex3-19VC 12 0.000379 95% CI for mean difference: (-0.10) < 2.151) + " + ( 0. 0. Test H0: µd = 0 versus H1: µd ≠ 0.10 P-Value = 0.151583 0. MTB > Stat > Basic Statistics > Paired t > Samples in Columns Paired T-Test and CI: Ex3-19MC.000835 0.151167 0.001311 12 = −1.8188 d=1 ( n∑ j =1 n xMicrometer.000417 0. j ) = 1 ⎡⎣( 0.295 3-21 .001621 0. so do not reject H0. Reject H0 if |t0| > tα/2.151 − 0.22 = 2.10 (|t0| = 1.000417 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 . tα/2. n1 + n2 − 2. There is no strong evidence to indicate that the two calipers differ in their mean measurements.Chapter 3 Exercise Solutions 3-19.Ex3-19VC N Mean StDev SE Mean Ex3-19MC 12 0. n1 + n2 − 2 = t0.000417) T-Test of mean difference = 0 (vs not = 0): T-Value = -1.152 ) ⎤⎦ = −0.005.001311 0.

500 SE Mean 2. Reject H0 if t0 > tα. In order to achieve this result. (b) χ2α/2.85 ≤ σ 2 ≤ (20 − 1)1. n −1 = t0.Chapter 3 Exercise Solutions 3-20.52 ⎤⎦ 1 = 42.19 3-22 .6 ml.571 95% Lower Bound 149. tα.n −1 (20 − 1)1.5 20 = 1.44 P 0.975. n-1 = χ20.91.4389 ( ) ( ) (t0 = 1.30 ≤ σ 2 ≤ 4. µ > 150.19 = 8.7291.700 StDev 11.1170. χ21-α/2.80 1. n −1. There is insufficient evidence to indicate that the mean strength is greater than 150 psi. χ21-α. n-1 = χ20. (b) n = 20. x = 752. α = 0. s = 1.5.083 3-21.95. n-1. t0 = ( x − µ ) S n = (153.7.7 − 150 ) 11. so do not reject H0.n −1 ≤ σ 2 ≤ (n − 1) S 2 χ12−α / 2.52 32.52 8. Reject H0 if χ20 < χ21-α.05 (a) Test H0: σ2 = 1 versus H1: σ2 < 1.5. (n − 1) S 2 χα2 / 2.025.91 1.19 = 32.19 = 10.75 > 10.05. s = 11.85. x = 153.05 Test H0: µ = 150 versus H1: µ > 150. (a) The alternative hypothesis H1: µ > 150 is preferable to H1: µ < 150 we desire a true mean weld strength greater than 150 psi. MTB > Stat > Basic Statistics > 1-Sample t > Summarized data One-Sample T Test of mu = 150 vs > 150 N 20 Mean 153.1170 χ 02 = ⎡⎣(n − 1) S 2 ⎤⎦ σ 02 = ⎡⎣(20 − 1)1.4389) < 1. The standard deviation of the fill volume is not less than 1ml.14 ≤ σ ≤ 2.254 T 1. H0 must be rejected in favor of the alternative H1.19 = 1. so do not reject H0. n = 20.7291.75 χ20 = 42. n-1 = χ20. α = 0.

4 752.0 752.25 Mean Median 752. ml 756 758 The plotted points do not fall approximately along a straight line.55 1.281321 0.2 (c) MTB > Graph > Probability Plot > SIngle Probability Plot of Pinot Gris Fill Volume (Ex3-21) Normal .Chapter 3 Exercise Solutions 3-21 (b) continued MTB > Stat > Basic Statistics > Graphical Summary Summary for Pinot Gris Fill Volume.37 0.00 95% C onfidence Interv al for M ean 751.27 95% C onfidence Interv al for M edian 752.00 753.172 M ean S tD ev V ariance S kew ness Kurtosis N 752.511 0.172 Percent 80 70 60 50 40 30 20 10 5 1 748 750 752 754 Fill Volume. 3-23 .6 1. so the assumption that battery life is normally distributed is not appropriate.538 20 0. ml (Ex3-21) A nderson-D arling N ormality Test 749 750 751 752 753 754 755 A -S quared P -V alue 0.95% CI 99 Mean StDev N AD P-Value 95 90 752.51 0.00 753.191843 20 M inimum 1st Q uartile M edian 3rd Q uartile M aximum 756 750.00 756.8 753.25 753.00 751.17 2.54 2.00 95% C onfidence Interv al for S tD ev 95% Confidence Intervals 1.83 753.

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. β. n = 4. What n is needed such that the Type II error.05. Let µ1 = µ0 + δ.96 − 5 4 3 ( ) ( ) ( ) ( ) = Φ (−1. α = 0.2933) = 0.6667 From Figure 3-7.9 0. then Φ − Zα / 2 − δ n σ is likely to be small compared with β. 3-46.96 − 5 4 3 − Φ −1. β = Φ Zα / 2 − δ n σ − Φ − Zα / 2 − δ n σ ( ) ( ) ( ) If δ > 0. the operating characteristic curve for two-sided at α = 0. Test H0: µ = 15 versus H1: µ ≠ 15.10? δ = µ1 − µ2 = 20 − 15 = 5 d = δ σ = 5 9 = 1. ( 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 . From Eqn. µ1 = 20.0. µ0 = 15.0848 − 0.3733) − Φ (−5.05. Check: β = Φ Zα / 2 − δ n σ − Φ − Zα / 2 − δ n σ = Φ 1. σ2 = 9.05 Assumed standard deviation = 3 Sample Target Difference Size Power Actual Power 5 4 0.Chapter 3 Exercise Solutions 3-22. is less than or equal to 0. So.915181 3-23.0000 = 0.

n2 . 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 H 0 if Z 0 > Zα / 2 3-25 . x1 − x2 Maximize: Z 0 = Subject to: n1 + n2 = N . x1 . x2 . x1 independent of x2 . σ 1 n1 + σ 2 n2 2 2 Since ( x1 − x2 ) is fixed. Assume µ1 = 2µ2 and let Q = ( x1 − x2 ) . 3-25.Chapter 3 Exercise Solutions 3-24. n1 . Given x ~ N .

688 − 1) 1 1000 = −9.005 = 2. at α = 0.645 Z 0 = ( x − λ0 ) λ0 n = ( 0.110 Test H0: λ = 0.645. x ~ Poi(λ). 3-26 . Each xi ~ POI(λ). so do not reject H0.0328 (|Z0| = 1. x = 688. (a) Wish to test H0: λ = λ0 versus H1: λ ≠ λ0. x = x/N = 3/5 = 0.6 − 0.110 − 0. 3-28. i =1 Using the normal approximation to the Poisson.15. x = x/N = 11/100 = 0.3162 (Z0 = 0. x ~ Poi(λ). n = 100.05. n ∑x i ~ POI( nλ ) .6 Test H0: λ = 0. ….05.Chapter 3 Exercise Solutions 3-26.5 5 = 0. Reject H0 if |Z0| > Zα/2. 3-27. x2. x = x/n = ~ N(λ. if n is large.5 versus H1: λ > 0. Z 0 = ( x − λ ) λ0 / n . Zα = Z0. xn. x = 11. n = 1000.15 100 = −1.688 Test H0: λ = 1 versus H1: λ ≠ 1. Reject H0: λ = λ0 if |Z0| > Zα/2 (b) x ~ Poi(λ). Zα/2 = Z0.5 ) 0.025 = 1. at α = 0.01. so reject H0. x = x/N = 688/1000 = 0. Zα/2 = Z0.5.3162) < 1.8663) > 1.5758.8663 (|Z0| = 9.0328) < 2.05 = 1. λ/n). at α = 0.15 ) 0. x = 3. so do not reject H0. Reject H0 if |Z0| > Zα. Select random sample of n observations x1. Reject H0 if Z0 > Zα.15 versus H1: λ ≠ 0.5758 Z 0 = ( x − λ0 ) λ0 n = ( 0.96 Z 0 = ( x − λ0 ) λ0 n = ( 0.96. n = 5.

7132 DF SS MS F P 2 3.630 0.00 3.5 4.4167 0. However.648 1.3167 0. Boxplots of data MTB > Graph > Boxplot > One Y. (b) MTB > Stat > ANOVA > One-Way > Graphs.60 4.5 125 160 C2F6 Flow (SCCM) 200 Gas flow rate of 125 SCCM gives smallest mean percentage uniformity.7600 (---------*----------) 160 6 4.20 4. (a) MTB > Stat > ANOVA > One-Way One-way ANOVA: Ex3-29Obs versus Ex3-29Flow Source Ex3-29Flow Error Total S = 0. so there is some evidence that gas flow rate affects the etch uniformity.32% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev -----+---------+---------+---------+---125 6 3.5 3.278 R-Sq = 32.0 4.2.0 3.824 3. 3-27 .6823) > (F0 = 3.5231 (----------*---------) 200 6 3.59 0. the P-value is just slightly greater than 0.053 15 7.05.80 Pooled StDev = 0.0 2.15 = 3.509 17 11.05.9333 0.59). so flow rate does not affect etch uniformity at a significance level α = 0.34% R-Sq(adj) = 23.05.8214 (----------*---------) -----+---------+---------+---------+---3.7132 (F0.Chapter 3 Exercise Solutions 3-29. With Groups Boxplot of Etch Uniformity by C2F6 Flow Etch Uniformity (%) 5.

5 -1.4 3.0 4.2 3.0 0.8 Fitted Value 4.6 3.2 4.5 0.0 3. 3-28 . 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 -0.4 Residuals are satisfactory. Residuals versus fits Residuals Versus the Fitted Values (response is Etch Uniformity (Ex3-29Obs)) 1.5 Residual 1. (d) MTB > Stat > ANOVA > One-Way > Graphs.Chapter 3 Exercise Solutions 3-29 continued (c) MTB > Stat > ANOVA > One-Way > Graphs.

5087 6 = 0.Chapter 3 Exercise Solutions 3-30.6 3 . Flow Rate 125 160 200 scale factor = Mean Etch Uniformity 3.9% MSE n = 0.0 3 . 125 and 200.8 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. say. The statistically significant difference between the mean uniformities can be seen by centering the t distribution between.3% 4.2 4 . 3-29 .5 4 .9 (1 6 0 ) 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 . and noting that 160 would fall beyond the tail of the curve.4% 3.

16% R-Sq(adj) = 19.7 107.05.7 (-----------*-----------) 20 3 1606.214 8 40933 5117 11 69567 R-Sq = 41.7 77. 3-30 .0 10. (a) MTB > Stat > ANOVA > One-Way > Graphs> Boxplots of data.0 (-----------*----------) 15 3 1586.0 52.5 No difference due to rodding level at α = 0. (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. Normal plot of residuals One-way ANOVA: Ex3-31Str versus Ex3-31Rod Source Ex3-31Rod Error Total S = 71.9 (-----------*-----------) 25 3 1500.53 DF SS MS F P 3 28633 9544 1.09% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev ----+---------+---------+---------+----10 3 1500.87 0.Chapter 3 Exercise Solutions 3-31.0 (-----------*----------) ----+---------+---------+---------+----1440 1520 1600 1680 Pooled StDev = 71.

3-31 .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.

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 .Chapter 3 Exercise Solutions 3-32.

497 (----------*----------) --------+---------+---------+---------+41.141 (----------*----------) 525 6 41.75 42. (a) MTB > Stat > ANOVA > One-Way > Graphs> Boxplots of data.152 1.700 0.553 R-Sq = 17.57% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev --------+---------+---------+---------+500 6 41.25 41.450 0.339 (----------*----------) 575 6 41.6 -0.194 (----------*----------) 550 6 41.Chapter 3 Exercise Solutions 3-33.258 2. 3-33 . (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 0.6 0.8 -0.2 0.00 Pooled StDev = 0.324 Temperature level does not significantly affect mean baked anode density. Normal plot of residuals One-way ANOVA: Ex3-33Den versus Ex3-33T Source DF Ex3-33T 3 Error 20 Total 23 S = 0.105 2.583 0.45 0.50 41.457 0.2 0.0 Residual 0.4 -0.8 Normality assumption is reasonable.89% R-Sq(adj) = 5.333 0.3238 SS MS F P 0.4 0.

50 -0. Recommend 500°C for smallest variability.50 500 525 550 Firing Temperature.25 0. select the temperature with the smallest variance.25 -0. 3-34 .Chapter 3 Exercise Solutions 3-33 continued (c) Boxplot of Baked Density by Firing Temperature 42. 500°C (see Boxplot). which probably also incurs the smallest cost (lowest temperature).25 41. MTB > Stat > ANOVA > One-Way > Graphs> Residuals versus the Variables Residuals Versus Firing Temperature (Ex3-33T) (response is Baked Density (Ex3-33Den)) 0. 3-34. More uniform anodes are produced at lower temperatures. deg C 575 Since statistically there is no evidence to indicate that the means are different.75 500 510 520 530 540 550 Temperature (deg C) 560 570 580 As firing temperature increases.50 Residual 0.00 -0.00 40.50 41. so does variability.75 40.75 41.00 Baked Desnity 41.

99 and 1.750 2.750 3.304 (----*---) 1.71 1.000 18 132.85 0.0 70.711 Orifice size does affect mean % radon release.826 (---*---) 1. % (Ex3-35Rad) 85 80 75 70 65 60 0.711 DF SS MS F P 5 1133.40 1.309 (---*---) 0.38 226.0 77.51 0.99 Smallest % radon released at 1.65% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev ----+---------+---------+---------+----0. Boxplot of Radon Released by Orifice Diameter Radon Released.750 2.0 Pooled StDev = 2.05.68 30.559 (---*---) 1.35 23 1265. (a) MTB > Stat > ANOVA > One-Way > Graphs> Boxplots of data One-way ANOVA: Ex3-35Rad versus Ex3-35Dia Source Ex3-35Dia Error Total S = 2.40 4 65.37 4 82.000 3.0 84.37 0.71 4 75.63 R-Sq = 89.Chapter 3 Exercise Solutions 3-35.000 1.062 (---*---) 0.000 2. at α = 0.55% R-Sq(adj) = 86.4 orifice diameters.02 4 71. 3-35 .02 Orifice Diameter 1.25 7.51 4 77.754 (---*---) ----+---------+---------+---------+----63.99 4 62.

5 -5. with larger variances at the larger diameters (1.37 0.0 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.40 Orifice Diameter (Ex3-35Dia) 1.5 0.99 The assumption of equal variance at each factor level appears to be violated.0 Residual 2.5 -5.0 -2.51 0.99). Residuals Versus the Fitted Values (response is Radon Released (Ex3-35Rad)) 5.0 Residual 2. Residuals Versus Orifice Diameter (Ex3-35Dia) (response is Radon Released (Ex3-35Rad)) 5. 1.5 0.5 0. 1.02 1. 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 -5.0 -2.Chapter 3 Exercise Solutions 3-35 continued (b) MTB > Stat > ANOVA > One-Way > Graphs> Normal plot of residuals.0 Residuals violate the normality distribution.0 0.02.5 5.40.71 1.0 -2. Residuals versus fits.0 Residual 2. 3-36 .

7733 0. Boxplots of data One-way ANOVA: Ex3-36Un versus Ex3-36Pos Source DF SS MS F P Ex3-36Pos 3 16.4066 − 0. 1 is different from 2.9267 0.3853 (------*------) 3 3 1.6522 = 1.3962 + 0.0 4.0484 3-37 .6522 = = 0.008 Error 8 5.3067 1.4636 (------*------) 2 3 1.5 6.217 0.66% R-Sq(adj) = 66. and 4.3167 0.4366 (------*------) 4 3 1.8076 There is a statistically significant difference in wafer position.6522 2 σˆ uniformity = σˆτ2 + σˆ 2 = 0.29 0. 3.220 5.0 Pooled StDev = 0.407 8.5 3. 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 5.652 Total 11 21.53% Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev --------+---------+---------+---------+1 3 4.3570 (------*------) --------+---------+---------+---------+1.8076 R-Sq = 75.3962 n 12 (c) σˆ 2 = MSE = 0.437 S = 0. (a) MTB > Stat > ANOVA > One-Way > Graphs.Chapter 3 Exercise Solutions 3-36.

5 2.0 -0.0 3.0 Residual 0.5 1.0 Residual 0.5 -1.0 -1.Chapter 3 Exercise Solutions 3-36 continued (d) MTB > Stat > ANOVA > One-Way > Graphs> Normal plot of residuals.5 The variability in residuals does appear to depend on the magnitude of predicted (or fitted) values.5 -1.0 -0.0 -1. 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.5 1 2 3 Wafer Position (Ex3-36Pos) 4 Both outlier residuals are from wafer position 1. but there are two very unusual observations – the outliers at either end of the plot – therefore model adequacy is questionable.0 4.5 1.0 1.5 0. Residuals Versus the Fitted Values (response is Uniformity ( Ex3-36Un)) 1. Residuals versus fits.5 3.0 2.5 0.5 Fitted Value--Film Thickness Uniformity 4. Residuals Versus Wafer Position (Ex3-36Pos) (response is Film Thickness Uniformity (Ex3-36Un)) 1.5 1. 3-38 .

**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

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 . Many possible solutions.

beginning and end of process are shown below.1 55.8 81.8 95. Yellow is nonvalue-added activity.9 g in et C m om itm t en 19 13. The next step would be to determine the causes of “Being Rude” and to work on eliminating those causes. Snooze No Check time Awake Yes 6:30am ? Get out of bed Arrive at work … 4-31☺.3 100. “Being Rude” represents the greatest opportunity to make an improvement. Defect Overeating Being Rude Not meeting commitments Missing class Etc. green is value-added activity.0 To reduce total count of defects.Chapter 4 Exercise Solutions 4-29☺.7 r he Ot 0 6 4. 4-11 . 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. Many possible solutions. including defect categories and counts.1 g in iss s as Cl e tM No 37 26. Many possible solutions. Example of a check sheet to collect data on personal opportunities for improvement.

0027) (1 − 0.0027.0027 x P( X <= x ) 0 0. Cumulative Probability Cumulative Distribution Function Binomial with n = 5 and p = 0.0779 0 ⎝ ⎠ α1 = 1 − Pr{0 of 30 beyond} = 1 − ⎜ Cumulative Distribution Function Binomial with n = 30 and p = 0.947363 m = 30 ⎛ 30 ⎞ 0 30 ⎟ (0.0526 ⎝0⎠ α1 = 1 − Pr{0 of 20 beyond} = 1 − ⎜ Cumulative Distribution Function Binomial with n = 20 and p = 0.0134 ⎝0⎠ MTB > Calc > Probability Distributions > Binomial. as the number of samples increases (m).0027 x P( X <= x ) 0 0.922093 m = 50 ⎛ 50 ⎞ 0 50 ⎟ (0.0027) = 0. the probability that at least one of the points is beyond the limits also increases.0027) = 0.0027)5 = 1 − 0.0027) (1 − 0.0027) = 1 − 0.1025 ⎝0⎠ α1 = 1 − Pr{0 of 50 beyond} = 1 − ⎜ Cumulative Distribution Function Binomial with n = 50 and p = 0.0027) (1 − 0.973326 m = 20 ⎛ 20 ⎞ 0 20 ⎟ (0.0027 x P( X <= x ) 0 0.0027) = 0.9733 = 0.0027)0 (1 − 0.986573 m = 10 ⎛10 ⎞ 0 10 ⎟ (0.873556 Although the probability that a single point plots beyond the control limits is 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 − ⎜ ⎟ (0.9866 = 0.0267 ⎝0⎠ α1 = 1 − Pr{0 of 10 beyond} = 1 − ⎜ Cumulative Distribution Function Binomial with n = 10 and p = 0.0027 x P( X <= x ) 0 0.0027 x P( X <= x ) 0 0.Chapter 4 Exercise Solutions 4-32☺. 4-12 .

4-13 . In fact. When the process mean µ and variance σ2 are unknown. the points used to estimate these sample statistics are not independent—they do not reflect a random sample from a population. 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. they must be estimated by sample means x and standard deviations s.Chapter 4 Exercise Solutions 4-33☺. However.

0 CL = 4.29 UCL R = D4 R = 2.0 2 29. 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. 2.2 x= 1 2 = = 34. a run of n consecutive points on one side of the center line is defined as 9 points.00 LCL x = x − A2 R = 34.72 CL x = x = 34. Several exercises in this chapter differ from those in the 4th edition. This can be changed under Tools > Options > Control Charts and Quality Tools > Define Tests.00 R chart for Bearing ID (all samples in calculations) X-bar Chart for Bearing ID (all samples in calculations) 41. A second exercise number in parentheses indicates that the exercise number has changed. 4.71) = 9. D4 = 2.0 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. This can be changed in dialog boxes or under Tools>Options>Control Charts and Quality Tools>Estimating Standard Deviation.00 m 24 R + R2 + " + Rm 3 + 4 + " + 2 R= 1 = = 4. (a) for n = 5.00 − 0.00 + 0.0 UCL = 9. it is out of control.577.0 6 33. The MINITAB convention for determining whether a point is out of control is: (1) if a plot point is within the control limits.96 CL R = R = 4.114. In particular.72 8 CL = 34.71 LCL R = D3 R = 0(4. 5-1. MINITAB uses pooled standard deviation to estimate standard deviation for control chart limits and capability estimates. MINITAB defines some sensitizing rules for control charts differently than the standard rules.115(4.29 4 31.71 LCL = 31.71 m 24 UCL x = x + A2 R = 34.5 + 34. A2 = 0.577(4. or (2) if a plot point is on or beyond the limits.71) = 0.0 12 39.71) = 31.0 LCL = 0 0 27.577(4.00 R x-bar 35.71) = 36. 3.96 UCL = 36. D3 = 0 x + x + " + xm 34. New exercises are denoted with an “☺”. 5-1 .0 12 10 15 37. An “*” indicates that the description has changed. it is in control. not 8.Chapter 5 Exercise Solutions Notes: 1.2 + " + 34.

σˆ x = R / d 2 = 4.29) = 0 + 1 − 0.326 = 1.25.Chapter 5 Exercise Solutions 5-1 (a) continued The process is not in statistical control.5 / 2. 15.0 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.65 R x-bar 35.0 12 39.65 ⎞ ⎡ ⎛ 40 − 33.99950 = 0. 12 and Sample No.52.07) + 1 − Φ (3.5. R = 4. 15 excluded) 41.52 15 37.0 4 2 27.93 ⎠ ⎣ ⎝ 1. LCL x = 31.0 LCL = 31.0 CL = 4.65.CL x = 33. LCL R = 0. Assuming an assignable cause is found for these two out-of-control points. (b) pˆ = Pr{x < LSL} + Pr{x > USL} = Pr{x < 20} + Pr{x > 40} = Pr{x < 20} + [1 − Pr{x < 40}] ⎛ 20 − 33.0 6 33.93 UCL x = 36.25 8 CL = 33.0 12 10 UCL = 9.06 29.50 31. the two samples can be excluded from the control limit calculations.65 ⎞ ⎤ = Φ⎜ ⎟ + ⎢1 − Φ ⎜ ⎟⎥ ⎝ 1.0 UCL =36.93 ⎠ ⎦ = Φ (−7. x is beyond the upper control limit for both Sample No.CL R = 4.00050 5-2 .06 UCL R = 9.65. 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.5.00 x-bar Chart for Bearing ID (samples 12. The new process parameter estimates are: x = 33. 15 excluded) R chart for Bearing ID (samples 12.

36 4. R = 6. σˆ X = R / d 2 = 6.00 -14 0 14 28 5.44 4.00 PPM Total 0.035) MTB > Stat > Quality Tools > Capability Analysis > Normal Process Capability Analysis of Ex5-2V LSL LSL Target USL Within Ov erall Process Data -50. Actual specs are 350 ± 5 V.00 PPM > USL 0.00 PPM Total 0.00 PPM Total 0.03545 3.49 . x = 10. or cycles. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Ex5-2V Sample M ean 15.00 PPM > USL 0.49 Overall Capability Pp PPL PPU Ppk Cpm -42 Observed Performance PPM < LSL 0.36 5.33.26 12 8 _ R=6.49 CPL 6.00000 * USL Sample Mean Sample N StDev(Within) StDev(Overall) Potential (Within) Capability Cp 5.23 4.0 U C L=14.5 _ _ X=10.0 2 4 6 8 10 Sample 12 14 16 18 20 16 Sample Range U C L=14. so the process is capable.5 LC L=5.00 5-3 . (b) n = 4.0 7.33 10.34 6.88 12.25 / 2. trends.00 PPM > USL 0.035 . 6σˆ 6(3. Overall Performance PPM < LSL 0.00000 10.62 50. 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 * 42 Exp.00 -28 Exp. runs.25. With xi = (observed voltage on unit i – 350) × 10: USLT = +50.059 = 3. LSLT = –50 USL − LSL +50 − (−50) Cˆ P = = = 5.12282 CPU Cpk CCpk 4.Chapter 5 Exercise Solutions 5-2.77 5.32500 80 3.

9 0 -20 LC L=-25.064 Percent 80 70 60 50 40 30 20 10 5 1 0.1 0 5 10 Ex5-2V 15 20 A normal probability plot of the transformed output voltage shows the distribution is close to normal.3 100 _ R=63. trends. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Ex5-3Dia U C L=47.113 80 0.53 Sample M ean 40 20 _ _ X=10. 5-4 .9 Mean StDev N AD P-Value 99 95 90 10. 5-3.704 0.33 3.5 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.Chapter 5 Exercise Solutions 5-2 continued (c) MTB > Stat > Basic Statistics > Normality Test Probability Plot of Ex5-2V Normal 99. or cycles.73 2 4 6 8 10 Sample 12 14 16 18 20 150 Sample Range U C L=134. runs.

79 PPM Total 574.30009 Potential (Within) Capability Cp 1.09 1.00 PPM Total 0.5 / 2.00000 * 100. so the process is capable.67 PPM Total 219.00 -60 Exp. Cˆ P = 6σˆ x 6(27.09 -30 0 30 60 1.22 CPL CPU Cpk CCpk 25.29384 1.90000 100 27.00 PPM > USL 0.09 1.48 5-5 .Chapter 5 Exercise Solutions 5-3 continued (b) σˆ x = R / d 2 = 63.326 = 27.30 PPM > USL 549.17 * 90 Exp.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) -100.35 1. Overall Performance PPM < LSL 5.81 PPM > USL 213.46 1. Within Performance PPM < LSL 24. LSL = –100 USL − LSL +100 − (−100) = = 1.3 (c) USL = +100.22 Overall Capability Pp PPL PPU Ppk Cpm -90 Observed Performance PPM < LSL 0.32 1.17 1.00000 10.22 .

5-6 .002368 0.062011 1 2 4 6 8 10 12 14 Sample 16 18 20 22 24 1 Sample Range 0.Chapter 5 Exercise Solutions 5-4.063893 0.0024 U C L=0.0018 0.0000 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 1.00 standard deviations from center line. One point more than 3.0640 U C L=0.00092 0. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). Test Failed at points: 22 TEST 5.0012 _ R=0. at points: 15 If graph is updated with new data.0006 0.0625 0. Test Failed at points: 22 Test Results for R Chart of Ex5-4Th TEST 1. One Test Failed * WARNING * * point more than 3. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Thickness (Ex5-4Th) Sample M ean 0.0630 0.0635 _ _ X=0.0620 LC L=0.00 standard deviations from center line.062952 0. the results above may no longer be correct.

22) and recalculate control limits.0620 1 2 4 6 8 10 12 14 Sample 16 18 20 22 24 1 0.0630 0.062945 0. With the revised limits.064408] 5-7 .061492.000823 /1.002118 0. failing tests on both the x and the R charts. Removing all three samples from calculation.693 = 0.063787 0. sample 14 is also out-of-control on the x chart. the new control limits are: Xbar-R Chart of Thickness (Ex5-4Th) (Samples 15.0000 LCL=0 2 4 6 8 10 12 14 Sample 16 18 20 22 24 (b) σˆ x = R / d 2 = 0.06295 ± 3(0. Assuming assignable causes are found. remove the out-of-control points (samples 15.0012 _ R=0.000823 0.0024 Sample Range UCL=0.0018 0.000486 (c) Natural tolerance limits are: x ± 3σˆ x = 0.Chapter 5 Exercise Solutions 5-4 continued The process is out-of-control. 0.0006 0.0640 1 Sample Mean UCL=0.0625 LCL=0.000486) = [0.062104 0. 14 removed from control limits calculations) 0.0635 _ _ X=0. 22.

31 Exp.06450 Sample Mean 0. 15.99 CPU 1.0636 0.00049 StDev(Overall) 0.33 PPM > USL 1814.06295 Sample N 66 StDev(Within) 0.07 Cpk 0.90 * 0.0628 0. and excluding samples 14.Chapter 5 Exercise Solutions 5-4 continued (d) Assuming that printed circuit board thickness is normally distributed.0015 − (−0.03 CPL 0.0015) = = 1.06150 Target * USL 0.55 PPM Total 5233.0620 0.99 CCpk 1.03 Overall Capability Pp PPL PPU Ppk Cpm 0.97 0.94 0.0632 0. Within Performance PPM < LSL 1467.0640 0. Overall Performance PPM < LSL 3419.70 PPM Total 2157.028 Cˆ P = 6σˆ x 6(0.00 PPM Total 15151.0624 0.00053 Potential (Within) Capability Cp 1. 15.000486) MTB > Stat > Quality Tools > Capability Analysis > Normal Process Capability Analysis of Thickness (Ex5-4Th_w/o) (Estimated without Samples 14.52 PPM > USL 0. 22) LSL USL Within Ov erall Process Data LSL 0.0644 Observed Performance PPM < LSL 15151.61 PPM > USL 689.0616 0.88 5-8 .90 0. and 22 from the process capability estimation: USL − LSL +0.52 Exp.

066 1.Chapter 5 Exercise Solutions 5-5.5 -1.0 LC L=-1. or cycles.0 0.5 _ _ X=-0.5 LC L=0.302 1 2 3 4 5 6 7 8 Sample 9 10 11 12 13 14 15 The process is in statistical control. 5-9 .5 LC L=0.830 1.0 U C L=0.0 Sample StDev U C L=1.0 -0. with no out-of-control signals. Estimate” select Sbar as method to estimate standard deviation.5 _ _ X=-0.0 -0.003 0. runs.0 1 2 3 4 5 6 7 8 Sample 9 10 11 12 13 14 15 The process is in statistical control.714 0.5 _ S =1.0 1.2 3.983 0.5 _ R=3. or cycles. runs.0 LC L=-0.686 4.003 0.990 1 2 3 4 5 6 7 8 Sample 9 10 11 12 13 14 15 Sample Range 6.037 Sample M ean 1. There is no difference in interpretation from the x − s chart.0 U C L=5.5 -1. trends. trends. Xbar-S Chart of Fill Volume (Ex5-5Vol) U C L=1. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-S (Ex5-5Vol) Under “Options. (b) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R (Ex5-5Vol) Xbar-R Chart of Fill Volume (Ex5-5Vol) Sample M ean 1.0 0.043 1 2 3 4 5 6 7 8 Sample 9 10 11 12 13 14 15 2. with no out-of-control signals.

Otherwise there are no runs.n −1 = 1.136 1. CL = s 2 = 1.0662 (15 − 1) ( 4. and then copying & pasting to an empty worksheet column (results in Ex5-5Variance). sample 5 would be within the limits. s = 1.010 / 2.0 0.0027 (not tabulated in textbook).542 LCL = s 2 (n − 1) χ12−(α / 2).0662 (15 − 1) ( χ 0. “Variables” is column with sample data (Ex5-5Vol).n −1 = 1.010. If the limits had been calculated using α = 0.330 MINITAB’s control chart options do not include an s2 or variance chart. and “By Variables” is the sample ID column (Ex5-5Sample).542 2.33 0.066 (15 − 1) ( 31. selecting only the variance column.010 / 2). Graph > Time Series Plot > Simple Control limits can be added using: Time/Scale > Reference Lines > Y positions Control Chart for Ex5-5Variance UCL = 2.32 ) = 2.15−1 ) = 1. 5-10 . In “Statistics” select “Variance”.0662 (15 − 1) ( χ12−(0.0 1.Chapter 5 Exercise Solutions 5-5 continued (c) Let α = 0.0 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.07 ) = 0. n = 15. Copy results from the session window by holding down the keyboard “Alt” key. or cycles.136 2 2 UCL = s 2 (n − 1) χα2 / 2.0662 = 1. To construct an s2 control chart. first calculate the sample standard deviations and then create a time series plot.5 CL = 1. Results are displayed in the session window.5 LCL = 0. and there would be no difference in interpretation from either the x − s or the x−R chart.5 s^2 (Variance) 2.15 −1 ) = 1. To obtain sample standard deviations: Stat > Basic Statistics > Store Descriptive Statistics. trends.066.

268 16.5420 16.204 5-11 . x = 16.6 Sample M ean U C L=16.475. (b) n = 5.2 16. or cycles.9940 2 4 6 8 10 Sample 12 14 16 18 20 U C L=1.326 = 0.475 0.50 0. σˆ x = R / d 2 = 0.268.25 0. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Net Weight (Ex5-6Wt) 16.00 0.475 / 2.Chapter 5 Exercise Solutions 5-6. trends.75 _ R=0. R = 0. runs.0 LC L=15.4 _ _ X=16.00 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.004 Sample Range 1.

and is approximately symmetrical with a bell shape.00 16.the histogram has one mode.27 0.0 16. 5-12 .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 15.257 <0.6 MTB > Graph > Probability Plot > Single (Ex5-6Wt) Probability Plot of Net Weight (Ex5-6Wt) Normal .2 Ex5-6Wt 16.005 80 70 60 50 40 30 20 10 5 1 0.4 16.9 Mean StDev N AD P-Value 99 Percent 95 90 16.00 Visual examination indicates that fill weights approximate a normal distribution .25 Ex5-6Wt 16.8 16.50 15.50 16.75 16.1 15. Points on the normal probability plot generally fall along a straight line.95% CI 99.75 17.2014 100 1.

20196 Potential (Within) Capability Cp 0.23 PPM > USL 16215.78) = 0. "Overall" Performance PPM < LSL 2458.96 5-13 .204) specifications.0 Exp.71 0.00 16.268 ⎞ pˆ lower = Pr{x < LSL} = Pr{x < 15.41 PPM Total 19902.71 * 16.94 0.7 − 16.7} = Φ ⎜ ⎟ = Φ (−2.82 .00 PPM > USL 0.73 PPM Total 18673. MTB > Stat > Quality Tools > Capability Analysis > Normal Under “Estimate” select Rbar as method to estimate standard deviation.23 PPM > USL 16215.71 CCpk 0.96 (e) ⎛ 15.2 16.5 − (−0.0027 0.204 ⎝ ⎠ The MINITAB process capability analysis also reports Exp.71 Cpk 0.73 PPM Total 18673. Process Capability Analysis of Net Weight (Ex5-6Wt) LSL USL Within Ov erall Process Data LSL 15.6 Exp.83 0.00 PPM Total 0. Within Performance PPM < LSL 2706.8 Observed Performance PPM < LSL 0.93 CPU 0.82 Overall Capability Pp PPL PPU Ppk Cpm 15.82 CPL 0.70000 Sample Mean 16.4 0.20 PPM > USL 17196.20421 StDev(Overall) 0. Overall Performance PPM < LSL 2458.26800 Sample N 100 StDev(Within) 0. so the process is not capable of meeting Cˆ P 6σˆ x 6(0.70000 Target * USL 16.Chapter 5 Exercise Solutions 5-6 continued (d) USL − LSL +0.5) = = 0.61 16.

MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-S (Ex5-2Vl) Xbar-S Chart of Output Voltage (Ex5-2V) Sample M ean 15. or cycles.91 Sample M ean 40 20 _ _ X=10.5 0.Chapter 5 Exercise Solutions 5-7. trends.11 2 4 6 8 10 Sample 12 14 16 18 20 60 Sample StDev U C L=52. runs.125 Sample StDev 6. runs.92 5.71 45 30 _ S =25.5 LC L=5.23 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.5 _ S =2. trends. 5-14 .0 4.73 12.0 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.0 U C L=14.0 1. 5-8.0 7. 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=46.9 0 -20 LC L=-25. or cycles.703 3.5 _ _ X=10.33 10.0 2 4 6 8 10 Sample 12 14 16 18 20 U C L=6.

00 73. the x control limits based on S will be slightly different than limits based on R.048 0. 5-15 .024 0. (b) The control limits on the x charts in Example 5-3 were calculated using S to estimate σ. or cycles.04914 Sample Range 0.000 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.01 _ _ X=74.02324 0. They will not always be the same.036 _ R=0.98777 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.012 0.00118 74. and in general.99 LC L=73. runs. (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. trends.Chapter 5 Exercise Solutions 5-9☺.01458 Sample M ean 74.

00 Exp.00118 ⎞ ⎡ ⎛ 74.59 1. Overall Performance PPM < LSL 0.05 − 73.00118 ⎞ ⎤ = Φ⎜ ⎟ + ⎢1 − Φ ⎜ ⎟⎥ 0. Within Performance PPM < LSL 0.16 pˆ = Pr{x < LSL} + Pr{x > USL} = Pr{x < 73.950 73.51 PPM Total 0.886) = 0 +1−1 =0 5-16 .05}] ⎛ 73. MTB > Stat > Quality Tools > Capability Analysis > Normal Under “Estimate” select Rbar as method to estimate standard deviation.95} + [1 − Pr{x < 74.009991) specifications.Chapter 5 Exercise Solutions 5-9 continued (c) σˆ x = R / d 2 = 0.00 PPM > USL 0.95} + Pr{x > 74. so the process is not capable of meeting USL − LSL 74.00999 StDev(Overall) 0.02324 / 2.63 1.67 Overall Capability Pp PPL PPU Ppk Cpm 1.01022 Potential (Within) Capability Cp 1.00 PPM Total 0.05000 Sample Mean 74.009991 ⎝ ⎠ ⎣ ⎝ ⎠⎦ = Φ (−5.009991 .009991 0.89 PPM Total 1.95000 Target * USL 74.28 PPM > USL 0.15 PPM > USL 0.010 74.71 CPU 1.00118 Sample N 125 StDev(Within) 0.668 Cˆ P = 6σˆ x 6(0.67 CPL 1.123) + 1 − Φ (4.326 = 0.95 = = 1.66 Exp. Process Capability Analysis of Inner Diameter (Ex5-9ID) LSL USL Within Ov erall Process Data LSL 73.025 74.965 73.05} = Pr{x < 73.59 * 73.980 73.67 1.95 − 74.995 74.05 − 74.040 Observed Performance PPM < LSL 0.63 Cpk 1.63 CCpk 1.

increasing the process mean. Test Failed at points: 35. Since this point and the three subsequent points plot above the upper control limit.01458 74. 39 TEST 5. 5-17 . 38.01 0.00118 74.02 1 5 5 _ _ X=74.04914 0.048 Sample Range U C L=74. 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. 39. Test Failed at points: 37.99 LC L=73.02324 0. 40 The control charts indicate that the process is in control.00 73. Test Failed at points: 38.024 0. 39.98777 4 8 12 16 20 Sample 24 28 32 36 40 U C L=0. 37.00 standard deviations from center line.Chapter 5 Exercise Solutions 5-10☺. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).036 _ R=0.000 LC L=0 4 8 12 16 20 Sample 24 28 32 36 40 Test Results for Xbar Chart of Ex5-10ID TEST 1. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). One point more than 3. an assignable cause has likely occurred. until the x -value from the 37th sample is plotted. 38.012 0. 40 TEST 6.

949(10) = 89.579 ⎠ or 0.Chapter 5 Exercise Solutions 5-11 (5-9).533) = 0. σˆ x ≈ 1. B5 = 0.433) = 1 − 0.483(4) = 41.49 LCL x = µ − Aσ x = 80 − 0.276 centerline x = µ = 80 UCL x = µ + Aσ x = 80 + 0. or 0. (e) First. Cˆ P = 6σˆ x 6(1.669(10) = 16.727 S UCL S = B6σ x = 1.534) = [35. 5-18 .016 LCL R = D3 R = 0(4) = 0 (b) natural tolerance limits: x ± 3σˆ x = x ± 3 ( R / d 2 ) = 40 ± 3(4 / 2.0057 . and A = 0.949.253 .932 x= i =1 = LCL x = x − A2 R = 40 − 0. so the process is not capable.004(4) = 8. R = i =1 = =4 m m 50 50 UCL x = x + A2 R = 40 + 0.276(10) = 2. ∑ xi = 2000.000005 ⎝ 1.068 UCL R = D4 R = 2. to reduce scrap and rework costs. σ x = 10 in-lb.0005%. ⎝ 1. 44.579 ⎠ ⎛ 47 − 40 ⎞ pˆ rework = Pr{x > USL} = 1 − Pr{x < USL} = 1 − Φ ⎜ ⎟ = 1 − Φ (4.579) (d) ⎛ 36 − 40 ⎞ pˆ scrap = Pr{x < LSL} = Pr{x < 36} = Φ ⎜ ⎟ = Φ (−2.949(10) = 70.51 centerline = c4σ x = 0.264.999995 = 0. 50 50 i =1 i =1 n = 6 items/sample.9727(10) = 9.57%. Second.056 .0 − (−5.736] (c) USL . n = 10. say.LSL +5. not 40.669.76 5-12* (5-10).69 LCL S = B5σ x = 0. B6 = 1. µ = 80 in-lb.483(4) = 38. reduce variability such that the natural process tolerance limits are closer to. m = 50 samples (a) 50 ∑ xi 50 ∑ Ri 2000 200 = 40.0) = = 1. ∑ Ri = 200. center the process at 41.

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

58) + Φ (−3. m = 25 samples (a) 25 x= ∑ xi i =1 m = 662.50 ⎞ ⎛ 25.71 R= i =1 = LCL x = x − A2 R = 26.577(0.90 − 26.155 ⎝ ⎠ ⎝ ⎠ = 0.155 pˆ = Pr{x > USL} + Pr{x < LSL} = 1 − Pr{x ≤ USL} + Pr{x < LSL} ⎛ 26.36) = 0.50 + 0.29 UCL R = D4 R = 2.40 ⎞ ⎛ 25.36) = 26.155 ⎝ ⎠ ⎝ ⎠ = 1 − 0.36 / 2.23) 0.00005 0. ∑ Ri = 9.155 0.87) = 1 − 0.50 ⎞ = 1− Φ ⎜ ⎟+Φ⎜ ⎟ = 1 − Φ (2.23) + Φ (−3.40 ⎞ pˆ = 1 − Φ ⎜ ⎟+Φ⎜ ⎟ = 1 − Φ (3. ∑ xi = 662.00 = 0.50. 25 25 i =1 i =1 n = 5.76 LCL R = D3 R = 0(0.90 − 26.50 25 25 ∑ Ri 9.155 0.99506 + 0.114(0.Chapter 5 Exercise Solutions 5-20 (5-17).36) = 26.00499 (c) ⎛ 26.326 = 0.36 m 25 UCL x = x + A2 R = 26.90 − 26.00062 = 0.36) = 0 (b) σˆ x = R / d 2 = 0.00.50 = 26.577(0.90 − 26.50 − 0.00124 5-28 .99938 + 0.

998) = 23.8338 6σˆ x 6(3. x = 104.14 LCL x = x − A3 S = 20.5 / 0.0 − 1.5) = 3.84134 = 0.326 = 3. n = 5.86 − 22 ⎞ = Φ⎜ ⎟ −Φ⎜ ⎟ = Φ (0. not all of the output will meet specification. n = 5.0 + 1.57926 5-22 (5-19). even if the mean is located at the nominal dimension.6 5 ⎝ ⎠ ⎝ ⎠ = 0.6 5 1. X ~ N . 100.30 / 2.0.998 and 6σˆ x = 6(3.13 LCL S = B3 S = 0(1.60 (b) UCL x = x + A3 S = 20.20) − Φ (−5. m = 50 samples (a) σˆ x = S / c4 = 1. So.5) = 22.14 − 22 ⎞ ⎛ 17.427(1. R = 9.84134 ⎞⎤ ⎡ ⎛ ⎞ ⎛ ⎞⎤ ⎟ ⎥ = 1 − ⎢Φ ⎜ 209 − 188 ⎟ − Φ ⎜ 191 − 188 ⎟ ⎥ ⎟⎥ ⎢⎣ ⎝ 6 4 ⎠ ⎝ 6 4 ⎠ ⎥⎦ ⎠⎦ 5-23 (5-20).427(1. 2(10) = 20.089(1.Chapter 5 Exercise Solutions 5-21 (5-18).9400 = 1.5) = 0 (c) Pr{in control} = Pr{LCL ≤ x ≤ UCL} = Pr{x ≤ UCL} − Pr{x ≤ LCL} ⎛ 22. x = 20. S = 1.5.5) = 17.86 UCL S = B4 S = 2.57926 − 0 = 0. USL − LSL +10 − (−10) Cˆ P = = = 0. 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 + 0.99 is larger than the width of the tolerance band.998) 5-29 . USL=110.79) 1.30. LSL=90 σˆ x = R / d 2 = 9.

µ = 10.606(2.5) = 0 5-30 .686(2.5 .5) = 15.5) = 2.5) = 0 (c) centerline S = c4σ x = 0. σ = 2.121(2.128(2.7979(2.Chapter 5 Exercise Solutions 5-24* (5-21).5) = 1.99 UCL S = B6σ = 2. n = 2.82 UCL R = D2σ = 3.30 LCL x = µ − Aσ x = 10 − 2.22 LCL R = D1σ = 0(2.70 (b) centerline R = d 2σ x = 1.52 LCL S = B5σ = 0(2.5) = 9. These are standard values.5) = 4.121(2.5) = 6. x (a) centerline x = µ = 10 UCL x = µ + Aσ x = 10 + 2.

56) = 9.577(4. so the process is not capable of meeting Cˆ P = 6σˆ x 6(1.Chapter 5 Exercise Solutions 5-25 (5-22).63 − 24 ⎞ ⎛ 17. (d) Pr{not detect} = Pr{LCL ≤ x ≤ UCL} = Pr{x ≤ UCL} − Pr{x ≤ LCL} ⎛ UCL x − µnew ⎞ ⎛ LCL x − µnew ⎞ ⎛ 22.56.56) = 17.96 (c) USL − LSL +5 − (−5) = = 0. n = 5.56) + Φ (−7.56) = 22.05938 − 0 = 0.96 5 ⎠ ⎝ σˆ x n ⎠ ⎝ σˆ x n ⎠ = Φ (−1.326 = 1. R = 4.56) = 0 (b) σˆ x = R / d 2 = 4. m = 25 samples (a) UCL x = x + A2 R = 20 + 0.37 UCL R = D4 R = 2. x = 20.96 5 ⎠ ⎝ 1.114(4.577(4.63 LCL x = x − A2 R = 20 − 0.64 LCL R = D3 R = 0(4.56) = 0.96) specifications.05938 5-31 .85 .37 − 24 ⎞ = Φ ⎜⎜ ⎟⎟ − Φ ⎜⎜ ⎟⎟ = Φ ⎜ ⎟−Φ⎜ ⎟ ⎝ 1.56 / 2.

Chapter 5 Exercise Solutions 5-26☺. Test Failed at points: 18 The process is out of control on the x chart at subgroup 18.00 standard deviations from center line.98 30 20 _ R=16.49 430 1 2 4 6 8 10 Sample 12 14 16 18 20 Sample Range 40 UCL=38.82 450 _ _ X=448.68 450 440 LCL=437. so these limits can be used for future production.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. 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. Excluding subgroup 18 from control limits calculations: Xbar-R Chart of TiW Thickness (Ex5-26Th) Excluding subgroup 18 from calculations UCL=461.88 Sample Mean 460 _ _ X=449. One point more than 3.00 standard deviations from center line.69 440 LC L=436. 5-32 . One point more than 3. Test Failed at points: 18 No additional subgroups are beyond the control limits.56 430 1 2 4 6 8 10 Sample 12 14 16 18 20 Sample Range 40 U C L=37.18 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.

059 = 8.74 / 2.672 80 70 60 50 40 30 20 10 5 1 0. 5-33 .13 (c) MTB > Stat > Basic Statistics > Normality Test Probability Plot of TiW Thickness (Ex5-26Th) Normal 99.111 80 0.68 σˆ x = R / d 2 = 16.1 420 430 440 450 Ex5-26Th 460 470 480 A normal probability plot of the TiW thickness measurements shows the distribution is close to normal.269 0.9 Mean StDev N AD P-Value 99 Percent 95 90 448.Chapter 5 Exercise Solutions 5-26 continued (b) Excluding subgroup 18: x = 449.7 9.

30 440 450 460 470 1.05 1.14 1. is estimated from the within-subgroup variation.08645 9.05 * 480 Exp.01 PPM > USL 306.00000 Target * USL 480. Cp = 1.09 1.18 The Potential (Within) Capability.00 PPM > USL 0.13) 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. or in other words.92 248.00 430 Exp. Within Performance PPM < LSL 194. Cˆ P = 6σˆ x 6(8.24 8.Chapter 5 Exercise Solutions 5-26 continued (d) USL = +30.17 PPM Total 1154. so the process is capable.13944 Overall Capability Pp PPL PPU Ppk Cpm 420 Observed Performance PPM < LSL 0. LSL = –30 USL − LSL +30 − (−30) = = 1. This is the same result as the manual calculation.23 .18 CCpk 1. 5-34 .38 PPM > USL PPM Total 53.00 PPM Total 0.24.18 CPU 1.68750 Sample N 80 StDev(Within) StDev(Overall) Potential (Within) Capability Cp 1. Overall Performance PPM < LSL 848. σ x is estimated using R .29 Cpk 1.24 CPL 1.00000 Sample Mean 448.

