HW 5

advertisement

4/9/2020

IENG 486

CH 7: 9, 10, 17, 25 D. H. JENSEN

PROBLEM 7.9 CONSIDER THE TWO PROCESSES SHOWN HERE (THE SAMPLE SIZE n = 5).

SPECIFICATIONS ARE AT 100  10. CALCULATE C

P

, C

PK

, AND C

PM

AND INTERPRET

THESE RATIOS. WHICH PROCESS WOULD YOU PREFER TO USE?

PROCESS A

X

A

= 100

S

A

= 3

_

PROCESS B

X

B

= 105

S

B

= 1

CONVERTING FROM s TO σ, AND COMPUTING C

P

FOR BOTH PROCESSES: SOLN.

C

C

A

P

P

A

 s c

4

USL

6

( 100

3

.

9400

LSL

10 )

6 ( 3

.

 3 .

19

( 100

19 )

,

10 )

B

 c s

4

20

19 .

14

.

9400

1

1 .

04

 1 .

06

C

P

B 

( 100  10 )  ( 100

6 ( 1 .

06 )

 10 )

20

6 .

36

 3 .

14

JUDGING FROM THE PROCESS CAPABILITY RATIO C

P

ALONE: PROCESS B HAS

THE HIGHER C

P

RATIO, AND THEREFORE THE BETTER POTENTIAL CAPABILITY.

COMPUTING C

PK

FOR BOTH PROCESSES:

C

PK

C

PK

A

C

PK

B

MIN

110

3 (

110

3 (

[ C

PL

3 .

19 )

1 .

 105

06

,

)

C

PU

]

100

9 .

MIN

10

57

1



3 .

5

18

 1 .

.

04

57

 LSL

3 

,

USL

3 

  



ASSESSING: PROCESS A HAS A C

P

RATIO THAT IS EQUAL TO ITS’ C

PK

RATIO,

INDICATING THAT THE PROCESS IS CENTERED AND MAY PRODUCE FEWER

DEFECTS.

COMPUTING C

PM

FOR BOTH PROCESSES:

C

PM

6

USL

 2 

( 

LSL

 T ) 2

C

C

PM

PM

A

B

6

6

( 100

3

( 100

.

19

1 .

06 2

10

2

10

)

)

(

(

100

(

( 100

100

105 

10

100 ) 2

10

100 )

)

2

)

20

19 .

14

20

30 .

67

1 .

04

0 .

652

1

6

4/9/2020

IENG 486

CH 7: 9, 10, 17, 25 D. H. JENSEN

PROBLEM 7.9 (CONT) ASSESSING: PROCESS A HAS A C

P

RATIO THAT IS EQUAL TO ITS’ C

PK

RATIO AND C

PM

RATIO, INDICATING THAT THE PROCESS IS CENTERED.

PROCESS B HAS A C

PM

RATIO THAT INDICATES THAT IT IS CENTERED TO THE

RIGHT OF THE SPECIFICATION TARGET. NOTE THAT FOR A CENTERED PROCESS

WITH TWO-SIDED C

P

= 1.10, THERE WOULD BE APPROXIMATELY 967 PARTS PER

MILLION DEFECTIVE (SEE TEXTBOOK TABLE 7-2), AND FOR A ONE SIDED C

P

=

1.60, THERE WOULD BE APPROXIMATELY 1 PART PER MILLION DEFECTIVE.

SINCE PROCESS A WOULD GENERATE A HIGHER NUMBER OF DEFECTIVE UNITS,

PROCESS B WOULD BE PREFERRED EVEN THOUGH IT IS NOT CENTERED WITHIN

THE SPECIFICATION TOLERANCE.

