Академический Документы
Профессиональный Документы
Культура Документы
INFORMS is the largest professional society in the world for professionals in the fields of operations research, management
science, and analytics.
For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Downloaded from informs.org by [150.212.46.144] on 03 December 2015, at 05:23 . For personal use only, all rights reserved.
informs
doi 10.1287/ited.1090.0041cs-a
2010 INFORMS
I N F O R M S
Transactions on Education
Case
Katz Graduate School of Business, University of Pittsburgh, 358 Mervis Hall, Pittsburgh, Pennsylvania 15260,
pmirchan@katz.pitt.edu
Introduction
The Situation
79
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Downloaded from informs.org by [150.212.46.144] on 03 December 2015, at 05:23 . For personal use only, all rights reserved.
80
Mirchandani: Case: The MotoTech Manufacturing Company: Process Control and Improvement
The Process
John Tagole is clearly proud of his success as he nishes giving The Famous Five a tour of his plant.
They too must be proud of his achievement (as he
is of theirs), because all ve of them have shown
up for this fact-nding trip. They ask the right questions, and John Tagole is impressed by the speed with
which they zero in on the problem. The diffusion
stage, The Famous Five conclude, is the problem with
the current process. Lets address the problems at
the diffusion stage rst, and then we will look at the
remaining stages of the process. John Tagole marvels: These people must have a sixth sense. Danika
Katzs new specs all deal with the diffusion stage. In
any event, he goes ahead and describes the diffusion
stage to The Famous Five.
In many process steps, the wafers are processed in
a batch, with several wafers processed together at the
same step. One such process is the diffusion step. As
many as two hundred wafers may be batch processed
together in a diffusion furnace. During this step, a
layer of material (e.g., silicon dioxide) is deposited
onto each wafer. The thickness of the deposit must be
tightly controlled at a specied target thickness. The
target thickness for this diffusion process is 3,000 .3
To ensure the desired functionality and reliability of
the chip, the deposit thickness must lie in the range
of 2,900 to 3,100 .
The Famous Five go into a huddle. John Tagole
is used to their loud arguments from his time back
at the GWPGSBbut his staff is not. He chuckles to
himself as he overhears worried staff members muttering about calling plant security.
After about an hour, The Famous Five emerge and
discuss their plan of action with John Tagole. They
3
Mirchandani: Case: The MotoTech Manufacturing Company: Process Control and Improvement
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Downloaded from informs.org by [150.212.46.144] on 03 December 2015, at 05:23 . For personal use only, all rights reserved.
81
Mirchandani: Case: The MotoTech Manufacturing Company: Process Control and Improvement
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Downloaded from informs.org by [150.212.46.144] on 03 December 2015, at 05:23 . For personal use only, all rights reserved.
82
Table 1
Pre_Improvement Data
Data and Initial Computations
Thickness
(in thousands of )
Sample no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Table 2
Sample
mean
Sample
range
3265
3273
3197
3215
3321
3253
3186
3255
3195
3205
3008
3191
3108
3005
3228
3080
3152
3144
3244
3062
3141
3179
3180
3195
2960
3175
3286
3005
3065
3173
3068
2919
3001
3250
3179
3109
2996
3097
3244
3174
3158
3200
3254
3172
2927
3145
3160
3098
3122
3103
3301
3007
3069
3029
3133
3308
3100
3025
3037
3137
3098
3051
3215
3212
3123
3062
3177
3090
2970
3021
3127
3104
3100
3042
3201
3125
3186
3122
3320
3177
3138
3173
3125
3131
3094
3151
3105
3010
3000
3063
3202
2993
3202
3091
3113
3207
3171
3173
3046
3096
3265
3138
3030
3075
3064
2979
2930
3230
3194
3227
3063
31877
32140
31857
31650
32803
32280
31653
31183
31550
31653
30667
31547
31053
31053
30783
30707
31277
30900
32513
30843
30930
31410
31627
31553
30190
31620
32543
30887
30523
31417
30740
29560
29840
32023
31590
31453
30337
01680
01480
00230
00930
01210
00770
00480
03280
00700
00740
00900
00690
00050
02960
02280
00170
01730
01510
01060
00380
01160
01700
00430
00970
00910
01190
00740
01330
00350
01020
00260
00600
00910
01230
00900
01270
00670
Thickness
(in thousands of )
Sample no.