Chapter 5 Exercise Solutions 5-27☺.68 450 440 LCL=437. 5-35 .00 standard deviations from center line. One point more than 3.74 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=461.88 Sample Mean 460 _ _ X=449.49 430 1 3 6 9 12 15 Sample 18 21 24 27 30 Sample Range 40 UCL=38.18 30 20 _ R=16. Test Failed at points: 18 The process continues to be in a state of statistical control.

nold = 4.128 ⎤ CL R = Rnew = ⎢ ⎥ Rold = ⎢ ⎥ (16.128 ⎤ LCL x = xold − A2(new ) ⎢ (16.059 ⎦ ⎢⎣ d 2(old) ⎥⎦ σˆ new = Rnew d 2(new ) = 9.128 ⎤ LCL R = D3(new) ⎢ ⎥ Rold = 0 ⎢ ⎥ (16.68.128 ⎤ UCL R = D4(new) ⎢ (16.96 ⎥ Rold = 3.74. 5-36 .Chapter 5 Exercise Solutions 5-28☺.880 ⎢ ⎣ 2.93 460 _ _ X=449.128 ⎤ UCL x = xold + A2(new ) ⎢ (16.68 + 1. and enter new parameter values.74) = 466.128 = 8. xold 449.13 MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Select Xbar-R options.74) = 0 ⎣ 2. R old = 16.96 Sample Range 30 20 10 _ R=9.059 ⎥⎦ ⎢⎣ d 2(old) ⎥⎦ ⎡ d 2(new ) ⎤ ⎡ 1.74) = 432.059 ⎦ ⎣⎢ d 2(old) ⎦⎥ ⎡ d 2(new ) ⎤ ⎡ 1.880 ⎢ ⎣ 2.17 ⎣ 2.74) = 29.059 ⎥⎦ ⎣⎢ d 2(old) ⎦⎥ ⎡ d 2(new ) ⎤ ⎡ 1. Parameters. Limits derived from N=4 subgroups Sample Mean 470 UCL=466. nnew = 2 ⎡ d 2(new ) ⎤ ⎡ 1.17 0 LCL=0 1 2 3 4 5 6 7 8 9 10 Sample The process remains in statistical control.17 1.44 ⎥ Rold = 449.68 450 440 LCL=432.74) = 9. Xbar-R Chart of TiW Thickness (Ex5-28Th) New subgroups with N=2.267 ⎢ ⎣ 2.059 ⎥⎦ ⎣⎢ d 2(old) ⎦⎥ ⎡ d 2(new ) ⎤ ⎡ 1.43 430 1 2 3 4 5 6 7 8 9 10 Sample UCL=29.68 − 1.92 ⎥ Rold = 449.

22 Sample Mean 460 _ _ X=449.70 5 0 LCL=0 3 6 9 12 15 Sample 18 21 24 27 30 The process remains in statistical control.68 450 440 LCL=437. with prior limits UCL=462. 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=462.00 standard deviations from center line. After finding assignable cause. so these limits can be used for future production.15 430 1 3 6 9 12 15 Sample 18 21 24 27 30 20 Sample StDev UCL=17.22 Sample Mean 460 _ _ X=449.15 430 1 2 4 6 8 10 Sample 12 14 16 18 20 20 Sample StDev UCL=17.Chapter 5 Exercise Solutions 5-29☺.68 450 440 LCL=437. Xbar-S Chart of Thickness (Ex5-27Th) 10 subgroups of new data. The process is out of control on the x chart at subgroup 18. Test Failed at points: 18 No additional subgroups are beyond the control limits.44 15 10 _ S=7.44 15 10 _ S=7. One point more than 3.70 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. 5-37 .

59 − 199 ⎞ ⎟−Φ⎜ ⎟ ⎝ 1.483(5) = 202.Chapter 5 Exercise Solutions 5-30 (5-23). 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 + 0.42 R= i =1 = LCL x = x − A2 R = 200 − 0. Even though the process is centered at nominal. ∑ xi = 6000.97 USL − LSL +5 − (−5) = = 0.02 LCL R = D3 R = 0(5) = 0 (b) σˆ x = R / d 2 = 5 / 2.25) − Φ (−1.97 6 ⎠ = Φ (4.04006 = 0. ∑ Ri = 150.483(5) = 197.97) The process is not capable of meeting specification.97 6 ⎠ ⎝ 1. (c) ⎛ 202.59 UCL R = D4 R = 2.42 − 199 ⎞ ⎛ 197. 30 30 i =1 i =1 n = 6.85 Cˆ p = 6σˆ x 6(1. the variation is large relative to the tolerance.534 = 1.004(5) = 10.95994 β − risk = Pr{not detect} = Φ ⎜ 5-38 .75) = 1 − 0.

µ1 = 92 k = ( µ1 − µ0 ) σ = ( 92 − 100 ) 6 = −1.577(3.95) = 101. n = 4.329 LCL x = x − A2 R = 104. L = 3.05 + 0.539 (c) UNTL = x + 3σˆ x = 104 + 3(1.114(3.539) = 99.95) = 8.37 5-32 (5-25).Chapter 5 Exercise Solutions 5-31 (5-24).37 ] = 0.95 UCL x = x + A2 R = 104.05.771 UCL R = D4 R = 2.539) = 108.38 5-39 .577(3.95) = 0 Sample #4 is out of control on the Range chart.566 LCL R = D3 R = 0(3. R = 3.579) = 106.95) = 106.935 UCL R = D4 R = 2.33) 4 ) ⎤ ⎣ ⎦ = 1 − ⎡Φ L − k n − Φ − L − k n ⎤ ⎣ ⎦ = 1 − [ Φ (5. σˆ x = R / d 2 = 3.34) ] = 1 − [1 − 0.62 LNTL = x − 3σˆ x = 104 − 3(1. So.579) = 101.66) − Φ (−0.350 LCL R = D3 R = 0(3.114(3.579 UCL x = x + A2 R = 104 + 0. excluding #4 and recalculating: x = 104.065 LCL x = x − A2 R = 104 − 0.577(3. µ0 = 100.33 Pr{detecting shift on 1st sample} = 1 − Pr{not detecting shift on 1st sample} = 1− β ( ) ( ) = 1 − ⎡Φ ( 3 − (−1. σ = 6.326 = 1. (a) x = 104.33) 4 ) − Φ ( −3 − (−1.579 / 2. R = 3.579) = 0 (b) Without sample #4.577(3.579) = 7.05 − 0.

997) + Φ (−5.8 30 m UCL x = x + A2 R = 20. 30 n = 5.49 UCL R = D 4 R = 2.Chapter 5 Exercise Solutions 5-32 continued (d) ⎛ 107 − 104 ⎞ ⎛ 99 − 104 ⎞ pˆ = 1 − Φ ⎜ ⎟+ Φ⎜ ⎟ = 1 − Φ (1.8) = 23.8.03 R= i =1 = LCL x = x − A2 R = 20.539 ⎠ ⎝ 1.539 ⎠ ⎝ 1.60) = 1 − 0.577(4. ⎛ 107 − 103 ⎞ ⎛ 99 − 103 ⎞ pˆ = 1 − Φ ⎜ ⎟+ Φ⎜ ⎟ = 1 − Φ (5.667 .0195 ⎝ 2.26 + 0.539 ⎠ Next work on reducing the variability.60) + Φ (−2.9953 + 0. ∑ xi = 607.667 ⎠ ⎝ 0.8) = 0 (b) σˆ x = R / d 2 = 4. i =1 30 ∑ Ri = 144.114(4.8 / 2.326 = 2.8) = 17.064 ⎠ 5-40 .064 ⎛ 16 − 20.9744 + 0. first center the process at nominal.0000 + 0.0006 = 0.25) = 1 − 0.147 LCL R = D 3 R = 0(4.0000 ⎝ 0.667 ⎠ 5-33 (5-26). if σˆ x = 0.26 30 m ∑ Ri 144 = 4.0000 = 0.8) = 10.539 ⎠ (e) To reduce the fraction nonconforming.0094 ⎝ 1.26 ⎞ pˆ = Pr{x < LSL} = Φ ⎜ ⎟ = Φ (−2.8 = 20.95) + Φ (−3.577(4.0047 = 0. ⎛ 107 − 103 ⎞ ⎛ 99 − 103 ⎞ pˆ = 1 − Φ ⎜ ⎟+ Φ⎜ ⎟ = 1 − Φ (2.997) = 1 − 1.064) = 0. then almost 100% of parts will be within specification.0262 ⎝ 1. m = 30 i =1 (a) m x= ∑ xi i =1 m = 607.26 − 0.

67 10 _ R=6. R Chart of Detent (Ex5-34Det) 20 1 15 Sample Range UCL=13. 5-41 . One point more than 3.00 standard deviations from center line.sample #12 exceeds the upper control limit on the Range chart. Estimate” select Rbar as method to estimate standard deviation. Test Failed at points: 12 Process is not in statistical control -.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). (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > R Under “Options.

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.38 6σˆ x 6(2.64 / 2.64 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. Test Failed at points: 12 (c) Without sample #12: σˆ x = R / d 2 = 5.3200 Cˆ P = = = 1. Without sample #12: USL − LSL 0.3220 − 0.326 = 2.42 × 0.42 (d) Assume the cigar lighter detent is normally distributed.00 standard deviations from center line.93 10 _ R=5.0001) 5-42 . R Chart of Detent (Ex5-34Det) Sample 12 Excluded from Calculations 20 1 Sample Range 15 UCL=11. Estimate” omit subgroup 12 and select Rbar.

25 5.00 standard deviations from center line.Chapter 5 Exercise Solutions 5-35 (5-28). 17. Estimate” use subgroups 1:11 and 13:15. Test Failed at points: 15. Failed at points: 24. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). Xbar-R Chart of Ex5-35Det Limits based on Samples 1-11. 16.5 6 6 2 2 0. and select Rbar. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 18. Failed at points: 24. 23.93 2 10 2 5 0 _ R=5. MTB > Stat > Control Charts > Variables Charts for Subgroups > R Under “Options. 22.45 5 2. 13. 20 TEST 6. Failed at points: 1. 17. Test Failed at points: 2. 20. 12.0 6 -2.05 1 12 14 Sample 16 18 20 22 24 1 15 UCL=11.2 LCL=-3.64 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. 2. 23 2.00 standard deviations from center line.5 1 -5.0 1 1 1 1 1 UCL=3. One point more than 3. One point more than 3. 24 Test Results for R Chart of Ex5-35Det TEST Test TEST Test 1. Failed at points: 12 2. 3. 9 points in a row on same side of center line. 19. 9 points in a row on same side of center line.0 2 4 6 8 10 20 Sample Range _ _ X=0. 25 5-43 . 20. 18. 13-15 1 1 Sample Mean 5. 13.

26 1 12 14 Sample 16 18 20 22 24 1 15 1 1 10 1 1 2 2 UCL=9.94 _ _ X=-0. No additional samples are out of control on the x chart.0 LCL=-3. Note that samples 1. 12 and 13 1 1 Sample Mean 5. Removing sample 3 gives Xbar-R Chart of Ex5-35Det Limits based on first 15 samples.0 2 4 6 8 10 20 Sample Range UCL=1. While the limits on the above charts may be used to monitor future production. 5-44 . 12.5 1 -5. (a) Given the large number of points after sample 15 beyond both the x and R control limits on the charts above.52 _ R=4. and 13 are out of control on the x chart. the fact that 6 of 15 samples were out of control and eliminated from calculations is an early indication of process instability.66 0.5 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.0 1 1 1 1 1 1 2. 2. the process appears to be unstable. excluding 1. sample 3 is also out of control on the x chart. If these samples are removed and the limits recalculated. 3.5 5 6 6 2 5 2 -2. 2.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. One point more than 3.08 1 12 14 Sample 16 18 20 22 24 1 UCL=15.55 Sample Mean 5.23 0.5 _ _ X=1.82 15 10 _ R=7.48 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. Test Failed at points: 12 5-45 .00 standard deviations from center line.Chapter 5 Exercise Solutions 5-35 continued (b) Xbar-R Chart of Detent (Ex5-35Det) All Samples in Calculations 1 1 1 UCL=5.0 2. 17 Test Results for R Chart of Ex5-35Det TEST 1.00 standard deviations from center line. 16. One point more than 3.0 -2. 12. 13.5 1 -5.0 2 4 6 8 10 Sample Range 20 LCL=-3.

0 2 4 6 8 10 Sample Range 20 1 12 14 Sample 16 18 20 22 24 1 15 UCL=14.48 10 _ R=6. 16.0 UCL=4.Chapter 5 Exercise Solutions 5-35 (b) continued Removing samples 1. 16. 12. 12.85 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. 13. 13.5 LCL=-2.94 2. 17. 17 excluded from calculations 1 1 1 1 Sample Mean 5. One point more than 3.00 standard deviations from center line. 13.5 _ _ X=0. 12. and 17 from calculations: Xbar-R Chart of Detent (Ex5-35Det) Samples 1.96 1 -5. Test Failed at points: 1. Test Failed at points: 12 5-46 . One point more than 3.99 0.00 standard deviations from center line. 16.0 -2. 20 Test Results for R Chart of Ex5-35Det TEST 1.

12. One point more than 3. 17. 5-47 . 17.00 standard deviations from center line.11 1 12 14 Sample 16 18 20 22 24 1 15 UCL=14.78 0.Chapter 5 Exercise Solutions 5-35 continued Sample 20 is now also out of control.0 -2. Xbar-R Chart of Ex5-35Det Samples 1. One point more than 3. and that the sources of variation need to be identified and removed.5 1 -5.0 2 4 6 8 10 Sample Range 20 LCL=-3. Test Failed at points: 12 Sample 18 is now out-of-control. with runs of points both above and below the centerline.66 2.74 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. 20 excluded from calculations 1 1 Sample Mean 5. 16. 18.5 _ _ X=0.0 1 1 1 UCL=4. 16. 13. 20 Test Results for R Chart of Ex5-35Det TEST 1. 12. This suggests that the process is inherently unstable. Test Failed at points: 1.24 10 _ R=6. Removing sample 20 from calculations. for a total 7 of the 25 samples. 13.00 standard deviations from center line.

326 = 0.500 ⎠ z = +2.758 µˆ = x = i =1 = 5-48 .400 / m ) d = (6.5121(0.i / mx i =1 σˆ y 10 y 2 i =1 y .i = 6.608 / 20) / 2.004(45.i y 2 2 (b) Want Pr{(x − y) < 0. (a) 20 10 i =1 i =1 n = 5. ∑ Ri = 1350.346 5-37 (5-30).32 = 0. m = 30 (a) m ∑ Ri 1350 = 45.870.300 σˆ x = Rx / d 2 = ∑ Rx .0 30 m σˆ x = R / d 2 = 45. Let z = x − y.0 / 2.09 − z ⎞ Φ −1 ⎜ ⎟ = Φ (0. 30 30 i =1 i =1 n = 6.006 ⎝ σˆ z ⎠ ⎛ 0.500 ⎠ ⎛ 0.09 − z ⎞ ⎜ ⎟ = −2.978 ( = R / d = (∑ R 20 ) d = (18.500) + 0.006. m y = 10.978 /10) / 2.5121 ⎝ 0.870 = 429.326 = 0. ∑ xi = 12. ∑ Rx .0) = 90. ∑ Ry .006) ⎝ 0.09 = 1.Chapter 5 Exercise Solutions 5-36 (5-29). mx = 20.18 R= i =1 = LCL R = D3 R = 0(45.500 ⎛ 0. Then σˆ z = σˆ x2 + σˆ y2 = 0.09 − z ⎞ Φ⎜ ⎟ = 0.0 30 m UCL R = D4 R = 2.42 + 0.534 = 17.608.09} = 0.i = 18.0) = 0 (b) m ∑ xi 12.

667) = 30.0516 = 0.63) = 1 − 0.9979 + 0.870 = 429.751 Cˆ p 6σˆ x 6(17.266(13. m = 30 (a) m ∑ Si 410 = 13. LSL = 440 .87) + Φ (−1. 30 30 i =1 i =1 n = 4.667 / 0.Chapter 5 Exercise Solutions 5-37 continued (c) USL = 440 + 40 = 480.758 ⎠ ⎝ 17.870.667 30 m UCL S = B4 S = 2.758) ⎛ 480 − 429 ⎞ ⎛ 400 − 429 ⎞ pˆ = 1 − Φ ⎜ ⎟+Φ⎜ ⎟ = 1 − Φ (2.667) = 0 (b) m ∑ xi 12. ∑ xi = 12.0537 ⎝ 17. ∑ Si = 410.0 30 m σˆ x = S / c4 = 13.834 µˆ = x = i =1 = 5-49 .9213 = 14.40 = 400 USL − LSL 480 − 400 = = 0.758 ⎠ (d) To minimize fraction nonconforming the mean should be located at the nominal dimension (440) for a constant variance.969 S= i =1 = LCL S = B3 S = 0(13. 5-38 (5-31).

LCL x = 96.2877 = 0.7049 5-41 (5-34). USL − LSL USL − LSL 202.9535 − 0. µ = 100. n = 5.807 8 5-40 (5-33).807 ( 8 4 ) = 88.005 / 2 = Z 0. 1 1 1 ARL1 = = = = 2.992 1 − β 1 − Pr{not detect} 1 − 0.000 0. (a) n = 4. centerline x = 100.50-197. 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 − 2.807 ( = µ − k (σ UCL x = µ + kσ x = µ + k σ x LCL x = µ − kσ x x ) ( 4 ) = 111.6658 5-42 (5-35).50 Cˆ P = = = = 0. UCL x = 104.7678 6σˆ x 6 (1. µ = 98.6658)3 = 0.Chapter 5 Exercise Solutions 5-39 (5-32). k = 3.772 n = 100 + 2. 5-50 .56) ⎟ −Φ⎜ ⎟ = Φ⎜ σx σx ⎝ ⎠ ⎝ ⎠ ⎝ 8 5 ⎠ ⎝ 8 5 ⎠ = 0.9213) 6 ( S c4 ) The process is not capable of meeting specifications.6658 1 − [Pr{not detected}]3 = 1 − (0.68) − Φ (−0.0025 = 2. 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.

3. LSL = 600 − 20 = 580 (a) USL − LSL USL − LSL 620 − 580 Cˆ P = = = = 1.0000. 0. 0. µ = 200.9974.0000 0. 0. 0.0000} Operating Characteristic Curve for n = 9.2000 1. 1.75.96 (10 4 ) = 190. 2. L = 3.5 0.25. USL = 600 + 20 = 620.3 1.6000 0.2266.088(10) = 20. 0.Chapter 5 Exercise Solutions 5-43 (5-36). 0. σ x = 10 (a) centerline S = c4σ = 0.9332.213 UCL S = B6σ x = 2.0.0 1.2 n = 200 + 1. n = 9.5.8 n ) = 200 − 1.0013.8 1.0 k 5-51 . (b) n = 9. 1.0 2.0 0.111 6σˆ x 6 (17.8000 0.05/ 2 = Z 0. 0. 0.970 ) 6 ( R d2 ) Process is capable of meeting specifications.0}. 0. β = Φ L − k n − Φ − L − k n ( ) ( ) for k = {0.5 2. n = 4.4000 0. 2.96 ( = µ − k (σ UCL x = µ + kσ x = µ + k σ x LCL x = µ − kσ x x ) ( 4 ) = 209.5. 1. L = 3 1.0000 beta 0.2000 0.9213(10) = 9.5.7734.5.88 LCL S = B5σ x = 0(10) = 0 (b) k = Zα / 2 = Z 0. β = {0.025 = 1.0. 0.82 / 2.0668.5 3.96 10 5-44 (5-37).

R = i =1 = =4 m m 30 30 UCL x = x + A2 R = 90 + 0.671 LCL S = 0.924(4) = 7.7 n = 600 + 2. n = 9.Chapter 5 Exercise Solutions 5-45 (5-38).59 / 2. 30 30 i =1 i =1 n = 7.576 ( = µ − k (σ UCL x = µ + kσ x = µ + k σ x LCL x = µ − kσ x x ) ( 9 ) = 610.059 = 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) = 0.419 UCL S = 1.479 (c) S = c4σˆ x = 0. ∑ xi = 2700. m = 30 (a) m ∑ xi m ∑ Ri 2700 120 = 90.9594(1.419) = 0.118(1.0000 = 0. µ = 600.3 n ) = 600 − 2.419(4) = 88.1587 5-52 .167 5-46 (5-39).005 = 2.696 LCL R = D 3 R = 0.01 k = Zα / 2 = Z 0. σˆ x = R / d 2 = 20.576 12 5-47 (5-40).704 = 1. ∑ Ri = 120.479) = 1.882(1.1587 + 1 − 1.419) = 2. σ x = 12.676 x= i =1 = LCL x = x − A2 R = 90 − 0.076(4) = 0. α =0.304 (b) σˆ x = R / d 2 = 4 / 2.419(4) = 91.324 UCL R = D 4 R = 1.576 (12 9 ) = 589.01/ 2 = Z 0.

667 6σˆ x 6(3) The process is not capable of producing all items within specification.1587 5-49 (5-42).576 3 ) ) ( n = 360 + 2. (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. (a) σˆ x = R / d 2 = 8.Chapter 5 Exercise Solutions 5-48 (5-41).30 ARL1 = 1 − β 1 − Pr{not detect} Pr{detect} 0.01/ 2 = Z 0. k = Zα / 2 = Z 0.91/ 2.970 = 3.5000 = 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) = 0.5000 ⎝ 3 9 ⎠ ⎝ 3 9 ⎠ (d) α = 0.005 = 2.576 3 ) 9 = 362.0013 + 1 − 0. 1 1 1 1 = = = = 6.01.576 9 = 357.576 ( UCL x = x + kσ x = x + k σˆ x ( LCL x = 360 − 2.9987 = 0.0000 − 0.0026 ⎝ 3 9 ⎠ ⎝ 3 9 ⎠ (b) USL − LSL +6 − (−6) Cˆ P = = = 0.424 5-53 .

try moving the center of the process from its current mean of 620 closer to the nominal dimension of 610.685) = 8. Also consider reducing the process variability.1056 (d) To reduce the fraction nonconforming.Chapter 5 Exercise Solutions 5-50 (5-43). (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) + 1 − Φ (1.0000 5-54 .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) = 0. (a) σˆ x = R / d 2 = 8.9213(4) = 3.8944 = 0.266(3.9772 (f) Pr{detect by 3rd sample} = 1 − Pr{not detect by 3rd sample} = 1 − (Pr{not detect})3 = 1 − (1 − 0.236 / 2.0000 = 0.0000 + 1 − 0.9772 + 1 − 1.059 = 4.000 (b) S = c4σˆ x = 0.9772)3 = 1.865 UCL S = B4 S = 2.351 LCL S = B3 S = 0(3.

827) = 711.41) + 1 − Φ (8.738 / 0.9515 = 1.Chapter 5 Exercise Solutions 5-51 (5-44).827 (b) UNTL = x + 3σˆ x = 706 + 3(1.827) = 700.9920 + 1 − 1.9920 (e) Pr{detect by 3rd sample} = 1 − Pr{not detect by 3rd sample} = 1 − (Pr{not detect})3 = 1 − (1 − 0. (a) µˆ = x = 706.0000 = 0. σˆ x = S / c4 = 1.1006 (d) Pr{detect on 1st sample} = Pr{x < LCL} + Pr{x > UCL} ⎛ LCL − µnew ⎞ ⎛ UCL − µnew ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ σx σx ⎝ ⎠ ⎝ ⎠ ⎛ 703.827 ⎠ ⎝ 1.8 − 702 ⎞ ⎛ 708.642) = 0.9920)3 = 1.31) = 0.827 ⎠ = Φ (−1.827 6 ⎠ = Φ (2.48 LNTL = 706 − 3(1.0000 5-55 .00.642) + 1 − Φ (1.9497 = 0.827 6 ⎠ ⎝ 1.0503 + 1 − 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 ⎜ ⎟ ⎝ 1.2 − 702 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 1.

3085 + 1 − 0.979 / 0.22 1 − β 1 − Pr{not detect} Pr{detect} 0.31) = 0.5) + 1 − Φ (2.661 4 ⎠ = Φ (−2.661 ⎠ ⎝ 8.0208 (d) Pr{detect on 1st sample} = Pr{x < LCL} + Pr{x > UCL} ⎛ LCL − µnew ⎞ ⎛ UCL − µnew ⎞ = Φ⎜ +1− Φ ⎜ ⎟ ⎟⎟ ⎜ σ ⎟ ⎜ σ x . σˆ x = S / c4 = 7.1355 (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 = 0.0104 + 1 − 0.15) + 1 − Φ (2.661 ⎠ = Φ (−1.Chapter 5 Exercise Solutions 5-52 (5-45).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) + 1 − Φ (2.3108 (e) ARL1 = 1 1 1 1 = = = = 3.1251 + 1 − 0.9977 = 0.new ⎝ ⎠ ⎝ ⎠ ⎛ 690 − 693 ⎞ ⎛ 710 − 693 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 12 4 ⎠ ⎝ 12 4 ⎠ = Φ (−0.83) = 0.new x .3108 5-56 .31) = 0.9213 = 8. (a) µˆ = x = 700.661 4 ⎠ ⎝ 8.9896 = 0.

14 _ X=16.11 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.06 0.02044 25 0.05 LC L=16.1684 16.02375 0.100 Ex5-53Wt 16.11 0.Chapter 5 Exercise Solutions 5-53 (5-46).342 Percent 80 70 60 50 40 30 20 10 5 1 16.9964) = 0.125 16.1052 16.02375 MTB > Stat > Basic Statistics > Normality Test Probability Plot of Weight (Ex5-53Wt) Normal 99 Mean StDev N AD P-Value 95 90 16.17 U C L=16.021055.04 __ M R=0. MR2 = 0.050 16.150 Visual examination of the normal probability indicates that the assumption of normally distributed coffee can weights is valid.08 U C L=0.021055 ⎠ 5-57 . σˆ x = 0.1052. x = 16.397 0.00003% ⎝ 0. %underfilled = 100% × Pr{x < 16 oz} ⎛ 16 − 16. MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Weight (Ex5-53Wt) Individual Value 16.08 16.02 0.1052 ⎞ = 100% × Φ ⎜ ⎟ = 100% × Φ (−4.07760 0.075 16.0420 2 4 6 8 10 12 14 O bser vation 16 18 20 22 24 M oving Range 0.

50 M oving Range 10.Chapter 5 Exercise Solutions 5-54(5-47).0 __ M R=3.712 15 0. MR2 = 3. the p-value is greater than 0.0 1 2 3 4 5 6 7 8 9 O bser vation 10 11 12 13 14 15 x = 53. 5-58 .21 2.84954. σˆ x = 2.217 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. MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Hardness (Ex5-54Har) U C L=61.05.0 7.27 50 45 LC L=44.72 1 2 3 4 5 6 7 8 9 O bser vation 10 11 12 13 14 15 U C L=10.21429 MTB > Stat > Basic Statistics > Normality Test Probability Plot of Hardness (Ex5-54Har) Normal 99 Mean StDev N AD P-Value 95 90 53.27 2.5 5.82 Individual Value 60 55 _ X=53.2667.5 LC L=0 0. indicating that it may be reasonable to assume that hardness is normally distributed.465 0.

trends. runs.Chapter 5 Exercise Solutions 5-55 (5-48).0 20 0. (c) µˆ = x = 2928.9 2800 2600 LC L=2534. σˆ x = 131.346.158 5-59 .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.319 0. MR2 = 148.1 M oving Range 480 360 240 __ M R=148.511 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. (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=3322.9. (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. or other patterns.9 3200 3000 _ X=2928.9 2 4 6 8 10 12 O bser vation 14 16 18 20 U C L=484.

5-60 . 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=3322.9 3200 _ X=2928.Chapter 5 Exercise Solutions 5-56 (5-49).1 Moving Range 480 360 240 __ MR=148. indicating that a shift in the mean may have occurred.9 2 4 6 8 10 12 14 Observation 16 18 20 22 24 UCL=484.2 120 LCL=0 0 2 4 6 8 10 12 14 Observation 16 18 20 22 24 All points are inside the control limits. However all of the new points on the I chart are above the center line.9 3000 2800 2600 LCL=2534.

480 Percent 80 70 60 50 40 30 20 10 5 1 40 45 50 Ex5-57aTh 55 60 The normality assumption is reasonable. MTB > Stat > Basic Statistics > Normality Test Probability Plot of Oxide Thickness (Ex5-57aTh) Normal 99 Mean StDev N AD P-Value 95 90 49.14 60 _ X=49. 5-61 .Chapter 5 Exercise Solutions 5-57 (5-50).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.85 50 40 LC L=34. (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=65.85 4.79 15 10 __ M R=5.55 30 3 6 9 12 15 18 O bser vation 21 24 27 30 M oving Range 20 U C L=18.338 0.534 30 0.

Test Failed at points: 34. several 4 of 5 beyond 1 sigma. One point more than 3.85 50 40 LCL=34. 37. 39. Failed at points: 38. it is clear that the process is out of control during this period of operation.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. Failed at points: 38 2.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. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL).14 _ X=49. even without use of the sensitizing rules. 40 We have turned on some of the sensitizing rules in MINITAB to illustrate their use. 9 points in a row on same side of center line. 38. 5-62 . Test Failed at points: 35. 39. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). and several 2 of 3 beyond 2 sigma on the x chart.00 standard deviations from center line. 40 5.79 15 10 __ MR=5. 39. However.55 30 4 8 12 16 20 Observation 24 28 32 36 40 Moving Range 20 UCL=18. There is a run above the centerline. 40 TEST 6.

5-63 .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.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.14 2 _ X=49.85 50 40 LCL=34. with Sensitizing Rules On 70 1 Individual Value 5 60 6 6 2 UCL=65.55 30 1 6 12 18 24 30 Observation 36 48 54 60 1 20 Moving Range 42 UCL=18.79 15 10 __ MR=5.

(a) The normality assumption is a little bothersome for the concentration data. Test Failed at points: 11 Test Results for MR Chart of Ex5-58C TEST 1.88 5 100 80 _ X=73.26 30 20 __ M R=11. One point more than 3.71 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. 5-64 . Test Failed at points: 17 The process is not in control. with two Western Electric rule violations. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).73 60 LC L=42. in particular due to the curve of the larger values and three distant values.00 standard deviations from center line.Chapter 5 Exercise Solutions 5-58 (5-51). (b) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR I-MR Chart of Concentration (Ex5-58C) Individual V alue U C L=104.59 40 3 6 9 12 15 18 O bser vation 24 27 30 1 40 M oving Range 21 U C L=38.

288 0.0 4.9 4.3 4.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 4.6 4.2 4.1 4. 5-65 .408 0.1567 30 0.4 Ex5-58lnC 4.5 4.7 The normality assumption is still troubling for the natural log of concentration.327 Percent 80 70 60 50 40 30 20 10 5 1 3. again due to the curve of the larger values and three distant values.

24 __ M R=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 4.2884 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 5.0 LC L=3. 5-66 .5154 M oving Range 0. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).4 _ X=4.36 0.1577 0.8689 3 6 9 12 15 18 O bser vation 21 24 27 30 1 U C L=0.6 4.00 standard deviations from center line. Test Failed at points: 17 The process is still not in control. There does not appear to be much difference between the two control charts (actual and natural log).7079 5 4. Test Failed at points: 11 Test Results for MR Chart of Ex5-58lnC TEST 1.12 0. One point more than 3.48 0. with the same to Western Electric Rules violations.2 4.8 U C L=4.

7 300 200 __ M R=122.9 500 2 4 6 8 10 12 O bser vation 14 16 18 20 1 M oving Range 400 U C L=400. 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).1 1000 _ X=909 750 LC L=582.00 standard deviations from center line.Chapter 5 Exercise Solutions 5-59☺.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 1.9 20 0. 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. There may also be an early indication of less variability in the later measurements. consider the process to be in a state of statistical process control.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. use all data for control limits calculations. 5-67 . 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=1235.672 0. Since neither of these measurements seems unusual. One point more than 3. For now.

which is reflected in the number of observations below the centerline of the moving range chart. 9 points in a row on same side of center line. Failed at points: 8 2.Chapter 5 Exercise Solutions 5-60☺. (b) Early measurements exhibit more variability than the later measurements. (a) MTB > Stat > Control Charts > Variables Charts for Individuals > I-MR Select I-MR Options.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. however the method of measuring is producing varying results—this is a chart of the measurement process. 9 points in a row on same side of center line. 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. Failed at points: 36. One point more than 3. 37 The velocity of light in air is not changing. meaning the method is producing gradually smaller measurements.1 1000 _ X=909 2 2 2 750 2 2 LCL=582. 38.7 Moving Range 400 300 200 100 2 __ MR=122. 39.9 500 4 8 12 16 20 Observation 24 28 32 36 40 1 UCL=400. 40 Test Results for MR Chart of Ex5-60Vel TEST Test TEST Test 1. Test Failed at points: 36. There is a distinct downward trend in measurements. 5-68 .00 standard deviations from center line. 37.

The data are skewed right.07 5.shaped curve to the plot points on a normal probability plot.0 2. Probability Plot of ln(Uniformity) (Ex5-61lnUn) Normal 99 Mean StDev N AD P-Value 95 90 2.3493 30 0.8 Ex5-61lnUn 3.626 0.546 30 1.005 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. (a) MTB > Stat > Basic Statistics > Normality Test Probability Plot of Uniformity Determinations (Ex5-61Un) Normal 99 Mean StDev N AD P-Value 95 90 15.6 The distribution of the natural-log transformed uniformity measurements is approximately normally distributed. as well as the Anderson-Darling test p-value. as evidenced by the “S”. so a compressive transform such as natural log or square-root may be appropriate.653 0. 5-69 .2 3.093 Percent 80 70 60 50 40 30 20 10 5 1 2.4 2.Chapter 5 Exercise Solutions 5-61☺.

00 0.5 2.00 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.25 0.75 0.5 3 6 9 12 15 18 O bser vation 21 24 27 30 U C L=1.146 M oving Range 1.5 3.351 0.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=3.50 __ M R=0. 5-70 .0 LC L=1.653 2.586 Individual V alue 3.720 1.0 _ X=2.

84 Ex5-62Pur 0.174 <0.Chapter 5 Exercise Solutions 5-62 (5-52).82 0. 5-71 . (a) MTB > Stat > Basic Statistics > Normality Test Probability Plot of Batch Purity (Ex5-62Pur) Normal 99 Mean StDev N AD P-Value 95 90 0.86 0.87 Purity is not normally distributed.78 0.79 0.85 0.01847 20 1.81 0.824 0.005 Percent 80 70 60 50 40 30 20 10 5 1 0.80 0.83 0.

σˆ x = 0.012 0.04987 M oving Range 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 1.01526 0.82 6 0.024 __ M R=0.80 LC L=0.86 6 0. One point more than 3.048 0.00 standard deviations from center line.84 _ X=0.8216 .824 .824 0.86459 5 0.78341 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. 20 The process is not in statistical control.78 2 4 6 8 10 12 O bser vation 14 16 18 20 U C L=0.0133 5-72 .0135 without sample 18: µˆ = 0. Test Failed at points: 19 TEST 6. (c) all data: µˆ = 0.036 0. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). Test Failed at points: 11. σˆ x = 0. Test Failed at points: 18 TEST 5. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).

08 16. control limits for both are essentially the same.04 __ M R=0.1052 16. 5-73 .1681 16.14 _ X=16.08 U C L=0. 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 16.07726 0.0423 2 4 6 8 10 12 14 O bser vation 16 18 20 22 24 M oving Range 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 5-53.05 LC L=16.02 0.02365 0.Chapter 5 Exercise Solutions 5-63 (5-53).06 0.11 16.17 U C L=16.

Chapter 5 Exercise Solutions 5-64 (5-54).27 50 LC L=45.0 __ M R=2.5 0.66 7.0 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. with no practical difference for this set of observations.13 Individual V alue 60 55 _ X=53.41 45 1 2 3 4 5 6 7 8 9 O bser vation 10 11 12 13 14 15 M oving Range 10.96 2. 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=61.5 5. 5-74 .0 U C L=9.

with no practical meaning for this set of observations.7 3200 3000 _ X=2928. 5-75 .2 M oving Range 480 360 240 __ M R=153.7 120 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.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.

One point more than 3. Test Failed at points: 34. Failed at points: 38. 40 TEST 6.03 1 6 12 18 24 30 Observation 36 48 54 60 1 20 Moving Range 42 UCL=18.05 50 40 LCL=36. 9 points in a row on same side of center line. Failed at points: 38 2. 40 5. 39.65 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 1.00 standard deviations from center line.46 15 10 __ MR=5. 39.07 2 6 _ X=51. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 38. Remove observation 38 and recalculate control limits. 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. 40 Test Results for MR Chart of Ex5-57cTh TEST 1. 39. 37. Test Failed at points: 41 Recall that observations on the Moving Range chart are correlated with those on the Individuals chart—that is. One point more than 3. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).00 standard deviations from center line. 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. Test Failed at points: 35. 5-76 .Chapter 5 Exercise Solutions 5-66 (5-56).

94 1 6 12 18 24 30 Observation 36 42 48 54 60 1 20 Moving Range UCL=18.77 50 40 LCL=35. Failed at points: 38. Failed at points: 38 2.60 2 6 _ X=50. 39.58 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. Test Failed at points: 41 5-77 .00 standard deviations from center line. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 38.Chapter 5 Exercise Solutions 5-66 (a) continued Excluding observation 38 from calculations: I-MR Chart of Oxide Thickness (Ex5-57cTh) Less Observation 38 -. One point more than 3.22 15 10 5 __ MR=5. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 39. 39. 40 5. 37. 40 TEST 6. 40 Test Results for MR Chart of Ex5-57cTh TEST 1. Test Failed at points: 34. 9 points in a row on same side of center line. One point more than 3.00 standard deviations from center line. Test Failed at points: 35.Average Moving Range Method 70 1 Individual Value 5 60 2 6 UCL=65.

40 Test Results for MR Chart of Ex5-57cTh TEST 1.Chapter 5 Exercise Solutions 5-66 continued (b) I-MR Chart of Oxide Thickness (Ex5-57cTh) All 60 Observations -.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. One point more than 3. Test Failed at points: 41 5-78 . 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 40 TEST 6. Test Failed at points: 35. 39. 39. Failed at points: 38 2. 40 5. One point more than 3.00 standard deviations from center line. 39. Failed at points: 38. 9 points in a row on same side of center line. Test Failed at points: 34.05 50 40 LCL=36.00 standard deviations from center line.83 2 6 _ X=51.16 15 10 5 __ MR=5.Median Moving Range Method 70 1 Individual Value 5 60 2 6 UCL=65.27 1 6 12 18 24 30 Observation 36 42 48 54 60 1 20 Moving Range UCL=18. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 38. 37.

One point more than 3. Test Failed at points: 34. 38 2. 38. One point more than 3. Failed at points: 38. 5-79 .Median Moving Range Method 70 1 1 Individual Value 5 60 2 6 2 UCL=64.61 6 _ X=50. 39. 40 Test Results for MR Chart of Ex5-57cTh TEST 1. Failed at points: 33. 33.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 -.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. 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. 37. 9 points in a row on same side of center line. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). Test Failed at points: 35. 39.93 1 6 12 18 24 30 Observation 36 42 48 54 60 1 20 Moving Range UCL=17. 40 TEST 6.00 standard deviations from center line. 39.00 standard deviations from center line.00 15 10 5 __ MR=5. 40 5.77 50 40 LCL=36. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).

018 12 _ X=10. (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.Chapter 5 Exercise Solutions 5-67 (5-57).630 σˆ x = S / c4 = 1.262 M oving Range 4 3 2 __ M R=1.128 = 1.305 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 = 1.549 StDev 1.157 (b) MTB > Stat > Basic Statistics > Descriptive Statistics Descriptive Statistics: Ex5-67Meas Variable Ex5-67Meas Total Count 25 Mean 10.079 2 4 6 8 10 12 14 O bser vation 16 18 20 22 24 U C L=4.549 10 8 LC L=7.682 5-80 .7979 = 1.342 / 0.305 /1.342 Median 10.

there are fewer observations to estimate the standard deviation.435 (e) As the span of the moving range is increased. This tends to be true for unstable processes.128 = 1.186 / 3.049 /1.693 = 1.137 2 4 6 8 10 12 14 Observation 16 18 20 22 24 UCL=4.137 (d) Average MR3 Chart: σˆ x = R / d 2 = 2.283 /1.262 Average MR19 Chart: σˆ x = R / d 2 = 5. For this example.549 10 8 LCL=7.406 Average MR20 Chart: σˆ x = R / d 2 = 5.192 Moving Range 4 3 2 __ MR=1.735 = 1. 5-81 .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. and the estimate becomes less reliable.283 1 LCL=0 0 2 4 6 8 10 12 14 Observation 16 18 20 22 24 σˆ x = R / d 2 = 1. σ gets larger as the span increases.689 = 1.961 12 _ X=10.598 / 2.210 Average MR4 Chart: σˆ x = R / d 2 = 2.059 = 1.

.0328230 Within 0. When compared to the s chart for all data.70 LCL=5.7566 5. Estimate” and choose R-bar method to estimate standard deviation I-MR-R (Between/Within) Chart of Vane Heights (Ex5-68v1.10 0.0143831 Between/Within 0. Ex5-68v5 Standard Deviations Between 0.07074 _ R=0.0358361 The Individuals and Moving Range charts for the subgroup means are identical.82 _ X=5. Ex5-68v5) Subgroup Mean UCL=5.8569 5. MTB > Stat > Control Charts > Variables Charts for Subgroups > I-MR-R/S (Between/Within) Select “I-MR-R/S Options. For this example.08 UCL=0.04 0.Chapter 5 Exercise Solutions 5-68 (5-58)...00 LCL=0 2 4 6 8 10 12 14 16 18 20 Sample Range 0...05 __ MR=0.1233 0.76 5.00 LCL=0 2 4 6 8 10 Ex5-68Cast 12 14 16 18 20 I-MR-R/S Standard Deviations of Ex5-68v1. .6563 MR of Subgroup Mean 2 4 6 8 10 12 14 16 18 20 UCL=0. the control schemes are identical. 5-82 . .03345 0.0377 0.. the R chart tells the same story—same data pattern and no out-of-control points.

. between castings or within a casting.7931 11.7226 11..79 U C L=11.03 0.70 1 1 2 4 6 8 10 Sample 12 14 16 18 20 U C L=0. Test Failed at points: 7 (b) Though the R chart is in control. .Chapter 5 Exercise Solutions 5-69 (5-59). If no assignable cause can be found for these out-of-control points.0611 0.09 _ R=0.. One point more than 3.1292 Sample Range 0. 9. 13. Since these are high precision castings... 17 TEST 5. 5-83 . Ex5-69d5) 1 11. we may want to consider treating the averages as an Individual value and graphing “between/within” range charts.. .. plot points on the x chart bounce below and above the control limits. (a) MTB > Stat > Control Charts > Variables Charts for Subgroups > Xbar-R Xbar-R Chart of Casting Diameter (Ex5-69d1. Ex5-69d5 TEST 1..73 LC L=11.00 LC L=0 2 4 6 8 10 Sample 12 14 16 18 20 Xbar-R Chart of Ex5-69d1. Test Failed at points: 5.82 1 Sample M ean 1 11. 7.76 _ _ X=11. Ex5-69d5 Test Results for Xbar Chart of Ex5-69d1. ..12 0.7579 11. This will lead to a understanding of the greatest source of variability. we might expect that the diameter of a single casting will not change much with location.06 0. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).00 standard deviations from center line.

Ex5-69d5 I-MR-R/S Standard Deviations of Ex5-69d1.0416 0.8 _ X=11.0437327 (d) We are taking several diameter measurements on a single precision casting.8685 11. Ex5-69d5 11.. It describes the measurement variability within a sample (variability in diameter of a single casting). . etc.. (e) The “within” chart is the usual R chart (n > 1).1292 0.10 _ R=0..05 0.1360 0. damage.05 LCL=0 0.0611 0.7 LCL=11...7579 11.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. Estimate” and choose R-bar method to estimate standard deviation MR-R/S (Between/Within) Chart of Casting Diameter (Ex5-69d1.. in the wax mold.9 Subgroup Mean UCL=11. Ex5-69d5 Standard Deviations Between 0..10 __ MR=0. Though the nature of this process leads us to believe that the diameter at any location on a single casting does not change much..00 LCL=0 2 4 6 8 10 Sample 12 14 16 18 20 I-MR-R/S (Between/Within) Chart of Ex5-69d1.0262640 Between/Within 0.. . we should continue to monitor “within” to look for wear. .00 2 4 6 8 10 12 14 16 18 20 Sample Range UCL=0. 5-84 ..0349679 Within 0.6472 MR of Subgroup Mean 2 4 6 8 10 12 14 16 18 20 UCL=0.

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

Chapter 5 Exercise Solutions 5-71 (5-61).04515 200 1.9 Mean StDev N AD P-Value 99 Percent 95 90 2. s = 0.1 1.333 <0.95 2.04515 UNTL = x + 3s = 2. with many repeated values.90 1.05 2. The natural tolerance limits (± 3 sigma above and below mean) are: x = 2.25 Although the p-value is very small.074 + 3(0.074 − 3(0.005 80 70 60 50 40 30 20 10 5 1 0. (a) MTB > Stat > Basic Statistics > Normality Test Probability Plot of Critical Dimensions (Ex5-71All) Normal 99.939 5-86 . the plot points do fall along a straight line.15 2.074. The wafer critical dimension is approximately normally distributed.209 LNTL = x − 3s = 2.04515) = 2.00 2.20 2.074 0.04515) = 1.10 Ex5-71All 2.

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 (Ex5-71p1.07 0.. for a total of 40 subgroups.00 LCL=0 4 8 12 16 20 24 28 32 Sample (Lot Number-Wafer Order) 36 40 The Range chart is in control.04 0.12 0.Chapter 5 Exercise Solutions 5-71 continued (b) To evaluate within-wafer variability.02 0.10 0.06 0.1480 0. 5-87 . .16 UCL=0.08 _ R=0. construct an R chart for each sample of 5 wafer positions (two wafers per lot number).. Ex5-71p5) 0.14 Sample Range 0.

00 LCL=1.1480 _ R=0. Ex5-71p5) Subgroup Mean Variability between wafers 2.07 0.1066 0.00 LCL=0 4 8 12 16 20 Sample 24 28 32 36 40 I-MR-R/S Standard Deviations of Ex5-71p1. The “within” chart (Range) is not required to evaluate variability between wafers.00 4 8 12 16 20 24 28 32 36 40 Sample Range 0. 5-88 .16 UCL=0.10 0.. 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.16 UCL=2. MTB > Stat > Control Charts > Variables Charts for Subgroups > I-MR-R/S (Between/Within) Select “I-MR-R/S Options.. Estimate” and choose R-bar method to estimate standard deviation I-MR-R/S (Between/Within) Chart of Crit Dim (Ex5-71p1.05 __ MR=0.0395043 Both “between” control charts (Individuals and Moving Range) are in control.0326 LCL=0 0..0255911 Within 0.0735 6 6 2. indicating that between-wafer variability is also in-control.Chapter 5 Exercise Solutions 5-71 continued (c) To evaluate variability between wafers.08 0. Ex5-71p5 Standard Deviations Between 0.. ..9868 MR of Subgroup Mean 4 8 12 16 20 24 28 32 36 40 UCL=0.1603 2. ..08 _ X=2.0300946 Between/Within 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 2.096 0. three charts are needed: (1) lot average.0394733 Within 0.1956 2.10 0.9515 2 4 6 8 10 12 14 16 18 20 MR of Subgroup Mean 0.15 _ R=0.05 LCL=0.0 LCL=1.0214 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.Chapter 5 Exercise Solutions 5-71 continued (d) To evaluate lot-to-lot variability.16 UCL=0.1500 0. (2) moving range between lot averages.0311891 Between/Within 0.0459 0.1 _ X=2.1706 0.2 UCL=2.0735 2.00 LCL=0 2 4 6 8 10 12 14 16 18 20 Sample Range UCL=0.08 __ MR=0. 5-89 .0503081 All three control charts are in control. indicating that the lot-to-lot variability is also in-control. and (3) range within a lot—the Minitab “between/within” control charts.

∑ Di = 117. not 8.00 standard deviations from center line.0585 − 0. This can be changed under Tools > Options > Control Charts and Quality Tools > Define Tests. it is considered to be out of control.1289 Proportion 0.0585 0. Test Failed at points: 12 6-1 . m = 20.16 1 0.06 0.0585(1 − 0.12 0. In particular. m m n = 100.0704 ⇒ 0 n 100 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Assemblies (Ex6-1Num) 0. Also fewer special cause tests are available for attributes control charts. 3.1289 n 100 LCL p = p − 3 p (1 − p ) 0. a run of n consecutive points on one side of the center line is defined as 9 points.0585) = 0.08 _ P=0.02 0. 2. 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.Chapter 6 Exercise Solutions Notes: 1.0585(1 − 0. If a plot point is within the control limits. we follow the MINITAB convention for determining whether a point is out of control.14 UCL=0. p = i =1 ∑ Di i =1 mn = 117 = 0.0585 + 3 = 0.0585) = 0. it is considered to be in control. MINITAB defines some sensitizing rules for control charts differently than the standard rules. For these solutions. If a plot point is on or beyond the control limits.0585 20(100) UCL p = p + 3 p (1 − p ) 0.0585 − 3 = 0. New exercises are denoted with an “☺”.04 0. 6-1.10 0.

1213 100 LCL p = 0.0537(1 − 0.Chapter 6 Exercise Solutions 6-1 continued Sample 12 is out-of-control.0537(1 − 0. Test Failed at points: 12 6-2 .0537) = 0.12 0.0537) = 0.0537 − 3 0. p = i =1 ∑ Di i =1 mn = 102 = 0.10 0.04 0.0537 + 3 0.00 standard deviations from center line.0537 − 0. so remove from control limit calculation: m m n = 100.06 0.08 _ P=0.0676 ⇒ 0 100 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Assemblies (Ex6-1Num) Sample 12 removed from calculations 0. ∑ Di = 102.16 1 0.14 UCL=0.1213 Proportion 0. m = 19.00 LCL=0 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-1Num TEST 1.02 0.0537 0.0537 19(100) UCL p = 0. One point more than 3.