PROBLEM 7.10 SUPPOSE THAT 20 OF THE PARTS MANUFACTURED BY THE PROCESS IN

EXERCISE 7-9 WERE ASSEMBLED SO THAT THEIR DIMENSIONS WERE

ADDITIVE; THAT IS, x = x

1

+ x

2

+ . . . + x

20

SPECIFICATIONS ON x ARE 2000  200. WOULD YOU PREFER TO PRODUCE THE

PARTS USING PROCESS A OR PROCESS B? WHY? DO THE CAPABILITY RATIOS

COMPUTED IN EXERCISE 7-7 PROVIDE ANY GUIDANCE FOR PROCESS

SELECTION?

SOLN. ASSUMING THAT WE CAN CHANGE THE CENTERING OF THE PROCESS, I WOULD

PREFER TO CONTINUE TO USE PROCESS B. WHILE PROCESS B WAS NOT

CENTERED, THE PROCESS IS TIGHT ENOUGH TO BE THE BEST CHOICE (FEWEST

% DEFECTIVE) AS ORIGINALLY ESTIMATED IN PROBLEM 7.9. SINCE THE

VARIANCES IN AN ADDITIVE ASSEMBLY ALSO ADD, THE PROCESS WILL

CONTINUE TO BE THE TIGHTER OF THE TWO.

2

6

NOTE: ALL OF THE GUIDANCE FOR THIS INFORMATION MAY BE DERIVED FROM THE

CAPABILITY RATIOS IN THIS EXERCISE. IT IS ALSO POSSIBLE TO COME TO

THE SAME CONCLUSIONS BY COMPUTING THE PROCESS FALLOUT (NON-

CONFORMING PROPORTION UNDER THE DISTRIBUTION CURVES), BUT IT

WOULD TAKE MORE EFFORT.

4/9/2020

IENG 486

CH 7: 9, 10, 17, 25 D. H. JENSEN

PROBLEM 7.17 THE FAILURE TIME IN HOURS OF 10 LSI MEMORY DEVICES IS SHOWN HERE.

PLOT THE DATA ON NORMAL PROBABILITY PAPER, AND IF APPROPRIATE,

ESTIMATE THE PROCESS CAPABILITY. IS IT SAFE TO ESTIMATE THE

PROPORTION OF CIRCUITS THAT FAIL BELOW 1200 HOURS?

1210 1275 1400 1695 1900

2105 2230 2250 2500 2625

SOLN. COMPUTING THE CHART VALUES FOR A NORMAL PROBABILITY PLOT:

4

5

6 j

1

2

3

7

8

9

10

(j - 0.5) / n

0.05

0.15

0.25

0.35

0.45

0.55

0.65

0.75

0.85

0.95

Zj

-1.6449

-1.0364

-0.6745

-0.3853

-0.1257

0.1257

0.3853

0.6745

1.0364

1.6449

Sorted Data

1210

1275

1400

1695

1900

2105

2230

2250

2500

2625

PLOTTING THE DATA:

2.00

1.60

1.20

0.80

0.40

0.00

-0.40

-0.80

-1.20

-1.60

-2.00

Normal Probability Plot

Sorted Data

SINCE THE POINTS AT EACH END OF THE PLOT DO NOT PASS THE FAT PENCIL

TEST, THE DATA ARE NON-NORMAL. NO FURTHER ANALYSIS IS SUPPORTED.

3

6

4/9/2020

IENG 486

CH 7: 9, 10, 17, 25 D. H. JENSEN

PROBLEM 7.25 TEN PARTS ARE MEASURED THREE TIMES BY THE SAME OPERATOR IN A GAGE

CAPABILITY STUDY. THE DATA ARE SHOWN HERE (SEE TABLE, BELOW).

A.)

SOLN.

DESCRIBE THE MEASUREMENT ERROR THAT RESULTS FROM THE USE OF THIS GAGE.