0230 20(150) UCL p = p + 3 p (1 − p ) 0.0230 − 0.00 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-2Num TEST 1.10 0. p = ∑ Di i =1 mn i =1 = 69 = 0.0597 n 150 LCL p = p − 3 p (1 − p ) 0.0230) = 0.0230 + 3 = 0.Chapter 6 Exercise Solutions 6-2.0230(1 − 0.02 LCL=0 0.023 0.0230(1 − 0. m = 20.08 Proportion 1 0.06 UCL=0.00 standard deviations from center line.0230) = 0. Test Failed at points: 9.0367 ⇒ 0 n 150 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Switches (Ex6-2Num) 1 0. One point more than 3.0230 − 3 = 0. m m n = 150. 17 6-3 . ∑ Di = 69.04 _ P=0.0597 0.

00 LCL=0 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-2Num TEST 1.0473 0. 17 6-4 .10 0.06 1 UCL=0. Test Failed at points: 1.0163 0. 9.04 0. One point more than 3.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 0.02 _ P=0.00 standard deviations from center line.08 Proportion 1 0.

One point more than 3. 17 excluded from calculations 1 0. ∑ Di = 36.0141 − 3 0.06 1 UCL=0.10 0.0141 + 3 0.0141 − 0.08 Proportion 1 0.0430 150 LCL p = 0.00 standard deviations from center line. Test Failed at points: 1.0141 17(150) UCL p = 0.02 _ P=0.04 0.0141 0. 17 6-5 . 9.0141(1 − 0.0141) = 0.00 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(1 − 0.0289 ⇒ 0 150 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Switches (Ex6-2Num) Samples 1. 9. m = 17.0141) = 0.0430 0.Chapter 6 Exercise Solutions 6-2 continued Also remove sample 1 from control limits calculation: m m n = 150.

08 _ P=0.06 0.14 UCL=0.1397 LCL1 = p − 3 p (1 − p ) n1 = 0.06 UCLi = p + 3 p (1 − p ) ni and LCLi = max{0.046. The Fraction Nonconforming for Day 5 should be 0. for n = 80: UCL1 = p + 3 p (1 − p ) n1 = 0.16 0.06) 80 = 0. NOTE: There is an error in the table in the textbook.1331 Proportion 0.02 0.06 + 3 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 0.10 0.12 0.06) 80 = 0.06(1 − 0.06 − 3 0.0797 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Units (Ex6-3Num) 0. m m m m i =1 i =1 i =1 i =1 m = 10.04 0. p = ∑ Di ∑ ni = 60 1000 = 0. ∑ Di = 60. ∑ ni = 1000.06(1 − 0.Chapter 6 Exercise Solutions 6-3. p − 3 p (1 − p ) ni } As an example.06 − 0. 6-6 .

03 0.0167 > 529.0167) 150 = 0.9 Select n = 530.0167 − 0. m = 20.0167) 150 = 0.00 2 4 6 8 10 12 Sample 14 16 18 20 The process appears to be in statistical control.0167 + 3 0.01667 0.0167) 2 (3) > 0. 6-7 . ∑ Di = 50.0314 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Units (Ex6-4Num) 0.Chapter 6 Exercise Solutions 6-4.01 LCL=0 0.02 _ P=0. p = ∑ Di mn = 50 20(150) = 0.0167(1 − 0.0167 − 3 0. (a) m m i =1 i =1 n = 150.0167(1 − 0. (1 − p) 2 n> L p (1 − 0.04 0.04802 Proportion 0. (b) Using Equation 6-12.05 UCL=0.0480 LCL = p − 3 p (1 − p ) n = 0.0167 UCL = p + 3 p (1 − p ) n = 0.

1228 0. Test Failed at points: 1.1228) 2500 = 0.1031 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Nonconforming Belts (Ex6-5Num) 0.1031 1 1 1 0. 17.1425 LCL = p − 3 p (1 − p ) n = 0. (a) UCL = p + 3 p (1 − p ) n = 0. 5.175 1 1 1 Proportion 0. One point more than 3.1228 + 3 0.050 1 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for P Chart of Ex6-5Num TEST 1.075 1 1 0.200 1 1 0.100 LCL=0. 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.1425 0.1228(1 − 0.1228(1 − 0. 3. Instead. 11. the process should be investigated for causes of the wild swings in p.00 standard deviations from center line.Chapter 6 Exercise Solutions 6-5. 2.125 _ P=0. 12. 15. 19. 6-8 .150 UCL=0.1228) 2500 = 0.1228 − 3 0. 16.

One point more than 3.976 LCL = np − 3 np (1 − p ) = 4 − 3 4(1 − 0.00 standard deviations from center line. UCL = np + 3 np (1 − p ) = 4 + 3 4(1 − 0.008) = 9.008) = 4 − 5.Chapter 6 Exercise Solutions 6-6. Test Failed at points: 6 6-9 .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 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > NP NP Chart of Number of Nonconforming Units (Ex6-6Num) 1 12 Sample Count 10 UCL=9.

Chapter 6 Exercise Solutions 6. One point more than 3.6 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=8.00 standard deviations from center line.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 1.39 8 6 4 __ NP=3. Test Failed at points: 6 Recommend using control limits from second chart (calculated less sample 6). 6-10 .

0044(1 − 0. 6-11 .0044(1 − 0. the UCL.04) = 2.Chapter 6 Exercise Solutions 6-7.624 6-8. n = 250.02 − 3 0. From the 6th sample. ∑ pˆ i = 0.0126 ⇒ 0 No.0594 ⇒ 0 Since pnew = 0.02 − 0.2) + POI(0.02) 50 = 0. Pr{detected by 3rd sample} = 1 – Pr{detected after 3rd} = 1 – (1 – 0.0794 LCL = p − 3 p (1 − p ) n = 0. use the Poisson approximation to the binomial with λ = npnew = 50(0. Pr{detect|shift} = 1 – Pr{not detect|shift} =1–β = 1 – [Pr{D < nUCL | λ} – Pr{D ≤ nLCL | λ}] = 1 – Pr{D < 50(0.278 where POI(⋅) is the cumulative Poisson distribution.0044 + 3 0.00.0440 = 0. p = UCL = p − 3 p (1 − p ) n = 0.0170.0044) 250 = 0.857 + 0.278)3 = 0. n = 50 UCL = p + 3 p (1 − p ) n = 0.02.2) = 1 – 0.04 < 0. ( pˆ 6 = 0. The data from the shipment do not indicate statistical control.135 = 0.0044 − 3 0. 0. p = 0.02(1 − 0.0794) | 2} + Pr{D ≤ 50(0) | 2} = 1 – POI(3.02 + 3 0.0170 10 m = 10.020) > 0.0044) 250 = 0.0044 i =1 10 UCL = p + 3 p (1 − p ) n = 0.0440.1 and n = 50 is "large".02) 50 = 0.02(1 − 0.0044 − 0.

21 0.503553 Assuming L = 3 sigma control limits.44154 0.2125) | p} − Pr{D ≤ 64(0) | p} = Pr{D < 13.10) 64 = 0.000000 β 0.44154 0.10(1 − 0.10(1 − 0.05 0.503553 Pr{D ≤ 0|p} 0.212 Pr{D ≤ 13|p} 0. p = 0.519279 0.996172 0.480098 0.1125 ⇒ 0 β = Pr{D < nUCL | p} − Pr{D ≤ nLCL | p} = Pr{D < 64(0.519279 0.10 0.598077 0.000000 0.000000 0.000000 0.000000 0. n = 64 UCL = p + 3 p (1 − p ) n = 0.037524 0.10 > 81 6-12 .215 0.2125 LCL = p − 3 p (1 − p ) n = 0.480098 0.10 + 3 0.6) | p} − Pr{D ≤ 0 | p} p 0.20 0.598077 0.22 0.10 − 0.999999 0.10) 2 (3) > 0.10) 64 = 0.Chapter 6 Exercise Solutions 6-9.10 − 3 0.001179 0. (1 − p ) 2 n> L p (1 − 0.994993 0.10.962475 0.

875) = 1 − 0. 6-11.16) 2 (3) > 0. p = 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. np = 16.00 (a) npnew = 20.030 Pr{detect by at least 3rd} = 1 – Pr{detected after 3rd} = 1 – (1 – 0.16 UCL = np + 3 np(1 − p ) = 16 + 3 16(1 − 0.16) = 5.5 − 20 ⎞ = 1− Φ ⎜ + Φ⎜ ⎟ ⎜ 20(1 − 0.970 + 0.10) = 81 ⎝δ ⎠ ⎝ 0.Chapter 6 Exercise Solutions 6-10.875) + Φ (−3.10.2) ⎟ ⎜ 20(1 − 0. p = 16 100 = 0.10 ⎠ 6-13 .50.0.0 > 15. assume k = 3 sigma control limits new − p = 0.00 LCL = np − 3 np (1 − p ) = 16 − 3 16(1 − 0.0873 (b) Assuming L = 3 sigma control limits.2) ⎟⎟ ⎝ ⎠ ⎝ ⎠ = 1 − Φ (1.10 = 0.000 = 0.10 δ =p new 2 2 ⎛k⎞ ⎛ 3 ⎞ n = ⎜ ⎟ p(1 − p) = ⎜ ⎟ (0.5 − 20 ⎞ ⎛ 5 − 0.16 > 47. (1 − p ) 2 n> L p (1 − 0.030)3 = 0. n = 100. n = 48 is the minimum sample size for a positive LCL. desire Pr{detect} = 0. so use normal approximation to binomial distribution.10)(1 − 0.20.20 − 0.16) = 27.25 So. p = 0.

161 − 0. LCL = 0 (a) np = 100(0.1 and n =100 is large.08(1 − 0.080 < 0.080) = 8 − 8.07494)4 = 0.07494 (d) Pr{detect shift by at most 4th sample} = 1 – Pr{not detect by 4th} = 1 – (0.20 0 − 0.1388 ⇒ 0 (b) p = 0.37) = 0. UCL = 0.07494 − 0 = 0. so use Poisson approximation to the binomial.Chapter 6 Exercise Solutions 6-12. 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(16.08.08) 100 ⎟⎟ ⎝ ⎠ ⎝ ⎠ = Φ (−1.004 where POI(⋅) is the cumulative Poisson distribution. so use the normal approximation to the binomial. (c) npnew = 100(0.99997 6-14 .8)] = 0 + [1 – 0. n = 100.08(1 − 0.44) − Φ (−7.20 = Φ⎜ −Φ⎜ ⎟ ⎜ 0.080) = 16.20) = 20 > 15.161. 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 ⎟⎟ ⎝ ⎠ ⎝ ⎠ ⎛ ⎞ ⎛ ⎞ 0.14 LCL = np − 3 np (1 − p ) = 8 − 3 8(1 − 0.08) 100 ⎟ ⎜ 0.080) = 8 UCL = np + 3 np(1 − p ) = 8 + 3 8(1 − 0.996] = 0. p = 0.

07(1 − 0.297)(0.533) + Φ (−4. 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 ⎟⎟ ⎝ ⎠ ⎝ ⎠ ⎛ ⎛ 0.297 (c) Pr{detect on 1st or 2nd sample} = Pr{detect on 1st} + Pr{not on 1st}×Pr{detect on 2nd} = 0.032 − 0.1) 400 ⎟ ⎜ 0.1) 400 ⎟⎟ ⎝ ⎠ ⎝ ⎠ = 1 − Φ (0.533) = 1 − 0.07 + 3 0.07) 400 = 0.297) = 0.06 2 2 ⎛k⎞ ⎛ 3 ⎞ n = ⎜ ⎟ p(1 − p) = ⎜ ⎟ (0.50 after a shift to pnew = 0.26 − 0.20 and L = 3 sigma control limits (1 − p ) 2 n> L p (1 − 0. δ =pnew − p = 0.07.108 − 0.20 = 0.10) = > 40.07) 400 = 0. so use the normal approximation to the binomial.20)(1 − 0.108 LCL = p − 3 p(1 − p ) n = 0.26.506 6-14.07(1 − 0. k = 3 sigma control limits.06 ⎠ 6-15 . p = 0.703 + 0. (a) p = 0.000 = 0.Chapter 6 Exercise Solutions 6-13.1(1 − 0.297 + (1 – 0.07 − 3 0.20 > 36 For Pr{detect} = 0. n = 400 UCL = p + 3 p(1 − p) n = 0.1 ⎞ 0.1 ⎞ = 1− Φ ⎜ + Φ⎜ ⎟ ⎜ 0.20) = 400 ⎝δ ⎠ ⎝ 0.032 (b) npnew = 400(0.1(1 − 0.20) 2 (3) > 0.

29 5 1 2 3 4 5 6 Sample 7 8 9 10 Test Results for NP Chart of Ex6-15Num TEST 1. n = 100. (a) m = 10.4(1 − 0.51 Sample Count 25 20 __ NP=16.4 15 10 LCL=5.51 LCL = np − 3 np (1 − p ) = 16. One point more than 3. Test Failed at points: 3 6-16 .4(1 − 0.4 UCL = np + 3 np (1 − p ) = 16.164) = 5.4 − 3 16.164.292 MTB > Stat > Control Charts > Attributes Charts > NP NP Chart of Number Nonconforming (Ex6-15Num) 1 30 UCL=27.164) = 27. np = 16. i =1 10 p = ∑ Di i =1 ( mn ) = 164 [10(100)] = 0.Chapter 6 Exercise Solutions 6-15.4 + 3 16.00 standard deviations from center line. 10 ∑ Di = 164.

13 1 2 3 4 5 6 Sample 7 8 9 10 Test Results for NP Chart of Ex6-15Num TEST 1.78 15 10 5 LCL=4.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. Test Failed at points: 3 6-17 .42 Sample Count 25 20 __ NP=14. One point more than 3.00 standard deviations from center line.

0662 LCL p = p − 3 p (1 − p ) n = 0.813 6-16.5 − 30 ⎞ ⎛ 4.08.03(1 − 0.5 − 30 ⎞ = 1− Φ ⎜ +Φ⎜ ⎟ ⎜ ⎜ 30(1 − 0.03(1 − 0.8903) + Φ (−5.3) ⎟⎟ 30(1 − 0.3) ⎟⎠ ⎝ ⎝ ⎠ = 1 − Φ (−0.03) 200 = 0. Pr{detect by at least 4th} = 1 – Pr{detect after 4th} = 1 – (1 – 0.30 is not too far from 0.0362 ⇒ 0 (b) pnew = 0.7544) = 1 − (0. use the Poisson approximation to the binomial.2745 + 0.03 − 3 0.08) < 0.03 − 0.000) = 0.7255)4 = 0.10 and n is large.50.9943 6-18 . the normal approximation to the binomial can be used.Chapter 6 Exercise Solutions 6-15 continued (b) pnew = 0. and n = 100 > 10. Since p = 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 ) ⎟⎠ ⎝ ⎝ ⎛ 25.7255 where POI(⋅) is the cumulative Poisson distribution.03 + 3 0.16) + POI(0.08)} + Pr{D ≤ 200(0) | 200(0.16) = 1 − 0.42 + 0. (a) UCL p = p + 3 p (1 − p ) n = 0.30.08)} = 1 − POI(13. 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(0.000 = 0.03) 200 = 0.187) + (0. Since (pnew = 0.0662) | 200(0.13 − 0.

417 + 0. np = 400(0.15) ⎟⎟ ⎝ ⎠ ⎝ ⎠ = 1 − Φ (−0. (a) m p = ∑ Di i =1 UCL LCL np np ( mn ) = 1200 [30(400)] = 0.15) = 60 > 15.210) + Φ (−5. so use the normal approximation to the binomial.10) = 22 (b) npnew = 400 (0.10) = 58 = np − 3 np (1 − p ) = 40 − 3 40(1 − 0.000 = 0.10.Chapter 6 Exercise Solutions 6-17. 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 + 0.583 6-19 .5 − 60 ⎞ ⎛ 22 − 0.5 − 60 ⎞ = 1− Φ ⎜ + Φ⎜ ⎟ ⎜ 60(1 − 0.39) = 1 − 0.15) ⎟ ⎜ 60(1 − 0.10) = 40 = np + 3 np (1 − p ) = 40 + 3 40(1 − 0.

004 where POI(⋅) is the cumulative Poisson distribution. This is a continuation of Exercise 6-17.10) = 10. 20) = 0.715 ≅ 2 6-20.381 ARL1 = 1/(1 –β) = 1/(1 – 0. 20) − POI(1.19) | 20} − Pr{D ≤ 100(0. NOTE: There is an error in the textbook.10) = 0.20.000 + 1 − 0. 6-19.1) ⎜ ⎟ = 100 ⎝ 0.19 − 0. Pr{type II error} = β = Pr{D < nUCL | λ} − Pr{D ≤ nLCL | λ} = Pr{D < 100(0. from 6-17(b). (c) pnew = 0. λ = np = 100(0. λ = npnew = 100(0.20) = 20.381 − 0. Pr{type I error} = Pr{ pˆ < LCL | p} + Pr{ pˆ > UCL | p} = Pr{D < nLCL | λ} + 1 − Pr{D ≤ nUCL | λ} = Pr{D < 100(0.1 ⎠ ⎝ UCL − p ⎠ 2 (b) Using the Poisson approximation to the binomial. not 6-18. 1 – β = 0.19) |10} = POI(0.996 = 0.10) + 1 − POI(19.Chapter 6 Exercise Solutions 6-18.1(1 − 0.01) |10} + 1 − Pr{D ≤ 100(0.616 ≅ 2 6-20 .583) = 1.01) | 20} = POI(18.000 = 0. β = 0.583 ARL1 = 1/(1 –β) = 1/(0.381 where POI(⋅) is the cumulative Poisson distribution.381) = 1. (a) UCL = p + 3 p (1 − p ) n 2 ⎛ ⎞ 3 3 ⎛ ⎞ n = p (1 − p ) ⎜ ⎟ = 0. Using the Poisson approximation to the binomial. from 6-18(c).

0.05005 0.0581] [0. 0.02 0.0221 ± 3 0.0221(1 − 0.0221 and control limits are at p ± 3 p (1 − p ) / ni [LCLi.0500] ni 100 150 200 250 MTB > Stat > Control Charts > Attributes Charts > P P Chart of Second Visit Required (Ex6-21Sec) 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.0216 / ni . 0.0662] [0.0533] [0.04 0. In those cases. and to plot Z i = ( pˆ i − 0. preferred in many cases. (a) For a p chart with variable sample size: p = ∑ i Di ∑ i ni = 83 / 3750 = 0. UCLi] [0.02213 0.0221) 0. 0.Chapter 6 Exercise Solutions 6-21. would be to construct standardized control limits with control limits at ± 3.05 UCL=0. (b) There are two approaches for controlling future production. 6-21 . The second approach.03 _ P=0.01 LCL=0 0.06 Proportion 0.0221) ni .07 0. calculate the exact control limits by p ± 3 p (1 − p ) / ni = 0.

and carefully examine points near the control limits. MTB > Stat > Control Charts > Attributes Charts > P P Chart of Second Visit Required (Ex6-21Sec) Limits based on average sample size (n=187) 0.01 0.Chapter 6 Exercise Solutions 6-22.03 _ P=0. Use a sample size of n = 187.02 0. however MINITAB accepts only integer values for n.02219 0.5.05451 0.04 0. MTB > Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Ex6-21Req Variable Ex6-21Req N 20 Mean 187. 6-22 .00 LCL=0 2 4 6 8 10 12 Sample 14 16 18 20 Process is in statistical control.5 Average sample size is 187.05 Proportion 0.06 UCL=0.

zi = ( pˆ i − p ) p (1 − p ) ni = ( pˆ i − 0.0221) 0. 6-23 .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.

0581.0533.05 0. UCL250 = 0. UCL200 = 0. CL = 0.Chapter 6 Exercise Solutions 6-24.00 2 4 6 8 10 12 Week 14 16 18 20 6-24 .01 0.07 Variable Ex6-24pi Ex6-24n100 Ex6-24n150 Ex6-24n200 Ex6-24n250 Ex6-24C L Ex6-24LCL 0.02 0. UCL150 = 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 0.0662.04 0.06 0.0221. LCL = 0 UCL100 = 0.03 0.

0895 = 0.05 < 0.05(1 − 0. LCL = 0.0895 where POI(·) is the cumulative Poisson distribution. The np chart is inappropriate for varying sample sizes because the centerline (process center) would change with each ni.8349 (b) CL = np = 400(0.12) + POI(7. use Poisson approximation to binomial.0399. UCL = 0. n = 100 ⎛ 1− p ⎞ 2 n>⎜ ⎟L ⎝ p ⎠ ⎛ 1 − 0.01 ⎠ > 891 ≥ 892 6-26. p = CL = 0.01 ⎞ 2 >⎜ ⎟3 ⎝ 0. 6-25 .8349 np (1 − p ) = 20 − 2.1. LCL = 0. n = 400.Chapter 6 Exercise Solutions 6-25.8349 20(1 − 0.12) = 1 − 1.0109) L = 2. UCL = 0.05) = 20 UCL = np + 2.64 |12} = 1 − POI(32. Pr{detect shift to 0.05) = 32.0191 (a) 0.8349 20(1 − 0.0000 + 0.05) 400 = 0.64 (c) n = 400 is large and p = 0. 6-27.03 on 1st sample} = 1 − Pr{not detect} = 1− β = 1 − [Pr{D < UCL | λ} − Pr{D ≤ LCL | λ}] = 1 − Pr{D < 32.0809 = 0.05) = 7.0500.36 |12} + Pr{D ≤ 7. p = CL = 0.36 LCL = np − 2.8349 np(1 − p) = 20 + 2.05 + L(0.01.05 + L 0.0809.

01 ⎞ 2 >⎜ ⎟2 ⎝ 0.15) 400 ⎟ ⎜ 0.Chapter 6 Exercise Solutions 6-28. 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 ⎟⎟ ⎝ ⎠ ⎝ ⎠ ⎞ ⎛ ⎞ ⎛ 0.0500 + L 0.01.15(1 − 0.01 ⎠ > 396 ≥ 397 (b) δ = 0.0962 − 0.99865 6-29.04 – 0.03 ⎠ 6-26 .15) 400 ⎟ ⎝ ⎠ ⎝ ⎠ = 1 − Φ (−3.15 0.15) = 60 > 15.01)(1 − 0.00) + Φ (−8. L = 2 (a) ⎛ 1− p ⎞ 2 n>⎜ ⎟L ⎝ p ⎠ ⎛ 1 − 0.0962 = 0.19) = 1 − 0.15(1 − 0.05) 400 L = 4.01 = 0.05(1 − 0.15 = 1− Φ ⎜ +Φ⎜ ⎟ ⎟ ⎜ 0.00135 + 0. (a) UCL = p + L p (1 − p ) n 0.01) = 44 ⎝δ ⎠ ⎝ 0.03 2 2 ⎛L⎞ ⎛ 2 ⎞ n = ⎜ ⎟ p(1 − p) = ⎜ ⎟ (0.24 (b) p = 15. p = 0. use normal approximation to binomial.000 = 0. λ = np = 400(0.0038 − 0.

6) = 0.005) |100(0.742 where POI(⋅) is the cumulative Poisson distribution. (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(0.948 = 0.070 where POI(⋅) is the cumulative Poisson distribution.04)} + 1 − Pr{D ≤ 100(0. 6-27 .Chapter 6 Exercise Solutions 6-30. 6) − POI(0.06)} − Pr{D ≤ 100(0.04)} = POI(0.744 − 0. 4) + 1 − POI(7.002 = 0. (b) Pr{type II error} =β = Pr{D < nUCL | np} − Pr{D ≤ nLCL | np} = Pr{D < 100(0. 4) = 0.075) |100(0.0050) |100(0.018 + 1 − 0.06) = POI(7.075) |100(0.

005 0.9000 0.0698 0.0000 0 0.0000 Pr{D<=0.5978 0.070 = 14.5|np} 1.2202 0.5.15 0.0000 1.5 1.7440 0.15 0.01 0.1 0.8000 0.2 0.07 0.0009 0. UCL=7.9306 0.0003 0. CL=4. LCL=0.0000 0.5000 0.02 0.xls : worksheet Ex6-30 p 0 0.3679 0.2000 0.5 1 2 3 4 5 6 7 8 9 10 12.3000 0.29 ≅ 15 from part (b).0742: ARL1 = 1/(1 –β) = 1/(1 – 0.3239 0. α = 0.0008 0.25 np 0 0.0000 1.8666 0. β = 0.0000 0.7000 Beta 0.03 0.08 0.0000 0.9383 0.0067 0.125 0.6000 0.2 0.05 0.5 |100 p} − Pr{D ≤ 0.742) = 3.04 0.5 |100 p} Excel : workbook Chap06.5987 0.9881 0.7415 0.06 0.1353 0.8636 0.1000 0.0000 beta 0.861 ≅ 4 6-28 .0000 0.0000 0.0001 0.3935 0.5|np} 1.5 15 20 25 Pr{D<7.1 0.0698 0.4526 0.0498 0.3238 0.09 0.070: ARL0 = 1/α = 1/0.9989 0.Chapter 6 Exercise Solutions 6-30 continued (c) β = Pr{D < nUCL | np} − Pr{D ≤ nLCL | np} = Pr{D < 100(0.3 p (d) from part (a).0025 0.2202 0.0183 0.0008 0.0050) |100 p} = Pr{D < 7.0180 0.25 0.6321 0.0750) |100 p} − Pr{D ≤ 100(0.0000 0.0000 0.4000 0.8599 0.05 0.0180 0.4530 0.9489 0.0000 0.0000 OC Curve for n=100.6065 0.

01 0.07 UCL=0.062 Proportion 0.0191 6-32. One point more than 3.02 0.09 1 0.02 − 3 0. n = 100.Chapter 6 Exercise Solutions 6-31.00 LCL=0 1 2 3 4 5 6 Sample 7 8 9 10 Test Results for P Chart of Ex6-31Num TEST 1.02(1 − 0.03 _ P=0.05 0.062 LCL = p − 3 p (1 − p ) n = 0.02 (a) UCL = p + 3 p (1 − p ) n = 0.02) 100 = 0. Test Failed at points: 4 Sample 4 exceeds the upper control limit.038 and σˆ p = 0.02 0. LCL = np − k np (1 − p ) > 0 np > k np (1 − p ) ⎛ 1− p ⎞ n > k2 ⎜ ⎟ ⎝ p ⎠ 6-29 . p = 0.00 standard deviations from center line.06 0.02) 100 ⇒ 0 (b) MTB > Stat > Control Charts > Attributes Charts > P P Chart of Number Nonconforming (Ex6-31Num) 0. p = 0.04 0.02(1 − 0.08 0.02 + 3 0.

505(1 − 0.213 LCL = np − 3 np (1 − p ) = 2. ∑ D = 50.505 UCL = np + 3 np (1 − p ) = 2.505 − 4.5 2 1 LCL=0 0 2 4 6 8 10 12 Sample 14 16 18 20 The process is in control. 6-30 . results are the same as for the p chart. p = 0.204 7 Sample Count 6 5 4 3 __ NP=2.505 + 3 2.0167) = 2.708 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > NP NP Chart of Numer of Nonconforming Units (Ex6-4Num) 8 UCL=7.0167) = 7. m = 20.Chapter 6 Exercise Solutions 6-33.0167 CL = np = 150(0. n = 150.

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 1.3 __ NP=307.77 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. 12. 20 Like the p control chart.1228) = 257. 2. 11.23 LCL = np − 3 np (1 − p ) = 307 − 3 307(1 − 0. 6-31 . CL = np = 2500(0.1228) = 307 UCL = np + 3 np (1 − p ) = 307 + 3 307(1 − 0. 15. Test Failed at points: 1. 5. many subgroups are out of control (11 of 20).1 300 LCL=257.Chapter 6 Exercise Solutions 6-34.1228) = 356. 16. 19. One point more than 3. 3. 17. indicating that this data should not be used to establish control limits for future production.00 standard deviations from center line.

040 0 -1 -2 LCL=-2.Chapter 6 Exercise Solutions 6-35.494 Individual Value 2 1 _ X=0.06) 0.414 -3 1 2 3 4 5 6 Observation 7 8 9 10 The process is in control.0564 / ni MTB > Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Standardized Fraction Nonconforming (Ex6-35zi) 3 UCL=2.06) / ni = ( pˆ i − 0.06 zi = ( pˆ i − 0.06) 0. p = 0. 6-32 . results are the same as for the p chart.06(1 − 0.

Test Failed at points: 13 No.36 UCL = c + 3 c = 2.36 + 3 2.Chapter 6 Exercise Solutions 6-36. The plate process does not seem to be in statistical control. One point more than 3. 6-33 .97 LCL = c − 3 c = 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 = 6.36 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. CL = c = 2.00 standard deviations from center line.969 6 5 4 3 _ C=2.

2622] [0.Chapter 6 Exercise Solutions 6-37.0 0.1527.8 _ U=0.7007 UCLi = u + 3 u ni = 0.153 0.0 2 4 6 8 10 12 Sample 14 16 18 20 Tests performed with unequal sample sizes 6-34 . 1.2926] [0.2133] MTB > Stat > Control Charts > Attributes Charts > U U Chart of Imperfections in Paper Rolls (Ex6-37Imp) 1.7007 / ni ni 18 20 21 22 24 [LCLi. 1.2 LCL=0.249 Sample Count Per Unit 1.2 1.1088. UCLi] [0.6 0.701 0.1881.7007 / ni LCLi = u − 3 u ni = 0.4 0. 1.7007 − 3 0.2361] [0. 1. CL = u = 0.4 UCL=1.1392.1653. 1.7007 + 3 0.2487] [0.

0 2 4 6 8 10 12 Sample 14 16 18 20 6-35 .2 LCL=0.55.2547 LCL = u − 3 u n = 0.7007 / 20.550 Average sample size is 20.2 1. CL = u = 0.7007 / 20. n = 20. however MINITAB accepts only integer values for n.151 0.1467 MTB > Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Ex6-37Rol Variable Ex6-37Rol N 20 Mean 20.4 0.55 UCL = u + 3 u n = 0.289 Sample Count Per Unit 1. MTB > Stat > Control Charts > Attributes Charts > U U Chart of Imperfections in Paper Rolls (Ex6-37Imp) with average sample size n=20 1.55 = 1.8 _ U=0.7007. and carefully examine points near the control limits.55 = 0.6 0.72 0. Use a sample size of n = 20.0 0.7007 − 3 0.7007 + 3 0.4 UCL=1.Chapter 6 Exercise Solutions 6-38.

zi = (ui − u ) u ni = (ui − 0.Chapter 6 Exercise Solutions 6-39.004 0 -1 LCL=-1.898 Individual Value 1 _ X=-0.7007 / ni MTB > Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Standardized Paper Roll Imperfections (Ex6-39zi) 2 UCL=1.7007) 0.906 -2 2 4 6 8 10 12 Observation 14 16 18 20 6-36 .

5 = 5.5 + 3 1. Use these limits to control future production.5 1 0 LCL=0 2 4 6 8 10 Sample 12 14 16 18 Process is in statistical control.5 UCL = c + 3 c = 1. c chart based on # of nonconformities per cassette deck CL = c = 1. 6-37 .17 LCL ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > C C Chart of Cassette Deck Nonconformities (Ex6-40Num) UCL=5.174 5 Sample Count 4 3 2 _ C=1.Chapter 6 Exercise Solutions 6-40.

59 + 3 8. One point more than 3.Chapter 6 Exercise Solutions 6-41.59. 11.59 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 1.59 = 17.384.00 standard deviations from center line. LCL = c − 3 c = 8.59 − 3 8.38 15 10 _ C=8. UCL = c + 3 c = 8. CL = c = 8. Test Failed at points: 10.59 ⇒ 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. 22 6-38 .

11. then re-calculate the control limits.17 + 3 6. 22 6-39 . One point more than 3.17 = 13. CL = c = 6. 22 excluded from calculations 25 1 Sample Count 20 1 1 1 15 UCL=13.17. Exclude subgroups 10. three subgroups exceed the UCL. LCL ⇒ 0 C Chart of Number of Nonconformities (Ex6-41Num) Samples 10.Chapter 6 Exercise Solutions 6-41 continued Process is not in statistical control. UCL = c + 3 c = 6. 11 and 22.00 standard deviations from center line.62. 15. 15. Test Failed at points: 10.62 10 _ 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. 11. Subgroup 15 will then be out of control and should also be excluded.

42 /1 = 27. 6-40 .35 LCL = u − 3 u n = 6 − 3 6 1 ⇒ 0 6-43. (b) The sample is n =1 new inspection units.65 The plot point.43 = 27. (a) The new inspection unit is n = 2500/1000 = 2.64 The plot point. A u chart of average nonconformities per inspection unit is appropriate.42 + 3 15.5 of the old unit.17) = 15.42 − 3 15. CL = nc = 2.00 CL = u = total inspection units (18 / 4) UCL = u + 3 u n = 6 + 3 6 1 = 13. is the total number of nonconformities found while inspecting a sample 2500m in length. c . CL = nc = 4(1.42 /1 = 3. total nonconformities 27 = = 6. and since n = 1.43 − 3 15.5(6.5) = 6 UCL = nc + 3 nc = 6 + 3 6 = 13.21 LCL = nc − 3 nc = 15.20 LCL = u − 3 u n = 15. A c chart of the total number of nonconformities per inspection unit is appropriate. (a) The new inspection unit is n = 4 cassette decks. total nonconformities 111 = = 15.42 CL = u = total inspection units (18 × 1000) / 2500 UCL = u + 3 u n = 15. A c chart of the total number of nonconformities per inspection unit is appropriate.43 UCL = nc + 3 nc = 15. is the average number of nonconformities found in 2500m.Chapter 6 Exercise Solutions 6-42. u .43 = 3. A u chart of average nonconformities per inspection unit is appropriate.43 + 3 15. this is the same as the total number of nonconformities.35 LCL = nc − 3 nc = 6 − 3 6 ⇒ 0 (b) The sample is n =1 new inspection units.

the same control limits may be used for future production with the new sample size.422 UCL = u + 3 u n = 0.2 0.422 + 3 0.0 LCL=0 2 4 6 8 10 Sample 12 14 16 (b) The process is in statistical control.2 1. However.422 0. with n = 4 transmissions in each inspection. (If this was a c chart for total nonconformities in the sample. CL = u = ∑ ui m = ( ∑ xi / n ) m = (27 / 4) 16 = 6. the control limits would need revision.4 0. (a) A u chart of average number of nonconformities per unit is appropriate.4 UCL=1.Chapter 6 Exercise Solutions 6-44. since this chart was established for average nonconformities per unit.) 6-41 .6 _ U=0.75 16 = 0.211 ⇒ 0 MTB > Stat > Control Charts > Attributes Charts > U U Chart of Manual Transmission Subassemblies (Ex6-44Num) 1.396 LCL = u − 3 u n = 0.422 − 3 0.0 0. (c) The new sample is n = 8/4 = 2 inspection units.8 0.396 Sample Count Per Unit 1.422 4 = 1.422 4 = −0.

0774. c = 16 Pr{x ≤ 21| c = 16} = 0.Chapter 6 Exercise Solutions 6-45. LCL = 10 6-47. (a) CL = c = 9 UCL = c + 3 c = 9 + 3 9 = 18 LCL = c − 3 c = 9 − 3 9 = 0 (b) c = 16. n = 4 CL = u = c / n = 4 / 4 = 1 UCL = u + 3 u n = 1 + 3 1/ 4 = 2. UCL = 21 Pr{x ≤ 10 | c = 16} = 0. 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 .5 LCL = u − 3 u n = 1 − 3 1/ 4 ⇒ 0 6-46. (a) CL = c = 4 UCL = c + 3 c = 4 + 3 4 = 10 LCL = c − 3 c = 4 − 3 4 ⇒ 0 (b) c = 4. Use the cumulative Poisson tables.9108.

96 7.923 0.030 0.4 − 1. As a comparison.062 0.015 0.00.95 u n = 1.6 = 13.05 u n = 1. As a comparison.905 LCL = u + z0.019 0.784 6-43 .972 0.976 for the c chart.020.4 10 = 2. UCL = 13 and LCL = 2.645 1. the normal distribution gives: UCL = u + z0.645 1. c = u × n = 18.055 0.980 and Pr{D < LCL} = 0.20 6-50.00 LCL = c − z0.980 u n = 6 + 2.054 6 3 = 3.020 u n = 6 − 2.954 0. the normal distribution gives: UCL = u + z0.Chapter 6 Exercise Solutions 6-48.96 7.4 + 1.032 0. and LCL = x/n = 7/10 = 0.4 10 = 0.6 − 1. Using the cumulative Poisson distribution with c = u n = 1.6 + 1. and LCL = x/n = 9/3 = 3.952 UCL = x/n = 20/10 = 2. u chart with u = 6.054 6 3 = 8.095 6-49.4(10) = 14: x 7 8 19 20 Pr{D ≤ x | c = 14} 0.0 and n = 3. the normal distribution gives UCL = c + z0.6} 0. Using the cumulative Poisson distribution: x 2 3 12 13 Pr{D ≤ x | c = 7.025 c = 7. Find limits such that Pr{D ≤ UCL} = 0. As a comparison.975 c = 7.983 UCL = x/n = 27/3 = 9. From the cumulative Poisson tables: x 9 10 26 27 Pr{D ≤ x | c = 18} 0.6 = 2.016 LCL = u + z0.70.

0} = 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. Pr{x ≤ 6 | c = 2.0.005)(0.Chapter 6 Exercise Solutions 6-51. (b) Pr{two consecutive out-of-control points} = (0.94 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. 6-52.00003 6-44 . (a) From the cumulative Poisson table. u chart with control limits based on each sample size: u = 7. So set UCL = 6.995.005) = 0. UCLi = 7 + 3 7 / ni .

22 |10.25) = 10.25 = 12.25 UCL = c + 3 c = 5.997 ] = 0. the normal distribution gives UCL = c + z0.5 = 3. c = 850 /100 = 8. (b) Let new inspection unit n = 2 water heaters CL = nc = 2(5.030 13 0.5) ] = 0.5 + 1.10.5 = 20. (a) Plot the number of nonconformities per water heater on a c chart.78 |10.949 14 0.5) + [1 − POI(20.02 6-54.5 + 3 10.78 (c) Pr{type I error} = Pr{D < LCL | c} + Pr{D > UCL | c} = Pr{D < 0.22 LCL = nc − 3 nc = 10.03 c = 8.973 LCL = 3 and UCL = 13.97 c = 8.5 = 13.5 − 3 10. For comparison.88 8. A c chart with one inspection unit equal to 50 manufacturing units is appropriate.5} + [1 − Pr{D ≤ 20.10.25 + 3 5. From the cumulative Poisson distribution: x Pr{D ≤ x | c = 8.5} 3 0.5 .12 LCL ⇒ 0 Plot the results after inspection of each water heater.5}] = POI(0. CL = c = ∑ D m = 924 /176 = 5.003 6-45 .5 = 0.Chapter 6 Exercise Solutions 6-53.000 + [1 − 0.5 UCL = nc + 3 nc = 10. approximately 8/day.88 8.5 − 1.98 LCL = c + z0.

43 0. Use the cumulative Poisson distribution to determine the UCL: MTB : worksheet Chap06.017 where POI(⋅) is the cumulative Poisson distribution.24 0.533 (a) 3-sigma limits are c ± 3 c = 0.89 0.00 1.723] (b) α = Pr{D < LCL | c} + Pr{D > UCL | c} = Pr{D < 0 | 0. 2. Use a c chart for nonconformities with an inspection unit n = 1 refrigerator.533}] = 0 + [1 − POI(2.0 average number of nonconformities/unit. u = 4. Desire α = 0. when in fact it is. c = 16 / 30 = 0. 0.mtw Ex6-55X 0 1 2 3 4 5 6 7 8 9 10 11 Ex6-55alpha 0.00 An UCL = 9 will give a probability of 0.02 0.99.533 ± 3 0.79 0.533) ] = 1 − 0.533} + [1 − Pr{D ≤ 2.99 of concluding the process is in control.09 0.95 0.99 1.Chapter 6 Exercise Solutions 6-55.983 = 0. 6-56.533 = [0.72 | 0.63 0. ∑ Di = 16 in 30 refrigerators. 6-46 .98 0.

1004 where POI(⋅) is the cumulative Poisson distribution.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) = 0. 0.0} − Pr{D ≤ 0 | 2.72 | 2.135 = 0. 2) − POI(0.372 ≈ 2 1 − β 1 − 0.533}] = 0 + [1 − POI(1. 2) − POI(0.993 | 2} − Pr{D ≤ 0 | 2} = POI(1.271 6-47 .541 6-57.1.0} = POI(2. 2) = 0.533 (a) c ± 2 c = 0.271 where POI(⋅) is the cumulative Poisson distribution. (d) ARL1 = 1 1 = = 2.993] (b) α = Pr{D < LCL | c } + Pr{D > UCL | c } = Pr{D < 0 | 0.533} + [1 − Pr{D ≤ 1.5414 where POI(⋅) is the cumulative Poisson distribution.533 + 2 0.406 − 0. (d) ARL1 = 1 1 = = 1.18 ≈ 2 1 − β 1 − 0. (c) β = Pr{D < UCL | c} − Pr{D ≤ LCL | c} = Pr{D < 1.6767 − 0.8996 = 0.993 | 0.1353 = 0.533)] = 1 − 0.533 = [0. c = 0.

6-60.708 LCL ⇒ 0 6-59.75(6) = 4. 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(10.997 = 0.5 + 3 4.5 = 10. u = 0. 1 inspection unit = 10 radios.5 UCL = c + 3 c = 4. c: nonconformities per unit.86 LCL ⇒ 0 6-61.Chapter 6 Exercise Solutions 6-58.5 average nonconformities/radio CL = c = u × n = 0.75 nonconformities/clock CL = c = u × n = 0.003 where POI(⋅) is the cumulative Poisson distribution. 4)] = 1 − 0. 1 inspection unit = 6 clocks. u = 0.5(10) = 5 UCL = c + 3 c = 5 + 3 5 = 11. L: sigma control limits nc − L nc > 0 nc > L nc n > L2 c 6-48 .

04 28 1. (b) Probability Plot of Transformed "Days-between-Homicides" (Ex6-62t27) Normal .95% CI 99 Mean StDev N AD P-Value 95 90 1. 6-49 .95% CI 99 Mean StDev N AD P-Value 95 90 12.2777th 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 0.238 0.25 12. (a) MTB > Graphs > Probability Plot > Single Probability Plot of Days-Between-Homicides (Ex6-62Bet) Normal . indicating that the normal distribution assumption is not reasonable.806 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.5635 28 0.Chapter 6 Exercise Solutions 6-62.572 <0.

95% CI 99 Mean StDev N AD P-Value 95 90 1.0 _ X=1.5 2.5 1.Chapter 6 Exercise Solutions 6-62 continued (c) Probability Plot of Transformed "Days-betwee-Homicides" (Ex6-62t25) Normal .4789 28 0.0 0. (d) MTB > Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Transformed Homicide Data (0.246 0.2777 root) (Ex6-62t27) 3. It is not very different from the transformed data in (b).807 Percent 80 70 60 50 40 30 20 10 5 1 0.5 1.806 1.0 3 6 9 12 15 18 Observation 21 24 27 6-50 .223 0.5 UCL=3.5 The 0.695 0.5 LCL=0.0 1.366 3.0 Individual Value 2.0 0.5 3.25th root transformation makes the data more closely resemble a sample from a normal distribution.0 Ex6-62t25 2.0 3.5 2.

which affects the rate at which homicides occur. with an identical pattern of points relative to the UCL. wrong package labeling). There are endless possibilities for collection of attributes data from nonmanufacturing processes.025 3. meaning that the days-between-homicides is approximately constant. say in population. etc.0 0. Or consider an accounting firm—errors in statements. mean and LCL. packaged incorrectly (wrong material. wrong part labeling.0 Individual Value 2. One could track the number of orders filled incorrectly (wrong parts.). too few/many parts.695 1.0 _ X=1. policy. the mean time between may get longer (or shorter) with plot points above the upper (or below the lower) control limit. law. (hopefully caught internally with a verification step). If a change is made.365 0. There is no difference in interpretation. Consider a product distribution center (or any warehouse) with processes for filling and shipping orders.25 root) (Ex6-62t25) UCL=3. 6-51 .5 1. invoiced incorrectly. errors in tax preparation.Chapter 6 Exercise Solutions 6-62 continued (e) I Chart of Transformed Homicide Data (0.0 3 6 9 12 15 18 Observation 21 24 27 Both Individuals charts are similar. 6-63. (f) The “process” is stable. workforce. etc.5 2.5 LCL=0.. etc.

86 20 15 LCL=15. if you’re looking for data on time-between-events that must be obtained from external sources (for example.Chapter 6 Exercise Solutions 6-64.77 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. representing an identical “area of opportunity” for each “sample”. One point more than 3. The “process characteristic” to be controlled is the rate of CAT scans. There are ways to address these difficulties. 6-52 . time-to-field failures). Collection of “start” time data may be facilitated by serializing or date coding product. However.94 30 25 _ U=25. MTB > Stat > Control Charts > Attributes Charts > U U Chart of CAT Scans (Ex6-65NSCANB) Sample Count Per Unit 40 1 35 UCL=35. it may be hard to determine with sufficient accuracy—both the “start” and the “end”. A u chart which monitors the average number of CAT scans per NYRSB is appropriate. it can usually be obtained reliably and consistently. Test Failed at points: 15 The rate of monthly CAT scans is out of control.00 standard deviations from center line. The variable NYRSB can be thought of as an “inspection unit”. the conditions of use and the definition of “failure” may not be consistently applied. Also. If time-between-events data (say failure time) is being sought for internally generated data. 6-65☺.

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 . (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=2476.5 2400 _ U=2303.0 2300 2200 LCL=2129. The “process characteristic” to be controlled is the rate of office visits. The variable NYRSE can be thought of as an “inspection unit”. A u chart which monitors the average number of office visits per NYRSB is appropriate.Chapter 6 Exercise Solutions 6-66☺. representing an identical “area of opportunity” for each “sample”.

10.0 2200 2100 LCL=2141. 6-54 . 14. 15 The phase 2 data appears to have shifted up from phase 1. Test Failed at points: 9.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. 13. The 2nd phase is not in statistical control relative to the 1st phase.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 1.0 2400 2300 _ U=2303.00 standard deviations from center line. 11. One point more than 3. 12.

are in statistical control.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. 6-55 .5 2700 _ U=2623.6 2400 9 10 11 12 Sample 13 14 15 Tests performed with unequal sample sizes The Phase 2 data. separated from the Phase 1 data.5 2600 2500 LCL=2450.

36 Cˆ pl = 3σˆ 3(1. σˆ = R d 2 = 1.965. Cˆ pu = 1.010) µˆ − LSL 74.13 7-2.65 − 20 = = 2. samples 12 and 15 are out of control.5. µˆ = x = 33. n = 5. R = 4. 7-1. µˆ = x = 74.001 = = 1.93) µˆ − LSL 33.023 2.65.010) ( ) Cˆ pk = min Cˆ pl . Cˆ pu = 1.17 Cˆ p 6σˆ 6(0.001.Chapter 7 Exercise Solutions Note: Several exercises in this chapter differ from those in the 4th edition.93) USL − µˆ 40 − 33.10 Cˆ pu = 3σˆ 3(1.000 ± 0.326 = 0. LSL = 20 USL − LSL 40 − 20 = = 1.13 Cˆ pu = 3σˆ 3(0. R = 0.65 = = 1. New exercises are denoted with an “☺”.93 USL = 40.010) USL − µˆ 74.965 = = 1.965 = = 1. σˆ = R d 2 = 0.035] USL − LSL 74.73 Cˆ p 6σˆ 6(1. A second exercise number in parentheses indicates that the exercise number has changed.035 − 74. and the new process parameters are used in the process capability analysis.20 Cˆ pl = 3σˆ 3(0. An “*” indicates that the description has changed.93) ( ) Cˆ pk = min Cˆ pl .010 SL = 74.035 = [73. 74.001 − 73.035 − 73. In Exercise 5-1.023.10 7-1 .

4128) 2 V= Since Cpm is greater than 4/3. Cˆ ) = 4.04 Cˆ pkm = Cˆ 1.375 − 0 = = 3.41 σˆ 3.Chapter 7 Exercise Solutions 7-3. Rx = 6.48 Cˆ pm = = = 1.48 Cˆ p = 6σˆ x 6(3. σˆ x = R d 2 = 6.04) Cˆ = min(Cˆ . USL x − µˆ 50 − 10.54 pk = = 0. T − x 0 − 10.4128 S 3.059 = 3. However. indicating that the process is not centered and is not achieving potential capability.25.34 pk pu pl This is an extremely capable process.34 Cˆ pu = 3σˆ x 3(3. µˆ = x = 10.04 Cˆ p 5.412 1 + ξˆ 2 7-2 .54 2 1+V 1 + (−3.62 Cˆ pl = 3σˆ x 3(3.04 USL x = [(350 + 5) − 350] ×10 = 50.43 1 + 3.04) µˆ − LSL x 10. ξˆ = µˆ − T 10.04) The process produces product that uses approximately 18% of the total specification band.375 = = 4. LSL x = [(350 − 5) − 350] × 10 = −50 xi = (obsi − 350) ×10 USL x − LSL x 50 − (−50) = = 5. this PCR does not tell where the mean is located within the specification band.375 = = −3.25 2. the mean µ lies within approximately the middle fourth of the specification band. with an estimated percent defective much less than 1 ppb.375. Note that the Cpk is less than Cp.375 − (−50) = = 6.

01 + 0. failing to meet the minimally acceptable definition of capable Cpk ≥ 1.01 2 2 ˆ 1 0. Cˆ ) = 1.00109 − 0 = = 0.22 p ˆ C pm = = = 1. USL − µˆ 0.00109.399) 2 V= Since Cpm is greater than 1.00109 Cˆ pu = = = 1.00273 Cˆ 1.35 3σˆ 3(0. the mean µ lies within approximately the middle third of the specification band.00273) µˆ − LSL 0.399 S 0.09 pk pl pu This process is not considered capable.01) Cˆ pl = = = 1.01 Cˆ p = = = 1. tolerances: 0 ± 0.00273 .00635.22 6σˆ 6(0.Chapter 7 Exercise Solutions 7-4.00273 Cˆ pkm = Cˆ 1. R = 0. x = 0.09 pk = = 1. n = 5.09 3σˆ 3(0.33 T − x 0 − 0.13 1+V 2 1 + (−0.01 USL − LSL 0.00273) The process produces product that uses approximately 82% of the total specification band.00109 = = −0. ξˆ = µˆ − T 0.00109 − (−0.399 + 1+ ξ 7-3 .01 − 0. σˆ x = 0.00273) Cˆ = min(Cˆ .399 σˆ 0.

117 ⎭ ⎣ 1.000000] = 0.49 3σˆ x 3(1.999996] = 0.476) ] = 0.117 ⎭ ⎣ 1.117 ⎭⎥⎦ ⎩ ⎩ = Φ (−8. σˆ x = s c4 = 1.9400 = 1.Chapter 7 Exercise Solutions 7-5.98 Potential: Cˆ p = 6σˆ 6(1.000000 + [1 − 1.953) ] = 0.000004 85 − 95 ⎫ ⎡ ⎧ ⎧ 105 − 95 ⎫⎤ pˆ Potential = Pr ⎨ z < ⎬ + ⎢1 − Pr ⎨ z ≤ ⎬ 1.49 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 ≤ ⎬ 1.117) (b) µˆ − LSL x 100 − (95 − 10) Cˆ pl = = = 4. Cˆ ) = 1.000000 7-4 .117) USL x − µˆ (95 + 10) − 100 Actual: Cˆ pu = = = 1. s = 1.429) + [1 − Φ (4.117 ⎭⎥⎦ ⎩ ⎩ = Φ (−13.48 3σˆ x 3(1.117 (a) USL − LSL (95 + 10) − (95 − 10) = = 2.05. µˆ = x = 100.0000 + [1 − 0.05 0.117) Cˆ = min(Cˆ .953) + [1 − Φ (8.

0000013 = 0.059 = 1. µˆ = x = 199.2941) ] = 0. n = 4.Chapter 7 Exercise Solutions 7-6☺.70 USL = 200 + 8 = 208.37 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 ≤ ⎬ 1.76 3σˆ 3(1.37 Actual: Cˆ pl = 3σˆ 3(1.0000191 If the process mean could be centered at the specification target.1176) + [1 − Φ (5.0000026 7-5 . (b) USL − µˆ 208 − 199 Cˆ pu = = = 1. Cˆ ) = 1.70 ⎭ ⎩ = 2 × 0. LSL = 200 – 8 = 192 (a) USL − LSL 208 − 192 = = 1.5 2. the fraction nonconforming would be: ⎧ 192 − 200 ⎫ pˆ Potential = 2 × Pr ⎨ z < ⎬ 1. σˆ x = R d 2 = 3.0000191 + [1 − 1] = 0.70) The process produces product that uses approximately 64% of the total specification band. R = 3.70) µˆ − LSL 199 − 192 = = 1.57 Potential: Cˆ p = 6σˆ 6(1.70 ⎭⎥⎦ ⎩ ⎩ = Φ (−4.70) Cˆ = min(Cˆ .5.70 ⎭ ⎣ 1.

7 − 35 = = 0.0169634 + [1 − 0.70 1+V 1 + (−0. LSL = 40 – 5 = 35 (a) USL − LSL 45 − 35 = = 0.5. Cpk. n = 2.5 1.7 Cˆ pu = = = 0. Cˆ ) = 0.216 ⎭ ⎣ 2.128 = 2. R = 2.7 ⎫⎤ ⎧ ⎧ = Pr ⎨ z < ⎬ + ⎢1 − Pr ⎨ z ≤ ⎬ 2.75 = = = 0. σˆ x = R d 2 = 2.7 ⎫ ⎡ 45 − 39.991615] = 0.12094) + [1 − Φ (2.216) µˆ − LSL 39.71 pk pl pu (c) V= Cˆ pm x − T 39.216) Cˆ = min(Cˆ .216 ⎭⎥⎦ ⎩ ⎩ = Φ (−2.216 USL = 40 + 5 = 45.7 − 40 = = −0.39170)] = 0.7.Chapter 7 Exercise Solutions 7-7☺.216) (b) USL − µˆ 45 − 39.71 Actual: Cˆ pl = 3σˆ 3(2.80 3σˆ 3(2.75 Potential: Cˆ p = 6σˆ 6(2.71 = 0. Cpm. 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 − 39.74 2 1+V 1 + (−0.135 s 2.216 Cˆ p 0. and Cpkm indicate that the process mean is very close to the specification target.135) 2 The closeness of estimates for Cp. µˆ = x = 39.025348 7-6 .135) 2 Cˆ pkm = Cˆ pk = 0.

756) = 0.13 ⎭ ⎩ ⎩ = Φ (−3.756) + 1 − Φ (3.13) (b) µˆ − LSL 75 − (80 − 8) Cˆ pl = = = 0.13 (a) USL − LSL 2(8) Potential: Cˆ p = = = 1. the fraction nonconforming would be: 35 − 40 ⎫ ⎧ pˆ Potential = 2 × Pr ⎨ z < ⎬ 2.02382 7-8 (7-6).26} = 2 × 0. S = 2. µˆ = 75.25 6σˆ 6(2.47 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.Chapter 7 Exercise Solutions 7-7 (d) continued If the process mean could be centered at the specification target. σˆ = Sˆ c4 = 2 0.216 ⎭ ⎩ = 2 × Pr{z < −2.000172 7-7 .9400 = 2.13) Cˆ = min(Cˆ .000086 + 1 − 0.03 Actual: Cˆ pu = 3σˆ 3(2.13) USL − µˆ 80 + 8 − 75 = = 2.47 3σˆ 3(2. Cˆ ) = 0.01191 = 0.13 ⎭ 2.999914 = 0.

699 3σˆ x 3(1.064) µˆ − LSL x 105 − (100 − 10) Cˆ pl = x = = 4. σˆ A = s A c4 = 3 0.00174 Process B µˆ = xB = 105.9400 = 3.045 Cˆ p = 6σˆ 6(3.045 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. σˆ B = sB c4 = 1 0. sB = 1.191) µˆ − LSL x 100 − (100 − 10) Cˆ pl = = = 1.191) USL x − µˆ (100 + 10) − 100 Cˆ pu = = = 1.045 3σˆ x 3(3.566 7-8 .191 Cˆ p 1. Assume n = 5 Process A µˆ = x A = 100.064 Cˆ p = USL − LSL 6σˆ = (100 + 10) − (100 − 10) = 3.99913 = 0.13) = 0.064) Cˆ pk = min(Cˆ pl .133 6(1.191 ⎭ 3.13) + 1 − Φ (3.191 USL − LSL (100 + 10) − (100 − 10) = = 1.Chapter 7 Exercise Solutions 7-9 (7-7).191 ⎭ ⎩ ⎩ = Φ (−3.566 Cˆ pu = 3σˆ x 3(1.00087 + 1 − 0.9400 = 1. s A = 3. Cˆ ) = 1.045 = = = 1.045 pk pl V= Cˆ pm pu x − T 100 − 100 = =0 s 3.064) USL x − µˆ x (100 + 10) − 105 = = 1.045 3σˆ x 3(3.191) Cˆ = min(Cˆ . Cˆ pu ) = 1.

7-10 (7-8).652 1+V 2 1 + (−4. 7-9 .191) 2 = 14.000001 Prefer to use Process B with estimated process fallout of 0.271 Process B: µˆ B = 20(105) = 2100.000000 + 1 − 0.064 Cˆ p 3.566 = Cˆ pk . A = 1.000001 instead of Process A with estimated fallout 0.699) = 0. σˆ B = 20σˆ 2 = 20(1. σˆ A = 20σˆ 2 = 20(3.098) + 1 − Φ (4.064 ⎭ ⎩ ⎩ = Φ (−14.699) 2 90 − 105 ⎫ ⎧ ⎧ 110 − 105 ⎫ pˆ = Pr ⎨ z < ⎬ + 1 − Pr ⎨ z ≤ ⎬ 1.133 Cˆ pm = = = 0.999999 = 0. B indicates that more parts from Process B are within ( ) ( ) specification than from Process A.699 s 1.758 Process B will result in fewer defective assemblies. For the parts Cˆ pk .Chapter 7 Exercise Solutions 7-9 continued x − T 100 − 105 V= = = −4.064 ⎭ 1.045 < 1. Process A: µˆ A = 20(100) = 2000.064) 2 = 4.001726.

0225 σˆ ⎭ ⎩ ⎩ ⎭ 7-10 .9975 − 0.289105 0.9975.0200 σˆ = p84 − p50 = 1.19 C pl = 3σˆ 3(0.0200 − 0. x ≈ p50 = 0.000 Ex7-9Wt 1.556) = 0.985 = = 0.02167 15 0. LSL = 0.323 0.025 1.0225) = 0.9975 = 0.950 0.0225) LSL − µˆ ⎫ 0.9975 ⎫ ⎧ ⎧ pˆ = Pr ⎨ z < ⎬ = Pr ⎨ z < ⎬ = Φ (−0.975 1.492 Percent 80 70 60 50 40 30 20 10 5 1 0.Chapter 7 Exercise Solutions 7-11 (7-9).985 − 0.9968 0. p84 = 1.050 A normal probability plot of the 1-kg container weights shows the distribution is close to normal.0225 6σˆ = 6(0.985 kg µˆ − LSL 0.1350 7-12☺. 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.

00905 − 19.05514 7-11 .00 Disk Height. x ≈ p50 = 19.00919) = 0.99 20.00919 6σˆ = 6(0.01 20.) Probability Plot of Disk Height (Ex7-13Ht) Normal 99 95 90 84 Percent 80 Mean StDev N AD P-Value 20.00 0.174 70 60 50 40 30 50 19.009242 25 0.Chapter 7 Exercise Solutions 7-13☺.02 A normal probability plot of computer disk heights 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.00905 20 20.00905 σˆ = p84 − p50 = 20.99986 = 0. mm 20.99986 10 5 1 19.99986 p84 = 20.515 0.98 19.

2 4.2 5 5 17.07) = 24.2 p84 = 17.401 0.2 = 4.27 σˆ = p84 − p50 = 17.27 − 13.27 10 10 15 20 Reimbursement Cycle Time.097 30 0. Days 25 A normal probability plot of reimbursement cycle times 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 = 13.Chapter 7 Exercise Solutions 7-14☺.07 6σˆ = 6(4.42 7-12 .) Probability Plot of Cycle Time (Ex7-14CT) Normal 99 95 90 84 Percent 80 Mean StDev N AD P-Value 13.340 70 60 50 40 30 50 20 1 13.

98 10 90 100 110 Response Time.98 σˆ = p84 − p50 = 110.) Probability Plot of Response Time (Ex7-15Resp) Normal 99 95 90 84 Percent 80 Mean StDev N AD P-Value 98.2) USL − µˆ ⎫ USL − µˆ ⎫ ⎧ ⎧ ⎧ 120 − 98.739) = 1 − 0.041017 7-13 .78 = 12.958983 = 0.463 0.58 C pu = 3σˆ 3(12.78 p84 = 110.78 = = 0.243 70 60 50 40 30 50 20 1 98.78 12.2 (b) USL = 2 hrs = 120 mins USL − µˆ 120 − 98.2) = 73. (a) x ≈ p50 = 98.2 6σˆ = 6(12.2 ⎩ ⎭ ⎩ ⎭ ⎩ ⎭ = 1 − Φ (1.78 ⎫ pˆ = Pr ⎨ z > ⎬ = 1 − Pr ⎨ z < ⎬ = 1 − Pr ⎨ z < ⎬ σˆ σˆ 12.78 5 70 80 110. minutes 120 130 A normal probability plot of response times shows the distribution is close to normal.27 40 0. MTB > Stat > Basic Statistics > Normality Test (Add percentile lines at Y values 50 and 84 to estimate µ and σ.Chapter 7 Exercise Solutions 7-15☺.98 − 98.

69) = 16.96 σˆ = p84 − p50 = 55. x ≈ p50 = 53. MTB > Stat > Basic Statistics > Normality Test (Add percentile lines at Y values 50 and 84 to estimate µ and σ.) Probability Plot of Hardness Data (Ex5-59Har) Normal 99 95 90 84 Percent 80 Mean StDev N AD P-Value 53.96 10 56 58 60 A normal probability plot of hardness data shows the distribution is close to normal.27 5 46 48 50 52 54 Hardness 55.27 = 2.27 p84 = 55.27 2.465 0.712 15 0.Chapter 7 Exercise Solutions 7-16 (7-10).14 7-14 .96 − 53.69 6σˆ = 6(2.217 70 60 50 40 30 50 20 1 53.

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. so it is not appropriate to estimate capability. 7-15 .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.Chapter 7 Exercise Solutions 7-17 (7-11).272 0.1 10 0.

395).Chapter 7 Exercise Solutions 7-18 (7-12).52 χ12−α .24 = 12.49 7-16 . n = 25.9303 = 1. 7-19 (7-13). n −1 0.95 2 = 33. n = 50 Cˆ = 1.n −1 = χ 0.n −1 ≤ C p n −1 1.n −1 = χ 0.40 39.52 ⎠ 1 − α = 0.24 = 39. LSL = 75. 1.5) (b) α = 0.26 ≤ C p 49 The company cannot demonstrate that the PCR exceeds 1.52 p 1 − α = 0.88 α = 0.12 2 1−α .975.36 χ12−α / 2.40 2 χα2/ 2.11 12. USL = 85.49 χ2 Cˆ p 1−α .9303 χ12−α .n −1 χα2/ 2.33 ⎞ χ = 49 ⎜ ⎟ = 37.42 This confidence interval is wide enough that the process may either be capable (ppm = 27) or far from it (ppm ≈ 16.80 ≤ C p ≤ 1.52 ⎝ 1.11 Cˆ p = 6σˆ 6(1.05 χ2 = χ2 1−α / 2.33 at a 95% confidence level.5 (a) USL − LSL 85 − 75 = = 1.025.95.n −1 ˆ ˆ Cp ≤ Cp ≤ Cp n −1 n −1 1. S = 1.36 ≤ C p ≤ 1.11 25 − 1 25 − 1 0.33 2 ⎛ 1.52 33.49 49 = 1.

63 ⎢1 − 1.6.6) µˆ − LSL x 97 − 90 Cˆ pl = x = = 1. n = 30.4287 ≤ C pk ≤ 0.Chapter 7 Exercise Solutions 7-20 (7-14).46 3σˆ x 3(1.960 ⎡ 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 + + 0.6) Cˆ pk = min(Cˆ pl .96 ⎥ 2 2 9(30)(0.63 (b) α = 0. LSL = 90 (a) USL x − µˆ x 100 − 97 Cˆ pu = = = 0.63 3σˆ x 3(1.63 ⎢1 + 1.8313 7-17 . USL = 100.63) 2(30 − 1) ⎦ ⎣ ⎣ 0. Cˆ pu ) = 0. S = 1.025 = 1.05 zα / 2 = z0. x = 97.96 ⎥ ≤ C pk ≤ 0.63) 2(30 − 1) ⎦ 9(30)(0.

05. USL = 2350.42 ⎢1 + 1.96 + + ⎥ ≤ C pk ≤ 0.42) 2(50 − 1) ⎦ ⎣ ⎣ 0. 7-20. Cˆ pu ) = 0.42) 2(50 − 1) ⎦ 9(50)(0. 7-23 (7-17).42 (b) α = 0. n = 30 ⎡ ⎡ 1 ⎤ 1 ⎤ Cˆ pk ⎢1 − zα / 2 ⎥ ≤ C pk ≤ Cˆ pk ⎢1 + zα / 2 ⎥ 2(n − 1) ⎦ 2(n − 1) ⎦ ⎣ ⎣ ⎡ ⎤ ⎡ ⎤ 1 1 0.96 ⎥ 2(30 − 1) ⎦ 2(30 − 1) ⎦ ⎣ ⎣ 0.5443 7-22 (7-16).97 Cˆ pl = x 3σˆ x 3(60) Cˆ pk = min(Cˆ pl .42 Cˆ pu = 3σˆ x 3(60) µˆ − LSL x 2275 − 2100 = = 0. σˆ I = 3.96 ⎥ ≤ C pk ≤ 0. x = 2275. zα / 2 = z0. σ OI = 0.025 = 1.96 ⎥ 2 2 9(50)(0.2957 ≤ C pk ≤ 0.42 ⎢1 − 1. from Ex.63.63 ⎢1+1.63 ⎢1 − 1. but it is not too far off. n = 50 (a) USL x − µˆ x 2350 − 2275 = = 0. s = 60.79 The approximation yields a narrower confidence interval.96. nominal = 2225. Cˆ pk = 0. LSL = 2100.Chapter 7 Exercise Solutions 7-21 (7-15). zα / 2 = 1.47 ≤ C pk ≤ 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 .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.

6% T USL − LSL 2(15) 7-19 .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. R = 2.07 25 _ _ X=21. σˆ Gauge = 2. and the x chart has a few out-of-control parts. (b) specs: 25 ± 15 6σˆ Gauge 6(2. 12.00 standard deviations from center line. x = 21.482) P = ×100 = ×100 = 49. The new gauge is more repeatable than the old one. Test Failed at points: 8. 15.53 20 U C L=9.15 Sample Range 8 6 4 _ R=2.8 20 1 15 2 4 6 8 1 1 10 Sample 12 14 16 18 LC L=16.8 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 1.8. One point more than 3. 20 The R chart is in control. (a) n = 2.Chapter 7 Exercise Solutions 7-24 (7-18).

Chapter 7 Exercise Solutions 7-25 (7-19). MTB > Stat > Control Charts > Variables Charts for Subgroups > X-bar R Xbar-R Chart of Part Mesaurements (Ex7-25All) 102.359 × 100 = 62. (b) x = 98.359 2 σˆ Total = 4.0 _ R=2.5 3. One point more than 3.359) = = = 0.0 _ _ X=98.921 4.693 = 1.0 U C L=5.717 2 2 2 σˆ Product = σˆ Total − σˆ Gauge = 4. R = 2.872 σˆ Product = 1.553 Sample M ean 100.00 standard deviations from center line. Test Failed at points: 2.3 1.3.695 (c) σˆ Gauge σˆ Total × 100 = 1.3592 = 2. 3 The x chart has a couple out-of-control points.0 1 U C L=100.5 96.847 1 1 2 3 4 5 6 7 8 9 10 Sample Sample Range 6.717 − 1.2 97.5% 4. σˆ Gauge = R d 2 = 2. LSL = 100 – 15 = 85 6σˆ Gauge P 6(1.272 T USL − LSL 115 − 85 7-20 .5 99.3 1.0 LC L=95.5 0. This indicates that the operator is not having difficulty making consistent measurements. and the R chart is in control.717 (d) USL = 100 + 15 = 115.2.0 LC L=0 1 2 3 4 5 6 7 8 9 10 Sample Test Results for Xbar Chart of Ex7-25All TEST 1.

70. R2 = 2.693 = 1.00 1.03.151 = 1.128 = 0. x2 = 49.87.17 n = 2 operators d 2 = 1.7% T USL − LSL 60 − 40 7-21 .191 (c) specs: 50 ± 10 6σˆ Gauge P 6(1.181 + 0.418 σˆ Measurement Error = 1.30 R = 2. (a) Excel : workbook Chap07.693 σˆ Repeatability = R d 2 = 2.00 n = 3 repeat measurements d 2 = 1.Chapter 7 Exercise Solutions 7-26 (7-20).xls : worksheet Ex7-26 x1 = 50.128 σˆ Reproducibility = Rx d 2 = 0. R1 = 1.191) = ×100 = × 100 = 35.151 (b) 2 2 2 ˆ2 ˆ2 σˆ Measurement Error = σ Repeatability + σ Reproducibility = 1.181 Rx = 0.17 1.

0 17.Chapter 7 Exercise Solutions 7-27 (7-21). 7-22 .7 20.0 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 1.0 _ R=1.154 (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. 12 Out-of-control points on R chart indicate operator difficulty with using gage. (a) σˆ Gauge = R d 2 = 1.128 = 1.00 standard deviations from center line.5 LC L=0 0.5 3.58 22.010 4.359 Gauge capability: 6σˆ = 8.533 1.5 _ _ X=20.0 1 1 2 3 4 5 6 7 8 Sample 9 10 Sample Range 6.82 1 15.0 U C L=23. Test Failed at points: 11. One point more than 3.5 LC L=17.0 11 12 1 1 13 14 15 U C L=5.533 1.

592 S = 0. 8.84 0.712 2 σˆ Operator = = 0. 5. Ex7-28Op Factor Ex7-28Part Ex7-28Op Type random random Factor Ex7-28Part 18. For the reduced model: ANOVA: Ex7-28Reading versus Ex7-28Part.861 R-Sq(adj) = 90.0149 = pn 20(2) MSP − MSP×O 62. MTB > Stat > ANOVA > Balanced ANOVA In Results. 7.2798 = on 3(2) 2 σˆ Part×Operator = The manual calculations match the MINITAB results. 17.308 Ex7-28Part*Ex7-28Op 38 27.65 1.1399 0.10).1400 ⇒ 0 n 2 MSO − MSP×O 1. 12.500 0.173 0. 6.2798 0. 14. select “Display expected mean squares and variance components” ANOVA: Ex7-28Reading versus Ex7-28Part. Levels 20 3 3. 3 Analysis of Variance for Ex7-28Reading Source DF SS MS Ex7-28Part 19 1185. 16. R-Sq = 95. 19.712 − 0.992 MSP×O − MSE 0. we can fit a reduced model without that term.8832 Error term 3 3 Expected Mean Square for Each Term (using unrestricted model) (3) + 6 (1) (3) + 40 (2) (3) 7-23 .33% Source Ex7-28Part Ex7-28Op Ex7-28Part*Ex7-28Op Error F 87. Since the Part × Operator term is not significant (α = 0.Chapter 7 Exercise Solutions 7-28☺. Ex7-28Op … 1 2 3 Source Ex7-28Part Ex7-28Op Error Variance component 10.617 1. 10.050 0.74% Variance component 10. Note the Part × Operator variance component is negative. 2.308 − 0. 13. 1.992 = = −0. 2. 20 Ex7-28Op Values 1.0149 -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 = 0.000 0. 15.712 2 σˆ Part = = 10. 4.995825 1 2 3 4 9.2513 0.992 Total 119 1274.72 P 0.391 Ex7-28Op 2 2.0106 0.712 Error 60 59. 11.425 62.391 − 0.

92 0.6470 1.09 95. and total gauge variability may also be found using: MTB > Stat > Quality Tools > Gage Study > Gage R&R Study (Crossed) Gage R&R Study .15 3.10 0.43 2.232 Gage R&R Source Total Gage R&R Repeatability Reproducibility Ex7-28Op Part-To-Part Total Variation VarComp 0.3083 0.2106 Total Variation 3.7178 P 0. repeatability.3908 1.59 MS 62.09 3.8832 = 0.43 2.000 0.000 0.861 Two-Way ANOVA Table Without Interaction Source Ex7-28Part Ex7-28Op Repeatability Total DF 19 2 98 119 SS 1185.93977 5.05 59.0305 Number of Distinct Categories = 4 %Study Var (%SV) 28.00 7-24 .Chapter 7 Exercise Solutions (a) 2 2 σˆ Reproducibility = σˆ Operator = 0.33842 20.7118 0.10 91.1050 /T = P USL-LSL 60 − 6 This gauge is borderline capable since the estimate of P/T ratio just exceeds 0.9454 n = = 0.10.02 7.3908 1.10310 0.8832 0.8938 σˆ Gauge = 0.10310 0.32 28.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 1185.2513 11.8832 (b) 2 2 2 σˆ Gauge = σˆ Reproducibility + σˆ Repeatability = 0.6186 Part-To-Part 3.98 100.91 100.20176 19.6386 Reproducibility 0.8380 0.173 0.8938 0.62 27.1451 %Contribution (of VarComp) 8.55 1274.0106 + 0. Estimates of variance components.50 1274.6186 Ex7-28Op 0.62 86.6724 Repeatability 0.9454 (c) 6 × σˆ Gauge 6 × 0.00 Study Var Source StdDev (SD) (6 * SD) Total Gage R&R 0.0106 2 2 σˆ Repeatability = σˆ Error = 0.59 MS 62.9917 F 87.0106 0.94541 5.3083 0.4814 P 0. reproducibility.6447 1.8832 F 70.0106 10.

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=24.55 _ _ X=22.39 20 LCL=20.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=3.757 30 25 2 _ R=1.23 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 .

(Also note that the MINITAB Gage R&R output indicates “Number of Distinct Categories = 4”. 2 2 σˆ Part = 10.94 DR 1 − ρˆ P 1 − 0.2513 = = 0.9198 2 σˆ Total 11. This is near the AIAG-recommended value of five levels or more.9198 SNR = 4.79 indicates that fewer than five distinct levels can be reliably obtained from the measurements.9198 = 23. this is also the number of distinct categories of parts that the gauge is able to distinguish) DR = 23.9772 = 0.94. 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 (2.79 1 − ρˆ P 1 − 0.9198 m = 1 + ρˆ P = 1 + 0.000) = 1 − 0. σˆ Total = 11. 7-30 (7-22).1451 n= SNR 2 ρˆ P 2(0.9198) = = 4. By this measure.1451 ρˆ P = 2 σˆ Part 10. but larger than a value of two (or less) that indicates inadequate gauge capability. the gauge is capable. exceeding the minimum recommendation of four.2513.0228 7-26 .Chapter 7 Exercise Solutions 7-29☺.

9} ⎛ 0.4330 7-32 (7-24).6.02 (0.4 σ y2 = σ 12 + σ 22 = 0.01) = 0.50 Pr{Nonconformities} = Pr{ y < LSL} + Pr{ y > USL} = Pr{ y < 0.4 ⎞ = Φ⎜ ⎟ +1− Φ ⎜ ⎟ ⎝ 0.02 (0.00181 σˆ Weight ≅ 0.01)(0.6) + 1 − Φ (1.42 ) Nonconformities will occur if y = x1 − x2 < 0.2 2 )(0.0 2 σ Volume ≅ µ L2σ H2 σ W2 + µ H2 σ L2σ W2 + µW2 σ L2σ H2 = 6.0(3. 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 = 6.25 ⎠ ⎝ 0.08(10)(20)(3) = 48 2 σˆ Weight ≅ d 2 ⎡⎣ µˆW2 σˆ L2σˆT2 + µˆ L2σˆW2 σˆT2 + µˆT2σˆW2 σˆ L2 ⎤⎦ = 0.22 )(0.9} = Pr{ y < 0.00) = 0. 0.25 ⎠ = Φ (−0.1} + 1 − Pr{ y ≤ 0.32 + 0.12 ) + 202 (0.32 ).6 = 0.9 − 0. x1 ~ N (20. 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 ] = 0.2743 + 1 − 0.32 ) ⎤⎦ = 0.32 )(0.42 = 0.4 ⎞ ⎛ 0.04252 7-27 . x2 ~ N (19.12 ) + 32 (0.25 σ y = 0.0061 7-33 (7-25).01)(0.9 µ y = µ1 − µ2 = 20 − 19.1} + Pr{ y > 0.01) + 3.1 or y = x1 − x2 > 0.0) = 72. 0.8413 = 0.082 ⎡⎣10 2 (0.1 − 0.0)(4.02 (0.01) + 4.01)(0.Chapter 7 Exercise Solutions 7-31 (7-23).

05 x) 2 and f ( x) = σ x2 = E ( x 2 ) − [ E ( x)] = 10. 1 (5 x − 2).1026 26 ⎜⎝ 4 2 3 2 ⎟⎠ 2 ⎣ 26 ⎦ s = (3 + 0.05(3.1001 7-35 (7-27).05(3. 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 ⎟ = 3.05µ x )(0.05)(0.1282) 2 = 0.3170) = 0.05( µ x )] = [3 + 0.Chapter 7 Exercise Solutions 7-34 (7-26).3170 2 µ s ≅ g ( x) = [3 + 0.1282)] = 9.1282) ] (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 .9629 2 ⎡ ∂ g ( x) ⎤ σ ≅⎢ ⎣ ∂ x ⎥⎦ 2 2 2 s σ x2 µx ⎡ ∂ (3 + 0.05)σ x2 = 2 [3 + 0.1026 − (3.05 x)2 ⎤ ⎥ =⎢ ∂x ⎢⎣ ⎥⎦ 2 σ x2 µx = 2(3 + 0.1282 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.

0.0022 + 0.5(−2.3002 = 0. 0.4002 ). 0.512) − 0.000005 ⎠ = 1 − Φ (−2.09 − µ y ⎫⎪ −1 Pr ⎨ z < ⎬ = Φ (0.989 1 ⎛ 2 − α ⎞ χ1−γ . ID ~ N (2.006) σ y ⎭⎪ ⎩⎪ 0.4 = 5.9964 7-38 (7-30).010 − 2. x2 ~ N ( µ2 .004.01 ⎞ 5.006 2 2 σ y2 = σ ID + σ OD = 0.09] = 1. α = 0.010.01 γ = 0.004 = 0. x1 ~ N ( µ1 .4 1 ⎛ 2 − 0.5 µ y = −[0.683) = 1 − 0.09} = 0.01 ⎠ 4 2 n≅ 7-29 .0036 = 0.346 7-37 (7-29).80 2 χ12−γ . 0.Chapter 7 Exercise Solutions 7-36 (7-28).09 − µ y = −2.0012 ) Interference occurs if y = ID – OD < 0 µ y = µID − µOD = 2.002236 Pr{positive clearance} = 1 − Pr{interference} = 1 − Pr{ y < 0} ⎛ 0 − 0.0022 ) and OD ~ N (2.512 0.20.3002 ) µ y = µ1 − µ2 σ y = σ 12 + σ 22 = 0.0012 = 0.000005 σ y = 0.4002 + 0.4 = χ 0.989 +⎜ = +⎜ = 299 ⎟ ⎟ 2 ⎝ α ⎠ 4 2 ⎝ 0.5 Pr{ y < 0.006 ⎧⎪ 0.006 ⎞ = 1− Φ ⎜ ⎟ ⎝ 0.

one-sided From Appendix VIII: K = 2.55 7-40 (7-32).102 ).208(10) = 372.12 ).162 7-41 (7-33). x ~ N (350. γ = 0. n = 10.95 . γ = 0. γ = 0. one-sided From Appendix VIII: K = 2.838(1) = 83. x ~ N (300. x ~ N (85.05. α = 0.10.838 x − KS = 85 − 1. n = 25.Chapter 7 Exercise Solutions 7-39 (7-31).355 UTL = x + KS = 300 + 2.779 1 ⎛ 2 − α ⎞ χ1−γ .05 γ = 0.10.90 .4 = 7. one-sided From Appendix VIII: K = 1.779 +⎜ = +⎜ = 77 ⎟ ⎟ 2 ⎝ α ⎠ 4 2 ⎝ 0. 7-30 .90 2 χ12−γ .4 = χ 0.208 UTL = x + KS = 350 + 2.95 . a natural tolerance interval would be the smallest to largest observations.355(10) = 323.10. α = 0. n = 20.08 7-42 (7-34).05 ⎞ 7. α = 0.102 ).05 ⎠ 4 2 n≅ After the data are collected. α = 0.4 1 ⎛ 2 − 0.

γ = 0. tα / 2. 0. x ~ N 0.1265] Part (a) is a tolerance interval on individual thickness observations.n −1 S ( n = 0.0003 ) 40 = [0.1264 ± 2.00032 ( ) (a) α = 0. 7-31 .1271] (b) α = 0. 0. 7-44 (7-36).39 = 2.1264 ± 2.445 TI on x : x ± KS = 0.05.1264. while in part (b) the interval refers to a parameter of a distribution (an unknown constant). the interval relates to individual observations (random variables).025.0003) = [0. part (b) is a confidence interval on mean thickness.05.023 0.1257.95. γ = 0. 0.1263. and two-sided From Appendix VII: K = 2. α = 0.95 log(1 − γ ) log(1 − 0.05.445(0.95) n= = = 59 log(1 − α ) log(1 − 0.n −1 = t0.Chapter 7 Exercise Solutions 7-43 (7-35).023 CI on x : x ± tα / 2. In part (a).05) The largest observation would be the nonparametric upper tolerance limit.

An “*” following the exercise number indicates that the description has changed. std dev = 25. 8-1.4345 No.5.5. k = 0. New exercises are denoted with an “☺”. The assignable cause occurred after observation 10 – 3 = 7. The point at which the assignable cause occurred can be determined by counting the number of increasing plot points. 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. H = 5σ = 5(25) = 125 (a) MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Molecular Weight (Ex8-1mole) target = 1050. µ0 = 1050. δ = 1σ. 8-1 . The estimate used for σ is much smaller than that from the data. σ = 25.Chapter 8 Exercise Solutions Several exercises in this chapter differ from those in the 4th edition. K = (δ/2)σ = (1/2)25 = 12.8421/1. (b) σˆ = MR2 d 2 = 38. A number in parentheses gives the exercise number from the 4th edition.128 = 34.

The assignable cause occurred after observation 10 – 3 = 7. 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.Chapter 8 Exercise Solutions 8-2. k = 0. MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Standardized Molecular Weight (Ex8-2std) target = 1050.5. 8-2 . std dev = 25.

Chapter 8 Exercise Solutions 8-3. (a) µ0 = 1050. the same as the CUSUM without a FIR feature.5] = 45 Using the tabular CUSUM.5.5/25 = 2. 8-3 .1045 − (1050 + 12. xi − ( µ0 − K ) + C0+ ⎤⎦ = max [ 0. σ = 25. H/2 = 125/2 = 62.5) + 62.5 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Molecular Weight (Ex8-1mole) FIR=H/2 = 62.5. in std dev units = 62.5. h = 5.5 (or 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. the process signals out of control at observation 10. C1+ = max ⎡⎣0.5 FIR = H/2 = 62. K = 12. k = 0.

5σ limits on the Individuals chart. there are no out-of-control signals.8 Individual Value 1200 _ X=1116.3 1100 1000 -3.6 120 80 __ MR=38. However there does appear to be a trend up from observations 6 through 12—this is the situation detected by the cumulative sum. 8-4 .5SL=0 2 4 6 8 10 12 Observation 14 16 18 20 Using 3.5SL=995.5-sigma limits +3.5SL=141.5SL=1236.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 3.8 40 0 -3.8 2 4 6 8 10 12 Observation 14 16 18 20 160 Moving Range +3.

h = 4.128 = 0.5)(5.02. σ = 0.77.5) 2 1 1 1 2 = + = = 0. b = h + 1.166 = 5.0 0 -0.936 δ * = 0.964 2(−0.05) = 0.2 LCL=-0.77. k = 0.964 ARL+0 = ARL−0 ≅ ARL0 = 1/ 0.3 UCL=0.0027 + − ARL0 ARL0 ARL0 742.77 + 1. In Exercise 8-4: µ0 = 8. ∆ + = δ * − k = 0 − 0.0027 = 371.3 2 4 6 8 10 12 14 Sample 16 18 20 22 24 There are no out-of-control signals. k = 1/ 2.05. and h=4.2385 (a) MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Can Weight (Ex8-4can) target = 8.02. µ0 = 8.5 = −0.5. ∆ − = −δ * − k = −0 − 0.936)] + 2(−0.0186957 /1. (b) σˆ = MR2 1.02. σ = 0.166 = 4. H = hσ = 4.936) − 1 = 742.2385 -0.0166 .2 0.77 (0. h = 4.Chapter 8 Exercise Solutions 8-4. so σ = 0.1 -0. k=1/2.05.128 = 0.5.5 exp[−2(−0.2385 Cumulative Sum 0.1 0.5)(5.5 = −0.48 8-5 .05 is probably not reasonable.77 0.

σ = 0.01. ∆ − = −δ * − k = −0 − 0.25 = −0.25)(9.25.6771 2(−0.05.01 + 1.176) − 1 = 741.3 Cumulative Sum 0.176 δ * = 0.4 0. 8-6 .84 The theoretical performance of these two CUSUM schemes is the same for Exercises 8-4 and 8-5.2 -0.01.2 0.02.166 = 8.1 0 0.01 UCL=0.0027 = 370.05. k = 0. H = hσ = 8.166 = 9.4005 -0.1 -0.01 (0.25 = −0. h = 8.4005 0.176)] + 2(−0.Chapter 8 Exercise Solutions 8-5.25. In Exercise 8-5: µ0 = 8.0 -0.0027 + − ARL0 ARL0 ARL0 741. ∆ + = δ * − k = 0 − 0.3 LCL=-0. h = 8. b = h + 1. h = 8.25.6771 ARL+0 = ARL−0 ≅ ARL0 = 1 0.05) = 0.25 exp[−2(−0. σ = 0.5 2 4 6 8 10 12 14 Sample 16 18 20 22 24 There are no out-of-control signals. k = 0.25.4 -0.02.4005 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Can Weight (Ex8-4can) target = 8.25)(9. µ0 = 8.02. k = 0.25) 2 1 1 1 2 = + = = 0.

05) = 0.0 0 -0. k = 0.00. Process was out of control at process start-up.1 -0.3 2 4 6 8 10 12 14 Sample 16 18 20 22 24 The process signals out of control at observation 20. H = h σ = 4. h = 4. h = 4.3 UCL=0.2385 -0.05. FIR in # of standard deviations = h/2 = 4. σ = 0.77.77 0.385 std dev.1 0. µ0 = 8.385 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Ex8-4can FIR = 2.77/2 = 2.00.Chapter 8 Exercise Solutions 8-6.2 LCL=-0.2385 FIR = H/2. target = 8. k = 1/2.2 0. 8-7 .77 (0.5.2385 Cumulative Sum 0.

h = 5 75 UCL=60. h = 5 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Temperature Readings (Ex8-7temp) target = 950.128 = 12. (a) σˆ = MR2 d 2 = 13.Chapter 8 Exercise Solutions 8-7. The assignable cause occurred after observation 12 – 10 = 2. 8-8 .5.16 (b) µ0 = 950. One point beyond control limits. k = 0. σˆ = 12.8 1 8 16 24 32 40 48 Sample 56 64 72 80 Test Results for CUSUM Chart of Ex8-7temp TEST.16.8 Cumulative Sum 50 25 0 0 -25 -50 LCL=-60. 13 The process signals out of control at observation 12. Test Failed at points: 12.7215 /1. k = 1/ 2.

629. 16. h = 5 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Bath Concentrations (Ex8-8con) target = 175. 32 The process signals out of control on the lower side at sample 3 and on the upper side at sample 12. 14.1 0 -100 -200 3 6 9 12 15 18 Sample 21 24 27 30 Test Results for CUSUM Chart of Ex8-8con TEST. σˆ = 5. k = 1/2. 13. 20. 24.128 = 5. k = 1/ 2. (a) σˆ = MR2 d 2 = 6. 22. 19. std dev = 5.629 (from a Moving Range chart with CL = 6. Test Failed at points: 12. 23.35 /1. h = 5 400 Cumulative Sum 300 200 100 UCL=28. 8-9 . 15. 25. 26. 17. 21. One point beyond control limits. 31.35) (b) µ0 = 175. 27. 29. 28. 30. 18.Chapter 8 Exercise Solutions 8-8.629. Assignable causes occurred after startup (sample 3 – 3 = 0) and after sample 8 (12 – 4).1 0 LCL=-28.

01. h = 8. Assignable causes occurred after startup (sample 2 – 2) and after sample 9 (16 – 7).25. Test Failed at points: 16. h = 8.7 0 0 Cumulative Sum LCL=-47.949 (from a Moving Range chart with CL = 6. One point beyond control limits. σˆ = 5. should be balanced by a larger control limit.01 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Viscosity Measurements (Ex8-9vis) target = 3200. 8-10 .01 UCL=47. k = 0.128 = 5. 17. k = 0.71) (b) µ0 = 3200.Chapter 8 Exercise Solutions 8-9.949.71/1. h = 8. to give longer in-control ARLs with shorter out-of-control ARLs. (a) σˆ = MR2 d 2 = 6. k = 0. 18 The process signals out of control on the lower side at sample 2 and on the upper side at sample 16.25. std dev = 5.949. (c) Selecting a smaller shift to detect.25.7 -100 -200 -300 -400 -500 4 8 12 16 20 Sample 24 28 32 36 Test Results for CUSUM Chart of Ex8-9vis TEST.

41. .2504 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Flow Width Data (Exm5-1x1. Test Failed at points: 40. σ = 0. and remains above the upper limit. 8-11 .5..14 5 = 0...50.5.250 -0. This CUSUM detects the shift in process mean earlier. K = kσ x = 0. The x -R chart shown in Figure 5-4 signals out of control at sample 43.2 1. k = 0.0313.0 Cumulative Sum 0.250 0. n = 5.8 0. One point beyond control limits. at sample 40 versus sample 43.4 UCL=0.2 0 0. Exm5-1x5 TEST.. 43. σ x = σ n = 0.4 4 8 12 16 20 24 Sample 28 32 36 40 44 Test Results for CUSUM Chart of Exm5-1x1. 42.14..6 0. k = δ 2 = 0. H = hσ x = 0.2 LCL=-0. .50. 44.0626 δ = 1.14.. µ0 = 1. h = 4 1. Exm5-1x5) target = 1.0 -0.Chapter 8 Exercise Solutions 8-10*. std dev = 0. 45 The CUSUM chart signals out of control at sample 40. h = 4.

5 5 Obs.68 3.9 10 OOC 7 30.52 4.xls : worksheet Ex8-11 mu0 = sigma = delta = k= h= 1050 25 1 sigma 0. 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 -0.848 2 0 0 0.44 5 OOC 7 14.N.94 3 0.52 3.12 2.90 1151 4.49 1008 -1.498 3 OOC 7 8.21 1095 1.35 6 OOC 7 16.66 9 OOC 7 26.52 -0.84 3.989 1 1. 8-12 .56 -0.2 -1.86 1 0.26 1167 4.07 1055 0.54 12 OOC 7 36.OOC? When? 0 0.04 3.2 -1.349 ( ) Excel file: workbook Chap08. Process variability is increasing.74 1188 5.88 13 OOC 7 one-sided lower cusum Si.84 3.36 1087 1.76 3.22 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 = 7.13 1125 3 2.71 1238 7.049 4 OOC 7 11.40 1128 3.68 1.52 5.Chapter 8 Exercise Solutions 8-11.65 4 0 0 0 0 1.07 1037 -0.631 1 2.48 1.05 1169 4.55 7 OOC 7 21.56 8 OOC 7 23.61 1163 4.78 11 OOC 7 33.84 one-sided upper cusum Si+ N+ OOC? When? 0 0 0 0 0 0 0 0 0 0.26 1139 3.8 1.50 1125 3 2.822 0.36 1050 0 -2.56 3.61 1146 3.29 1064 0.15 2 0.38 1146 3. Vi = | yi | − 0.57 1 1.738 2 5.

46 31 19 OOC 0 182 1.822 0.9540 1.86 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 = 0.27 23.0199 2.5 5 one-sided upper cusum xi yi vi Si+ N+ OOC? When? 0 160 -2.9540 1.00 19.2633 3.45 9 OOC 0 195 3.5528 3.65 20.31 13 5 OOC 0 154 -3.7304 3.59 14 OOC 0 186 1. Vi = | yi | − 0.32 28.31 26.7764 1.3 8 OOC 0 186 1.9081 3.34 20 OOC 0 one-sided lower cusum Si.349 ( ) Excel file : workbook Chap08.68 6 OOC 0 158 -3.36 20.06 22.62 22.4410 3.56 10.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 1.4869 2.7106 0.9081 3.18 15.19 4 OOC 0 153 -3.0199 2.3093 2.62 17.04 18 OOC 0 185 1.N.37 17 OOC 0 189 2.16 30.84 31.6294 (from Exercise 8-8) 1 sigma 0.62 3. Process variability has been increasing since start-up.65 23.0199 2.2435 0.32 1.56 11 OOC 0 184 1.0000 -2.129 3 OOC 0 151 -4.8 7 OOC 0 162 -2.32 12 OOC 0 175 0.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 5.946 2 150 -4.Chapter 8 Exercise Solutions 8-12.822 1 158 -3.6646 2.6646 2.5987 1.47 13 OOC 0 192 3.55 16 OOC 0 190 2.74 15 OOC 0 197 3.68 7.05 23 10 OOC 0 179 0. 8-13 .

7) 2 1 1 1 1 1 = + = + = 0.5 − 0.2 = 0.166) − 1 ARL+1 = = 25.166 = 9. two-sided cusum with k = 0.166)] + 2(0.005 = 215.3)(9.040 = 25.166 exp[−2(−0.005 + − ARL0 ARL0 ARL0 430. Standardized.2)(9.Chapter 8 Exercise Solutions 8-13.7)(9.2 = −0.3 ∆ − = −δ * − k = −0.166 exp[−2(0.3)(9.166)] + 2(−0.5 ∆ + = δ * − k = 0. 767 ARL1 = 1/ 0.556 ARL0 = 1/ 0.040 + − ARL1 ARL1 ARL1 25.2)(9.2 ∆ − = −δ * − k = −0 − 0.166) − 1 ARL+0 = ARL−0 ≅ = 430.2 b = h + 1.2 = −0.556 2(−0.2 = −0.166 = 8 + 1.2) 2 1 1 1 2 = + = = 0.166)] + 2(−0.166) − 1 ARL−1 = = 381.166 = 9.023 381.2 and h = 8 In control ARL performance: δ* = 0 ∆ + = δ * − k = 0 − 0.023 2(0.02 8-14 .7 b = h + 1.3) 2 exp[−2(−0.166 = 8 + 1.7)(9.23 Out of control ARL Performance: δ * = 0.5 − 0. 767 2(−0.

H = hs = 5(5.976.5 (5. h = 5 K = ks = 0. The upper CUSUM is used to detect upward shifts in the level of the process. there is no option to display a single-sided chart. k = 0. The assignable cause occurred at start-up (2 – 2).95238) = 29.762 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.Chapter 8 Exercise Solutions 8-14. µ0 = 3150. s = 5.95238) = 2.95238. The process signals out of control on the upper side at sample 2. 8-15 .5.

7 (from a Moving Range chart with CL = 122.5 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.128 = 108.7) = 54.5(108. The CUSUM chart reflects a consistent run above the target value 734.6 /1.5. σˆ = MR2 d 2 = 122. displayed a distinct downward trend in measurements.5. There is a distinct signal on both charts. 8-16 . of either a trend/drift or a shit in measurements.Chapter 8 Exercise Solutions 8-15☺.7) = 543. k = 0. from virtually the first sample.5.35 H = hσˆ = 5(108.5 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Light Velocity (Ex5-60Vel) target = 734. starting at about sample 18. The outof-control signals should lead us to investigate and determine the assignable cause. h = 5 K = kσˆ = 0.6) µ0 = 734. with a centerline at x = 909 .

7 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Light Velocity (Ex5-60Vel) lambda = 0.7SL=667.7 900 EWMA 850 800 +2.8 750 _ _ X=734.5. σ = 108.7SL=801. L = 2. and also indicates the slight downward trend starting at about sample 22. 734. CL = µ0 = 734.Chapter 8 Exercise Solutions 8-16☺.5.2 4 8 12 16 20 24 Sample 28 32 36 40 The EWMA chart reflects the consistent trend above the target value.1. λ = 0.1. L = 2. 8-17 .5 700 -2.7.

7SL=1065. 8-18 .6 2 4 6 8 10 12 Sample 14 16 18 20 Process exceeds upper control limit at sample 10. (a) λ = 0.7 1140 1120 EWMA 1100 1080 +2.1.1 (2 − 0. the width of the control limits also increases. the same as the CUSUM chart.49.31.69] (b) λ = 0. L = 3 limits = µ0 ± Lσ λ (2 − λ ) = 10 ± 3(1) 0.10. 8-18 (8-16). L = 2.5] As λ increases.1.7SL=1034.4 (2 − 0.2) = [9.2. L = 2. λ = 0.4.11] (c) λ = 0.1. L = 3 limits = µ0 ± Lσ λ (2 − λ ) = 10 ± 3(1) 0.2 (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). UCL = 1065.4) = [8.1) = [9.5. L = 3 limits = µ0 ± Lσ λ (2 − λ ) = 10 ± 3(1) 0.4 _ _ X=1050 1060 1040 -2.11. LCL = 1034.7. CL = µ0 = 1050. σ = 25.

2.06 EWMA 8.2. LCL = 7.07.Chapter 8 Exercise Solutions 8-19 (8-17).02 8.96 2 4 6 8 10 12 14 Sample 16 18 20 22 24 The process is in control.04 _ _ X=8. UCL = 8. CL = µ0 = 8.08 UCL=8.02.00 7.97 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Can Weight (Ex8-4can) lambda = 0. L = 3 8.98 LCL=7. 8-19 .9700 7. λ = 0. L = 3.05. Assume σ = 0.02 8.0700 8.