EACH OF THE SETS OF MEASUREMENTS IN THE DATA TABLE HAS SUMMARY

STATISTICS (MEAN AND RANGE) COMPUTED, AND THE GRAND MEAN AND

AVERAGE RANGE ARE ALSO COMPUTED:

R = 2.30

X = 98.2

4

6

PART

NUMBER

1

2

3

4

1

100

MEASUREMENT

2

101

95

101

96

93

103

95

3

100

97

100

97

X

100.3

95.0

101.3

96.0

_

R

1

4

3

2

8

9

10

5

6

7

98

99

95

100

100

100

98

98

97

99

100

98

96

98

98

98

97

99

97.3

98.3

96.7

99.0

99.0

99.0

2

3

2

2

1

3

NOTING THAT n = 3, THE CENTER AND UPPER AND LOWER CONTROL LIMITS FOR

THE X-BAR CHART ARE COMPUTED, AND THE DATA ARE PLOTTED:

‗ _

UCL = X + A

2

R = 98.2 + (1.023)(2.30) = 100.6

CL = X = 98.2

‗ _

LCL = X - A

2

R = 98.2 - (1.023)(2.30) = 95.8

102.00

100.00

98.00

96.00

94.00

92.00

90.00

1

X-Bar Chart (Gage Capability)

2 3 4 5 6 7 8 9 10

THE LACK OF OUT-OF-CONTROL POINTS (ONLY TWO!) INDICATES THAT THE

SYSTEM IS NOT VERY CAPABLE OF DETECTING DIFFERENCES IN THE PARTS.

4/9/2020

PROBLEM 7.25 (CONT.)

AGAIN USING n = 3, THE CENTER AND UPPER AND LOWER CONTROL LIMITS FOR

THE R – CHART ARE COMPUTED, AND THE DATA ARE PLOTTED:

_

UCL = D

4

R = (2.575)(2.30) = 5.92

_

CL = R = 2.30

_

LCL = D

3

R = (0)(2.30) = 0

7.00

6.00

5.00

4.00

R- Chart (Gage Capability)

3.00

2.00

1.00

0.00

1 2 3 4 5 6 7 8 9 10

THE LACK OF OUT-OF-CONTROL POINTS (NONE!) IN THIS CHART INDICATES

THAT THE OPERATOR IS CAPABLE IN USING THE GAGE APPROPRIATELY.

B.)

SOLN.

IENG 486

CH 7: 9, 10, 17, 25 D. H. JENSEN

ESTIMATE TOTAL VARIABLITY AND PRODUCT VARIABILITY.

FROM THE 30 INDIVIDUAL MEASUREMENTS IN THE DATA TABLE, THE

STANDARD DEVIATION IS COMPUTED, GIVING THE TOTAL VARIATION:

TOTAL

 2 .

17 , SO  2

TOTAL

 2 .

17  2  4 .

71

5

6

SINCE:

GAGE

R d

2

2 .

30

1 .

693

 1 .

36 , SO  2

GAGE

 1 .

36  2  1 .

85

C.)

SOLN.

GIVEN:

 2

TOTAL

  2

PRODUCT

  2

GAGE

, THEN 

PRODUCT

  2

TOTAL

  2

GAGE

WHAT PERCENTAGE OF TOTAL VARIABLITY IS DUE TO THE GAGE?

TOTAL VARIABILITY IS:

GAGE

TOTAL

1

2 .

.

36

17

 0 .

627 ,  OR  62 .

7 %

4 .

71  1 .

85  1 .

69

4/9/2020

IENG 486

CH 7: 9, 10, 17, 25 D. H. JENSEN

PROBLEM 7.25 (CONT.)

D.) IF SPECIFICATIONS ON THE PART ARE AT 100  15, FIND THE P/T RATIO FOR THIS

GAGE. COMMENT ON THE ADEQUACY OF THE GAGE.

SOLN. THE PRECISION TO TOLERANCE RATIO IS:

6 

USL

GAGE

 LSL

( 100 

6

15

(

)

1 .

36

(

)

100  15 )

8 .

16

30

 0 .

27

SINCE THIS RATIO IS MUCH GREATER THAN 0.10, THE GAGE IS INADEQUATE.

6

6

Download