UCL = 8.00 7.01 8. L = 2.1. λ = 0.04 EWMA 8.02 8.05 8. CL = µ0 = 8.03 _ _ X=8. L = 2.99 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Can Weight (Ex8-4can) lambda = 0.1.02 8.99 -2.05087 8.05.Chapter 8 Exercise Solutions 8-20 (8-18).05.02. 8-20 .7.7SL=8.7 +2.98913 2 4 6 8 10 12 14 Sample 16 18 20 22 24 The process is in control. LCL = 7. There is not much difference between the control charts. Assume σ = 0.7SL=7.

Chapter 8 Exercise Solutions 8-21 (8-19).53.7SL=957. Test Failed at points: 12.53 EWMA 955 _ _ X=950 950 945 -2. MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Temperature Readings (Ex8-7temp) lambda = 0. 8-21 . One point beyond control limits.7 960 +2. 12 and 13.47 1 8 16 24 32 40 48 Sample 56 64 72 80 Test Results for EWMA Chart of Ex8-7temp TEST. LCL = 942.1. L = 2.1. UCL = 957. CL = µ0 = 950. L = 2.7.7SL=942.47. λ = 0. σˆ = 12.16 . but not flagged on chart). 13 Process is out of control at samples 8 (beyond upper limit.

4.76 930 1 8 16 24 32 40 48 Sample 56 64 72 80 Test Results for EWMA Chart of Ex8-7temp TEST. Test Failed at points: 70 With the larger λ. L = 3 970 UCL=968. as compared to the chart in the Exercise 21 (with the smaller λ) which signaled out of control at earlier samples.Chapter 8 Exercise Solutions 8-22 (8-20). MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Temperature Readings ( Ex8-7temp) lambda = 0.24 EWMA 960 _ _ X=950 950 940 LCL=931. CL = µ0 = 950.24. the process is out of control at observation 70.4. One point beyond control limits. λ = 0.16 . UCL = 968. 8-22 . L = 3. σˆ = 12.76. LCL = 931.

70 170 3 6 9 12 15 18 Sample 21 24 27 30 Process is out of control. LCL = 172. 8-23 .05.Chapter 8 Exercise Solutions 8-23 (8-21).70. UCL = 177.05. The data is grouped into three increasing levels.634 . CL = µ0 = 175.594 is too far from the process target of µ0 = 175 for the process variability.6SL=172.6.30 _ _ X=175 175 -2. MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Bath Concentrations (Ex8-8con) lambda = 0.30. L = 2. The process average of µˆ = 183. σˆ = 5.6SL=177. λ = 0. L = 2.6 190 EWMA 185 180 +2.

L = 2.ov o v .7SL=18.n.4 .ep.2777th.1 t. a properly designed EWMA chart is very robust to the assumption of normally distributed data.a r .8 l.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.Chapter 8 Exercise Solutions 8-24☺.5 10.ec.b.and 0. The EWMA chart also does not signal out of control.1.7 20.Sep ep .0 20 23 25 r .9 26 .Ju Ju Ju Ju l.0 _ _ X=12.n.4 t. λ = 0. Individuals control charts of 0.23 5.9 22 24 t.7 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Homicide Data (Ex6-62Bet) lambda = 0.5 EWMA 15.y .7SL=6.0 +2.un n. 8-24 .25th-root transformed data showed no out-of-control signals. As mentioned in the text (Section 8.0 7.27 17. L = 2.5 10 r .b.5 -2.7 24 28 .7 16 16 22 25 l -6 l.8 19 .4-3).1.Oc Oc O c ct.25 12.p ay y .2 25 28 29 n.n .

L = 2.7SL=3203.1.7SL=3196. µ0 = 3200.7 MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Viscosity (Ex8-9vis) Target=3200. σˆ = 5.68 _ _ X=3200 3200 -2.7 3205 +2. λ = 0.95 (from Exercise 8-9). 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. sigma=5. L=2.95.1.Chapter 8 Exercise Solutions 8-25 (8-22). lambda=0.

4 1000 2 4 6 8 10 12 Sample 14 16 18 20 Test Results for Moving Average Chart of Ex8-1mole TEST. w = 6. 13. CL = 1050. std dev = 25 1200 Moving Average 1150 1100 UCL=1080. 19. 15. target value = 1050. 16. UCL = 1080. 18. σ = 25. Test Failed at points: 10.6. 14. LCL = 1019. µ0 = 1050. 17. 8-26 . the same result as for Exercise 8-1.Chapter 8 Exercise Solutions 8-26 (8-23). 11. One point beyond control limits. 20 Process is out of control at observation 10.6 _ _ X=1050 1050 LCL=1019.4 MTB > Stat > Control Charts > Time-Weighted Charts > Moving Average Moving Average Chart of Molecular Weight (Ex8-1mole) w = 6. 12.

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

is the same for each part type. c1130 = 64. with CL = 0. a standardized CUSUM could be used to detect smaller deviations from the target value.13 c1055 = 13.00 12.13 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. c1261 = 12.25. The chart would be designed so that δ. Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Rx9-4Def Rx9-4Def 1055 1130 1261 1385 4610 8611 13.Chapter 9 Exercise Solutions 9-3. LCL = −3.67. 9-4.67 26. In a short production run situation. c8611 = 50. Note: In the textbook. in standard deviation units. The standardized variable ( yi . 9-3 . j − µ0. c1385 = 26.63 4.00. the 4th part on Day 246 should be “1385” not “1395”.67 50. Set up a standardized c chart for defect counts. The plot statistic is Z i = ( ci − c ) c.25 64.63. c4610 = 4.67. UCL = +3. j ) / σ j (where j represents the part type) would be used to calculate each plot statistic.

There are many values of x max and x min that are outside the control limits. so the process is out-of-control. Excel : Workbook Chap09.00 Ex9-5XbUCL 51.00 Ex9-5Xbmin 53.023 0 2.000 Group Xbar Control Chart 61.Chapter 9 Exercise Solutions 9-5. not just a specific one.00 45. 9-4 .988 2.00 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 57.00 59.00 Ex9-5Xbmax Xbar 55.338 4 heads 3 units 1.379 50.596 6.017 0.00 47.00 Ex9-5XbLCL 49. The assignable cause affects all heads.574 55.xls : Worksheet Ex9-5 Grand Avg = Avg R = s= n= A2 = D3 = D4 = Xbar UCL = Xbar LCL = R UCL = R LCL = 52.

Chapter 9 Exercise Solutions 9-6. Excel : Workbook Chap09. a process change may have caused output of this head to be different from the others. 9-5 .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.

128 0 3.248 7. 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 9-5.000 Group Control Chart for Individual Obs.Chapter 9 Exercise Solutions 9-7.158 4 heads 2 units 1. Obs. 70 Individ.267 58. Obs.988 2.xls : Worksheet Ex9-7A Grand Avg = Avg MR = s= n= d2 = D3 = D4 = Xbar UCL = Xbar LCL = R UCL = R LCL = 52.727 47. 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. (a) Excel : Workbook Chap09. 9-6 .050 0.

128 0 3.267 58.Chapter 9 Exercise Solutions 9-7 continued (b) Excel : Workbook Chap09.248 7. 65 Individ.158 4 heads 2 units 1.050 0. 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. 9-7 .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. Obs. indicating a potential process change.988 2.

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. Ex9-7X4) U C L=56.. Xbar-S Chart of Head Measurements (Ex9-7X1.988 52 50 LC L=49..816 2 4 6 8 10 Sample 12 14 16 18 20 U C L=4.159 Sample M ean 56 54 _ _ X=52. 9-8 .415 Sample StDev 4 3 _ S =1.

. 9-9 ... .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..988 52 50 LC L=49.6 2. Xbar-S Chart of Head Measurements (Ex9-7X1..816 3 6 9 12 15 Sample 18 21 24 Sample StDev 4.4 _ S =1.0 LC L=0 3 6 9 12 15 Sample 18 21 24 27 30 Test Results for S Chart of Ex9-7X1.415 3. Ex9-7X4) U C L=56.159 Sample M ean 56 54 _ _ X=52. One point more than 3.2 0.. Test Failed at points: 27. 29 Only the S chart gives any indication of out-of-control process.00 standard deviations from center line. Ex9-7X4 TEST 1.8 27 1 30 1 U C L=4..948 1.

548 ) [6(0. R = 0.Chapter 9 Exercise Solutions 9-8.83 PCR Stat > Control Charts > Variables Charts for Subgroups > R Chart R Chart of Range Values ( Ex9-8R. . Ex9-8Rdum4) 0.005 UCL=0..326 = 0. σˆ = R / d 2 = 0..55025.00227 / 2.002 0. 9-10 .000976 n = ( USL-LSL ) 6σˆ = ( 0.000 2 4 6 8 10 12 Sample 14 16 18 20 The process variability.00227 0.002270 n=5 x = 0. as shown on the R chart is in control.55025 0.001 LCL=0 0. Ex9-8R Variable Ex9-8Xbar Ex9-8R Mean 0.552 − 0. Stat > Basic Statistics > Display Descriptive Statistics Descriptive Statistics: Ex9-8Xbar.003 _ R=0.004 0.000976)] = 6..004800 Sample Range 0.00227.

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

2 -0.42 -0.70 2.29 … Stat > Time Series > Partial Autocorrelation rtial Autocorrelation Function for Molecular Weight Measurements (Ex9-16mo (with 5% significance limits for the partial autocorrelations) 1. the 5th column.2 0.4 -0.6 0.48 … The decaying sine wave of the ACFs combined with a spike at lag 1 for the PACFs suggests an autoregressive process of order 1. AR(1).0 -0.0 0.Chapter 9 Exercise Solutions 9-16.6 -0. (a) Stat > Time Series > Autocorrelation Autocorrelation Function for Molecular Weight Measurements (Ex9-16mole) (with 5% significance limits for the autocorrelations) 1.92 0.033132 -0.4 0.2 0.8 Autocorrelation 0.658253 -0.0 2 4 6 8 10 Lag 12 14 16 18 Autocorrelation Function: Ex9-16mole Lag 1 2 3 4 5 ACF 0. Note: In the textbook.16 49.8 0.6 0.37 1.055640 T 5.0 2 4 6 8 10 Lag 12 14 16 18 Partial Autocorrelation Function: Ex9-16mole Lag 1 2 3 4 5 PACF 0.2 -0.30 0.74 49.96 -0.220536 0.110802 -0.0 -0.81 44.4 0.658253 0. the 5th row should be “2000” not “2006”.6 -0.0 Partial Autocorrelation 0. 9-20 .072562 -0.84 48.70 -0.23 LBQ 33.4 -0.039599 T 5.8 -1.8 -1.29 -0.373245 0.105969 0.

14. 40. 11. Failed at points: 12. 14. Failed at points: 7. 13. 34. 7. 36. 12. 14. Failed at points: 8.9 2000 6 1975 6 2 1950 2 5 2 1 1 1 1 7 1 1 LCL=1953. 8 points in a row more than 1 standard deviation from center line (above and below CL). 13. 15 3. 8. UCL = 2049. 32. 33. Failed at points: 6.97 1. 14.Chapter 9 Exercise Solutions 9-16 continued (b) x chart: CL = 2001. 11. 15. LCL = 1953 σˆ = MR d 2 = 17. 69 2. 32. 8. 10. with big swings and very few observations actually near the mean. 12.93 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Molecular Weight Measurements (Ex9-16mole) 1 1 2050 UCL=2048. 35.00 standard deviations from center line. 36. 15. 35. 13. 9-21 . 12. 6 points in a row all increasing or all decreasing. 53 5. One point more than 3.128 = 15. 13. violating many runs tests. 9. 37 The process is out of control on the x chart.7 8 6 6 6 6 6 2025 Individual Value 3 _ X=2000. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 9 points in a row on same side of center line. 16. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL).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. 37 8. Failed at points: 12. Failed at points: 7. 31. 70 6.

698 605.1 0.02 Mean 2003. One point more than 3.3 0.364 256.9 0.19 Constant 605.693 4 30929.100 1800. There are no other violations of special cause tests.756 3 32083.000 0.196 Relative change in each estimate less than 0.197 5 30898.697 606.6979 0.0 50 Individual Value 25 _ X=-0.00 standard deviations from center line.4 -75 1 7 14 21 28 35 42 49 Observation 56 63 70 Test Results for I Chart of Ex9-16res TEST 1.693 613.1 0.942 1 41717.0010 Final Estimates of Parameters Type Coef SE Coef T AR 1 0.400 1200.000 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Residuals from Molecular Weight Model (Ex9-16res) 1 UCL=58.843 2 35687.956 7 30897.494 8 30897.196 2. Test Failed at points: 16 Observation 16 signals out of control above the upper limit.550 900. 9-22 .675 650.Chapter 9 Exercise Solutions 9-16 continued (c) Stat > Time Series > ARIMA ARIMA Model: Ex9-16mole Estimates at each iteration Iteration SSE Parameters 0 50173.6 0.4 0.82… P 0.7 0.998 6 30897.698 605.0 0.0852 8.250 1500.1 0.21 7.7 0 -25 -50 LCL=-59.

Chapter 9 Exercise Solutions 9-17. Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Residuals from Molecular Weight Model (Ex9-16res) mu0 = 0.5. Let µ0 = 0.9 Cumulative Sum 50 0 0 -50 LCL=-97. The residuals are in control. h = 5.5. h = 5 100 UCL=97.9 -100 1 7 14 21 28 35 42 Sample 49 56 63 70 No observations exceed the control limit. 9-23 . k = 0. δ = 1 sigma. k = 0.

1 and L = 2.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=-0.7SL=11. Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Residuals from Molecular Weight Model (Ex9-16res) lambda = 0.5 and h = 5).71 0 -5 -10 -2. Let λ = 0. L = 2.7SL=-12.7 (approximately the same as a CUSUM with k = 0. 9-24 .1.7 +2.

046 1946.842 2048.97 1.1)).93 Excel : Workbook Chap09.004 No 1921.084 2012.128 1940.057 1982.700 No 1922.217 2074.654 CL UCL 2000.415 2026. Molecular Weight EWMA Moving Center-Line Control Chart for Molecular Weight 2150.533 2031.384 1995.046 1946.145 No 1996.000 2100.014 1938.731 1984.255 No 1934.848 1993.502 1970.415 2026.393 -0.000 2050.816 1986.722 1996.524 2044.806 1968.859 2029.057 1982.479 2017.281 2065.947 2044.677 No 1969.0762 0.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. To find the optimal λ.749 2092.608 2016.000 1850.304 2017.000 1950.000 1750.09 … P 0.947 2044.574 1954.128 = 15.665 No 1898.9238 σˆ = MR d 2 = 17.574 1954.502 1970.0762 = 0.128 1940.780 No 1899.65 Constant -0.613 No 1978.582 1947.521 0.920 No 1948.480 No 1916.1181 0.1.014 1938. 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.467 1946.930 λ = 1 – MA1 = 1 – 0.282 1964.000 2000.731 1984.269 1994.040 above UCL … Xt.722 1996.772 No 1907.282 1964.929 No 1936.806 1968.211 2.000 1800.376 2002.479 2017.930 1988. Stat > Time Series > ARIMA ARIMA Model: Ex9-16mole … Final Estimates of Parameters Type Coef SE Coef T MA 1 0.326 No 1892.000 1900.582 1947.644 LCL OOC? 1953.467 1946.244 No 1898.1) (= EWMA = IMA(1. fit an ARIMA (0. 9-25 .Chapter 9 Exercise Solutions 9-19.842 2033. No.212 No 1890.

6 0.4 0.23 LBQ 57.4 -0.004498 -0.6 -0.0 Partial Autocorrelation 0.177336 -0.86 144.36 99.46 1.37 3.10 1.46 4.04 -0.8 -1.095134 -0.2 -0.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) 1.03 150.0 2 4 6 8 10 12 14 16 18 20 22 24 Lag Autocorrelation Function: Ex9-20conc Lag 1 2 3 4 5 ACF 0.8 Autocorrelation 0.05 2. 9-26 .4 -0.4 0.0 0.0 -0.746174 0. AR(1).38 127.6 0.77 -0.6 -0.0 -0.520417 0.12 … Stat > Time Series > Partial Autocorrelation Partial Autocorrelation Function for Concentration Readings (Ex9-20conc) (with 5% significance limits for the partial autocorrelations) 1.746174 0.95 -1.58 … The decaying sine wave of the ACFs combined with a spike at lag 1 for the PACFs suggests an autoregressive process of order 1.2 -0.2 0.158358 T 7.0 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 0.2 0.390108 0.8 0.8 -1.238198 T 7.635375 0.

88. 89. 42. 42. 97. 100 TEST 5. 41. 68. 36. 21. 88. 17. 30. 44 The process is out of control on the x chart. Test Failed at points: 15.00 standard deviations from center line. Test Failed at points: 8. 37. 22. 44. 23. 71. 20. 41. 42. 34. 40. 15. 44. Test Failed at points: 10.64 1. 96. 66. 89. 9-27 . 94. 43. 38. 98. 66. 16. 41. 95 TEST 6. Test Failed at points: 15. 19. 39.227 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. 14. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 69. One point more than 3. 41. 69. 95 TEST 2. 43.68 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. 68. 29. 95. 12. 43. 99. 9 points in a row on same side of center line. 93. Test Failed at points: 11. 88. 65. 94. 40. 36.01 11 1 11 30 40 50 60 Observation 70 80 90 100 Test Results for I Chart of Ex9-20conc TEST 1. 29.34 1 20 _ X=200. 36. 86. 28. 43. 18.Chapter 9 Exercise Solutions 9-20 continued (b) σˆ = MR d 2 = 3. 37. 73.128 = 3. 93. with big swings and very few observations actually near the mean. 12. 94. 13. 86. 99 TEST 8. 8 points in a row more than 1 standard deviation from center line (above and below CL). 34. 87. 37. 22. 39. 72. 42. 21. 10. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 39. 38. 89. violating many runs tests. 21. 38. 40.

however this is not unlikely for a dataset of 100 observations.05 0 -5 -10 LCL=-13.76 Mean 200. 14 points in a row alternating up and down. Consider the process to be in control.0669 11. 9-28 .20 Constant 50.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 0.000 0.657 … P 0.7493 0.000 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Residuals from Concentration Model (Ex9-20res) 15 UCL=13.4155 120.73 -15 1 10 20 30 40 50 60 Observation 70 80 90 100 Test Results for I Chart of Ex9-20res TEST 4.122 1.62 Individual Value 10 5 4 _ X=-0.1734 0. Test Failed at points: 29 Observation 29 signals out of control for test 4.

0 2 4 6 8 10 12 14 Lag 16 18 20 22 24 9-29 .4 0.8 0.0 -0.6 0.2 -0.0 -0.6 -0.8 Autocorrelation 0.8 -1.0 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) 1.6 -0.2 0.0 Partial Autocorrelation 0.8 -1.2 0.4 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) 1.2 -0.4 -0.0 0.6 0.

9 Mean StDev N AD P-Value 99 95 Percent 90 -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 0.407 0.343 80 70 60 50 40 30 20 10 5 1 0. 9-30 . PACF and normal probability plot indicates that the residuals are normal and uncorrelated.1 -15 -10 -5 0 Ex9-20res 5 10 Visual examination of the ACF.

δ = 1 sigma.5.Chapter 9 Exercise Solutions 9-21. Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Residuals from Concentration Model (Ex9-20res) mu0 = 0. h = 5 UCL=22.79 Cumulative Sum 20 10 0 0 -10 -20 LCL=-22. Let µ0 = 0. 9-31 . k = 0.79 1 10 20 30 40 50 60 Sample 70 80 90 100 No observations exceed the control limit. k = 0. h = 5.5. and the AR(1) model for concentration should be a good fit. The residuals are in control.

Chapter 9 Exercise Solutions 9-22.051 0 -1 -2 -2.5 and h = 5).7 3 +2. L = 2.7 (approximately the same as a CUSUM with k = 0. Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Residuals from Concentration Model (Ex9-20res) lambda = 0.7SL=2.773 2 EWMA 1 _ _ X=-0.1. Let λ = 0.874 -3 1 10 20 30 40 50 60 Sample 70 80 90 100 No observations exceed the control limit. 9-32 . The residuals are in control.1 and L = 2.7SL=-2.

691 212.047 211. No.979 below LCL 181. xt CL UCL = LCL = The control chart of concentration data signals out of control at three observations (8.243 201.813 198.494 208.7055 σˆ = MR d 2 = 3. fit an ARIMA (0.243 201.418 200.15 … P 0.882 λ = 1 – MA1 = 1 – 0. To find the optimal λ.139 193.1)). Stat > Time Series > ARIMA ARIMA Model: Ex9-20conc … Final Estimates of Parameters Type Coef SE Coef T MA 1 0.366 201.418 200.685 192.562 191.128 = 3.182 0 0 0 0 0 0 0 0 0 … Xt.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 200.010 202.329 193.64 1. 90).506 211.813 198.23 CL 200.xls : Worksheet Ex9-23 lamda = 0.863 190.139 193.660 191.2945 = 0.003 0.558 188.02 Constant -0.239 198.820 203.920 208.0975 3.2945 0.1.660 191.099 209.132 188.010 202.0452 0.432 3.863 UCL = LCL = OOC? 209. 9-33 .544 190.3034 -0.239 198.227 Excel : Workbook Chap09.366 201. 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.144 192.341 200.737 190.924 211.1) (= EWMA = IMA(1.Chapter 9 Exercise Solutions 9-23.825 202. 56.825 202.458 184.

43 0.373763 T 8.47 -1.31… Stat > Time Series > Partial Autocorrelation Partial Autocorrelation Function for Temperature Measurements (Ex9-24temp) (with 5% significance limits for the partial autocorrelations) 1. 9-34 .2 -0.737994 0.66 -0.6 0.71 LBQ 77.31 211.66 4.047106 -0.78 -1.143236 0.078040 -0.8 -1. PACF graph suggest order 1.865899 -0.8 0.4 -0.25 133.0 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 0.0 -0.592580 0.6 0.2 -0. (a) Stat > Time Series > Autocorrelation Autocorrelation Function for Temperature Measurements (Ex9-24temp) (with 5% significance limits for the autocorrelations) 1.865899 0.4 0.0 -0.0 0.0 2 4 6 8 10 12 14 16 18 20 22 24 Lag Autocorrelation Function: Ex9-24temp Lag 1 2 3 4 5 ACF 0.2 0.4 -0.4 0.36 1.112785 T 8.67 3.86 196.0 Partial Autocorrelation 0.8 Autocorrelation 0.13… Slow decay of ACFs with sinusoidal wave indicates autoregressive process.489422 0.2 0.6 -0.94 170.13 2.8 -1.Chapter 9 Exercise Solutions 9-24.6 -0.

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. 16. 20. 18. 17. 18. Test Failed at points: 4. 25. 22. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 19. … Process is out of control. 3. violating many of the tests for special causes. 26. 22.00 standard deviations from center line. 23. 23. Test Failed at points: 65. 23. 3. 16. 39. 4. 22. 32. 19. 24.81 22 22 2 2 500 490 5 8 2 1 _ X=506. 8 points in a row more than 1 standard deviation from center line (above and below CL). 71 TEST 5. 26. 18. Test Failed at points: 2. Test Failed at points: 20. 9 points in a row on same side of center line. 24.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. 33. Test Failed at points: 1. 37. 18. The temperature measurements appear to wander over time. One point more than 3. … TEST 6. 24. … TEST 3. 5. 21. 24. 21. 19. 25. 21. 23.52 6 6 5 6 66 5 52 2 2 2 1 1 1 11 1 5 11 1 LCL=491. 22. 28. 24. 9-35 . 20. … TEST 8. 20. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 19. 38. 21. 21. 34. 6. 22. 27. Test Failed at points: 17. 6 points in a row all increasing or all decreasing. 23. … TEST 2. 36. 17. 27.

80 1 1 10 20 30 40 50 60 Observation 70 80 90 100 Test Results for I Chart of Ex9-24res TEST 1.985 … P 0.0480 18. One point more than 3.3794 0.23 20 5 Individual Value 10 _ X=0. The chemical process is in control.67 Constant 52. Test Failed at points: 94 TEST 5. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).8960 0. 9-36 .12 Mean 503. and observation 71 fails Test 5.727 6.22 0 -10 -20 LCL=-21.00 standard deviations from center line. The residuals do not exhibit cycles in the original temperature readings.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. and points are distributed between the control limits.000 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Residuals from Temperature Model (Ex9-24res) UCL=22. Test Failed at points: 71 Observation 94 signals out of control above the upper limit.7263 72.000 0.

5.69 30 Cumulative Sum 20 10 0 0 -10 -20 -30 LCL=-36. 9-37 .69 -40 1 10 20 30 40 50 60 Sample 70 80 90 100 No observations exceed the control limits. indicating the process is in control.Chapter 9 Exercise Solutions 9-25. h = 5 40 UCL=36. The residuals are in control. This is the same conclusion as applying an Individuals control chart to the model residuals. MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Residuals from Temperature Model (Ex9-24res) k = 0.

22 0.7SL=4.5 _ _ X=0.Chapter 9 Exercise Solutions 9-26. L = 2. This is the same conclusion as applying the Individuals and CUSUM control charts to the model residuals.0 1 10 20 30 40 50 60 Sample 70 80 90 100 No observations exceed the control limits. indicating the process is in control.76 EWMA 2.1. MTB > Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of Residuals from Temperature Model (Ex9-24res) lambda = 0.7SL=-4.33 -5.0 -2.0 +2. 9-38 . The residuals are in control.5 -2.7 5.

647 498.10 … P 0.6784 -0.267 509.227 below LCL 476. To find the optimal λ.0794 = 0.128 = 5.558 498. similar to the two out-of-control signals on the Individuals control chart (71.722 509.232 482.940 467.355 483.850 513.9206 σˆ = MR d 2 = 5.520 492.647 498.655 518.972 506. 69) and the lower limit (1.560 524.Chapter 9 Exercise Solutions 9-27.75 1.917 λ = 1 – MA1 = 1 – 0.173 0 0 0 0 0 0 0 0 0 EWMA Moving Center-Line Chart for Temperature Xt.558 498.070 487.812 489.232 482.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. fit an ARIMA (0.0975 (from a Moving Range chart with CL = 5.265 521.0711 0.525 498.813 507.758 OOC? 491.429 494.265 483.921 sigma^ = zt 506.1)).520 492.098 UCL 506. 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.0794 0.679 491. 9-39 .267 509. 58.1019 0.972 506. 94). Stat > Time Series > ARIMA ARIMA Model: Ex9-24temp … Final Estimates of Parameters Type Coef SE Coef T MA 1 0.812 489.465 LCL 521.1.037 CL 5.940 514.105 504.137 479.465 509.974 494.363 503. 94).363 503.1) (= EWMA = IMA(1.520 474.75) Excel : Workbook Chap09. xt CL UCL LCL A few observations exceed the upper limit (46.429 494.438 0.78 Constant -0.

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

17 -2.8 -1.0 -0.6 0.4 0.78 41. 9-41 .41 32.16 25.071963 T 4.8 Autocorrelation 0.94 -0.0 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. There is a serious problem with autocorrelation in viscosity readings. (a) Stat > Time Series > Autocorrelation Autocorrelation Function for Viscosity Readings (Ex9-29Vis) (with 5% significance limits for the autocorrelations) 1.29 41.85 r1 = 0.2 0.049610 -0.54 LBQ 25.4 -0.0 0.Chapter 9 Exercise Solutions 9-29.41 -2.283150 -0.49.22 -0.264612 -0.6 -0.2 -0. indicating a strong positive correlation at lag 1.494137 -0.

11 5 Individual Value 35 6 6 7 30 _ X=28. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 59. 9-42 . Test Failed at points: 40. 8 points in a row more than 1 standard deviation from center line (above and below CL). Test Failed at points: 2. 63.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. 64. 60.03 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. Test Failed at points: 38. 38. 75 TEST 7. 86 TEST 6.57 25 6 20 6 55 LCL=20. 86. The viscosity measurements appear to wander over time. Test Failed at points: 22 TEST 8. Test Failed at points: 64 Process is out of control. violating many of the tests for special causes. 15 points within 1 standard deviation of center line (above and below CL).00 standard deviations from center line. 58. 92 TEST 5. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL).

569 MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM CUSUM Chart of Viscosity (Ex9-29Vis) target = 28.5.Chapter 9 Exercise Solutions 9-29 continued (c) Let target = µ0 = 28. k = 0.24 -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.569.24 10 0 0 -10 LCL=-14. 9-43 . h = 5 20 Cumulative Sum UCL=14.

759 EWMA 30 _ _ X=28.569 29 28 27 -2.15. There are wide swings in the plot points and several are beyond the control limits.7 32 31 +2. 9-44 .7SL=26.380 26 25 1 10 20 30 40 50 60 Sample 70 80 90 100 The process is not in control.Chapter 9 Exercise Solutions 9-29 continued (d) MTB > Stat > Control Charts > Time-Weighted Charts > CUSUM EWMA Chart of Ex9-29Vis lambda = 0.7SL=30. L = 2.

367 20.937 20.xls : Worksheet Ex9-29 lambda = l Xi 0 1 2 3 4 5 6 7 8 9 10 … 29.922 below LCL 9.000 10.665 26.4839 0.1)).482 26.940 18.464 18.500 26.841 41.007 27.1007 -1.898 33.356 25.760 29.05 P 0.479 29.441 42.560 27.1579 σˆ = MR d 2 = 3.0231 0.464 18.Chapter 9 Exercise Solutions 9-29 continued (e) To find the optimal λ.75) Excel : Workbook Chap09.069 35.873 39.1579 0.207 35.898 33.785 31.122 17.330 31. Xi CL UCL LCL A few observations exceed the upper limit (87) and the lower limit (2.025 35.984 18.243 0 0 0 0 0 0 0 0 0 EWMA Moving Center-Line Chart for Viscosity Xt. No. fit an ARIMA (0.1.750 30.000 40.1) (= EWMA = IMA(1. 85). 37. 9-45 .527 26.144 1.980 25.680 33.330 19.128 = 2.909 29.57 Constant 0.330 31.527 26.21 1.022 38.756 24.85 LCL OOC? 19.452 37.8457 (from a Moving Range chart with CL = 5.962 λ = 1 – MA1 = 1 – (– 0.030 32.158 sigma^ = CL UCL 28.000 30.298 32.000 20.000 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 Obs.482 26. Stat > Time Series > ARIMA ARIMA Model: Ex9-29Vis … Final Estimates of Parameters Type Coef SE Coef T MA 1 -0.550 Zi 28.298 32.000 0. 55.665 26.000 31.328 2.909 29.120 0. Viscosity 50.1579) = 1.785 37.479 29.788 22.

54 Mean 28. 19.3278 62.4581 … P 0.000 0. 21.79 AR 2 -0. with a good distribution of points between the control limits and no patterns. Otherwise the process appears to be in control.0922 -4.0923 7.72 Constant 20.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 0.000 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Residuals from Viscosity AR(2) Model (Ex9-29res) 10 UCL=9.04 0 7 7 -5 LCL=-9.6514 0.60 Individual Value 5 7 7 7 _ X=-0. 20.7193 0. 22 The model residuals signal a potential issue with viscosity around observation 20.4349 0.000 0.5017 0. 15 points within 1 standard deviation of center line (above and below CL). 9-46 . Test Failed at points: 18.68 -10 1 10 20 30 40 50 60 Observation 70 80 90 100 Test Results for I Chart of Ex9-29res TEST 7.

01/hr or 1/λ = 100hr.0705 h λ h2 = 0. a4 = $100/hr g = 0.210.79/hr (b) n = 3. h = 1.xls : worksheet Ex9-30a n = 5. a2 = $0.472) − Φ (−7.27 −λh λh 1− e ( ) E(L) = $3. hopt = 1.895 E(L) = $3.0 a1 = $0.4992 τ≅ − 2 12 α e−λh α ≅ = 0. a'3 = $5. α = 0.05hr/sample.0705 − 0. δ = 2.0027 ⎛ ( 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 (−1.10/unit.027. 1 − β = 0. α = 0. λ = 0.50. k = 3. kopt = 2.231.472) = 0.6098/hr 9-47 . D = 2hr (a) Excel : workbook Chap09.0000 = 0.Chapter 9 Exercise Solutions 9-30. a3 = $2.00.50/sample.

2498 τ≅ − 2 12 2 12 0.5 0.01392/hr 9-48 .0027 α e−λh α ≅ = = 0.0027. a4 = $100/hr g = 0.0705 h λ h 2 0.01(0. D = 2hr (a) Excel : workbook Chap09.5. a3 = $25.0027 ≅ = = 0. δ = 2.918 E(L) = $4.0 a1 = $0.05hr/sample.01(0.50/sample.0000 = 0. β = 0.368.01(1) 1− e ( ) E(L) = $4. hopt = 1.00207.472) − Φ (−7. a'3 = $50.10/unit. λ = 0.5) 1− e ( ) E(L) = $4.54 −λh λ h 0.xls : worksheet Ex9-31 n = 5.472) = 0. α = 0.01/hr or 1/λ = 100hr.52 ) = − = 0.27 −λh λ h 0.080. α = 0. h = 1.0027 ⎛ ( 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.Chapter 9 Exercise Solutions 9-31. kopt = 3. α = 0.98/hr (c) n = 5. a2 = $0. 1 − β = 0.4992 τ≅ − 2 12 2 12 α e−λh α 0. k = 3. k = 3.01(12 ) = − = 0.0705 − 0. h = 0.12/hr (b) n = 5.0705 h λ h 2 1 0.

hopt = 2.xls : worksheet Ex9-32 D0 = 2hr.16/hr 9-33. α = 0.52 ) = − = 0. k = 3. D1 = 2hr V0 = $500.xls : worksheet Ex9-33 λ = 0.01(0.54 −λh λ h 0.240. k = 3.Chapter 9 Exercise Solutions 9-32. α = 0. kopt = 2. h = 1.0027 β = Φ k − δ n − Φ −k − δ n ( ) ( ) = Φ ( 3 − 1 5 ) − Φ ( −3 − 1 5 ) = Φ (−1. Excel : workbook Chap09.0000 = 0. ∆ = $25 n = 5.39762/hr 9-49 .0 a1 = $2/sample a2 = $0.775 h λ h 2 0.01(0.472) − Φ (−7.5. α = 0.2498 τ≅ − 2 12 2 12 0.5) 1− e ( ) E(L) = $16.0705 E(L) = $13.8218083 E(L) = $10. 1 − β = 0.489018.05 hr/sample D = 1 hr (a) n = 5.01/hr or 1/λ = 100hr δ = 2.775 − 0.472) = 0.5 0. Excel : workbook Chap09.50/unit a'3 = $75 a3 = $50 a4 = $200/hr g = 0.17/hr (b) n = 10.0027 α e−λh α ≅ = = 0.025091. β = 0. h = 0.0027.

00064 2.0015 1 2 3 4 5 6 7 8 9 10 11 12 Ex9-34Sample n = 5.00028) = 1.0035 − 3(0. Graph > Time Series Plot > With Groups Time Series Plot of Ex9-34Xb USL = 1.00028 CL = R = 0.00196 x chart at tool reset: CL = USL − 3σ = 1.00272 LCL = CL − 3σ x ( ) = 1. R = 0.326 = 0.Chapter 9 Exercise Solutions 9-34. σˆ = R d 2 = 0.00234 + 3 0.00064) = 0.00234 UCL = CL + 3σ x = 1.00028 5 ) = 1.0020 LSL = 1.00064.00028) = 1.00266 − 3 ( 0.00228 9-50 .00304 LCL = CL − 3σ x ( ) = 1.114(0.0035 Ex9-34Reset A fter Before Ex9-34Xb 1. It is good practice visually examine data in order to understand the type of tool wear occurring.0015 1.0015 + 3(0.00064.0030 1.00028 5 = 1.0025 1.0035 1.00028 5 ) = 1. LCL = 0 x chart initial settings: CL = LSL + 3σ = 1.00266 (maximum permissible average) UCL = CL + 3σ x = 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.00234 − 3 ( 0.00028 5 = 1.00266 + 3 0.00135. UCL = D4 R = 2.

1268 3.5211 0.1850 0 14. so many of these exercises have been solved in Excel.0451 1. Phase 2 T 2 control charts with m = 50 preliminary samples.1850 0 14.948) = 14.1850 0 14. Let α = 0.6338 14. p .mn − m − p +1 mn − m − p + 1 2(50 + 1)(25 − 1) = F0.9127 22. 10-1 .1850 0 14.1268 0.7042 0.1690 0.8169 0. 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. p (m + 1)(n − 1) UCL = Fα .1850 0 14.00 20.Chapter 10 Exercise Solutions Note: MINITAB’s Tsquared functionality does not use summary statistics.00 30.1850 0 14.00 10.2.001.0817 2.1850 0 14.0676 1.00 T^2 40.00 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 14.5408 13. 10-1.1199 50(25) − 50 − 2 + 1 = ( 2448 1199 ) (6.00 50.1850 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. p = 2 characteristics.6901 0.186 LCL = 0 Excel : workbook Chap10.xls : worksheet Ex10-1 Sample No.00 0.1690 0.0423 0.1850 0 14.4254 10. n = 25 sample size.0282 0.001.1268 3.

0808 0.00 0.1271 11.2 0 -0.3719 0.1189 1.6574 6.4 -0.2706 0.5 5 3 3. n = 10 sample size.mn − m − p +1 mn − m − p + 1 3(30 + 1)(10 − 1) = F0.1 0.3 0 -0.3 0.6 -1 -0.2 -0. p (m + 1)(n − 1) UCL = Fα .4 -0.5279 0.2 3.4249 17.2 0. Phase 2 T 2 control limits with m = 30 preliminary samples.7 0.4249 17.1399 1.5741 2.5 0.4249 17.00 10.7 0.0528 0.00 T^2 15.1 0.6 3 2.4249 17.6844 0.2372 2.1 4.9 -1 0.4249 17.8 11 3. Let α = 0.4249 17.2 0.00 5.8 diff1 diff2 diff3 0.0684 0.3 2.1 1.001.4249 17.4249 17.3966 1.1216 0.0793 20.2 0.8 3 2.9 3 12 3.3 -0.3 3.1887 0.5 0. p = 3 characteristics.8525 matrix calc t2 = n * calc UCL = LCL = OOC? 17.9 15 3.1 3 2.2 3.2 0.4249 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 25.2 0.4 4 2.2 3 8 3 3.7 4 3 13 4.6 3.6922 0.00 20.0397 10.00 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.xls : worksheet Ex10-2 Sample No.2 0.6 -0.0852 0.4 3 2.001.1 -0.3122 3. xbar1 xbar2 xbar3 1 3.8 6 4 4.4249 17.2 -0.8 4 2.5 -0.8084 1.425 LCL = 0 Excel : workbook Chap10.4249 17.579) ⎝ 268 ⎠ = 17.7 9 2.8 4.8692 8.9 3.1 0.7 3 2 3.6 1.4 0.0593 10.1 0 0.2 10 2 2.1880 1.2 14 3.8 0.6 2.1 3.3.7927 1.3 2.4 0.7 0.5 -0.4249 17.2 0 1 1. p .3990 0.4249 17.2 1.4249 17.Chapter 10 Exercise Solutions 10-2.5 -0.5 7 3.268 30(10) − 30 − 3 + 1 ⎛ 837 ⎞ =⎜ ⎟ (5.5932 0.8 0.8800 0. 10-2 .4 0.7 3.4249 17.

the chi-square formula will be used for the upper 2 = 13.001.2 Excel : workbook Chap10.0676 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 60.001.2676 6.8150 0 13.Chapter 10 Exercise Solutions 10-3.8150 0 13.8150 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 0.0000 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 14.8150 0 13. Since population parameters are known.0761 1.7042 0.3380 13.0000 T^2 40.8150 0 13.1690 0.0000 0.8150 0 13. p = χ 0.8169 0.6901 0.8150 0 13.1268 0.816 control limit: UCL = χα2 .0817 2.9014 2. Phase 2 T 2 control limits with p = 2 characteristics.0676 1.0000 20.8150 0 13.0000 10.2028 5.1268 3. Same results as for parameters estimated from samples.1127 52.0000 50.6901 0.1690 0.5211 0.8150 0 13.8150 0 13. Let α = 0.2535 6.0000 30.0451 1.0704 0. 10-3 .0423 0.0282 0.5408 13.8150 0 13.xls : worksheet Ex10-3 Sample No.4254 10.6338 0.8150 0 13.9127 22.8169 0.8150 0 13.1268 3.8150 0 13.6901 0.

0397 10.2 14 3.8 4.xls : worksheet Ex10-4 Sample No. xbar1 xbar2 xbar3 1 3.2660 16. 10-4 .2706 0.1 4. p = χ 0.2660 16.2 10 2 2.7 9 2.2372 2.5 5 3 3.266 control limit: UCL = χα2 .2660 16.4 3 2.4 0.3 2.1271 11.2660 16.5741 2.2660 16.2660 16.7 3 2 3.7 0.5 -0.2660 16.00 T^2 15.6922 0.2660 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 25.6844 0.6 -1 -0.4 -0.0528 0.5 0.8692 8.1887 0.9 3.1 0. the chi-square formula will be used for the upper 2 = 16.3990 0.2 3 8 3 3.8525 matrix calc t2 = n * calc UCL = LCL = OOC? 16.2 0.8 0.3 2.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sample Number Process is out of control at sample 13.4 0.0852 0.3122 3.2660 16.2 0.001.2660 16.7 0.2660 16.3966 1.1189 1.2660 16.8 11 3.1880 1. Phase 2 T 2 control limits with p = 3 characteristics.0684 0.4 -0.8 6 4 4.1 3 2.2660 16.8 4 2.00 0.8 0.1399 1.2 -0.2 0 -0.7927 1.1 3.8 3 2.6 1.0593 10.001.2 0 1 1.5 7 3.2 3.5 -0.6574 6.3719 0.2 0.3 Excel : workbook Chap10.9 15 3.00 20.6 3 2.3 0. Same as results for parameters estimated from samples.0793 20.3 -0.2 3.2 0.1 0.2 -0.5279 0.2 0.6 -0. Let α = 0.5 0.6 3.4 0.6 2.2 1.8 diff1 diff2 diff3 0.1216 0.00 5.1 -0.2660 16.00 10.1 0.1 0 0.2660 16.3 0 -0.9 3 12 3.4 4 2.7 4 3 13 4.2 0.8084 1.7 0. Since population parameters are known.Chapter 10 Exercise Solutions 10-4.5932 0.7 3.9 -1 0.0808 0.8800 0.5 -0.1 1.3 3.

407 3.7331 18. Excel : workbook Chap10.5042 20.223 3.005. (c) Quality characteristics.548 The Phase II UCL is almost 30% larger than the chi-square limit. 1.548) = 18.005 (a) Phase II limits: p (m + 1)(n − 1) UCL = Fα .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. p = 6.107 3.3527 21.107 18.005.005.Chapter 10 Exercise Solutions 10-5.9447 717 718 719 720 721 722 8616 8628 8640 8652 8664 8676 3.6 = 18.7328 18. 10-5 .107 3.338 3.107 3.294 3.107 3.209 UCL 23. m = 30 preliminary samples.882 LCL = 0 (b) 2 chi-square limit: UCL = χα2 .240 3.3184 20.7332 18. p = 6 characteristics.9650 20.6.7324 … 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.8820 22.1095 19. n = 3. α = 0.55 30(3) − 30 − 6 + 1 ⎛ 372 ⎞ =⎜ ⎟ (3.7325 18.mn − m − p +1 mn − m − p + 1 6(30 + 1)(3 − 1) = F0.01(18. n = 3 sample size. Find "m" such that exact Phase II limit is within 1% of chi-square limit.733. Samples size.531 3.7337 18. p = χ 0. α = 0.5920 20. p .531) ⎝ 55 ⎠ = 23.107 3.263 3.

p .106 3. n = 5.294) ⎝ 115 ⎠ = 21.548) = 18. p = χ 0. 1.174 3.7324 18.7376 18.105 3.189 3. p = 6.105 3.115 30(5) − 30 − 6 + 1 ⎛ 744 ⎞ =⎜ ⎟ (3. (c) Quality characteristics. p = 6 characteristics.8641 19.005.240 3.548 The Phase II UCL is almost 15% larger than the chi-square limit.3087 20.6685 19.01(18.105 18.155 3.7330 18.7318 … 411 preliminary samples must be taken to ensure that the exact Phase II limit is within 1% of the chi-square limit.Chapter 10 Exercise Solutions 10-6.5237 19. Samples size.733.1422 19. 10-6 .4119 19.294 3.6. Find "m" such that exact Phase II limit is within 1% of chi-square limit.005.7424 18.005.309 (b) 2 chi-square UCL: UCL = χα2 .5692 20.mn − m − p +1 mn − m − p + 1 6(30 + 1)(5 − 1) = F0. m = 30 preliminary samples.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. α = 0.164 3. n = 5 sample size.3232 390 400 410 411 412 9384 9624 9864 9888 9912 1555 1595 1635 1639 1643 3.149 UCL 21. α = 0. Excel : workbook Chap10.005 (a) Phase II UCL: p (m + 1)(n − 1) UCL = Fα .6 = 18.209 3.

Samples size. n = 3 sample size.677 2.005 (a) Phase II UCL: p (m + 1)(n − 1) UCL = Fα .742 2.3259 34.3808 28. α = 0.188) = 25. n = 3.799 2.7953 30. m = 25 preliminary samples.1991 31. Find "m" such that exact Phase II limit is within 1% of chi-square limit.326 (b) 2 chi-square UCL: UCL = χα2 . 1.005. p = 10.005.897 2. p = 10 characteristics.4398 25.629 UCL 39.4401 25. Excel : workbook Chap10.188 The Phase II UCL is more than 55% larger than the chi-square limit.530 2.530 2.4394 … 988 preliminary samples must be taken to ensure that the exact Phase II limit is within 1% of the chi-square limit.41 25(3) − 25 − 10 + 1 ⎛ 520 ⎞ =⎜ ⎟ (3. α = 0.4024 29.4405 25.4940 28.Chapter 10 Exercise Solutions 10-7.657 2. p = χ 0.10.440.530 2.0154 27.10 = 25. (c) Quality characteristics.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.mn − m − p +1 mn − m − p + 1 10(25 + 1)(3 − 1) = F0.101) ⎝ 41 ⎠ = 39. 10-7 .641 2.530 25. p .01(25.101 2.8549 28.005.704 2.4399 25.530 2.7246 986 987 988 989 990 19740 19760 19780 19800 19820 1963 1965 1967 1969 1971 2.

n = 5. (c) Quality characteristics.4419 25. α = 0.529 2.529 25.5595 1840 171 2.529 2.Chapter 10 Exercise Solutions 10-8.529 2. 1. n = 5 sample size.578 26.005.188 The Phase II UCL is more than 25% larger than the chi-square limit.4405 25.10 = 25.529 2.623 27. Samples size.594 27.4408 25.4141 3040 291 2.767) ⎝ 91 ⎠ = 31.4413 25.01(25. p = χ 0.005 (a) Phase II UCL: p (m + 1)(n − 1) UCL = Fα . p = 10 characteristics.xls : worksheet Ex10-8 m 25 35 45 55 65 75 85 95 105 num denom F UCL 1040 91 2.529 2.606 27.6251 1440 131 2.005.689 29.625 (b) 2 chi-square UCL: UCL = χα2 .4394 544 preliminary samples must be taken to ensure that the exact Phase II limit is within 1% of the chi-square limit.188) = 25. Find "m" such that exact Phase II limit is within 1% of chi-square limit. Excel : workbook Chap10.1011 3440 331 2.572 26. 10-8 .767 31.440. m = 25 preliminary samples.8495 2640 251 2.4399 25.mn − m − p +1 mn − m − p + 1 10(25 + 1)(5 − 1) = F0. α = 0.648 28.5335 540 541 542 543 544 545 21640 21680 21720 21760 21800 21840 2151 2155 2159 2163 2167 2171 2.4967 2240 211 2.8651 3840 371 2.91 25(5) − 25 − 10 + 1 ⎛ 1040 ⎞ =⎜ ⎟ (2.585 26.6812 4240 411 2. p = 10. p .005.10.

4 = 13.788) ⎝ 41 ⎠ = 35. n = 3 sample size.360 10-10.41 25(3) − 25 − 10 + 1 ⎛ 520 ⎞ =⎜ ⎟ (2.01. p .10.7 1 0. p .Chapter 10 Exercise Solutions 10-9.7 1 0.7 0.277 10-9 . Assume α = 0. p = 10 quality characteristics.7 1 ⎦ (b) 2 UCL = χα2 .7 ⎥ ⎢ ⎥ ⎣0.41 25(3) − 25 − 10 + 1 ⎛ 480 ⎞ =⎜ ⎟ (2.7 ⎤ ⎢0.788) ⎝ 41 ⎠ = 32.7 0.01.638 Phase II UCL: p (m + 1)(n − 1) UCL = Fα .xls : worksheet Ex10-10 (a) ⎡ 1 0. m = 25 preliminary samples.7 0.mn − m − p +1 mn − m − p + 1 10(25 − 1)(3 − 1) = F0.01.7 0.10.7 ⎥ ⎥ Σ=⎢ ⎢0.mn − m − p +1 mn − m − p + 1 10(25 + 1)(3 − 1) = F0.7 0.01. p = χ 0. Phase I UCL: p (m − 1)(n − 1) UCL = Fα . Excel : workbook Chap10.7 0.

5⎦ ⎣ 0 ⎦ ⎠ ⎣ 0.689 Investigate variables 1 and 3. d1 = 12.5⎥ ⎢ 0⎥ ⎟ ⎥ ⎜⎢ ⎥ ⎢ ⎥⎟ 0.635 2 No.300 T(3)2 = 14. an out-of-control signal is generated.5⎤ ⎡ 0⎤ ⎞ ⎜ ⎟ 1 0.7 ⎤ −1 ⎛ ⎡3.5⎤ ⎡0 ⎤ ⎞′ ⎡ 1 0.635 T(1)2 = 15.313 2 2 2 2 T(1) = T(2) = T(3) = T(4) = 15. Second.277 ) .5⎥ ⎢ 0 ⎥ ⎟ ⎢ 0. since all di are smaller than χ 0.7 1 0.01.7 0.1 . an out-of-control signal is generated.277) .7 1 ⎦ ⎜⎝ ⎣3.7 0.µ (1) )′ Σ (1) ( y (1) .5⎤ ⎡ 0 ⎤ ⎞′ ⎡ 1 ⎜⎢ ⎥ ⎢ ⎥⎟ ⎢ 3.806 ( 0.5⎦ ⎣ 0⎦ ⎟⎠ ) > ( UCL = 13.µ ) ⎛ ⎡3. since the standardized observations are equal (that is.7 ⎤ −1 ⎛ ⎡3.7 0.5⎤ ⎡ 0⎤ ⎞ ⎜ ⎟ ⎜ ⎟ = 1⎜ ⎢⎢3.5 0 0.7 ⎥ ⎜ ⎢3. all variables had the same shift). (d) −1 T(1)2 = n ( y (1) .7 = 1⎜ ⎢ ⎥ − ⎢ ⎥ ⎟ ⎢ ⎜ ⎢3.7 0. no variable is identified as a relatively large contributor.493 2 χ 0.313 = 0.5⎥ ⎢ 0⎥ ⎟ ⎦ ⎝⎣ ⎦ ⎣ ⎦⎠ ⎝⎣ ⎦ ⎣ ⎦⎠ ⎣ = 15.5⎥⎥ ⎢⎢ 0⎥⎥ ⎟ − 0. d 2 = 6.313 di = T 2 − T(2i ) d1 = d 2 = d3 = d 4 = 15.1 = 6.1 = 6.7 0.5⎥ ⎢0 ⎥ ⎟ ⎢ 0.7 ⎥⎥ ⎜ ⎢⎢3. Since T 2 = 15.µ )′ Σ −1 ( y .µ (1) ) ⎛ ⎡3.694.Chapter 10 Exercise Solutions 10-10 continued (c) T 2 = n ( y .7 1 ⎥ ⎜ ⎢3.7 1 0. this information does not assist in identifying which a process variable shifted.806 − 15.7 ⎜⎜ ⎢ ⎥ ⎢ ⎥ ⎟⎟ ⎢ ⎝ ⎣3.5⎥⎥ − ⎢⎢0 ⎥⎥ ⎟ ⎢⎢ 0.01. 10-10 . d3 = 13.479.806 Yes. (f) 2 χ 0. (e) Since (T 2 = 28. d 4 = 2.979.800 2 T(4) = 25.280) > (UCL = 13.585 2 T(2) = 21.7 0. First.5⎥⎥ − ⎢⎢ 0⎥⎥ ⎟ ⎜ ⎢3.7 ⎥⎥ ⎜ ⎢⎢3.7 = 15.590.01.

815 ) . d3 = 6.8 1 ⎥⎦ (b) 2 UCL = χα2 .8 0.000.111.778.154 > ( UCL = 7.3 = 7.05.154 Yes. This is confirmed since none of the di’s exceeds the UCL.538 2 T(2) = 5.841 T(1)2 = 5.1 = 3. d 2 = 1.05.815 (c) T 2 = 11.094 Since an out-of-control signal was not generated in (e).8⎤ Σ = ⎢⎢ 0.000.538 T(3)2 = 4.841 T(1)2 = 11. ( ) (d) 2 χ 0. d1 = 1.8 1 0.xls : worksheet Ex10-11 (a) ⎡ 1 0.376 T(3)2 = 5.05.815) .8 0. d3 = 2. d 2 = 8. d1 = 0. Excel : workbook Chap10. an out-of-control signal is not generated. an out-of-control signal is generated. (e) Since (T 2 = 6.154 Variables 2 and 3 should be investigated. it is not necessary to calculate the diagnostic quantities. Since T 2 = 11.8⎥⎥ ⎢⎣ 0.043 2 T(2) = 2.Chapter 10 Exercise Solutions 10-11. 10-11 .000. p = χ 0.1 = 3.538) > (UCL = 7. (f) 2 χ 0.444.

016 0.104 88.001 −0.101 −0. Excel : workbook Chap10. S 2 = ⎢⎢ −0.607 21.256 43.xls : worksheet Ex10-12 m = 40 ⎡ 4.016 0.012 7.158 10-12 .395 −0.003 0.282 3.207 S1 3.256 ⎤ ⎡ 1.561⎤ ⎡121.016 5.071 ⎦ ⎣ −0.740 80.339 0.950 ⎥ .xls : worksheet Ex10-13 m = 40 ⎡ 4.003 −0. S1 = ⎢⎢ −0.150 S2 2.720 0.104] .392 1.xls : worksheet Ex10-14 xbar xbar1 xbar2 10. S1 = ⎢ ⎥ ⎣ −0.001 0.641 V'V 133.950 587.395 ⎤ x′ = [15.001 ⎦ 10-13.256 0.016 ⎤ x′ = [15.740 67.003 0.339 0.305 5.307 1.001 ⎦ ⎡121.101 −0.Chapter 10 Exercise Solutions 10-12.526 ⎥⎦ ⎢⎣ 43.780 80. S2 = ⎢ V′V = ⎢ ⎥ ⎥ ⎣ −0.014 ⎥⎥ ⎢⎣ 5. Excel : workbook Chap10.553 −0.012 ⎥⎥ ⎢⎣ −0.071 0.125] .000 ⎥⎦ 10-14.003⎤ .561 0.440 −0. Excel : workbook Chap10.440 −0.720 ⎤ ⎢ ⎥ V′V = ⎢ −0.305 3.599 ⎥⎦ ⎡ 1.014 27.553 −0.392 1.256 0.

9231 -0.9231 3. Excel : workbook Chap10.75 1 0. This gives an ARL1 between 7.5 delta = 0.Chapter 10 Exercise Solutions 10-15.75 0.9231 -0.17.75 1 1 1 1 1 y' Sigma-1 = 0.75 0.75 0.231 delta = 1.22 and 12.308 0.0769 1 1 1 1 From Table 10-3.308 0.73 13.9231 3.53 ARL1 = Select λ = 0.xls : worksheet Ex10-15 p= mu' = Sigma = y' = 4 0 0 0 0 1 0.75 0.75 0.1 0.87 UCL = H = 12.75 0.75 0.9231 -0.9231 -0. select (lambda.0769 -0.308 0.0769 -0.9231 -0. 10-13 .1 with an UCL = H = 12.75 0.75 0.9231 y= -0. H) pair that closely minimizes ARL1 1 1.9231 -0.9231 -0.109 ARL0 = 200 Sigma-1 = 3.2 lambda = 12.9231 3.9231 -0.73.308 y' Sigma-1 y = 1.75 1 0.0769 -0.17 6.

26.9 0.432 7.432 7.568 -2.9 0.9 0.65 Select λ = 0.xls : worksheet Ex10-16 p= mu' = Sigma = y' = 4 0 0 0 0 1 0.432 -2.105 with an UCL = H = 15.432 7.9 1 0.60.60 7.270 y' Sigma-1 y = 1.9 1 0.105 0.5 lambda = 0.18 UCL = H = 15.Chapter 10 Exercise Solutions 10-16. H) pair delta = 1 1.081 delta = 1.9 0.9 0.568 1 1 1 1 From Table 10-4.568 -2.270 0.26 16.432 -2.568 -2.040 ARL0 = 500 Sigma-1 = y= 7.432 -2.432 -2.432 -2.9 0.03 ARLmin = 14.9 0.270 0.270 0.9 0.432 -2.9 1 1 1 1 1 y' Sigma-1 = 0. This gives an ARLmin near 14.9 0.432 -2. select (lambda.432 -2. 10-14 .432 -2. Excel : workbook Chap10.

20.08 UCL = H = 10.3 lambda = 8. H) pair that closely minimizes ARL1 1 1 1.64 9.8 0.65.20 5.Chapter 10 Exercise Solutions 10-17.49 5.65 9.5 delta = 0.7778 -2.2222 -2. Excel : workbook Chap10.556 y' Sigma-1 y = 1.054 ARL0 = 200 Sigma-1 = 2.xls : worksheet Ex10-17 p= mu' = Sigma = y' = 2 0 0 1 0.7778 y= 1 1 From Table 10-3.2 0.2 0.5 1. 10-15 .15 10. This gives an ARL1 between 5.556 0.48 ARL1 = Select λ = 0.65 10. select (lambda.2 with an UCL = H = 9.111 delta = 1.2222 2.1 0.49 and 10.8 1 1 1 y' Sigma-1 = 0.

10 TEST 5. 37. One point more than 3.Chapter 10 Exercise Solutions 10-18. 10. the y2 values for Observations 8. Test Failed at points: 8. (a) Note: In the textbook Table 10-5.00 standard deviations from center line. 37. 9. 40 TEST 6.12 5 6 95 _ X=91. 103. 40 TEST 8. and 107. 39. Test Failed at points: 5. Test Failed at points: 9. Test Failed at points: 40 10-16 . 35. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). 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. 38. 39.25 90 6 85 6 6 5 5 5 5 LCL=82. 11. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 10. 12. 8 points in a row more than 1 standard deviation from center line (above and below CL). 36. 11. and 10 should be 100.38 80 4 8 12 16 20 24 Observation 28 32 36 40 Test Results for I Chart of Tab10-5y2 TEST 1. 9.

Tab10-5x2.5 _ X=-0. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 21.Chapter 10 Exercise Solutions 10-18 continued (b) Stat > Regression > Regression Regression Analysis: Tab10-5y2 versus Tab10-5x1.6 Tab10-5x2 + 0.00 standard deviations from center line.435 Tab10-5x3 + 0.57 5 5 Individual Value 5. 25 TEST 6..57 4 8 12 16 20 24 Observation 28 32 36 40 Test Results for I Chart of Ex10-18Res TEST 1.0.215 Tab10-5x9 Stat > Control Charts > Variables Charts for Individuals > Individuals I Chart of Regression Model Residuals (Ex10-18Res) 7.11.2 Tab10-5x5 + 0.1 Tab10-5x7 + 10.0 2. Test Failed at points: 7.0 5 LCL=-6.5 1 1 UCL=6.192 Tab10-5x4 . The regression equation is Tab10-5y2 = 215 . Test Failed at points: 19. 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.73 Tab10-5x6 + 6.666 Tab10-5x1 . 10-17 . 18 TEST 5.00 0.. One point more than 3.0 -2.9 Tab10-5x8 . . 4 out of 5 points more than 1 standard deviation from center line (on one side of CL).3.5 -5.

8 -1. 10-18 . while the ACF for the residuals suggests a random process.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) 1.2 -0.4 -0.8 -1.4 0.2 0.6 0.6 -0.0 0.8 Autocorrelation 0.8 Autocorrelation 0.0 0.6 0.4 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 -0.2 -0.6 -0.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) 1.2 0.0 -0.4 -0.

3 69.7288 X X X 3 59.9 1.83020 X X X X 5 73..5 71.3 3.8 69.8 61..Chapter 10 Exercise Solutions 10-19. .1 5.9 13.4 1. Different approaches can be used to identify insignificant variables and reduce the number of variables in a regression model.0273 X X 4 72.7 3. ..1 24.0 4.5 36..4 24.5 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 43.9 1.8 60. Tab10-5x2.6 58.0 0.0 19.1665 X X 3 61.6 0.0 0.5660 X X X X X 6 67.5 0.9 2.1 0.83914 X X X X X X 7 80.6147 X X X 4 64.83550 X X X X X 6 79.3 29.5 76.6921 2 55.4 4.7 48.0171 X X 3 66.8 7.8 6.95522 X X X 5 79. Stat > Regression > Best Subsets Best Subsets Regression: Tab10-5y1 versus Tab10-5x1.3 1.1 4.6 52.6816 X 1 35.2 60.6200 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 y2.9 60.8 34.9 3.7 3..4 5.9751 X 2 50.1799 X 3 67.8 56.2 3.8160 X X 4 64.3087 1 31.2 5.4 3.0 3. Tab10-5x2..5208 X X X X 5 65.1 12.5 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 36. This solution uses MINITAB’s “Best Subsets” functionality to identify the best-fitting model with as few variables as possible.1 3.4378 X 2 62.9 1. 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 .6 60..4 3.0 52.1 3.0 4.4 60. a regression model of input variables x1. x8.9 60.8 76. x3..8 62.6526 X X X X 6 67.0 6.6387 X X X 5 67..1 34.9 6..8 0.1 41.7 62.9 2.8 64. 10-19 .7 8. 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 .6 0.3 3.7 18.3 0.95201 X X X 4 72..84292 X X X X X X … Best Subsets Regression: Tab10-5y2 versus Tab10-5x1...1 68. x4.5 1. and x9 maximize adjusted R2 (minimize S) and minimize Mallow’s C-p.5 24.1 13.5684 X X X X X 7 67.6149 X X X X X X 7 67.2 4..1 52.2 4.0760 X 2 55.5 4.2 3.93966 X X X X 6 79.3 76.9 4.9 13..83693 X X X X X 7 80.9 76..1 75.

The regression equation is Tab10-5y2 = 244 .47 4. The regression equation is Tab10-5y1 = 819 + 0.689 F 26.000 R-Sq(adj) = 76.31 -3. .41 12.075 0.02438 0.65 R-Sq = 79.424 T 28.35 P 0.0.1034 3.6367 114.64 Tab10-5x8 + 115 Tab10-5x9 Predictor Constant Tab10-5x1 Tab10-5x3 Tab10-5x4 Tab10-5x8 Tab10-5x9 Coef 818.23 P 0.236 Tab10-5x9 Predictor Constant Tab10-5x1 Tab10-5x3 Tab10-5x4 Tab10-5x8 Tab10-5x9 Coef 244.124 Tab10-5x3 .005 0.000 0.431 Tab10-5x1 .000 10-20 .75 3.3 R-Sq = 67.52081 SE Coef 123.12396 -0.056 0.2 Tab10-5x8 .51 -3.43080 -0..03530 0.025 R-Sq(adj) = 62..001 0.176 Tab10-5x4 + 11.47 Total 39 1303.3441 0. Tab10-5x3.4 -0.08113 0.80 0.4540 0.7604 23.84 3.47 -2.09146 2.10 5.0915 Tab10-5x4 + 2. .7% Analysis of Variance Source DF SS Regression 5 882.5% Analysis of Variance Source DF SS Regression 5 90.85 P 0.Chapter 10 Exercise Solutions 10-19 continued Stat > Regression > Regression Regression Analysis: Tab10-5y1 versus Tab10-5x1..198 0.03 Residual Error 34 421.001 0.9% MS 176.98 -1.6 0.7 S = 3.6329 0.0.5% MS 18.14 0.. Tab10-5x3.50 T 1.81 S = 0.000 Regression Analysis: Tab10-5y2 versus Tab10-5x1.098 0.633 Tab10-5x1 + 0.0.1758 11.001 0.000 0.175 -235.225 100.1497 0.454 Tab10-5x3 + 0.001 0.70 3.03 1.40 F 14.830201 SE Coef 29.990 Residual Error 34 23.40 P 0.434 Total 39 114.

586 2 __ M R=0. 11 Test Results for MR Chart of Ex10-19Res1 TEST Test TEST Test 1.105 1 _ X=-0. 9 points in a row on same side of center line. Failed at points: 25 2.105 4 8 12 16 20 24 O bser vation 28 32 36 40 1 M oving Range 3 U C L=2. Failed at points: 10. Failed at points: 26 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=2.00 standard deviations from center line.000 0 2 -1 2 -2 LC L=-2.00 standard deviations from center line.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. Failed at points: 11 10-21 . 9 points in a row on same side of center line. One point more than 3. One point more than 3.

10-22 . no out-of-control observations) and the subset regression model (observation 25 is OOC). 21. Test Failed at points: 26 For response y1. 4 out of 5 points more than 1 standard deviation from center line (on one side of CL). Test Failed at points: 7.00 standard deviations from center line. there is not a significant difference between control charts for residuals from either the full regression model (Figure 10-10.00 0 -4 5 LC L=-6.52 -8 4 8 12 16 20 24 O bser vation 28 32 36 40 1 M oving Range 8 U C L=8. 18 TEST 5. 25 TEST 6. One point more than 3.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. Test Failed at points: 19. there is not a significant difference between control charts for residuals from either the full regression model (Exercise 10-18.00 standard deviations from center line.02 6 4 __ M R=2. 2 out of 3 points more than 2 standard deviations from center line (on one side of CL). 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).52 5 4 _ X=-0. Test Failed at points: 7. For response y2.45 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. One point more than 3.

5 +2.4 0.0 -2. 22 The EWMA control chart for residuals from the response y1 subset model has no out-ofcontrol signals.5 -1. However the chart for y2 residuals still indicates a problem beginning near observation 20.0 -0.0 -0.2 0. Test Failed at points: 21.1 and L = 2.0 0.Chapter 10 Exercise Solutions 10-20.4 -2.347 -1.5 4 8 12 16 20 24 Sample 28 32 36 40 Test Results for EWMA Chart of Ex10-19Res2 TEST.7SL=0.7SL=-0. 10-23 .7SL=1.7SL=-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.0 1.435 0.1 -0.000 0.1 _ _ X=-0.3 EWMA 0.347 EWMA 1.000 0.435 4 8 12 16 20 24 Sample 28 32 36 40 EWMA Chart of y2 Regression Model Residuals (Ex10-19Res2) 2.5 _ _ X=-0. Use λ = 0. Stat > Control Charts > Time-Weighted Charts > EWMA EWMA Chart of y1 Regression Model Residuals (Ex10-19Res1) 0.5 +2.3 -0.2 -0. One point beyond control limits.7.

015 Cumulative 0.99211 -1.27825 -2.41131 -2.65608 2.82157 0.67046 -0.31493 0.12420 0.832 0.253 0.11243 0. (a) Stat > Multivariate > Principal Components Note: To work in standardized variables in MINITAB.56081 -0.0118 0.49153 0.60340 0.580 0.94470 -1.257 0.01299 -1.580 0.061 0. select Correlation Matrix.08767 1.19166 -0.03929 0.11787 -1.17849 Ex10-21z3 0.47259 -0. Ex10-21X4 Eigenanalysis of the Correlation Matrix Eigenvalue 2.42176 -0.104 Principal Component Scores Ex10-21z1 0.17976 -1.33164 0.685 -0.801 PC4 0.444 PC2 -0.36089 0.6088 0.534 -0.58815 1.286 0.152 0.60867 -0.718 -0.64094 1.20327 -0.32286 0.02496 1.62545 -0.21893 -1.3181 1.29428 0.21950 2.36528 0. Ex10-21X3.083 0.23100 0.13529 -1.14662 Ex10-21z2 -0.12378 -1.23823 -0.99498 0. Principal Component Analysis: Ex10-21X1.83902 3.0613 Proportion 0.985 1. select Storage and enter columns for Scores.87917 -2.14246 -0.Chapter 10 Exercise Solutions 10-21. Note: To obtain principal component scores.330 0.29168 0.387 PC3 0.794 0.19734 0.594 0.19592 0.50471 -0.30494 -0.607 0.09129 -0.20787 0.29514 1.64480 -0.31445 -0.10423 -0.334 -0. Ex10-21X2.94763 0.70241 -0.20488 0.000 Variable Ex10-21X1 Ex10-21X2 Ex10-21X3 Ex10-21X4 PC1 0.86838 10-24 .46958 0.

xls : worksheet Ex10-21. Ex10-21z2.9 (c) Note: Principal component scores for new observations were calculated in Excel.5 Ex1 0 -2 1 z2 all -3. Ex10-21z3 Principal Component Scores -2 0 2 2 0 Ex10-21z1 -2 2 0 Ex10-21z2 -2 1.7 Ex10-21z3 -0.Chapter 10 Exercise Solutions 10-21 continued (b) Graph > Matrix Plot > Simple Matrix of Plots Matrix Plot of Ex10-21z1.0 -0. Ex10-21z3all Principal Component Scores -3.0 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.7 1. Graph > Matrix Plot > Matrix of Plots with Groups Matrix Plot of Ex10-21z1all.5 0. See Excel : workbook Chap10. Ex10-21z2all.0 Ex10-21Obs New Original 4 Ex1 0 -2 1 z1 all 0 -4 2.5 2. the majority of new observations are clearly different from the original observations.0 -0.5 -2 0 2 -0. 10-25 .9 0.

579 0.727 0.108 0.089 0.407 0.662 -0.156 -0.035 0.467 -0. select Storage and enter columns for Scores.3292 1. 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 . (c) Graph > Matrix Plot > Simple Matrix of Plots Matrix Plot of Ex10-22z1.349 0.014 0.089 0.199 0.0730 1.232 0.2542 0.050 0.6129 0.579 -0.152 -0.021 0.090 -0.022 -0.529 0.357 0.419 0.372 0.120 0.000 0.0520 Proportion 0.032 -0.022 0.602 -0.322 0.874 0.332 -0.1973 0.3121 0.166 -0.035 0.277 0.018 0. Ex10-22x3. (a) Stat > Multivariate > Principal Components Note: To work in standardized variables in MINITAB.262 0.651 -0.425 -0.256 0.575 0.632 0.409 -0.124 -0.513 -0.723 0.123 0.436 -0.368 0.204 -0.323 0. ….200 0.135 0.406 0.003 1. Ex10-22z2.133 (b) 72.240 0.006 0.163 -0.193 -0.947 PC6 -0.145 -0.115 PC4 -0.912 PC5 0.431 -0.162 0.022 0.0287 0. Ex10-22x2. select Correlation Matrix.028 0.244 -0. Ex10-22x9 Eigenanalysis of the Correlation Matrix Eigenvalue 3. Principal Component Analysis: Ex10-22x1.105 0.127 -0.238 PC3 -0.068 0.074 -0.175 0.1407 2.349 0.267 0.261 -0.099 -0.7% of the variability is explained by the first 3 principal components.148 0.975 PC7 0.068 0.465 0.448 PC2 0.297 0.202 -0.997 PC8 0.262 0.Chapter 10 Exercise Solutions 10-22.508 0.188 0.230 0.471 0.525 -0.380 0. Note: To obtain principal component scores.000 PC9 0.247 0.117 Cumulative 0.844 Variable Ex10-22x1 Ex10-22x2 Ex10-22x3 Ex10-22x4 Ex10-22x5 Ex10-22x6 Ex10-22x7 Ex10-22x8 Ex10-22x9 PC1 -0.048 0.

indicating that the process is not in control. 10-27 .xls : worksheet Ex10-22. Ex10-22z2all. Graph > Matrix Plot > Matrix of Plots with Groups Matrix Plot of Ex10-22z1all. See Excel : workbook Chap10. 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.Chapter 10 Exercise Solutions 10-22 continued (d) Note: Principal component scores for new observations were calculated in Excel.

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 .Chapter 11 Exercise Solutions 11-1.

179 -6.0 -9.0 -20 6.375 -18.000 25.0 -18.625 4.025 -22.0 L = +10 10 4.Chapter 11 Exercise Solutions 11-2.318 -7.359 45 46 47 48 49 50 22 -9 3 12 3 12 9 -31 12 9 -9 9 8.969 5.0 -40 10.982 -8.000 -3.0 -9.700 1.000 0.0 12.975 -10.134 no yes no yes no no no no yes 0.625 10.975 -1.975 -13. The chart with λ = 0.0 -10.000 0.975 -13.688 4.923 10.975 -13.975 -13. Excel : workbook Chap11.625 12.800 10.854 0.984 0.000 0.000 0.0 0 2.0 -30 8.0 0.625 7.375 -22.127 -6.xls : worksheet Ex 11-2 0 0.0 -9.2 gives SS = 9780 and average deviation from target = 1.975 … SS = Average = 21468 17.625 19.000 11.375 0.000 0.3 10 0.000 0.000 -13.766 9.993 no no no no no no 0.0 30 -6.690 Bounded Adjustment Chart for Ex 11-2 50 -12.8 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 20.0 0.975 -1.76.975 -13.975 -0.0 L = -10 -10 Adjustment Scale -4.3 exhibits less variability and is closer to target on average.24 6526.975 -10.0 40 -8.000 5.0 20 -2.000 0.560 6.375 -18.375 -18.715 -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. 11-12 .000 0.0 -9.375 -18.

440 6.5 0.400 3.958 7.5 -6.3.175 no yes no yes no no no no no 46 47 48 49 50 -9 3 12 3 12 -31 12 9 -9 9 -20.5 -11.264 7.4 10 0.813 -3.061 -7.5 -8.022 -5.400 13. but is further from target on average than for the chart with λ = 0.288 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 -11.25 0.4 exhibits less variability.5 -11.91 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 = 0.5 -8.5 -11. Excel : workbook Chap11.880 0.5 … SS = Average = 21468 17.5 -11.037 -4.5 0.800 12.xls : worksheet Ex 11-3 Target yt = lambda = L= g= 0 0.Chapter 11 Exercise Solutions 11-3.24 5610. 11-13 .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 6.440 3.

7 -27.0 20.9 -23.0 10.3 1.0 20.0 -5.2 -1.0 -5.9 5.5 -4.6 1.7 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 = 0.2 0.9 5.8 3.6 -22.4 1.0 -5.3 -2.3 -27.24 5495.xls : worksheet Ex 11-4 T= lambda = g= 0 0.Chapter 11 Exercise Solutions 11-4.0 20.8 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 16.7 -0.6 -0.9 0.7 -6.8 -3.0 5.1 14. 11-14 .7 -4.3 -7.0 … SS = Average = 21468 17.4.3 -22.3 -13.0 -20.4 -4.0 -1.5 -7.0 -9.0 -19.7 45 46 47 48 49 50 22 -9 3 12 3 12 9 -31 12 9 -9 9 11.7 5.3 -2.3 -4.0 -14.7 -8.9 -6. Excel : workbook Chap11.2 -1.

72 147.500 1.47 179.217 0.39 162.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 0 5 10 15 20 25 Vm/V1 11-15 .370 0.000 0.70 1.43 201.002 1.000 1.500 2.02 136. Excel : workbook Chap11.929 1.943 Variogram for Ex 11-5 2.60 151.500 0.11 175.Chapter 11 Exercise Solutions 11-5.000 m 1.104 1.029 1.53 138.195 1.

0 0.55 48.012158 0.28 0.334961 0.0 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 0.68 1.90 0.59 54.299822 0.164698 0.066173 T 3.58 1.33 0.41 63.6 0.Chapter 11 Exercise Solutions 11-5 continued MTB : Chap11.44 67.86 63.6 -0.2 -0.8 -1.4 -0.71 62. 11-16 .05 0.0 -0. The slow decline in the sample ACF also indicates the data are correlated and potentially nonstationary.228540 0.01 2.80 41. so the observations are correlated and there may be some mild indication of nonstationary behavior.59 0.87 67.149321 0.mtw : Yt Stat > Time Series > Autocorrelation Function Autocorrelation Function for Data in Table 11-1 (Yt) (with 5% significance limits for the autocorrelations) 1.03 55.345327 0.440855 0.389094 0.39 27.75 Variogram appears to be increasing.12 2.26 LBQ 10.45 1.85 1.440819 0.316478 0.4 0.54 1.325056 0.8 Autocorrelation 0.31 16.2 0.14 32.

9 180.0 230.0 -20.323. t 1 2 3 4 5 6 7 8 9 10 Orig_out Orig_Nt Adj_out_t Adj_Obs_t+1 Cum_Adj 215.8 161.2 1.4 4.7 220.5 49 50 SS = Average = Variance = 11.0 193.4 1.0 30.1 13.0 290.0 195.9 200.8 0.5 208.0 -15.9 320.2 160.5 Unadjusted 1.0 15.3 2.0 Adjustment Scale Integral Control for Ex 11-6(a) 20.Chapter 11 Exercise Solutions 11-6.6 -13.0 110.8 176.5 24.0 129.0 170.0 -10.9 12.5 3.1 177.4 62.2 1.2 T= lambda = g= Obs. (a) and (b) Excel : workbook Chap11.9 -39.8 162.0 5.8 0. t Orig_out Adj_out_t Adj_Obs_t+1 Significant reduction in variability with use of integral control scheme.0 140.7 3.7 191.4 Adjusted 1.0 12.8 -20.2 0.0 63.4 0.xls : worksheet Ex 11-6a 200 0.0 -5.4 191.0 200.3 2.1 216.0 187.0 195.1 4.8 … 145.0 0.0 185.2 7.0 10.0 1 6 11 16 21 26 31 36 41 46 Obs.7 180.1 -3.0 260.5 176.3 -6.5 -7. 11-17 .0 1.510.7 4.8 -15.8 -1.3 467.7 187.7 0.3 192.4 170.818.3 -4.0 80.0 -0.5 192.871.

4 Adjusted 1.0 162.323.0 9.6 -2.9 183.0 3.8 -20.4 3.2 and λ = 0.3 467.3 -4. t Orig_out Adj_out_t Adj_Obs_t+1 Variances are similar for both integral adjustment control schemes (λ = 0.5 Unadjusted 1.0 -40.0 1 6 11 16 21 26 31 36 41 46 Obs.0 195.9 18.7 0.4 176.0 4.5 176.4).5 7.8 1.0 290.8 185.7 … 129.1 3.8 161.0 140.0 30.0 -20.871.0 7.5 5.5 216.7 2.5 -7.0 -30.9 208.9 164.0 Adjustment Scale Integral Control for Ex 11-6(c) 40.0 230.Chapter 11 Exercise Solutions 11-6 continued (c) Excel : workbook Chap11.8 7.0 170.0 195.0 30.xls : worksheet Ex 11-6c 200 0.2 T= lambda = g= Obs.0 198.0 5.0 SS = Average = Variance = -15.0 10.6 -13.5 13. t 1.5 24.4 18.0 Orig_out Orig_Nt Adj_out_t Adj_Obs_t+1 Cum_Adj 215.0 20.995.3 -6.1 176.6 7.5 192.0 195. 11-18 .5 50.5 -7.7 223.2 180.0 80.0 110.4 175.9 -39.0 2.0 0.0 8.0 -10.4 1.1 187.8 -0.8 0.4 191.6 8.8 193.0 10.0 320.0 260.0 189.0 -0.888.1 21.9 2.0 66.0 6.0 200.4 1.

840 191.7 216.364 1.872 161.600 -14. Excel : workbook Chap11.0 5.941 162.0 230.716 181.733 -1.459 -10.656 |EWMA_t|>L? Adj_Obs_t+1 Cum_Adj no no no no no no yes no no 0.000 0.000 0.8 -20 196 -0.900 -5.000 -0.969 -5.xls : worksheet Ex 11-7 200 0.8 SS = Average = Variance = 165.0 -15.0 290.5 -7 186.2 145.300 -2.000 0.412 185.000 0.0 15.5 191.716 193.000 0.2 Target yt = lambda = L= g= Obs.3 7 11.633 187.9 -39.000 0.000 0.0 10.000 0.323.0 Adjustment Scale Bounded Adjustment Chart for Ex 11-7 20.516 48.000 0.699 174.0 0.000 -8.051 172.696 176.717 0.4).516 48.516 48.0 200.000 48.0 1 6 11 16 21 26 31 36 41 46 Obs.265 182.000 0.0 140.870 216.000 6. 11-19 .304 467.0 -10.516 178.733 -3.9 126.0 -0.2 and λ = 0.000 0.3 -6 185. t 1 2 3 4 5 6 7 8 9 10 Orig_out Orig_Nt Adj_out_t EWMA_t 215.0 110.516 48.177 176.0 -5.632. 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 = 0.9 176.253 180.0 0.8 -15.516 … 122.4 162.0 0.057 -7.0 0.000 6.0 30.300 -4.2 133.5 1.8 0 195.0 129.9 193.000 0.016 -12.233 0.0 -20.233 6.5 24.233 yes no no no no 5.7 320.0 170.6 -13.233 6.0 80.702 -7.5 46 47 48 49 50 -7 3.2 12 1.1 176.3 -4.0 260.Chapter 11 Exercise Solutions 11-7.

000 0.0 170.9 126.326 -4.784 170.000 0.0 30.486 176.5 181.758 188.0 -8.967 -0.513 1.758 … 122.323.758 58.4 162.0 230.693 -7.720 -11.680 191.000 0.0 -0. t 1 2 3 4 5 6 7 8 9 10 Orig_out Orig_Nt Adj_out_t EWMA_t 215.8 0 195.0 290.0 -32.81 -9.000 0.86 Bounded Adjustment Chart for Ex 11-8 350.600 -23.5 24.900 -8.0 1 6 11 16 21 26 31 36 41 46 Obs.258 1.0 110.9 -39.3 -6 185.0 140.692 162.083 189.658 184.0 320.467 no no no no no 0.000 12.0 Adjustment Scale -16.758 58.773.5 -7 192. Excel : workbook Chap11.000 0.467 12.849 -10.3 -4.5 191.000 0.872 161.0 40.4 15 1.xls : worksheet Ex 11-8 200 0.300 -4.0 50.467 12.000 0.2 145.0 24.0 -24.000 12.000 58.488 185.0 200.256 176.000 1.000 -14.0 0.758 58.0 0. 11-110 .000 0.0 48.8 -20 196 -1.000 0. 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.0 16.9 176.0 0.0 260.573 216.2 133.0 32.7 216.3 7 11.0 8.0 0.967 -2.0 -40.958 203.000 0.775 187.Chapter 11 Exercise Solutions 11-8.467 0.0 80.8 -15.0 129.1 176.758 58.5 46 47 48 49 50 SS = Average = Variance = -7 3.013 180.000 0.2 T= lambda = L= g= Obs.304 467.6 -13.000 0.000 0.0 -48.300 -8.821 |EWMA_t|>L? Adj_Obs_t+1 Cum_Adj no no no no no no yes no no 0.958 191.9 199.

8 -0.0 42 1.7 49 50 23 26 3.262 78.6 -1.0 -13.0 -4.0 53.Chapter 11 Exercise Solutions 11-9.6 0.1 48.0 58 -1.6 -3.3 -1.7 0.32 Unadjusted 109.6 55.0 0.0 74 -3.2 22.4 -0.4 0.2 1.0 26 3.0 34 2.8 -0.2 -1.7 22.0 10.7 -5.4 65.51 Significant reduction in variability with use of an integral control scheme.0 10 1 6 11 16 21 26 31 36 41 46 Obs.0 43.8 -5.0 66 -2.0 11.0 50 0.0 3.8 -3.0 -11.9 0.4 223.4 54. 90 -5.0 -1. (a) and (b) Excel : workbook Chap11.520 44.0 45.3 -1.0 18 4.629 46.0 -0.0 82 -4.6 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 8.0 14.0 Adjustment Scale Integral Control for Ex 11-9 (a) 5.0 -1.0 58.7 -0.7 64.2 50.0 -9.2 51.xls : worksheet Ex 11-9a T= lambda = g= 50 0. t Orig_out Adj_out_t Adj_Obs_t+1 11-111 .9 … SS = Average = Variance = Adjusted 108.6 -1.

11-112 .Chapter 11 Exercise Solutions 11-9 continued (c) Excel : workbook Chap11.4 -1.8 0.5 -2.4 223.0 42.0 52.0 56 0.0 -3.0 58.6 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 8.8 0.520 44.819 47.0 -3.3 -6.0 50.9 69 14.0 63.2 56 -13.4.5 1.0 54 -4.4 -0. with a process average slightly closer to the target of 50.0 48.833 56.3 -2.0 -1. as compared to λ = 0.5 114.0 3.9 -0.0 55.xls : worksheet Ex 11-9c T= lambda = g= 50 0.0 46.8 -8.2.40 There is a slight reduction in variability with use of λ = 0.5 45 -9.0 66 10.1 -0.4 … 49 50 SS = Average = Variance = 23 26 109.8 27.0 54.5 53.2 55 -11.51 3.0 -0.4 26.4 1.0 63.3 -5.0 -2.0 -2.6 56 11.3 -9.1 -3.

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

00 3.0 10 5.0 50 0.00 0.19 109.00 2.600 54 -4 54.0 42 1.00 0.32 0. 11-114 . but different variances (bounded adjustment variance is larger).00 0.362 107.48 31.880 56 -13 54.79 43.0 26 3.00 0.00 -2. Excel : workbook Chap11.00 13.731 56 0 56.00 2.00 5.00 3.79 15.00 0.00 no yes no yes no 0.00 -2.00 1.664 56 11 56.00 0.72 98 -6.0 Adjustment Scale Bounded Adjustment Chart for Ex 11-10 6.822 45.0 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.0 18 4.00 0.00 0.79 17.268 55 -11 53.080 45 -9 45.620 121.0 90 -5.48 15.600 69 14 67.19 … 24 18 20 23 26 46 47 48 49 50 8 -6 2 3 3 37.00 0.0 58 -1.0 82 -4.00 -2.00 0.0 0.709 -2.40 0.2 4 1.79 38.0 66 -2.904 |EWMA_t|>L? Adj_Obs_t+1 Cum_Adj no no no no no yes no no no 0.Chapter 11 Exercise Solutions 11-10.00 -2.842 -4. with similar means and sums of squares. 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 1.00 1.4 223.520 44.00 0.515 -1.0 34 2.0 0.0 74 -3.xls : worksheet Ex 11-10 50 0.00 2.48 35.00 -2.585 66 10 66.00 0. t Orig_out Adj_out_t EWMA_t Adj_Obs_t+1 Nearly the same performance as the integral control scheme.19 17.6 Target yt = lambda = L= g= Obs.51 SS = Average = Variance = -2.505 -5.

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

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

2. 2 1.0 S = 7.7 8.7 1.08% R-Sq(adj) = 94.8 Total 17 16150.84 0.26483 R-Sq = 96.0 14450.0 273. 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 . do not select this option.3 633. since the factor levels in this experiment are categorical. Click on “Graphs”.3 66. Select the response (Brightness).3 466.000 Ex12-1Phosphor 2 933.10). and click “OK” twice. verify that the selected terms are Glass Type. Ex12-1Phosphor Factor Ex12-1Glass Ex12-1Phosphor Type fixed fixed Levels 2 3 Values 1. and their interaction. The option to plot residuals versus variables is for continuous factor levels.Chapter 12 Exercise Solutions 12-1 continued Select Stat > DOE > Factorial > Analyze Factorial Design. Click on “Storage”. Click “OK”.26 0.004 Ex12-1Glass*Ex12-1Phosphor 2 133.79 0. General Linear Model: Ex12-1Bright versus Ex12-1Glass.0 14450.10). Glass type (A) and phosphor type (B) significantly affect television tube brightness (P-values are less than 0. 3 Analysis of Variance for Ex12-1Bright. click “OK”.3 133. Phosphor Type. using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Ex12-1Glass 1 14450.3 933.44% No indication of significant interaction (P-value is greater than 0.3 52. select “Residuals Plots : Four in one”.318 Error 12 633. select “Fits” and “Residuals”. then click on “Terms”.

Click on “Scale”. The plot of residuals versus observation order is not meaningful since no order was provided with the data. select Graph > Individual Value Plot > One Y with Groups. the residuals should be re-examined. then click “OK” twice.Chapter 12 Exercise Solutions 12-1 continued Visual examination of residuals on the normal probability plot. select the “Reference Lines” tab. To plot residuals versus the two factors. 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 . and enter “0” for the Y axis. histogram. If the model were re-fit with only Glass Type and Phosphor Type. 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. and versus fitted values reveals no problems.

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 . 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. there appears to be more variability in results with phosphor type 2. however. To remove the jitter.Chapter 12 Exercise Solutions 12-1 continued Note that the plot points are “jittered” about the factor levels.

and click “OK” twice. select the response (Brightness) and both factors (Glass Type and Phosphor Type). 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. Select “Interaction Plot” and click on “Setup”. 12-6 . Final selected combination of glass type and phosphor type depends on the desired brightness level.Chapter 12 Exercise Solutions 12-1 continued Select Stat > DOE > Factorial > Factorial Plots.

0 12-7 .000 Ex12-1Phosphor 1 2 3 Individual 95% CIs For Mean Based on Pooled StDev -----+---------+---------+---------+---(--*-) (--*-) -----+---------+---------+---------+---240 260 280 300 Mean 260.0 272.3 52.7 1.0 280.5.000 273. Two-way ANOVA: Ex12-1Bright versus Ex12-1Glass.004 Interaction 2 133. Ex12-1Phosphor Source DF SS MS F P Ex12-1Glass 1 14450.0 S = 7.333 256.08% R-Sq(adj) = 94.0 264. To obtain results which match the output in the textbook’s Table 12.0 14450.79 0.265 R-Sq = 96.3 466. The DOE functionality was selected to illustrate the approach that will be used for most of the remaining exercises.667 Individual 95% CIs For Mean Based on Pooled StDev -------+---------+---------+---------+-(-------*-------) (-------*-------) (-------*-------) -------+---------+---------+---------+-256.0 273.000 Ex12-1Phosphor 2 933.44% Ex12-1Glass 1 2 Mean 291.26 0. select Stat > ANOVA > Two-Way.318 Error 12 633.667 235.3 66.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.8 Total 17 16150. and complete the dialog box as below.84 0.7 8.

Cutting Angle Estimated Effects and Coefficients for Life (coded units) Term Effect Coef SE Coef T Constant 413.62 12. Both approaches would achieve the same result. This solution uses the first approach. (a) To analyze the experiment. The design and data are in the MINITAB worksheet Ex12-2.483 0. change number of replicates to “2”.41 2.199 Based on ANOVA results. Metal Hardness.62 12.25 -5. B. Select “Designs”.25 42.25 -59. Cutting Speed will also be retained to maintain a hierarchical model.98 Cutting Speed*Metal Hardness* -34.81 Metal Hardness*Cutting Angle -24.36% P 0.41 -4.MTW.96 P 0.000 0.74 Metal Hardness 84. and click “OK”.020 ** 0.41 -0. 12-8 . leave factor types as “Numeric” and factor levels as -1 and +1.41 -1. then enter the data.41 -0. The worksheet is in run order. Select “Terms” and verify that all terms (A. Another approach would be to create a worksheet containing the data.25 -12.12 12.09 1.009 ** 0.662 0.199 R-Sq(adj) = 72.75 -17. AB.008 0. Select Stat > DOE > Factorial > Create Factorial Design.13 12. Factorial Fit: Life versus Cutting Speed. a full factorial model is not necessary. and change the Number of factors to “3”.10.40 Cutting Angle 71.001 ** 0. AC. one approach to solving this exercise is to create a 23 factorial design in MINITAB. select Stat > DOE > Factorial > Analyze Factorial Design.12 12.41 33.56% 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 6.81 8. Select “Factors”. and click “OK” twice.41 0. a reduced model in Metal Hardness. Since the standard order (Run) is provided.357 0. Based on P-values less than 0.40 Cutting Angle S = 49. enter the factor names. to change to standard order (and ease data entry) select Stat > DOE > Display Design and choose standard order.25 9.75 35. BC. Cutting Angle. and Cutting Speed*Cutting Angle is more appropriate.6236 R-Sq = 85. C.45 Cutting Speed*Cutting Angle -119.37 12.13 12.41 3. ABC) are included. highlight full factorial.Chapter 12 Exercise Solutions 12-2.30 Cutting Speed 18.89 Cutting Speed*Metal Hardness -11.014 0. then define a customer factorial design. Leave the design type as a 2-level factorial with default generators.88 12.

88 12. we see that the low level of cutting speed and the high level of cutting angle gives good results regardless of metal hardness. for an average predicted life of 552.5 1 266.008 0.04 0.47 3.8988 R-Sq = 79.001 1.38 Cutting Angle 71.000 0.Chapter 12 Exercise Solutions 12-2(a) continued Factorial Fit: Life versus Cutting Speed.480 0.0 380.0 -1 1 -1 Cutting Speed Cutting Angle -1 Longest tool life is at A-.001 R-Sq(adj) = 72.25% 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 6.5 405.75 35.5. 12-9 . Select Stat > DOE > Factorial > Factorial Plots. (c) From examination of the cube plot.73 Metal Hardness 84.85 P 0.25 9. Cube Plot (data means) for Life Exercise 12-2(b) 552.425 (b) The combination that maximizes tool life is easily seen from a cube plot.47 33.015 0.47 -4.62 12.47 0.12 12.006 0.13 12.65% P 0.25 42. Choose and set-up a “Cube Plot”. Cutting Angle Estimated Effects and Coefficients for Life (coded units) Term Effect Coef SE Coef T Constant 413. B+ and C+.13 12.12 Cutting Speed 18.25 -59.5 351.78 S = 49.5 391.0 1 Metal Hardness 446.0 512.88 Cutting Speed*Cutting Angle -119. Metal Hardness.47 2.74 22.

Residuals versus fitted values plot shows that the equal variance assumption across the prediction range is reasonable. 12-10 . 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. B. To save residuals to the worksheet. and for residuals plots choose “Normal plot” and “Residuals versus fits”. C. select “Storage” and choose “Residuals”. Select “Graphs”. To find the residuals. AC) are included. select Stat > DOE > Factorial > Analyze Factorial Design. Select “Terms” and verify that all terms for the reduced model (A.Chapter 12 Exercise Solutions 12-3.

50 4-Way Interactions 1 47.91 Pure Error 16 462.63 852.629 0. 12-11 .722 0.77 Syrup to Water -1. Factorial Fit: Total Score versus Sweetener.629 * * * * * 0.9504 -0.094 0.30 Syrup to Water*Temperature -0. select “Terms” and verify that all terms are selected.64 Residual Error 16 462.438 1.344 0.062 -4.9504 -2.688 33.500 0.531 47.9504 1.844 0.. .625 213.69 Carbonation -2.267 0.656 0. To preserve model hierarchy. carbonation.000 0.15 3-Way Interactions 4 405.9504 -0.188 -0.969 0.563 -0. sweetener*syrup to water.219 0.9504 -0.9504 -1.91 Total 31 1967.469 0. two-factor interactions: sweetener*carbonation.062 2.Chapter 12 Exercise Solutions 12-4.10 and select terms with P-value less than 0.469 0.9504 2. as well as lower-order terms included in the significant terms (main effects: syrup to water.53 47.379 0.938 -0.28 1.36 Sweetener*Temperature -2. along with interactions involving both sweetener and the ratio of syrup to water.188 -1. Syrup to Water.281 0.500 28.9504 0. syrup to water*temperature).344 0.MTW.47 Sweetener*Carbonation*Temperature -0.10.344 0.9504 -1.16 7.000 0.37 2-Way Interactions 6 199.07 Sweetener*Syrup to Water 4. Use an α = 0.47 P 0.50 462.9504 -4.125 101.13 405.9504 0. type of sweetener is dominant.53 1.025 0.001 0.094 0.313 -0.15 Syrup to Water*Carbonation -0.388 0. sweetener*syrup to water*carbonation.177 0.688 0.938 1.594 0.771 0.9504 -0.048 0.10 Carbonation*Temperature 1. and then enter the data.531 0.218 P 0.41 Temperature 3. sweetener*syrup to water*temperature). the reduced model will contain the significant terms (sweetener.9504 -0. Create a 24 factorial design in MINITAB. temperature.49 Syrup to Water*Carbonation* -0.687 0.9504 192.69 199.28 3.187 -2. syrup to water*carbonation. sweetener*temperature.50 462.015 0.49 Temperature Sweetener*Syrup to Water* 2.031 0. The design and data are in the MINITAB worksheet Ex12-4..923 0.938 -0.9504 2.031 0. Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef SE Coef T Constant 182.688 -1.688 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. Since there are two replicates of the experiment.055 0.218 From magnitude of effects.14 Sweetener*Carbonation 0.500 28.9504 2.31 Sweetener -9.73 Sweetener*Syrup to Water*Temperature 4. Select Stat > DOE > Factorial > Analyze Factorial Design.89 Sweetener*Syrup to Water*Carbonation -5.781 0.

30 Syrup to Water*Temperature -0.53 7..37 Sweetener*Temperature -2.9244 -0.80 2-Way Interactions 5 176.09 0.188 -1.020 P 0.344 0.585 12-12 .9244 -0.88 27.73 Pure Error 16 462.764 0.9244 -2.188 -2.18 Syrup to Water*Carbonation -0.50 28.313 -0. .344 0.094 0.06 195.9244 2.005 0.9244 -1.688 0..50 462.011 0.81 Sweetener*Syrup to Water*Temperature 4.38 21.094 0.000 0.062 -4.20 Sweetener*Carbonation 0.162 0.062 2.71 Carbonation -2. Syrup to Water.000 0.656 0.281 0.9244 -0.531 0.13 Sweetener*Syrup to Water 4.920 0.45 Temperature 3.16 7.344 0.969 0.486 0.63 852.9244 2.91 Total 31 1967.91 35.54 Analysis of Variance for Total Score (coded units) Source DF Seq SS Adj SS Adj MS F Main Effects 4 852.306 0.046 0.001 0. Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef SE Coef T Constant 182.563 -0.34 Lack of Fit 4 84.688 2.38 84.781 0.594 0.9244 0.188 -0.938 1.29 3-Way Interactions 2 391.06 391.9244 197.90 Syrup to Water -1.91 176.10 Sweetener*Syrup to Water*Carbonation -5.251 0.040 0.63 213.9244 -4.88 546.9244 2.15 Residual Error 20 546.9244 -1.38 1.47 P 0.Chapter 12 Exercise Solutions 12-4 continued Factorial Fit: Total Score versus Sweetener.714 0.73 Sweetener -9.688 -1.031 0.

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

12-14 . From examination of the above table.313 -0.688 2.094 0.938 1.688 0.188 -2.594 0.Chapter 12 Exercise Solutions 12-6.486 0.344 0.344 0.9244 -4.000 0. if the confidence interval includes zero.9244 197.9244 -1.046 0.656 0.162 0.9244 2. Syrup to Water.062 2. factors A. Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef SE Coef T Constant 182. the factor is not significant.9244 0. Factorial Fit: Total Score versus Sweetener.18 Syrup to Water*Carbonation -0.344 0.563 -0.54 P 0.920 0. and a value greater than approximately |2| would be considered significant. Also.9244 -0.020 The ratio of the coefficient estimate to the standard error is distributed as t statistic.9244 -1.9244 2.20 Sweetener*Carbonation 0.251 0.062 -4.90 Syrup to Water -1.094 0.031 0.714 0..040 0.9244 -2.10 Sweetener*Syrup to Water*Carbonation -5.281 0.531 0. Select “Terms” and verify that all terms for the reduced model are selected.37 Sweetener*Temperature -2. .45 Temperature 3.73 Sweetener -9.9244 -0..30 Syrup to Water*Temperature -0.71 Carbonation -2.9244 2.764 0.188 -0. and ABD appear to be significant. ABC.688 -1. Select Stat > DOE > Factorial > Analyze Factorial Design.969 0.13 Sweetener*Syrup to Water 4.781 0.000 0. D.81 Sweetener*Syrup to Water*Temperature 4.9244 -0. AB.011 0.188 -1.

500 Sweetener*Temperature -6.000 0.250 0. .500 Carbonation*Temperature 1.125 Carbonation 0.500 -5.250 Temperature Sweetener*Syrup to Water* 3.250 Syrup to Water -0.10 Factorial Fit: Total Score versus Sweetener. choose the normal effects plot. Then select “Graphs”. Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef Constant 183. Alpha = . Since there is only one replicate of the experiment.000 -1.750 Sweetener*Syrup to Water 4.MTW.750 Sweetener*Syrup to Water*Temperature 4.5 12-15 .500 -3.875 Carbonation*Temperature … Normal Probability Plot of the Effects (response is Total Score. and set alpha to 0.500 2.125 Syrup to Water*Carbonation -1.500 -1.250 -3.10) 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 = 4.375 Temperature 5.875 Syrup to Water*Temperature -3. Create a 24 factorial design in MINITAB.Chapter 12 Exercise Solutions 12-7.000 0.125 Sweetener*Carbonation*Temperature 0. The design and data are in the MINITAB worksheet Ex12-7..125 Syrup to Water*Carbonation* -2. Select Stat > DOE > Factorial > Analyze Factorial Design..750 1. Syrup to Water. select “Terms” and verify that all terms are selected.750 0.000 Sweetener*Carbonation 1.000 2. and then enter the data.750 -0.250 2.500 Sweetener*Syrup to Water*Carbonation -7.250 -0.625 Sweetener -10.

5 0. only factor A (sweetener) is significant.75 P 0.750 55.75 778.865 -2.750 55.59% Analysis of Variance for Total Score (coded units) Source DF Seq SS Adj SS Adj MS F Main Effects 1 441.0 Sweetener 0.0 FittedValue Residuals Versus Sweetener (responseisTotal Score) Residual 10 0 -10 -1.0 182.250 1.5 185. This is not serious enough to warrant concern. 12-16 .45822 R-Sq = 36.0 187.75 778.15% R-Sq(adj) = 31.63 Pure Error 14 778.Chapter 12 Exercise Solutions 12-7 continued From visual examination of the normal probability plot of effects.500 -5. Factorial Fit: Total Score versus Sweetener Estimated Effects and Coefficients for Total Score (coded units) Term Effect Coef SE Coef T P Constant 183.865 98. as well as in the predicted values.625 1.00 7.00 441.0 -0.014 S = 7.5 1.93 Residual Error 14 778.000 Sweetener -10.0 There appears to be a slight indication of inequality of variance for sweetener.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. Re-fit and analyze the reduced model.63 Total 15 1219.82 0.000 441.5 190.48 0.

or days. 12-9.Chapter 12 Exercise Solutions 12-8. 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. The ABCD interaction is confounded with blocks. 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 .

select “Terms” and verify that all main effects and interaction effects are selected.MTW. (a) Create a 25-1 factorial design in MINITAB.8475 12-18 .8375 -0.7175 Cat/React -1. choose the normal effects plot.0600 Cat/React*React pH 0. Then select “Graphs”.7025 -0. and set alpha to 0.1062 Normal Probability Plot of the Effects (response is Color.9125 -0.1200 0.2925 0. Select Stat > DOE > Factorial > Analyze Factorial Design..2725 -0.4350 0.5750 Solv/React*Temp -0.3513 Solv/React*Cat/React 1.2125 0. and then enter the data.10. The design and data are in the MINITAB worksheet Ex12-10. Estimated Effects and Coefficients for Color (coded units) Term Effect Coef Constant 2. Cat/React.Chapter 12 Exercise Solutions 12-10.4562 Solv/React*React Purity -1.1462 Cat/React*React Purity 0.2725 React pH -0.2300 -0.5450 2.1363 React Purity 4. Since there is only one replicate of the experiment.4187 Temp*React pH -0.1625 0.3650 -0.. .7700 Solv/React 1.7325 Temp -0.4275 0.1825 React Purity*React pH 0. Factorial Fit: Color versus Solv/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 = 0.1500 0.6150 Solv/React*React pH 0.4650 -0.2138 Cat/React*Temp 0.0812 Temp*React Purity -0. Alpha = .

770 0.5 0.0 React Purity 0.4147 6.545 2.63 82.5 1. only factor D (reactant purity) is significant.52 38.000 (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. Re-fit and analyze the reduced model.751 Pure Error 14 38.000 React Purity 4.48 0.52 38.000 S = 1.272 0.52 2.4147 5.15 P 0.0 Residual plots indicate that there may be problems with both the normality and constant variance assumptions.20% R-Sq(adj) = 65.0 -0.751 Total 15 121.03 Residual Error 14 38. Factorial Fit: Color versus React Purity Estimated Effects and Coefficients for Color (coded units) Term Effect Coef SE Coef T P Constant 2.68 0. 12-19 .52 2.93% Analysis of Variance for Color (coded units) Source DF Seq SS Adj SS Adj MS F Main Effects 1 82.628 30.Chapter 12 Exercise Solutions 12-10 (a) continued From visual examination of the normal probability plot of effects.65876 R-Sq = 68.63 82.

Chapter 12 Exercise Solutions 12-10 continued (c) There is only one significant factor. Cube Plot (data means) for Color 3. D (reactant purity). the 2nd and 3rd largest effects in absolute magnitude are A (solvent/reactant) and B (catalyst/reactant).865 -1 1 -1 React Purity -1 Solv/React 12-20 .865 -2.875 4. so this design collapses to a onefactor experiment.385 1.715 1.425 1 0. A cube plot in these factors shows how the design can be collapsed into a replicated 23 design. or simply a 2-sample t-test. Looking at the original normal probability plot of effects and effect estimates. The highest color scores are at high reactant purity.005 5. the lowest at low reactant purity.795 1 Cat/React 6.

913 A:Temp*D:Time -1.138 D:Time -0.638 … 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. 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.762 B:Matl1 -5.175 -2. (a) and (b) Select Stat > DOE > Factorial > Analyze Factorial Design.Chapter 12 Exercise Solutions 12-11. then define the experiment using Stat > DOE > Factorial > Define Custom Factorial Design.138 A:Temp*B:Matl1 1. the complete defining relation is: I = ACE = BDE = ABCD.675 -0. Enter the factor levels and yield data into a MINITAB worksheet.275 1. select “Terms” and verify that all main effects and twofactor interaction effects are selected. Factorial Fit: yield versus A:Temp.275 1.238 A:Temp -1. Since there is only one replicate of the experiment.337 E:Matl2 2. B:Matl1.525 -0. 12-21 .825 0. The design and data are in the MINITAB worksheet Ex12-11. E:Matl2 Estimated Effects and Coefficients for yield (coded units) Term Effect Coef Constant 19.MTW.587 C:Vol 2.275 -0. D:Time. C:Vol.

Since there is only one replicate of the experiment. (d) Select Stat > DOE > Factorial > Analyze Factorial Design.8 – 23.9 – 16. E:Matl2 … … Analysis of Variance for yield (coded units) Source DF Seq SS Adj SS Adj MS Main Effects 5 79.1) = ¼ (7.2 +23.5 – 16.9 16.8 + 18.913 4.2 + 15.1) = –1.8 + 18.525 [AB] = AB + BCE + ADE + CD = ¼ (+23.913 9. and set alpha to 0.5 + 16.825 This are the same effect estimates provided in the MINITAB output above.4 – 16.956 Residual Error 0 * * * Total 7 89. C:Vol. Then select “Graphs”.9 -16.826 79.8 18. B:Matl1.965 2-Way Interactions 2 9.4 – 16.4 16.5 16.10.826 15.3) = 1. D:Time.8 + 23.8 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. select “Terms” and verify that all main effects and twofactor interaction effects are selected.1) = ¼ (–6. The other main effects and interaction effects are calculated in the same way.1 [A] = A + CE + BCD + ABDE = ¼ (–23. choose the normal effects plot.2 15.739 … F * * P * * 12-22 . Factorial Fit: yield versus A:Temp.2 23.2 – 23.2 +15.

Alpha = .88 0.175 -2.8682 22.025 Residual Error 6 36.0 Lenth's PSE = 2.030 Pure Error 6 36. Then select “Graphs”. 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.56 53.10) 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 -7.030 Total 7 89.5 -5.16 0.56 53.8682 -2.18 6.18 6.000 B:Matl1 -5. Select “Terms” and select “B”.Chapter 12 Exercise Solutions 12-11 (d) continued Normal Probability Plot of the Effects (response is yield. Factorial Fit: yield versus B:Matl1 Estimated Effects and Coefficients for yield (coded units) Term Effect Coef SE Coef T P Constant 19.561 8.7375 Although none of the effects is significant at 0.587 0.0 -2.10.18 36. and select the “Normal plot” and “Residuals versus fits” residual plots. Re-fit a reduced model containing only the B main effect.025 … Analysis of Variance for yield (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 1 53.18 36. and pool the remaining terms to estimate error. Select Stat > DOE > Factorial > Analyze Factorial Design.98 0.74 … 12-23 .238 0.5 5.0 Effect 2.5 0.

If no issues can be identified. The run should be investigated for any issues which occurred while running the experiment.0 Residual 2.0 -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 5.0 Residual plots indicate a potential outlier.5 0.

00 3.85 0.63 0. and set alpha to 0. select “Terms” and verify that all main effects and twofactor interaction effects are selected.953 -4.386 A*B 22.005 * B 10.95 0. and then enter the data. Create a 24 factorial design in MINITAB.53 0.555 B*C -20.000 A -37.Chapter 12 Exercise Solutions 12-12. D Estimated Effects and Coefficients for Mole Wt (coded units) Term Effect Coef SE Coef T P Constant 837.953 1.014 2-Way Interactions 6 4000 4000 666.953 0.953 -0. C.953 2.765 C*D 7.50 1.79 0.953 0.151 Residual Error 5 1250 1250 250. Re-fit and analyze a reduced model containing A.953 -2.036 * A*C -2.75 3.50 -18.87 0.50 3.953 0.00 3.26 0. B.053 * B*D 2. BC) are significant.00 -15.7 2.25 3.67 0.262 C -30.10) 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. and BC.75 3.00 -10.75 3. The design and data are in the MINITAB worksheet Ex12-12.953 211.25 3. Since this is a single replicate of the experiment.95 0.013 * D -7.25 3.0 Total 15 15100 Normal Probability Plot of the Standardized Effects (response is Mole Wt. (a) Select Stat > DOE > Factorial > Analyze Factorial Design.953 -0.50 3.74 0.953 -3. The main effect B must be kept in the model to maintain hierarchy.00 5.50 3. B. AB.85 0. Alpha = .MTW.5 9. Then select “Graphs”. C. Factorial Fit: Mole Wt versus A.10.32 0.50 -1. choose the normal effects plot.386 … Analysis of Variance for Mole Wt (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 4 9850 9850 2462.50 -3.50 11.765 A*D 5.00 2.00 3.32 0. 12-25 .

Factorial Fit: Mole Wt versus A. AB. AB. C Estimated Effects and Coefficients for Mole Wt (coded units) Term Effect Coef SE Coef T P Constant 837. and BC.008 * B*C -20.001 * A*B 22.172 C -30. 12-26 .400 246.00 3. BC”.50 3. C.50 11.400 3.41 0.004 Residual Error 10 1850.00 -15.34 0. C. Select “Terms” and select “A.0 Lack of Fit 2 250. B.51 0.Chapter 12 Exercise Solutions 12-12 continued (b) Select Stat > DOE > Factorial > Analyze Factorial Design.0 Total 15 15100.400 1.400 -2.00 3.400 -4.3 17.5 9.015 * … Analysis of Variance for Mole Wt (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 3 9625.80 0.75 3.94 0.559 Pure Error 8 1600.400 -5.0 1850.0 185.000 A -37.0 3625.0 125.31 0.0 250. A.50 -18.00 5.0 1812.00 3.47 0.25 3. B.0 1600. and select the “Normal plot” and “Residuals versus fits” residual plots.0 9625.000 2-Way Interactions 2 3625.0 200.30 0.63 0. Then select “Graphs”.000 * B 10.0 … The same terms remain significant.0 0.0 3208.00 -10.

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. however neither plot reveals a major problem with the normality and constant variance assumptions. 12-27 .

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

2 Curvature 1 8820 8820 8820. a better strategy is to add runs that would enable estimation of the quadratic effects.527 1.60 0. Although one could pick a “winning combination” from the experimental runs.50 -18.0 Total 19 26920 … (b) The test for curvature is significant (P-value = 0.527 0.521 99.527 -0. which is a test of significance for terms not included in the model (in this exercise.Chapter 12 Exercise Solutions 12-13 (a) continued When looking at results in the ANOVA table.52 0.25 9.234 2-Way Interactions 6 4000 4000 666. (In MINITAB.75 9. the next test to consider is the “curvature” test.268 A*C -2.822 Residual Error 9 13070 13070 1452.321 B*D 2.57 0. If lack of fit is not significant.13 0.00 8. If this test is significant.081 B 10.00 9.00 2. and lack of fit (if available) and use this as the basis for testing for significant effects.25 0. which is a test of significance for the pure quadratic terms. This approach to sequential experimentation is presented in Chapter 13.50 11.703 … Analysis of Variance for Mole Wt (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 4 9850 9850 2462.527 0.898 A*D 5.39 0.05 0. and some terms need to be added to the model. pure error.00 9. then it is reasonable to pool the curvature. the first test to consider is the “lack of fit” test.5 1.898 C*D 7.7 0.) Factorial Fit: Mole Wt versus A.0 0.50 -1.004).50 -3.527 -1.50 3.25 9.18 0.13 0.612 C -30.00 9.25 9.799 B*C -20.39 0.527 0.00 -10.97 0.50 9.150 D -7.000 A -37.00 5. 12-29 .0 16.46 0. this is accomplished by not including center points in the model. D Estimated Effects and Coefficients for Mole Wt (coded units) Term Effect Coef SE Coef T P Constant 848.703 A*B 22.527 -1. If lack of fit is significant.26 0.004 * Lack of Fit 5 1250 1250 250. no further statistical analysis should be performed because the model is inadequate.52 0. the three-factor and four-factor interactions). B.00 -15. the model is not correctly specified.50 1.75 9.70 0. If tests for both lack of fit and curvature are not significant.527 -1.527 -0. C.75 9.915 Pure Error 3 3000 3000 1000.527 0.

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. and at least some 2-factor interactions are aliased with each other. so this is a resolution IV design. Main effects are clear of 2-factor interactions. this problem is of concern.Chapter 12 Exercise Solutions 12-14. The source of interest for any combined main and 2-factor interaction effect would be in question. From Table 12-23 in the textbook. 12-30 . Since significant 2-factor interactions often occur in practice. A lower resolution design would have some 2-factor interactions and main effects aliased together.

000 2-Way Interactions 3 10610.250 0.29 0.600 40. and set alpha to 0. choose the normal effects plot. The design and data are in the MINITAB worksheet Ex12-15.7621 0. 0 = center point). (a) Select Stat > DOE > Factorial > Analyze Factorial Design.6223 97.000 Residual Error 4 18.800 36.0 4. DO NOT include the center points in the model (uncheck the default selection).83 944.6 5.33 Total 11 28184.700 23. Select “Terms” and verify that all main effects and two-factor interactions are selected.7621 52.400 0.MTW.33 0.Chapter 12 Exercise Solutions 12-15.6 18.70 761.850 0.61 1. Enter the factor levels and resist data into a MINITAB worksheet.433 0.000 * A*D -2.000 * D -2.3 17555.10.200 0.57 0.7621 -1.0 Normal Probability Plot of the Standardized Effects (response is Resist.31 0. Also.759 C 80.6 5. Then select “Graphs”. C.3 4388.7621 -1.51 0.000 -1. Alpha = .12 0.260 … Analysis of Variance for Resist (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 4 17555.338 Pure Error 3 13. Factorial Fit: Resist versus A.6 4.1 10610.300 0.13 0.500 -0.550 0.0 13. including a column indicating whether a run is a center point run (1 = not center point.65 Curvature 1 5.1 3536. D Estimated Effects and Coefficients for Resist (coded units) Term Effect Coef SE Coef T P Constant 60. Then define the experiment using Stat > DOE > Factorial > Define Custom Factorial Design.88 0.000 A 47.7621 47.510 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 . B.000 * B -0.76 0.000 0.400 -1.72 0.7621 31.100 0.30 0.7621 -0.190 A*B 1.

8007 29.6537 92.11 0. C.000 * C 80.6 5.Chapter 12 Exercise Solutions 12-15 continued Examining the normal probability plot of effects. and select the “Normal plot” and “Residuals versus fits” residual plots.8007 50.0 41.000 * A*C 72.46 0.000 * … Analysis of Variance for Resist (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 2 17543.7 2066.7 10599.70 23. Select Stat > DOE > Factorial > Analyze Factorial Design.3 17543. so continue with analysis. (b) Factorial Fit: Resist versus A.3 8771.89 0.80 36.44 0.85 0.43 0.0 Curvature is not significant (P-value = 0. 12-32 .6 1710.43 0.30 0.000 A 47. the main effects A and C and their twofactor interaction (AC) are significant.4 35.6 1. and AC.000 2-Way Interactions 1 10599. Select “Terms” and select “A.1 Curvature 1 5. (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. Then select “Graphs”.40 0. AC”.0 5.7 10599. Re-fit and analyze a reduced model containing A.8007 45. C Estimated Effects and Coefficients for Resist (coded units) Term Effect Coef SE Coef T P Constant 60.1 Total 11 28184.79 0.327).4 5.60 40.33 0.327 Pure Error 7 35.6 5. C.

12-33 . The concern with variance in the predicted resistivity indicates that a data transformation may be needed.0 The normal probability plot of residuals is satisfactory.0 Residual 2.5 0.5 5.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 -5.0 -2.

70 70 . Ex13-1x1 y = 75 + 10x1 + 6x2 1. the initial experimental layout must be created in MINITAB or defined by the user. therefore some information required by MINITAB is not included.1 ≤ x2 ≤ 1 x2 6 = = 0.5 1. (a) Graph > Contour Plot Contour Plot of Ex13-1y vs Ex13-1x2.6 x1 10 ∆x1 = 1 ∆x2 = 0.0 -0.5 0.0 (b) yˆ = 75 + 10 x1 + 6 x2 -0.65 65 . The MINITAB instructions provided for the factorial designs in Chapter 12 are similar to those for response surface designs in this Chapter.85 > 85 Ex13-1x2 0.0 Ex13-1y < 60 60 .6 13-1 .0 − 1 ≤ x1 ≤ 1.0 Ex13-1x1 0. The Excel data sets contain only the data given in the textbook.75 75 .5 -1.0 -1.Chapter 13 Exercise Solutions Note: To analyze an experiment in MINITAB.80 80 .5 0. 13-1.

(a) This design is a CCD with k = 2 and α = 1.20 −15 1.Chapter 13 Exercise Solutions 13-2.0 2 βˆ1 ∆x1 = = = −0. and set ∆x2 = 1. βˆ2 = −15.13 βˆ ∆x −15 1.0 13-3.0 2 ∆x3 = 2 βˆ 3 βˆ2 ∆x2 = 3 = −0.5. i = 1. 2. The design is not rotatable.3 select x2 with largest absolute coefficient. yˆ = 50 + 2 x1 − 15 x2 + 3 x3 − 1 ≤ xi ≤ +1. 13-2 .

013 Interaction 1 2.2941 x2 2.5 83. Then define the experiment using Stat > DOE > Response Surface > Define Custom Response Surface Design.9231 x1*x2 -0.688 10.314 0. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 160.577 7.09 0. x2 The analysis was done using coded units.618 4. including a column indicating whether a run is a center point run (1 = not center point.461 -3.8546 x2*x2 6.5 5.2 0.590 0.082 x1*x2 -1.5 501.0 486.5 15.2 14467.1 173.6 Lack-of-Fit 3 15.769 0.5 9. Response Surface Regression: y versus x1.0 162.461 2.875 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS F P Regression 5 30583.03 0.088 0.4 6116.7 73. two-factor interactions.2 28934. and quadratic terms are selected.000 x2 3.704 0.875 Residual Error 6 501.0 Total 11 31084.03 0.273 0.991 Pure Error 3 486.423 7.868 4.0 823.18 0.285 -0.3 2.017 x2*x2 15.000 x1 -87.8682 x1 -58.555 35.MTW.000 Linear 2 28934.3 0.4118 x1*x1 -10. Select “Terms” and verify that all main effects.0 1647.4 30583.704 -18.7500 13-3 . 0 = center point). Select Stat > DOE > Response Surface > Analyze Response Surface Design. The design and data are in the MINITAB worksheet Ex13-3.441 4.000 Square 2 1647.85 0.9 … Estimated Regression Coefficients for y using data in uncoded units Term Coef Constant 160.Chapter 13 Exercise Solutions 13-3 continued (b) Enter the factor levels and response data into a MINITAB worksheet.164 0.3 2.471 x1*x1 -24.

5.5 -1 0 x1 1 x1 0 x2 -1 1 From visual examination of the contour and surface plots.471).7 13-4 .Chapter 13 Exercise Solutions 13-3 continued (c) Stat > DOE > Response Surface > Contour/Surface Plots Contour Plot of y vs x2.5 and letting x2 (temperature) range from −1.49384 x2 = -0.6101 0.5 1.5 x1 = 1. such as cost.5 to + 1.0 -0. The level for temperature could be established based on other considerations. it appears that minimum purity can be achieved by setting x1 (time) = +1.217615 y = 49.0 x2 0.0 240 180 y 120 60 1 0 -1 -1. x1 Surface Plot of y vs x2.50) + 750 = 825 Time = 15 x2 + 30 = 15(−0.5 40 50 75 100 125 150 175 200 y < > 40 50 75 100 125 150 175 200 225 225 -1. x1 1. (d) Temp = 50 x1 + 750 = 50(+1. A flag is planted at one option on the contour plot above.22) + 30 = 26. The range for x2 agrees with the ANOVA results indicating that it is statistically insignificant (P-value = 0.

70 > 70 Ex13-4x2 1 0 -1 -2 -2 -1 0 Ex13-4x1 1 2 yˆ max = 70.1.1x2 .0 + 1.6 (b) ∂ yˆ ∂ (69.6 x1 + 1.1 − 0.3x1x2 2 Ex13-4y < 56 56 .0 + 1.66 66 .3(0.60 60 .4) = 0.58 58 . Ex13-4x1 y = 69.012 at x1 ≈ +0.3x1 x2 ) = ⇒0 ∂ x1 ∂ x1 = 1.6x1 + 1.1x1^2 .9.68 68 .7) = 0.3 x1 = 0 ∂ x2 x1 = −13.1 − 2.569 13-5 .2 x22 + 0. x2 ≈ +0.2x2^2 + 0.Chapter 13 Exercise Solutions 13-4.9 (−15.62 62 .6 − 2 x1 + 0.3 x2 = 0 ∂ yˆ = 1.64 64 .1x2 − 1x12 − 1.885)] (−2. Graph > Contour Plot Contour Plot of Ex13-4y vs Ex13-4x2.885 x2 = [−1.4 x2 + 0.

00928 Lack-of-Fit 3 0. Select “Terms” and verify that a full quadratic model (A.020 Interaction 1 0. select Stat > DOE > Response Surface > Analyze Response Surface Design.000 x1*x1 -0.0790 0. B2.0790 x1*x2 0.01090 1.7273 x1 0.142 0.0550 0.580 0.070 x1*x2 0.291 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS F P Regression 5 2. The design is rotatable.2980 x2 -0. Leave the design type as a 2-factor.MTW.03706 -2.000 Linear 2 2.291 Residual Error 7 0.00807 Total 12 2.Chapter 13 Exercise Solutions 13-5.35 0.03228 0.012 x2*x2 -0.132 0.2980 0. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 13.01210 0.1249 0.371 0. To analyze the experiment.06499 0.19 0.377 Pure Error 4 0.22628 … Estimated Regression Coefficients for y using data in uncoded units Term Coef Constant 13.16128 0. one approach to solving this exercise is to create a two-factor response surface design in MINITAB.01563 1.43226 46.1249 x2*x2 -0. Select Stat > DOE > Response Surface > Create Response Surface Design.13355 0.03424 8. The worksheet is in run order. central composite design.03271 0.06678 7. (a) The design is a CCD with k = 2 and α = 1.0550 13-6 .54 0.4071 x1*x1 -0. (b) Since the standard order is provided.000 x2 -0.703 0.06499 0.04309 318.56 0.01210 1. and enter a custom alpha value of exactly 1. Select “Designs”. The design and data are in the MINITAB worksheet Ex13-5.04818 1. highlight the design with five center points (13 runs).03271 0. AB) is selected.000 Square 2 0.03424 -11.01563 2.30 0.13355 0.03706 -3.7273 0.00781 108. then enter the data. B.889 0. A2.01210 0. to change to standard order (and ease data entry) select Stat > DOE > Display Design and choose standard order.4. x2 The analysis was done using coded units.000 x1 0.4071 0. Response Surface Regression: y versus x1.16128 2.41421).03228 0.4 (the rotatable design is α = 1.

0 -0. Stat > DOE > Response Surface > Contour/Surface Plots Contour Plot of y vs x2.5 -1.25 14. x1 1.25 12.50 13. Lower = 10.0 x1 0. Target = 20.25 13.0 12. From the plots and the optimizer.4 will maximize viscosity.25 14.4 and setting x2 between -1 and -1. or using MINITAB’s Response Optimizer.5 0.5 y 13.50 12.50 12.75 13. 13-7 .00 14.00 12. setting x1 in a range from 0 to +1. and Weight = 7.0 13.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.25 12.00 0.75 13.0 x1 0 x2 -1 1 Stat > DOE > Response Surface > Response Optimizer In Setup.00 12.5 1 0 -1 -1.0 y < > 12.0 12.75 14.50 13.00 13.0 -0.5 1. x1 0.5 x2 Surface Plot of y vs x2.75 14.00 13.25 13. let Goal = maximize.

01 31. select Stat > DOE > Response Surface > Analyze Response Surface Design.94 6. Select “Designs” to establish three levels for each factor.064 … Analysis of Variance for y1 Source DF Seq SS Adj SS Adj MS F P Regression 9 1248237 1248237 138693 23.0283 x1*x3 75. it must also be defined using Stat > DOE > Response Surface > Define Custom Response Surface Design.3844 x3*x3 -29.030 0.001 Residual Error 17 98498 98498 5794 Total 26 1346735 … Estimated Regression Coefficients for y1 using data in uncoded units Term Coef Constant 327.94 0.97 3.000 Linear 3 1090558 1090558 363519 62.25 0.453 0.4256 x3 131.6237 x1 177.983 0.000 x1 177.003 x2*x3 43.0011 x2 109.03 21.4656 x1*x1 32.481 x3*x3 -29.94 7.502 Interaction 3 143461 143461 47820 8. Estimated Regression Coefficients for y1 Term Coef SE Coef T P Constant 327.Chapter 13 Exercise Solutions 13-6. Since the runs are listed in a patterned (but not standard) order.97 3.MTW. Select “Terms” and verify that a full quadratic model is selected.720 0.43 17.38 31.06 31. Response Surface Regression: y1 versus x1.317 x2*x2 -22. x2.62 38.935 0. Select Stat > DOE > Factoriall > Create Factorial Design.82 0.008 x1*x3 75.5833 13-8 .08 1.099 0.000 x2 109.94 9. one approach to solving this exercise is to create a general full factorial design in MINITAB.328 0. The design is a full factorial of three factors at three levels. The design and data are in the MINITAB worksheet Ex13-6.363 x1*x2 66.08 -0.0056 x2*x2 -22.47 21.435 0.0578 x1*x2 66.74 0.866 0.00 17. Change the design type to a general full factorial design. (a) To analyze the experiment.4708 x2*x3 43.58 21.08 -0. x3 The analysis was done using coded units.005 0.000 x3 131. and select the number of factors as “3”.47 17.000 x1*x1 32. then select “Factors” to specify the actual level values.76 8. In order to analyze this experiment using the Response Surface functionality.000 Square 3 14219 14219 4740 0. and then enter the data.97 1.

156 x3 29.31 1.1917 x1*x1 4.629 data in uncoded units 13-9 .719 12.074 0.082 12.030 0.528 10.7 27170.99 Residual Error 17 32650.192 10.81 0.550 x1*x3 5.012 x1*x1 4.280 x2 15.3 7319.9 … Estimated Regression Coefficients for y2 using Term Coef Constant 34.1978 x2*x2 -1.82 Interaction 3 3408.89 0. x2.59 P 0.89 0.691 x2*x3 14.942 x3*x3 16.Chapter 13 Exercise Solutions 13-6 continued (b) To analyze the experiment.281 … Analysis of Variance for y2 Source DF Seq SS Adj SS Adj MS Regression 9 27170. Select “Terms” and verify that a full quadratic model is selected.3 21957.65 0.57 3. Estimated Regression Coefficients for y2 Term Coef SE Coef T P Constant 34.116 0.0 3408. Response Surface Regression: y2 versus x1.33 2.97 Linear 3 21957.564 0. select Stat > DOE > Response Surface > Analyze Response Surface Design.33 1.198 17.09 Square 3 1805.0 1135.0825 F 1.483 0.202 0.1083 x2*x3 14.235 0.815 0.65 1.890 22.89 -0.136 x1 11.8896 x1 11.3189 x3*x3 16.7192 x1*x3 5.33 1.2 1920.319 17.5 1805.65 0.817 x2*x2 -1.2 32650.323 10.5 601.7794 x1*x2 7.108 12.60 Total 26 59820. x3 The analysis was done using coded units.938 0.5278 x2 15.3233 x3 29.404 0.7 3018.113 0.361 x1*x2 7.31 0.610 0.779 17.826 0.

0 -1. set low = 400 and high = 600.0 x1 0.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.5 -1.0 -0.0 Overlaid Contour Plot of y1.0 y1 400 600 y2 0 80 0.5 x2 Hold Values x3 0 0.0 -0. Stat > DOE > Response Surface > Overlaid Contour Plot After selecting the responses. 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 x1 0.5 x2 Hold Values x3 -1 0.0 13-10 . select the first two factors x1 and x2.5 0. Overlaid Contour Plot of y1. y2 1. y2 1.5 -1.0 y1 400 600 y2 0 80 0.0 -1.0 -0. Since the goal is to hold y1 (resistivity) at 500.5 0.0 -0.5 1.5 1. Select “Contours” to establish the low and high contours for both y1 and y2.

y2 1.5 -1. the predicted resistivity mean is 495. Lower = 400. and Upper = 80.5 1.16 and standard deviation is 44. set Goal = Minimize. for y1 set Goal = Target.5 x2 Hold Values x3 1 0. 13-11 . Upper = 600.0 y1 400 600 y2 0 80 0. Leave all Weight and Importance values at 1.0.0 x1 0. The graph below represents one possible solution.75.4. x2 = 0.0 -1.3 and x3 = -0.0 -0. Target = 500. For y2.Chapter 13 Exercise Solutions 13-6 (c) continued Overlaid Contour Plot of y1.0 Stat > DOE > Response Surface > Response Optimizer In Setup.5 0. At x1 = 1. Target = 0.0 -0.

C.Chapter 13 Exercise Solutions 13-7.MTW. 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 . and then define the experiment using Stat > DOE > Factorial > Define Custom Factorial Design. The design and data are in the MINITAB worksheet Ex13-7. (a) The defining relation for this half-fraction design is I = ABCD (from examination of the plus and minus signs). All main effects are clear of 2-factor interactions. but some 2-factor interactions are aliased with each other. Enter the factor levels and response data into a MINITAB worksheet. B. Stat > DOE > Factorial > Analyze Factorial Design Factorial Fit: Mean versus A. 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. D.

0248 0.02021 3.1194 0.8090 0.2806 0.98 0.29 0.0456 0.0596 -0.501 Residual Error 32 0.0319 0.36769 18.0148 0. B.45224 23.0912 0. D.0198 0.01994 -5.000 B -0.02021 0.37800 0.62707 0.48 0.02021 -1.023 3-Way Interactions 3 0.83846 0. C.1210 0.83846 1.02021 -0.000 Residual Error 42 0. B.04726 0.05400 2.631 A*D*E -0.0396 0.02021 1.26 0.1529 0.02021 377.000 B -0.335 A*B*E 0.01960 Total 47 2.8908 13-13 .0115 0.6271 0.84 0.8090 1.000 A 0.1638 -0.469 A*C 0.01742 0.02021 -1.99 0.02021 -0.1742 0.41 0.0021 0.2421 0.0010 0.000 A*B -0.0765 0.01960 Pure Error 32 0.81 0.2806 0.0329 -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.0229 -0.000 C -0.0637 0.000 2-Way Interactions 1 0. E Estimated Effects and Coefficients for Height (coded units) Term Effect Coef SE Coef T P Constant 7.027 E -0.2387 -0.58 0.76 0.89 0.976 A*D -0.05 0.000 … Analysis of Variance for Height (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 4 1.000 A 0.01994 382.62707 0.0496 -0.03 0.959 A*C*E 0. D.62707 0.89078 The reduced model for mean: Factorial Fit: Height versus A.02021 2.73 0.01994 6.0765 0.1742 0. E Estimated Effects and Coefficients for Height (coded units) Term Effect Coef SE Coef T P Constant 7.01908 Lack of Fit 10 0.229 D 0.000 B*E 0.01994 2.0819 0.02021 -5.1210 0.02021 5.47 0.0819 0.575 A*E 0.07 0.71 0.124 B*E 0.80 0.01960 Total 47 2.76 0.02021 0.01575 0.04726 0.78 0.11 0.1529 0.421 D*E 0.01994 3.0196 0.1194 0.0012 0.62707 0.0098 0.0298 0.51 0.0165 0.554 Pure Error 32 0.02021 0.02021 0.02021 -0.6271 0.37800 0.0296 -0.91 0.23 0.0913 0.8012 0.2387 -0.05 0.031 E -0.57 0.02021 -4.71 0.000 2-Way Interactions 7 0.28060 14.1638 -0.2421 0.150 … Analysis of Variance for Height (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 5 1.99 0.001 C*E -0.0456 0.6256 0.000 D 0.8012 0.6256 0.01994 -4.0006 0.

D.01812 40 -0.05 0.03625 0.13875 0.10 -0.01313 A DE 95 0.06812 5 CE -0. Alpha = .026806 6.02437 -0.05 0.23 0. D.81 0.004225 Total 15 0.05688 0.58 0.01625 0.12625 -0.06125 0.88 0.31909 … 13-14 .000 A 0.03125 0.02625 0.01375 -0. C.01625 3.13403 0.074256 17.21937 Normal Probability Plot of the Effects 0.01375 -0.19 0.07701 0.12625 -0.011 2-Way Interactions 1 0.06812 0.07426 0.01625 -0.02625 0. B.02125 -0.003 Residual Error 8 0. B.03380 0.06937 0.06312 99 0.01313 0.88 0.008 B -0.050625 The reduced model for range: Factorial Fit: Range versus A.00375 -0.01625 -3.003 … Analysis of Variance for Range (coded units) Source DF Seq SS Adj SS Adj MS F P Main Effects 5 0.13625 -0.00687 0.05688 (response is Range.13403 0. C.01625 -4.07426 0.02188 70 -0.13875 0.077006 18.03062 90 A -0.03380 0.01625 1.00 0.07701 0.42 0.00188 30 20 0.04875 0.005 C 0.13625 -0.03062 0.06312 0.34 0.003 A*D*E 0.10 0.443 D 0.01625 0.21937 0.Chapter 13 Exercise Solutions 13-7 continued (c) The full model for range: Factorial Fit: Range versus A.003 3-Way Interactions 1 0.00812 B 10 -0.15 Effect 0.06125 0.01625 13.27 0.15 -0.683 C*E -0.11375 0.11375 0. E Estimated Effects and Coefficients for Range (coded units) Term Effect Coef SE Coef T P Constant 0.01688 60 50 0.01562 1 0. E Effects and Coefficients for Range (coded units) Effect Coef 0.01625 4.01063 0.03375 -0.50 0.096 E -0.04375 0.06937 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 = 0.00687 80 0.50 0.20) -0.

45 0.00438 CE 10 -0.000 A 0.07149 -0.06 -0.06 0. along with factors B (heating time) and A (furnace temperature) are significant.04 0.020438 0.00 0.11744 (response is StdDev.01997 0.007559 2.504 D 0.07148 -0. 13-15 .02 0.007559 -0.13 2-Way Interactions 1 0.03574 0.07149 -0.00953 30 -0.14 0.01768 0.08 0.36 3-Way Interactions 1 0.0232179 The reduced model for standard deviation: Factorial Fit: StdDev versus A.06 0.0233690 25.06259 0.00329 -0.04 -0.007559 15.01093 50 40 0.004 0.00165 20 0.01768 90 A -0. C.03822 0.03574 0.007559 -4.007559 -4.07148 -0.00528 A DE 95 0.11744 0.73 0.00684 -0.70 0.01556 0.007559 5.08 -0.07643 0.020438 0.00877 0.06259 0.663 C*E -0.00684 -0.00778 1 -0. Alpha = . B.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 = 0.73 0.03536 0. D.023369 0.34 0.023369 0.001 … Analysis of Variance for StdDev (coded units) Source DF Seq SS Adj SS Adj MS F Main Effects 5 0.00770 70 60 -0.092869 P 0. E Estimated Effects and Coefficients for StdDev (coded units) Term Effect Coef SE Coef T P Constant 0.03574 5 B -0.0009142 Total 15 0.007559 4.00528 0.00234 0.01057 0.00468 -0.02 0.00999 Effect 0.041748 0. D.007314 0.001 A*D*E 0.01906 0.00342 D 80 0. Factors C and E are included to keep the models hierarchical.07643 0.00342 0.007314 0.001 C 0. B.03129 99 -0.001 For both models of variability.03129 0.54 0.56 Residual Error 8 0. E Effects and Coefficients for StdDev (coded units) Effect Coef Normal Probability Plot of the Effects 0.0083496 9.0204385 22.02185 -0.007559 0. interactions CE (transfer time × quench oil temperature) and ADE=BCE.041748 0.047 E -0.Chapter 13 Exercise Solutions 13-7 (c) continued The full model for standard deviation: Factorial Fit: StdDev versus A. C.003 B -0.20) 0.01057 0.01540 0.03574 0.03536 0.001 0.

5 1.0 0 Residuals Versus D 0 -2 -0.0 A Residuals Versus C (response is Range) (response is Range) 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 -2.0 -0.0 0 9 10 11 12 13 14 15 16 0 -0.0 -2 -1.0 0.0 B 0.5 Residuals Versus C (response is Height) 2 -1.5 0.0 E 13-16 .2 Fitted Value 0.0 E Standardized Residual Standardized Residual For range: Residual Plots for Range Histogram of the Residuals 3.5 1.0 A 0.5 0.5 0.5 1.5 -1.0 1.5 Residuals Versus B Standardized Residual Standardized Residual -2 0.5 1.0 2 -1.5 1.5 0.5 Standardized Residual 0.5 0.00 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 -1.75 Fitted Value 8.0 -0.5 1 Observation Order 1 -1.5 2 0 -2 1.0 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 0.5 0.0 C Residuals Versus E (response is Height) (response is Height) Standardized Residual Standardized Residual 1.0 D 0.0 10 15 20 25 30 35 40 45 Observation Order -0.50 7.5 1.0 0.5 1 0 -1 1.4 Residuals Versus the Order of the Data 1 0 -1 1 2 3 4 5 6 7 8 (response is Range) -1.5 0.5 0.0 -0.0 -0.5 0.0 -1.0 1.0 0.0 1.0 -0.5 -1.5 -1 Residuals Versus B -1 -1.0 0.5 Standardized Residual Residuals Versus A Residuals Versus the Fitted Values (response is Height) 2 0 -2 7.0 1 0 -1 -1.0 B 0.0 1.0 -0.0 D 0.0 -1.5 0.0 0.0 (response is Height) 0 -0.0 -1.0 C Residuals Versus E (response is Range) (response is Range) Standardized Residual Residuals Versus D 1 0 -1 -0.0 0.0 -1.5 1.5 2 1.5 0.

5 0.2 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. Normal probability plot of residuals support normality assumption.5 Residuals Versus C (response is StdDev) 2 -1.5 0.0 -0.0 A 0.5 1.0 B 0.0 D 1.0 1. A resolution V design can be generated with E = ± ABCD.5 2 1. Plots of residuals versus each factor indicate that the variance may be different at different levels of factor D.0 Residuals Versus E (response is StdDev) Standardized Residual Standardized Residual 0. (e) This is not the best 16-run design for five factors.0 E Mean Height Plot of residuals versus predicted indicates constant variance assumption is reasonable. Plots of residuals versus each factor shows that variance is less at low level of factor E.5 0.1 0.0 0 Residuals Versus D -0.0 1.5 0. then none of the 2-factor interactions will be aliased with each other.0 C 1.0 Residuals Versus A Residuals Versus the Fitted Values (response is StdDev) 2 0 -2 0. Range Plot of residuals versus predicted shows that variance is approximately constant over range of predicted values.0 0.0 -0.5 -2 (response is StdDev) 0 -2 0.5 1.0 -0.0 2 0 -2 -1.0 0.5 -1.0 2 -1.5 0.0 0.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. 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.5 Standardized Residual 2.5 -1.5 Observation Order Residuals Versus B Standardized Residual Standardized Residual -2 0. Standard Deviation Residuals versus predicted plot and residuals normal probability plot support constant variance and normality assumptions.0 (response is StdDev) 0 -0.0 9 10 11 12 1 3 14 15 16 -0. 13-17 .

4 1 Y 0. h 14 51 04 X -0 .15 PO E(f re e 0.49 and minimum standard deviation of 0.0 0 E : o il te m p = 0 .00 -0.12A – 0.Chapter 13 Exercise Solutions 13-8. so Variance of Free Height = (–0.0 0 B: h ea t tim e 0.0 0 D: h o ld ti m e = 0 .0 0 -1 . 99 A ctu a l Fa cto rs C: tra n s tim e = 0 .50 f r ee h eight: 7. and optimal settings could be identified from visual examination of both plots.077B)2 + σ2 Assume (following text) that σ2E = 1 and σˆ 2 = MSE = 0.00 0.081B + 0.00 A: fu rn te m p 13-18 .63 + 0. and then contour plots in factors A. B. These contour plots could be compared with a contour plot of Mean Free Height. and D could be constructed using the Graph > Contour Plot functionality. Factor E is hard to control (a “noise” variable).02 For the current factor levels.99.5 0 -1. This approach is fully described in the solution to Exercise 13-12. Overlay Plot DES IG N-E XP ERT P lo t 1. Using equations (13-6) and (13-7) the mean and variance models are: Mean Free Height = 7.12 + 0.12 + 0. Free Height Variance could be calculated in the MINITAB worksheet.056 at A = –0.046D Variance of Free Height = σ2E (–0.00 -0 . 49 59 4 POE( free height): 0.077B)2 + 0.50 0 .50 1. 55 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 igh7.44 and B = 0. The overlaid contour plot below (constructed in Design-Expert) shows one solution with mean Free Height ≅ 7.02 .

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 .046)2 + (–0.99. B. Assume σ D2 = σ E2 = 1 .Chapter 13 Exercise Solutions 13-9.63 + 0. The overlaid contour plot below (constructed in Design-Expert) shows one solution with mean Free Height ≅ 7.12A – 0. Factors D and E are noise variables.1 52 23 3 X -0.0 0 E : o il te m p = 0 .50 f r ee h eight: 7.02 : Variance of Free Height = (0. Using equations (13-6) and (13-7).5 0 -1.9 9 A ctu a l Fa cto rs C: tra n s tim e = 0 . and D could be constructed using the Graph > Contour Plot functionality. 55 0.02 For the current factor levels. and then contour plots in factors A.00 0.42 and B = 0.081B Variance of Free Height = σ2D (+0. the mean and variance are: Mean Free Height = 7. Overlay Plot DES IG N-E XP ERT P lo t 1.12 + 0.0 0 B: h ea t tim e 0.0 0 D: h o ld ti m e = 0 .16 Y 0 . 42 POE( free height): 0. and optimal settings could be identified from visual examination of both plots. These contour plots could be compared with a contour plot of Mean Free Height.00 A: fu rn te m p 13-19 .4 95 01 PO E(f re e 0h.50 0 .077B)2 + 0.12 + 0.00 -0 .077B)2 + σ2 Using σˆ 2 = MSE = 0.50 and minimum standard deviation of Free Height to be: A = – 0. Free Height Variance could be calculated in the MINITAB worksheet.00 -0.0 0 -1 . This approach is fully described in the solution to Exercise 13-12.50 1.046)2 + σ2E (–0.

195 2.001 0. These values are used in the Excel and MINITAB data files.74 1217. The design and data are in the MINITAB worksheet Ex13-10.081 3.95 Square 3 1217.33 9. 70.227 uncoded units 13-20 .699 20.764 0.069 0.513 57.Chapter 13 Exercise Solutions 13-10.801 1.305 2.067 Total 19 2636.425 3.29 277.710 -2. 70.8862 x3*x3 0. one approach to solving this exercise is to create a general full factorial design in MINITAB.000 x1*x1 -14.49 Interaction 3 292. 100. Since the runs are listed in a patterned (but not standard) order.000 0.08 Residual Error 10 137.000 0. 100.730 0.764 -5.458 7.359 1.29 2499.6250 P 0.3613 x3 -6.805 0.723 23.710 2.74 405.000 x2 2.38 292.073 x2*x3 -13.027 0.MTW.95 … Estimated Regression Coefficients for y using data in Term Coef Constant 87.66 13.053 x1*x3 -7.17 329. x2.132 3.192 0. Stat > DOE > Response Surface > Analyze Response Surface Design Response Surface Regression: y versus x1.17 989. x3 The analysis was done using coded units. 83.008 0.689 5.914 29.8279 x2 1. 78. 80.000 x1 9.466 2. 68. 65.175 0.000 x3*x3 2. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 87.205 x3 -10.764 -8.000 0.176 1.3589 x1 5.305 2.8750 x1*x3 -2.66 137.446 x1*x2 8.33 45.766 Lack-of-Fit 5 92.04 Pure Error 5 45.689 1. 80.6250 x2*x3 -4.0578 x2*x2 -7. The correct values are: 66.794 0. and then enter the data.17 Linear 3 989. 60.005 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS F Regression 9 2499.7759 x1*x2 2.000 x2*x2 -22.33 92.356 0.689 -6. 75.526 0.33 18. 90.38 97. 85. 91.710 -3. 87. 82.289 1. 63. Note: Several y values are incorrectly listed in the textbook.0509 x1*x1 -5. 88.

Chapter 13 Exercise Solutions 13-10 continued Reduced model: Response Surface Regression: y versus x1.004 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS Regression 8 2490.587 0.0 -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) 5.0 -5.40 24.0 -2.660 1.685 0.61 2490.905 0.006 1.371 0.303 Lack-of-Fit 6 101.131 0.660 5.263 69.0 -2 -1 0 x2 1 2 -2 -1 0 x3 13-21 .07 604.0 -2.036 0.523 2.458 Residual Error 11 146.44 7.000 0.78 45.0 5.660 -6.000 0.994 1.000 x2 2.067 Total 19 2636.523 2.34 146.00 16.000 x1*x2 8.0 2.048 x1*x3 -7.534 Interaction 3 292.081 3.95 … F 23.704 -5. x2.176 1. x3 The analysis was done using coded units.229 0.000 x1*x1 -14.801 1.0 -2.379 0.000 x2*x2 -22.647 2.5 -5.647 -3.195 x3 -10.5 -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 0.000 x1 9.33 P 0.5 0.17 989.289 1.834 Pure Error 5 45. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 87.329 0.000 0.34 13.38 97.07 1209.425 3.5 Residual Residual 0 x1 0.33 9.723 Square 2 1209.067 x2*x3 -13.257 Residual Plots for y Normal Probability Plot of the Residuals (response is y) 5 90 Residual 50 10 5.00 101.132 3.38 292.5 2.647 -2.33 45.327 Linear 3 989.86 0.704 -8.5 0 -5.61 311.0 0 2.17 329.

0 90.0 50. x1 40. x2 = 0. x1 y < 70.0 100.292.Chapter 13 Exercise Solutions 13-10 continued Stat > DOE > Response Surface > Contour/Surface Plots Contour Plot of y vs x2.0 60.0 50. x1 y < > Hold Values x3 0 40. Weight = 1.0 80.0 80.0 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. 13-22 .0 1 x2 Surface Plot of y vs x2.0 70. x1 Surface Plot of y vs x3.0 90.0 80. Upper = 120.0 60. and x3 = −1.807. Importance = 1 One solution maximizing growth is x1 = 1.0 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. Lower = 60. Predicted yield is approximately 108 grams.0 > 1 Hold Values x2 0 70.682.0 80.0 70.0 100.

000 Residual Error 7 154. Stat > DOE > Response Surface > Analyze Response Surface Design Response Surface Regression: y versus x1.038 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS Regression 5 315.58 0. Since the runs are listed in a patterned (but not standard) order.60 46.00 F 2.712 1.017 13-23 .58 123.Chapter 13 Exercise Solutions 13-11.209 x1*x2 6.463 1.058 Lack-of-Fit 3 139.60 315.119 Linear 2 48. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 41.80 3.00 144.660 -1.86 1.878 0.02 48.555 0.80 6.09 2.388 0.076 x2*x2 2.102 0.038 12. and then enter the data.616 0.40 22.000 2.011 Square 2 123.80 14. x2 The analysis was done using coded units.00 144.100 19.60 63. The design and data are in the MINITAB worksheet Ex13-11.409 x1*x1 3.274 x2 1.457 1.660 0.383 0.40 154.788 Interaction 1 144.53 P 0.085 0.700 Total 12 470.348 2.60 139.58 61.000 x1 -1.781 1.781 2.970 1.200 2.02 24.186 0.534 Pure Error 4 14.MTW. one approach to solving this exercise is to create a general full factorial design in MINITAB.128 0.

Upper = 55. x1 1.0 -0. Weight = 10. to achieve predicted filtration time of 46.0 x1 0.0.0 1 0 -1 x1 -1. to achieve predicted filtration time of 36.5 0 x2 -1 1 1. (b) Goal = Target.3415 and pressure = -0.4142.4109 and pressure = -1.5 x2 Surface Plot of y vs x2.5 0. Importance = 1 Recommended operating conditions are temperature = +1. Lower = 42.0 (a) Goal = Minimize. 13-24 .Chapter 13 Exercise Solutions 13-11 continued Contour Plot of y vs x2.0 40 45 50 55 0.7.0 60 y -0. Upper = 50. Importance = 1 Recommended operating conditions are temperature = +1. Target = 46.0785. x1 y < > 40 45 50 55 60 60 70 0. Target = 0. Weight = 1.5 50 40 -1.

8333 3.945 x1*x2 8.38 97.448 -1.003 x2*x2 -21.116 1.232 0.38 292.6250 P 0.75 Residual Error 8 135.448 -3.1517 x1*x2 2.3333 x1 5.94 226.001 0.021 0.22 90.1250 1.115 uncoded units The coefficients for x1z and x2z (the two interactions involving the noise variable) are significant (P-values ≤ 0. The design and data are in the MINITAB worksheet Ex13-12.34 Linear 3 789.8908 x2*x2 -7.7783 2.MTW Stat > DOE > Response Surface > Analyze Response Surface Design Response Surface Regression: y versus x1.Chapter 13 Exercise Solutions 13-12.361 -6.8279 x2 1.178 0.56 16.496 0.458 5.1317 4.804 0.53 Square 3 953. 13-25 .222 0.4147 2.116 0.56 135. x2.105 13.28 263.6250 x2*z -4.067 Total 17 2170.945 Lack-of-Fit 3 90.1517 2.256 z -6.001 0.092 15.209 0.001 x2 2.50 … Estimated Regression Coefficients for y using data in Term Coef Constant 87.33 9.29 953.873 1.764 18.074 3.109 x2*z -7.003 x1*x1 -13.000 x1 9.8750 x1*z -2.1250 x1*x1 -4.681 51.7192 z*z 0.32 Pure Error 5 45.975 0.3613 z -6.8333 3.8013 1.29 317.28 789.75 Interaction 3 292.000 z*z 0.968 0.361 -4.94 2034.116 0.2894 1.013 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS F Regression 9 2034.001 0.3333 1. so there is a robust design problem.072 0.873 5.10). Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 87. z The analysis was done using coded units.22 30.33 45.084 x1*z -4.455 -4.

767 5.31 22.462 0.64 15.5 1. z The analysis was done using coded units.548 0.019 -7.0 13-26 .5 -5.132 3.47 P 0.000 0.172 Residual Plots for y Normal Probability Plot of the Residuals Residual 10 5.28 789.361 1.5 0 -5.357 Linear 3 789.373 -4.000 0.64 135.49 0.008 … Analysis of Variance for y Source DF Seq SS Adj SS Adj MS Regression 8 2034.86 2034.31 90.86 254.20 476.0 -2.50 … F 16.000 x2 2.62 6.000 0. x2.5 0.558 0.002 x1*x1 -13.778 2.767 1.0 -0.0 -2 -1 0 x2 1 2 -1.675 0.000 x1 9.5 Residual Residual 0 x1 0.882 2.125 1.912 0.019 -4.0 -5.001 x2*x2 -21.0 5.415 2.760 3.Chapter 13 Exercise Solutions 13-12 continued Reduced model: Response Surface Regression: y versus x1.20 953.308 -3.801 1.072 Lack-of-Fit 4 90.33 9.541 56.013 2.38 97.5 2.296 0.092 Square 2 953.33 45.208 0.602 Interaction 3 292.095 0.370 0.289 1.578 Pure Error 5 45.088 x2*z -7.0 5 Residual 3.000 x1*x2 8.308 -1.28 263. Estimated Regression Coefficients for y Term Coef SE Coef T P Constant 87.46 31.88 17.5 -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 0.0 -2.0 Residuals Versus x1 Residuals Versus the Fitted Values 99 -2.0 2.458 Residual Error 9 135.760 3.0 Frequency (response is y) 5 90 1.227 z -6.5 0.066 x1*z -4.5 0.0 z 0.067 Total 17 2170.0 0 2.0 -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) 5.38 292.

x1 y < > -1 -1 Contour Plot of sqrt{Vz(y(x.0 80.63x2)2 + σˆ 2 = (–6.0 80.0 80.36 + 5.072 This equation can be added to the worksheet and used in a contour plot with x1 and x2.MTW.0 1 x2 50.0 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 = –1.63x1 – 4. There are other combinations that would work. 13-27 .63x1 – 4.0 70.0 70. assume σz2 = 1: Variance of Yield = σz2 (–6.86x12 – 7.63x2)z For the mean yield model.11 and x2 = – 0.0 80. x1 1 30.13 – 2.1692 50.0 40.0 1 -1 Contour Plot of y vs x2.308367 y = 90.0 50.0 80.393500 x2 = 0.0 90.10 10 .69x22 For the variance model.0 50. x1 x2 x1 = -0.0 70. (Refer to MINITAB worksheet Ex13-12.109756 x2 = -0.36x2 – 4.2054 0 x1 40.0 40.) Contour Plot of y vs x2.708407 x2 = -0.13 – 2.13 – 2.83x1 + 1.0 60.0 50.0 70.12 > 12 0 40. Comparison with the contour plot for variability shows that growth greater than 90 with minimum variability is achieved at approximately x1 = – 0.0198 -1 x1 = 0.0 60. set z = 0: Mean Yield = 87.z)]} vs x2.0 60.0 70.36x2 – 4.0 60.31 (mean yield of about 90 with a standard deviation between 6 and 8).293782 y = 90.63x2)2 + 15.555312 y = 90.Chapter 13 Exercise Solutions 13-12 continued yPred = 87.0 60.0 80.86x12 – 7.0 Hold Values z 0 0 1 1 y < > y < > 30.0 90.0 50.z)]} < 6 6 8 8 . x1 sqrt{Vz(y(x.0 1 x2 x2 -1 0 x1 Contour Plot of y vs x2.0 60.36 + 5.0 Hold Values z -1 0 x1 = -0.69x22 + (–6.63x1 – 4.83x1 + 1.

z ) = ∑ γ i zi + ∑ ∑ δ ij xi z j . z ) = ∑ γ i zi + ∑ ∑ δ ij xi z j + ∑ ∑ λij zi z j + ∑ θi zi2 . 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. then i =1 i =1 j =1 ( V [ y (x. z ) r = ∑ γ + ∑ ∑ δ ui xu + ∑ ∑ λij ( zi + z j ) . and u =1 ∂zi r If h(x. 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. z ) = ∑ γ i zi + ∑ ∑ δ ij xi z j + ∑ ∑ λij zi z j . i =1 i =1 j =1 i< j =2 r k r ∂h(x. z ) = ∑ γ i + ∑ ∑ δ ui xu + ∑ ∑ λij ( zi + z j ) + 2∑ θ i zi . 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. r k r r r i< j =2 i =1 If h(x. 13-28 . z ) = γ i + ∑ δ ui xu .Chapter 13 Exercise Solutions 13-13. 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. 13-14. and i =1 i =1 i i =1 u =1 i< j =2 ∂zi r then ∑ r ⎡ k r ⎤ V [ y (x. r k k ∂h(x. then i =1 i =1 j =1 r ∂h(x.

05324 0.01547 0.000 0.95121 0.27943 0.020 0.40 0.55528 0.09994 0.70382 0.21807 0.080 0. c=1 1.007 0.06725 0.04428 0.63633 0.03379 Type-B OC Curve for n=50.96353 0.36417 0. The BINOMDIST.020 0.28895 0.92528 0.070 0.00 Pr{acceptance} 0.73577 0. Solutions are in the Excel workbook Chap14.20 1.100 f(d=0) 0.99540 0.19000 0.81840 0.90475 0.002 0.04533 0.91056 0.090 0.30556 0.77831 0.20249 0.Chapter 14 Exercise Solutions Note: Many of the exercises in this chapter are easily solved with spreadsheet application software.00 0.005 0.26986 0.19556 0.001 0.009 0.14467 0.98998 0.22339 0.050 0.030 0.12649 0.040 0.95190 0.02863 Pr{d<=c} 0.97387 0.xls. and graphing functions in Microsoft® Excel were used for these solutions.04761 0.07694 0.080 0.12947 0.74015 0.00515 f(d=1) 0.08271 0.010 0.060 0.006 0.40048 0. p 0.09066 0.100 0.99881 0.02656 0.60 0.37160 0.80 0.040 0.120 0.98274 0.008 0. HYPGEOMDIST.00896 0.004 0.140 p 14-1 .93910 0.33721 0.60501 0.16434 0.24807 0.003 0.27060 0.060 0.12989 0.66924 0. 14-1.20 0.86051 0.

c=2 1.09057 0.140 p 14-2 .20 1.27065 0.60 0.27341 0. p 0.05661 0.020 0.60577 0.007 0.009 0.34920 0.07572 0.90479 0.00079 0.00000 Type-B OC Curve for n=100.36973 0.00592 0.090 0.99225 0.08118 0.01312 0.02579 0.04755 0.100 0.12185 0.050 0.040 0.00208 0.060 0.100 0.008 0.01978 0.030 0.26899 0.120 0.010 0.18486 0.22515 0.67669 0.09880 0.92063 0.98590 0.14498 0.07029 0.33068 0.36603 0.070 0.99888 0.66978 0.00194 0.006 0.00000 Pr{d<=c} 0.00531 0.Chapter 14 Exercise Solutions 14-2.060 0.44789 0.05347 0.00 0.36773 0.01687 0.99649 0.005 0.00 Pr{acceptance} 0.00205 0.004 0.00008 0.40 0.200 f(d=0) 0.03319 0.00388 0.41978 0.13262 0.36120 0.22281 0.00000 f(d=2) 0.00030 0.99985 0.00000 f(d=1) 0.01127 0.49536 0.20 0.040 0.00162 0.54782 0.97730 0.81857 0.95327 0.16531 0.04144 0.002 0.74048 0.93796 0.00476 0.00003 0.11826 0.080 0.00071 0.14707 0.30441 0.16404 0.96641 0.020 0.000 0.080 0.00449 0.80 0.001 0.40492 0.003 0.03116 0.23214 0.00024 0.14419 0.00895 0.01627 0.

000 0.080 0.10 (c) Based on values for α and β.100 0.10 (b) Type-B OC Curve for N=5000. (a) Type-A OC Curve for N=5000.00 0.140 p Pa (p = 0.05 Pa (d = 375) = 0.100 0. 14-3 .20 0. c=1 1.060 0.00 Pr{acceptance} 0.040 0. either is appropriate.10133.040 0. or α ≅ 0.075) = 0.800 0. n=50.9521.9521.200 0. or α ≅ 0.000 Pr{acceptance} 0.140 p Pa (d = 35) = 0.10133. or β ≅ 0.Chapter 14 Exercise Solutions 14-3.60 0.120 0.400 0.600 0. or β ≅ 0.120 0.020 0.80 0. c=1 1.007) = 0.000 0.060 0. the difference between the two curves is small.080 0.020 0.40 0.05 Pa (p = 0. n=50.000 0.

10 From the binomial nomograph. select a sampling plan of n = 300 and c = 12.12238. 1 − α = 1 − 0.10. 1 − α = 1 − 0. 1 − α = 1 − 0. p1 = 0. p2 = 0.99. p1 = 0.05 = 0.01 = 0. 14-4 .95.Chapter 14 Exercise Solutions 14-4. β = 0. p2 = 0. the sampling plan is n = 80 and c = 7. select n = 35 and c = 1.05.05 = 0.06. p1 = 0. β = 0.10 From the binomial nomograph. 14-5.04786 and β = 0. β = 0. p2 = 0. resulting in actual α = 0.02. 14-6.95.10 From the binomial nomograph.15.01.

00000 -0.99729 0.00000.00046 and Pa(N = 10.00000 0.0500 0.02395 0.0000 1.0250 0.00938 0.00000 Different sample sizes offer different levels of protection.0010 0.0003 1. but pays for the high probability of rejecting acceptable lots like those with p = 0.08151 0.0000 0.0700 N1 = n1 = pmax = cmax = binomial Pr{d<=10} 1.05 by Pa(N = 5.00000 1. Pa(p = 0.Chapter 14 Exercise Solutions 14-7.00000 0.0020 0.55910 0.99991 0.0000 1.99999 0.9903 0.0015 0.0300 0.18221 0.0000 0. LTPD = 0.99359 0. For N = 5.00030 0.58304 0. Pa(p = 0.0000 0.4409 0.8852 0.00001 -0.0001 0.0000 0.99903 0.0000 1.0004 0. while for N = 10.182.00000 0.03328 0.0000 0.99850 0.00000 1.98676 0.00600 -0.00001 0. the consumer is protected from a LTPD = 0.9667 0.0000 0.00000 1.0200 10 N2 = n1 = pmax = cmax = binomial Pr{reject} Pr{d<=20} 0.0000 1.00027 -0.00006 -0.00001 0.8178 0.00000 0.0010 0.00263 -0.0600 0.11479 0. 14-5 .05 p 0.025.0040 0.294.99994 0. Also.11183 0.00000 10000 1000 0.0050 0.000) = 0.00967 0.99972 0.9997 1.000) = 0.00046 0.0000 0.7060 0.0200 20 Pr{reject} 0.99959 0.0060 0.0090 0.0001 1.0132 0.0027 0.0000 0.0000 difference 0.00000 0.0000 1.0100 0.00000 0.4170 0.025) = 0.9995 0.0400 0.000.0030 0.00000 0.99999 0.00046 0.000.0000 0.01175 0.0070 0.025) = 0.00000 0.0200 0.00000 1.00000 5000 500 0.0064 0.0000 1.29404 0.00095 -0.0080 0.00000 0.

0500 0.2204 0.38898 0.00292 0. 14-6 .0002 0.73230 0.9994 1.0100 0.10543 0.9868 0.0080 0.0008 0.00000 1.11858 0.0900 0.97238 0.1000 0.7569 0.0000 1.9938 0.28210 0.0305 0.3500 N1 = n1 = pmax = cmax = binomial Pr{d<=0} 0.02758 0.2446 0.56992 0.49484 0. p 0.0300 0.05374 0.93236 0.5052 0.80432 0.0006 0.98157 0.9618 0.0004 0.7961 0.0200 0.0889 0.75558 0.0276 0.00136 0.2677 0.01 0 N2 = n1 = pmax = cmax = binomial Pr{reject} Pr{d<=1} 0.75536 0. For example.20391 0.61157 0.1957 0.98767 0.6110 0.2444 0.9463 0. at p = 0.2959 0.0062 0.9971 0.05741 0.0184 0.98610 0.0245 0.01315 0.9986 0.9724 0.3000 0.0123 0.9990 0.0000 0.3884 0.03815 0. depending on the lot size.97550 0.94552 0.0030 0.01.93982 0.77958 0.0000 This plan offers vastly different protections at various levels of defectives.91107 0.4654 0.7323 for N = 1000.7179 0.1702 0.00063 0. and Pa(p = 0.0411 0.86924 0.8946 0.0700 0.99382 0.0000 0.00000 1.4301 0.0676 0.1308 0.82981 0.8814 0.4689 0.01) = 0.88316 0.0400 0.8531 0. Pa(p = 0.9246 0.01 1 Pr{reject} 0.0040 0.0139 0.24312 0.0020 0.00000 1000 32 0.70407 0.00000 5000 71 0.0000 1.0060 0.01) = 0.0050 0.53107 0.53457 0.0602 0.9426 0.65622 0.Chapter 14 Exercise Solutions 14-8.00099 0.85608 0.0545 0.95886 0.0070 0.1168 0.0800 0.0010 0.0090 0.1897 0.00622 0.07541 0.3438 0.81033 0.00002 0.96946 0.1439 0.4949 for N = 5000.14688 0.2000 0.0600 0.

n = 35.020 AOQ 0.00 0.000 0.05 0.005 0.20 0. N = 2. c=1 2500 2000 ATI 1500 1000 500 0 0. c = 1.15 0.25 0.25 p AOQ Curve for n=35.015 0.Chapter 14 Exercise Solutions 14-9.010 0.0234 ATI Curve for n=35.05 0. c=1 0.00 0.30 p 14-7 .15 0.20 0.025 0.000 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 = 0.10 0.10 0.

c=2 1.030 0.0000 3000 (a) OC Curve for n=150.060 0.05840 0.0028 181 0.000 0.0003 2985 0.0009 151 0.00523 0.004 0.91092 0.045 0.0019 160 0.00036 0.0000 3000 0.42093 0.0009 2948 0.100 Pa=Pr{d<=2} 0.99951 0.070 0.030 0.007 0.009 0.0053 328 0. c = 2 p 0.0067 491 0.010 0.95991 0.27341 0.0048 2517 0.010 0.0000 2999 0.80948 0.008 0.090 0.00009 0.015 0.0087 AOQL 1265 0.025 0.020 0.03292 0. N = 3000.40 0.20 0.070 0.080 0.0037 215 0.0065 2221 0.00142 0.16932 0.10098 0.93769 0.035 0.006 0.001 0.80 0.60884 0.84615 0.0001 2996 0.00 Pr{accept} 0.Chapter 14 Exercise Solutions 14-10.0080 1800 0.00002 AOQ ATI 0.97716 0.0014 2906 0. n = 150.050 0.003 0.0077 693 0.0034 2712 0.99646 0.005 0.020 0.002 0.01815 0.98927 0.040 0.080 p 14-8 .0022 2834 0.00 0.0061 404 0.60 0.0046 264 0.040 0.060 0.0072 588 0.88019 0.050 0.

0070 AOQ 0.060 0.020 0.070 0.0090 0.0087 0.0010 0.0030 0.050 0.0000 0.010 0.060 0.030 0.0100 0.030 0.000 0.000 0.0020 0.080 p (c) ATI Curve for n=150.Chapter 14 Exercise Solutions 14-10 continued (b) AOQ Curve for n=150.010 0.040 0. c=2 3500 3000 2500 ATI 2000 1500 1000 500 0 0.020 0.040 0.080 p 14-9 .050 0.0050 0.070 0.0080 0.0040 0. c=2 AOQL ≅ 0.0060 0.

0300 0.98957 0.68921 0. n = 50.00129 0.99794 0.99657 0.1600 0.80 0.2000 Fraction defective.0030 0.90015 0.0010 0.09614 0.00015 0.00526 0.67671 0. c = 2 p Pa=Pr{d<=1} Pr{reject} 0.00109 0.0800 0.00048 0.0070 0.99474 0.11173 0.1010 0.0500 0.32329 0.20 0.0080 0.00000 1.22597 0.99891 0.1200 0. c=2 1.99952 0.0600 0.40 0.0020 0.89236 0.41625 0.1400 0.10764 0.0050 0.0100 0.0000 0.1040 0.0090 0.0400 0.0060 0.83946 0.0200 0.00758 0.00 Probabilty of Acceptance.09985 0.1000 0.99985 0.88827 0.00 0. Pa 0.1020 0.99871 0.31079 0.0700 0.0800 0.16054 0.00002 0.45947 0.0400 0.58375 0.00206 0.0040 0.1030 0.99242 0.18920 0.99998 0. (a) N = 5000.89632 0.1000 0.10368 0.1050 0.00000 OC Curve for n=50.81080 0.90386 0.07843 0.1800 0.2000 0.90745 0.3000 0.54053 0.0200 0.01043 0. p 14-10 .0600 0.09255 0.00343 0.01382 0.77403 0.0900 0.92157 0.98618 0.Chapter 14 Exercise Solutions 14-11.60 0.

99996 1.29011 0.0090 0.81791 0. yielding Pa = 1 – 0. select n = 20.15164 0.0050 0.87842 0.Chapter 14 Exercise Solutions 14-11 continued (b) p = 0.83459 0.11372 = 0. with acceptance number c = 0.90.03925 0.23424 0.85160 0.00000 Pr{reject} 0.12158 0.00000 14-11 .70989 0.0080 0.0300 0.0030 0.16541 0.84836 0.00080 0. (c) A zero-defects sampling plan. quality improvement begins with the manufacturing process control.88660 0.1030 will be rejected about 90% of the time. not the sampling plan.0600 0.96075 0.1000 0.01153 0. Generally.99920 0.0700 0.0800 0.0040 0.66761 0. The OC curve for this zero-defects plan is much steeper. p 0.33239 0.0070 0.0200 0.86893 0.0400 0.07703 0.76576 0.98019 0.54379 0.5000 Pa=Pr{d<=0} 0.35849 0.18209 0.09539 0.14840 0.18869 0.0500 0.00004 0.0100 0.90461 0.45621 0.2000 0.98847 0.44200 0. (d) From the nomograph.94168 0.4000 0. will be extremely hard on the vendor because the Pa is low even if the lot fraction defective is low.05832 0.0020 0.88638 ≈ 0.81131 0.0060 0.64151 0.0900 0.13107 0.0010 0.3000 0.92297 0.55800 0.01981 0.11340 0.

40 0. c = 2} = 0.00206 ATIc =0 = n + (1 − Pa )( N − n) = 20 + (0.005.00206)(5000 − 50) = 60 The c = 2 plan is preferred because the c = 0 plan will reject good lots 10% of the time. p (e) Pr{reject | p = 0.00 0.20 0. Pa 0.80 0. c=0 1.09539)(5000 − 20) = 495 ATIc = 2 = 50 + (0.0500 0.2500 Fraction defective. c = 0} = 0.09539 Pr{reject | p = 0.60 0.1000 0.Chapter 14 Exercise Solutions 14-11 (d) continued OC Curve for n=20.0000 0.00 Probability of acceptance.2000 0. 14-12 .005.1500 0.

0001 0.0000 0.5974 0.0002 0.d2<=1} Pr{d1=6.0000 0.0020 0.0000 0.9737 0.040 0.0002 0.0514 0.0521 0.025 0.0098 0.0000 0.030 0.9216 0.0001 0.0000 0.0627 0.0152 0.020 0.0791 0.0209 0.8883 0.0238 0.0000 0.005 0.60 0.0138 0.0190 0.5405 0.0098 0. n2=100.d2=0} 0.00 0.4162 0.0793 0.6078 0.0120 0.0030 0.1605 0.9237 0.0567 0.0708 0.0001 0.d2<=3} Pr{d1=4.0331 0.0002 0.115 0.0000 0.3610 0.0568 0.5237 0.0011 0.0004 0.0161 0.040 0.0011 0.140 0.1892 0.0142 0.6767 0.060 0.20 0.0784 0.0000 0.d3<=2} Pr{d1=5.0000 PaII Pa 0.0073 0.40 0.0001 0.0193 0.1128 0.090 0.0000 0.0004 0.0522 0.150 0.0019 0.0025 0.1117 0.8928 0.075 0.0021 0.2770 0.0051 0.0767 0.140 0.8706 0.0000 0.0142 Primary and Supplementary OC Curves for n1=50.0023 0.9979 0.0000 0.0001 0.080 0.80 Pr 0.7342 0.0073 0.7452 0.3922 0.0011 0. c1 = 2.0029 0.9862 0.0513 0.0000 0.4763 0.0000 0.0690 0.0330 0.0629 0.7560 0.0000 0.2260 0.0000 0.1294 0.065 0.8108 0.020 0.8395 0.2333 0.4595 0.0006 0.0002 0.0000 0.9661 0.0002 0.0238 0.0339 0.6770 0.0019 0.0024 0.0000 0.0444 0.110 0.0212 0.0004 0. c2 = 6 P d1 = PaI PrI 0.070 0.8296 0.0000 0.000 0.0112 0.0000 0.035 0.050 0.9373 0.9487 0.0011 0.0166 0.0019 0.3108 0.0001 0.0029 0. c1=2.0013 0.130 0.0000 0.0009 0.0000 0.2548 0.6892 0.0004 0.7740 0.0820 0.5838 0.3848 0.0001 0.3233 0.160 p.0035 0.2658 0.120 0.0001 0.055 0.9779 0.3274 0.0002 0.120 0.0001 0.Chapter 14 Exercise Solutions 14-12.0221 0.4494 0.0111 0.0120 0.0818 0.00 0.9414 0.0000 0.100 0.5207 0.0002 0.0221 0.0692 0.0008 0.0842 0.0339 0.0001 0.0165 0.0763 0.045 0.060 0.100 0.9982 0. n1 = 50. c2=6 1.010 0.0844 0.0000 0.1635 0.0129 0.9858 3 4 5 6 Pr{d1=3. proportion defective Pr{accept on 1st sample} Pr{reject on 1st sample} Pr{accept lot} 14-13 .9999 0.0000 0.0001 0.0001 0.0028 0.0001 0.0001 0.0000 0.0015 0.080 0.0029 0.0442 0.0000 0.0072 0.6390 0.0000 0.0707 0. n2 = 100.0000 0.

025 0.01. s = 0.325 1.95.2054.Chapter 14 Exercise Solutions 14-13. Acc = 1.034 3.055 … 0.199 … 2. (b) Three points on the OC curve are: p1 = 0. (a) p1 = 0.0397 n.104 -0.080 2.5621 h1 + h2 0.364 1.2054 p2 = 0.193 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 = 49.820 -0.2054 = = 0.10. Pa = h2 1.064 -0. Pa = β = 0.0397 X A = −0.890 0. h2 = 1. Pa = 1 − α = 0.0414.040 2.969 1.009 1.1 − α = 1 − 0.074 3. h1 = 0. Rej = 4.929 0.000 2.95 p = s = 0. p2 = 0.9389 + 0.049 XR 1.10.9389.9389 + 1.859 -0.120 2. β = 0.0397n n 1 2 3 4 5 … 20 21 22 23 24 25 … 45 46 47 48 49 50 XA -0.404 … 2.015 0.153 3.780 -0.0397.159 2.994 3.899 -0.2054 + 0.850 0.144 -0.285 1.01.10 k = 1. X R = 1.245 1.740 … -0.05 = 0.10 14-14 .113 3.

406 1. β = 0.342 0.15.474 0.0660 X A = −1.991 2.3399 = = 0.0660.670 … 2.10 k = 0.309 4.572 4.659 2.375 4.0660n n 1 2 3 4 5 … 20 21 22 23 24 25 … 45 46 47 48 49 50 XA -0.95 p = s = 0.507 4.10 14-15 .472 1.02.638 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 = 49.057 2.408 0.925 1.3399 p2 = 0. h1 = 1.15.923 2.441 4.978 -0. s = 0.0436.0660n.05 = 0.989 … 4.9369. Pa = β = 0.714 … 0.02. Acc = 1 and Rej = 4.791 2.276 0.846 -0. (a) p1 = 0.538 1. p2 = 0. Pa = h2 1.540 0.857 2.1 − α = 1 − 0. h2 = 1.604 1.725 2.255 XR 1.123 2. Pa = α = 0.780 -0.0436 + 0. (b) p1 = 0.0436 + 1.5622 h1 + h2 1.3399.95. X R = 1.Chapter 14 Exercise Solutions 14-14.606 … 1.912 -0.189 2.3399 + 0.

000. Ac = 3. n = 50. Ac = 2. N = 10. n = 80. AQL = 1% General level I Normal sampling plan: Sample size code letter = H. Re = 3 Reduced sampling plan: n = 50. Re = 2 Tightened sampling plan: Sample size code letter = J. Ac = 0. n = 20.Chapter 14 Exercise Solutions 14-15. Ac = 0. n = 125. Re = 4 Tightened sampling plan: n = 125. Ac = 0. Ac = 1. General inspection level II. Ac = 0.10%. N = 3000. Re = 1 Tightened: n = 200. Sample size code letter = L Normal: up to letter K. AQL = 1% General level II Sample size code letter = K Normal sampling plan: n = 125. AOQ = [ Pa × p × ( N − n)] [ N − Pa × (np) − (1 − Pa ) × ( Np) ] 14-16. Re = 2 14-18. N = 3000. Ac = 1. Re = 1 14-16 . Re = 1 Reduced: up to letter K. n = 50. Re = 2 Reduced sampling plan: Sample size code letter = H. AQL = 0. Ac = 1. Re = 4 14-17.

Re = 4 (b) N = 5000 n= c= p 0.0300 0.6767 0.8580 0.9202 0. AQL = 0.00 0.9970 0.0395 0.0070 0.60 0.20 0.0101 0. Ac = 2.2351 0. AQL=0. II.0000 0. (a) N = 5000.65%. proportion defective Normal Tightened Reduced 14-17 .0211 0.0900 0.9992 0. Pr{acceptance} 0.7309 0.9967 0.9771 0.8800 0.0023 0.0593 0.8340 0.0200 0.4315 0. Ac = 1.9922 0.1472 0. General level II.0018 0.8092 0.80 0.9667 0.7838 0.0861 0.65% 1.0700 0.8653 0.0000 0.00 Pa.5230 0.0090 0.0090 0.0433 0.0100 0.9999 0.0010 0.9886 0.40 0. Sample size code letter = L Normal sampling plan: n = 200. Re = 4 Tightened sampling plan: n = 200.1000 p.0047 0.0030 0.0000 0.0125 0.0400 0.0004 0. Ac = 3. Re = 3 Reduced sampling plan: n = 80.3038 0.Chapter 14 Exercise Solutions 14-19.1654 0.0200 0.0022 OC Curves for N=2000.0800 0.9389 0.0600 0.9529 0.9989 0.0500 0.8377 0.9911 0.0000 0.9220 0.0080 0.0600 0.9163 0.9588 0.0400 0.0020 0.9469 0.0003 0.0800 0.0001 0.0000 tightened 200 2 Pa=Pr{d<=2} 0.0060 0.0000 reduced 80 1 Pa=Pr{d<=1} 0.8922 0.0040 0.1000 normal 200 3 Pa=Pr{d<=3} 0.8916 0.0050 0.9756 0.0001 0.9813 0.

013 0.003 0.010 0.020 p.0002 0.4361 0.0005 0.008 0.25% n = 490. without replacement. AOQL = 0.0046 0.0015 0.0068 0.012 0.2% p 0.0002 0. A more precise solution would use the hypergeometric distribution to represent this sampling plan of n = 490 from N = 2000.0001 0.0021 0.017 0.000 0. N = 2000.20 500 0 0.0150 0.0664 0. Note that this solution uses the cumulative binomial distribution in a spreadsheet formulation. c=2 1.0001 0.018 0.012 0.0102 0.9235 0.012 p.5564 0.011 0.0001 0.002 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 0.010 0.0007 0.010 0.016 0.2489 0. LTPD = 1%.0220 0.0014 0.6875 0.006 0.008 0.015 0.0006 0.0464 0.40 0. proportion defective 0.0018 0.c=2.008 0.016 0.007 0.00 2000 0.018 0. Pr{acceptance} 2500 ATI 1500 1000 0.0012 0.1320 0.000 0.n=490.020 0.80 Pa.2886 0.014 0.005 0.002 0.Chapter 14 Exercise Solutions 14-20. c = 2.0008 0. p = 0.0021 AOQL 0.0000 OC Curve for N=2000.00 0.0016 0.009 0.8165 0. 14-18 .014 0.002 0. n=490.3330 0.0018 0.001 0.006 0.006 0.008 0.004 0. proportion defective The AOQL is 0.016 0.0942 0.004 0.014 0.60 0. AOQL=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 0.21% ATI Curve for N=2000.21%.004 0.0020 0.1827 0.9864 0.019 0.018 0.0010 0.0003 0.0321 0.

250 p let N = 50. n=65.160 Fraction defective.060 0.180 0. average process fallout = p = 0.000 ATI Probability of acceptance.00000 50. n=65. Pa OC Curve for Dodge-Romig. 001 Pa = Binom(3. n = 65.20000 0.040 0.120 0. c = 3.200 0.080 0.000 0.Chapter 14 Exercise Solutions 14-21.000. 65. AOQL = 3%.100 0.000 0.50% defective (a) Minimum sampling plan that meets the quality requirements is 50.200 0 0.80000 40.3% 14-19 .60000 0.20000 0.001.100 0.00000 0. the average inspection required will be 82 units.150 0.40000 20. Dodge-Romig single sampling.140 0.99967 ATI = n + (1 − Pa )( N − n) = 65 + (1 − 0. (b) ATI Curve for N=50. p 0.000 30.000 10. c=3 60.000 0.005) = 0.99967)(50. (c) LTPD = 10.050 0.000 1.020 0. 001 − 65) = 82 On average.000 0.001 ≤ N ≤ 100. 0. c=3 1. if the vendor’s process operates close to process average.

25% n = 46. AOQL = 3%. AOQL = 3%.3% (b) c 3 d =0 d =0 Pa = ∑ binomial(n.9958)(8000 − 65) ≈ 98 (c) N = 8000. 0. 0.9998)(8000 − 46) ≈ 48 14-20 . c = 3. p ≤ 1% n = 65. LTPD = 10.0025) = 0.01) = 0. LTPD = 11.9958 ATI = n + (1 − Pa )( N − n) = 65 + (1 − 0.9998 ATI = n + (1 − Pa )( N − n) = 46 + (1 − 0.Chapter 14 Exercise Solutions 14-22.6% c 2 d =0 d =0 Pa = ∑ binomial(n. p ≤ 0. p) = ∑ b(65. (a) N = 8000. p ) = ∑ b(46. c = 2.

70 g/cm3. β = 0.040 0.160 0.600 Pr{accept} p1 p Pr{accept} 0. Accept the lot if ⎡⎣ Z LSL = ( x − LSL ) S ⎤⎦ ≥ 1.100 0.73 − 0.400 0.100 0.140 0. k = 1.05 × 10−2 = 2.060 0.7 1.200 p Pa{p = 0.820 0.988 0.945 0. (b) x = 0.730 0.02.Chapter 15 Exercise Solutions 15-1.7.70 ) 1.030 0.38 (from nomograph) 15-1 .000 0. p1 = 0.10 = 0.150 0.020 0.8571⎤ ≥ 1.020 0. LSL = 0.05 (a) From the variables nomograph.10.05} ≈ 0.7 ⎣ ⎦ Accept the lot.005 0.800 0. S = 1. 1 – α = 1 – 0.xls : worksheet Ex15-1 From the variables nomograph at n = 35 and k = 1.040 0.560 0.05 × 10−2 ( ) ⎡ Z LSL = ( 0.050 0.010 0.190 0.025 0.080 0.7: p2 OC Curve for n=25.900 1-alpha 0.001 0.200 0.180 0. the sampling plan is n = 35.190 0.070 0.90.050 beta 0.000 0. k=1. Calculate x and S.73. p2 = 0.016 0.400 0.7 .000 0. (c) Excel workbook Chap15.120 0.

wiley.93 A reduced sampling (nreduced = 20. N = 7000. β = 0. for a given AOQL = Pa pm (N – n) / N (where pm is the value of p that maximizes AOQ).5%. Inspection level II. Now.3 15-3. . σ = 5 p1 = 0. n = 120 and k = 2. For example. Suppose n is arbitrarily specified. Finally. The equations do not change: AOQ = Pa p (N – n) / N and ATI = n + (1 – Pa) (N – n).1 − α = 1 − 0.10 From variables nomograph. standard deviation unknown Assume single specification limit .95. and special Romig tables are usually employed.com/college/montgomery 15-5. ATI is found from the above equation.80. LSL = 150. AQL = 1.05 = 0. n and k will define Pa. Accept the lot if ⎡⎣ Z LSL = ( x − 150 ) S ⎤⎦ ≥ 2. knormal = 1.51) can be obtained from the full set of tables in MIL-STD-414 using Table B-3. for a specified process average.Form 1.Chapter 15 Exercise Solution 15-2.3.02. The table required to do this is available on the Montgomery SQC website: www. kreduced = 1. Inspection level IV From Table 15-1 (A-2): Sample size code letter = M From Table 15-2 (B-1): n = 50. Then a k can be found to satisfy the AOQL equation. we know n and k are related. ktightened = 1. and generally involves some trial-and-error search. because both Pa and pm are functions of n and k. 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. Repeat until the n and k that minimize ATI are found. Calculate x and S. Under MIL STD 105E. No convenient mathematical method exists to do this. p2 = 0. The design of a variables plan in rectifying inspection is somewhat different from the attribute plan design.005. 15-4.

1527 .005 n x2 = 0.000 Assume inspection level IV.00. inspection level II.282 0. x = 255. N = 100. 15-8. LSL = 225psi. N = 500.05 x A − x1 = Φ (1 − α ) σ n x A − 0. S = 10 ⎡⎣ Z LSL = ( x − LSL ) S = ( 255 − 225 ) 10 = 3.95 = 0.005 n n ≈ 9 and the target x A = 0.14 Assume normal sampling is in effect.Chapter 15 Exercise Solution 15-6. sample size code letter = O Normal sampling: n = 100. k = 2.000 ⎤⎦ > 2.645 0. k = 1.15 Tightened sampling: n = 7. k = 1.15. σ = 0.145. AQL = 4% Sample size code letter = E Assume single specification limit Normal sampling: n = 7. so accept the lot.005 g/cm3 x1 = 0. k = 2.15 = +1.33 15-7. AQL = 1%.145 = −1.00 Tightened sampling: n = 100.1 − α = 1 − 0. β = 0.10 x A − x2 = Φ(β ) σ n x A − 0.

09 (k = 1.95. p1 = 1% = 0. Re = 6 Tightened: n = 200.000 Assume inspection level II: sample size code letter = L Normal: n = 200. β = 1 – 0. c = 2 The sample size is slightly larger than required for the variables plan (a). N = 5.08) Reduced: n = 20.69) σ known allows smaller sample sizes than σ unknown. p2 = 8% = 0. M = 1. p1 = 0.com/college/montgomery AQL = 1%. σ = 0.Chapter 15 Exercise Solution 15-9. Re = 5 The sample sizes required are much larger than for the other plans. target = 3ppm.95.08 (a) 1 – α = 0. Variables sampling would be more efficient if σ were known.93) Tightened: n = 50. Re = 4 Reduced: n = 80.01. p2 = 0. β = 0. σ unknown Double specification limit.10.90 = 0.wiley.71 (k = 2. M = 4. . assume inspection level IV From Table A-2: sample size code letter = M From Table A-3: Normal: n = 50.10 From the nomograph. (d) AQL = 1%. Ac = 3.00 (k = 1.01. the sampling plan is n = 30 and k = 1. Ac = 2.8. (c) 1 – α = 0.10ppm. N = 5000. M = 1. Ac = 5.08 From nomograph (for attributes): n = 60. (b) Note: The tables from MIL-STD-414 required to complete this part of the exercise are available on the Montgomery SQC website: www.

00 0 0.12 0. .00 Pa 0.Chapter 15 Exercise Solution 15-10.2 p single I=1 I=2 I=5 I=7 Compared to single sampling with c = 0.80 0.08 0.18 0.60 0.14 0. chain sampling plans with c = 0 have slightly less steep OC curves.16 0.20 1.04 0.02 0. Excel workbook Chap15.20 0.xls : worksheet Ex15-10 OC Curves for Various Plans with n=25.40 0. c=0 1.06 0.1 0.

6 0.00 0.0343 0.2708 0.0600 0.0030 0.0981 0.0060 0.14 0.0090 0.5239 0. Pa 0.9083 0.1000 0.9992 0.2000 0. average process fallout = 0.06 0. c=0 1.4761 0.3773 0.0917 0.12 Fraction Defective.9379 0.0 Probability of Acceptance.8063 0.0489 0.0080 0.0050 0.2013 0.0020 0.2750 0.02 0.7292 0.0300 0.2 0.3000 Pa Pr{reject} 0.0000 OC Chart for n=32.9685 0. p 0.000.0100 0.8619 0.0010 0.18 0.0500 0. c = 0 Excel workbook Chap15.9306 0.20 .2512 0.1204 0.08 0.9511 0.0070 0.9019 0.0008 0.0 0.7488 0.1752 0.0400 0.6227 0.Chapter 15 Exercise Solution 15-11.0315 0.0621 0.001.10% = 0.8518 0.0200 0.8 0.10 0.1482 0.8248 0.1381 0.0900 0.7250 0.7733 0.0040 0.04 0.0800 0.0700 0.8796 0.9657 0.xls : worksheet Ex15-11 (a) p 0.4 0.0694 0.1937 0. n = 32.16 0.7987 0. N = 30.2267 0.0000 1.

0310)(0.32) = 0. n) = P(1. n) = P (0. p = 0.0310 Pa = 0. but the average number inspected is much smaller.9958. (d) Pa = 0.001 Pa = P(0. n) + P(1.9685)(30000 − 32) = 976 (c) Chain-sampling: n = 32. c = 0.9685 + (0. n)]i P(0.Chapter 15 Exercise Solution 15-11 continued (b) ATI = n + (1 − Pa )( N − n) = 32 + (1 − 0. i = 3. the Pa for chain sampling is slightly larger. there is little change in performance by increasing i. ATI = 32 + (1 − 0.9967 ATI = 32 + (1 − 0.9685)3 = 0.9685 P(1.9958)(30000 − 32) = 158 . n)[ P(0.9967)(30000 − 32) = 131 Compared to conventional sampling.32) = 0.

0016 0.40 0.4096 0.8000 0.50 p 0.3456 0.9999 0.0000 P(1.70 0.7164 0.8853 0.9000 0.8756 0.0900 0.1000 0.5000 0.0010 0.2401 0.1993 0.0200 0.9950 0.0388 0.00 0.10 0.0040 0.3000 0.20 0.20 0.0256 0.4000 0.0500 0.6857 0.xls : worksheet Ex15-1 p 0.6000 0.0081 0.7000 0.0756 0.c=0 1.1536 0.7385 0.9500 P(0.2000 0.1095 0.6561 0.90 1.7807 0.4) 0.2492 0.8145 0.8080 0.0800 0.7481 0.0600 0.Chapter 15 Exercise Solution 15-12.0753 0.50 0.9613 0.0001 0.30 0.70 Pa 0.60 0.0625 0.0016 0.4096 0.1304 0.00 .4116 0.90 0.0005 Pa 0.1715 0.0100 0.30 0.0626 0.2458 0.0256 0.0001 0.0000 OC Curve for ChSP-1 n=4.9606 0.4) 0. i = 3 Excel workbook Chap15.10 0.0081 0.8423 0.9815 0.60 0.0256 0. c = 0.0300 0.9960 0.2500 0.4377 0.1296 0.9224 0.0036 0.00 0.9072 0.80 0.0700 0.7732 0.2713 0. n = 4.2916 0.40 0.80 0.00 0.2252 0.

6355E+04 33.0029 1.0000 2.0000 1.7692E+60 1000.8953 1.0200 0. Average process fallout.0000 0.88584 + 0.8863 3.4619E+07 400.0000 0.0000 1.0000 0.7594 1.0000 2.2451E+09 285.4444 0.6788E+02 800.0000 3.3439E+10 6666.8328 1.0000 3.0080 0. AFI = 0.1279 4.0401E+03 66666.6628 2.7143 0.0008 4.0000 0. n = 6 If c = 0. N = 500.2222 0.2222 0. accept.3652E+54 1111.9985 1.0000 f = 1/100 and i = 1302 u v Pa 2.7032 1.5731E+13 200.5729E+18 3333.0000 5.3.0000 0.1848E+03 4000.0666.4121E+27 2222.6667 0.8648E+02 133.6298 2.2037E+02 333.5035E+02 2000. Pa = 0.0015 9.0700 0.0020 0.9502E+03 2000.0000 1.5471E+07 20.0032 2.1116 1.0000 0.2024.6667 0.5492E+42 1428.2765E+03 50000.0037E+03 5000.4676E+06 25.0561 2.0000 0.0003 3.4286 0.0000 3. f = 1/10 and i = 550: u = 855.0000 0.2804E+15 4000.6331E+04 33333.0000 9.0900 0.2867E+08 333.3371E+03 66.9316 8.1604E+03 57.0000 0.0070 0.0040 0. Pa = 0.6790E+03 100000.0400 0.0600 0.3333 0.666.9301 1.3984E+04 44.0000 3.3333 0.95264 15-14.8608 1.0010E+04 40000.0000 0.4040 4.6667 0.5714 0.8192E+07 10000.7161E+04 28571.8885 1.5536 2.6712 2.8284E+02 222.0000 .Chapter 15 Exercise Solution 15-13.3934 4.0000 0.8953 2.1429 0.6195E+36 1666.0024 3.0000 1.0001 1.8571 0.0000 0.6667 0.1111 0. 6) + P(1.0000 0.0000 0.0000 3.0000 0.1429 0. v = 6666.0015 0.02} Pa = P(0.3373E+02 10000.0000 0.5930E+04 1111.0000 6.9936E+21 2857.0000 1.0035E+16 166.0384 2.4904 2.0150 0.6175E+02 1000.6944 2.7715 1.1111 0.0238E+04 1250.1088 1. p = 0.0000 0. v = 66.7157E+05 666.7998 1.3395E+06 14285.0000 0.0000 f = 1/10 and i = 550 u v Pa 7.0000 0.0000 1.0000 6.3333 0.0000 8.2222 0.0000 1.6667 0.5127 4.0000 0.0381 1.0000 0. accept if i = 4. f = ½ and i = 140: u = 155.0000 0.0000 0.2151 7.0000 0.3449E+03 80.0100 0.0000 0.4085E+11 250.1000 f = 1/2 and i = 140 u v Pa 1.198% would be: 1.0000 0.9429 p 0.0000 2.7927 1.0106 4.3333 0.3467E+06 500.0300 0.5984E+02 250.6667 0. 6)]4 = 0.8106E+02 571. f = 1/100 and i = 1302: u = 4040.0010 0.0000 2.0000 0.0000 0.3333 0.5602E+03 50.7143 0.1111 0.8816E+02 500.5444 2.0025 0.3526 7.1111 0.9590E+02 100.4774 2.7.0652 2.0060 0.8863 1.7586E+48 1250.3262E+13 5000.5523.6921E+05 28.0066 1.3638E+05 20000.0000 3.0000 0.0000 0.0255E+24 2500.7431E+02 666.3972E+03 1666.2222 0.2222 0.4066E+03 3333.1429 0.15% = 0.0318E+21 125.5912E+04 25000.0000 2.4286 0. Need to find Pa{p = 0. AFI = 0.6619E+03 1428.4383E+07 11111.0000 1.000.3909E+02 285.0000 0.7410E+23 111.8675E+04 22222.0000 0.88584) 4 = 0.0350 0.3333 0.0015 and q = 1 – p = 0.0000 0.0000 0.7143 0.0839E+02 200.0500 0.5714 0.0250 0.0090 0.2202 1.2749 6.0000 0.7.1766 1.0162E+03 2500.0277 4.2128E+12 222.0030 0.10847(0.3333 0.4329E+03 2222.7266 1.6266E+04 40.6667 0.5553E+02 6666.0000 0.6667 0. v = 1333.0001 4.4400 3.0346E+02 400.0002 2.2131E+05 16666.5714 0.3333 0.0000 0.8266 1.5592E+02 1333. f = ½ and i = 140 2.0000 0. 6)[ P(0.0852E+18 142.6020 2.0800 0. Three different CSP-1 plans with AOQL = 0.6667 0.9739 4.0035 0.9562E+02 444.4676E+26 100.0050 0. If c = 1.3521E+06 12500.0179E+30 2000. Pa = 0.915.2222 0. f = 1/100 and i = 1302 15-15.0251E+06 22. f = 1/10 and i = 550 3.9429 6.0045 0.4444 0.6795E+03 2857. AFI = 0.5056E+04 1000.0135 7.530.6667 0.0450 0.0000 0.

4925 Prefer Plan B over Plan A since it has a lower Pa at the unacceptable level of p.0375} = 0. . CSP-1 with AOQL = 1.5165 and Pa{p = 0.90% Plan A: f = 1/5 and i = 38 Plan B: f = 1/25 and i = 86 15-17.6043 Plan B: AFI = 0.5272 and Pa{p = 0.0375} = 0. Plan A: AFI = 0.Chapter 15 Exercise Solution 15-16.

Sign up to vote on this title

UsefulNot useful- 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)
- Ch02 Solution
- Applied+Multivariate+Statistical+Analysis+6E【课后习题答案】
- Methods, Standards and Work Design, (11th Ed) Solution Manual
- charts_QP-104
- Exercises Solutions
- Design And Analysis Of Experiments, 5Th Edition (Douglas C Montgomery)
- Facilities planning Manual
- Production and Operation Analysis
- SQC CH6
- Facilities Planning - James Tompkins
- 1 Introduction to Statistical Quality Control
- Assignment #3
- 510 Textbook
- Engineering Economy 7th Edition Solution Manual [Blank & Tarquin]
- Introduction to Statistical Quality Control
- Introduction to Statistical Quality Control
- tugas5ppmNadPrils
- Solution Manual -Quality Control 5th Edition Montgomery