Sample
mean
Sample
range
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
3170
3067
3819
3216
3156
3141
3226
3247
3164
3012
3209
3125
3111
3306
3170
3011
3137
3178
3105
3044
3164
3008
3236
3011
3170
3304
3037
3075
3255
3273
3108
3001
2978
3116
3022
3171
3146
3898
3253
3200
3019
3236
3250
3051
3260
3002
3132
3023
3119
3203
3197
3031
3003
3068
3005
3058
3025
3091
2982
3176
3243
2968
3174
3245
3067
3132
2987
2890
3110
3007
3229
3178
3309
3310
3067
3162
3250
3233
3136
3202
3174
3134
3177
3133
3210
3076
3039
3193
3249
3010
3197
3053
3271
2959
3098
3064
2959
3027
3241
3000
2904
3064
2952
3115
2997
31900
31303
36753
32597
31410
31073
32373
32433
31170
31580
31283
31303
31037
31860
31943
30947
30690
31247
31407
30197
31397
30287
31993
29840
31480
32037
29880
30920
32470
31133
30480
30173
29400
31137
30087
00590
01110
05890
00940
01330
01430
00240
00170
01130
02480
02070
00090
01540
01870
00400
01860
01060
01900
01810
00390
01390
00450
01800
00520
00780
02400
00780
01470
00140
02730
02280
00770
00880
00060
00250
31326
01165
Mean
Post_Improvement Data
Data and Initial Computations
Thickness
(in thousands of )
Sample no.
1
2
3
4
5
6
7
8
9
10
Thickness
(in thousands of )
Sample
mean
Sample
range
Sample no.
Sample
mean
Sample
range
3039
3041
3047
3074
3081
3018
3061
3075
3042
3083
3011
3071
3089
3047
3109
3148
3093
3080
3067
3133
3095
3049
3102
3019
3010
3030
3094
3074
3019
3057
30483
30537
30793
30467
30667
30653
30827
30763
30427
30910
00840
00300
00550
00550
00990
01300
00330
00060
00480
00760
38
39
40
41
42
43
44
45
46
47
3032
3157
3073
3068
3059
3057
3003
3025
3046
3060
3105
3079
3049
3041
3024
3109
3032
3051
3105
3133
3111
3014
3056
3042
2970
3068
3046
3066
3074
3121
30827
30833
30593
30503
30177
30780
30270
30473
30750
31047
00790
01430
00240
00270
00890
00520
00430
00410
00590
00730
Mirchandani: Case: The MotoTech Manufacturing Company: Process Control and Improvement
83
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Downloaded from informs.org by [150.212.46.144] on 03 December 2015, at 05:23 . For personal use only, all rights reserved.
Table 2
Sample no.
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Sample
mean
Sample
range
3057
3086
3084
3122
3132
3096
3048
3067
3014
3089
3094
3031
3051
3086
3045
3039
3047
2961
3044
3051
3071
3101
3073
3096
3075
3046
3079
3119
3040
3040
3056
2987
3044
3103
3040
3004
3006
3066
3139
3086
3018
3026
2980
3030
3076
3053
2982
3046
3093
2998
3071
3085
3098
3030
3105
3021
3087
3033
3048
3126
3057
3108
3062
3038
3005
3106
3058
3104
3036
3034
3099
3015
3027
3137
3074
3007
3103
3040
3049
3046
3135
30937
30490
30703
30703
30557
30887
30693
30717
30267
30443
30550
30920
30650
30693
30357
30177
30587
30173
30413
30567
30637
30670
30580
30690
30697
30633
30813
00620
00650
00470
00890
01450
00820
00550
00680
00580
00830
00890
01080
00350
00860
00190
00590
00690
01150
00260
01550
00280
00940
01050
00560
00360
00520
01050
Thickness
(in thousands of )
Sample no.
Sample
mean
Sample
range
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
3108
3068
3037
3047
3008
3058
3075
3042
3042
3072
3031
3023
3043
3032
3094
3017
3070
3063
3073
3078
3062
3078
3046
3152
3035
3007
3105
3060
3056
3052
3059
3017
3133
3048
3020
2981
3105
3077
3073
2991
3046
3099
2970
3120
3075
3129
3113
3069
3073
3094
3108
3007
3074
3067
3004
3095
3043
3111
3041
3034
3053
3075
3085
3060
3080
3044
2970
3057
3086
3109
3002
3051
3079
3053
3127
30743
30600
30570
30567
30213
30707
30450
30953
30437
30420
30217
30677
30683
30550
30550
30357
30463
30300
30930
30873
30643
30807
30647
30927
30853
01010
00980
00370
00200
00480
00370
00580
00910
00070
00520
00720
00820
00420
00410
01030
00290
01290
00930
00470
00340
01270
00620
00330
00990
00920
30613
00677
Mean
Mirchandani: Case: The MotoTech Manufacturing Company: Process Control and Improvement
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Downloaded from informs.org by [150.212.46.144] on 03 December 2015, at 05:23 . For personal use only, all rights reserved.
84
to
to
to
to
2,880
2,900
3,120
3,140
Rework cost
per unit ($)
125
50
50
125
Thus, the unit contribution for a wafer that originally had a thickness of 3,110 gets reduced to
$150 $50 = $100 because of rework. Any wafers with
a thickness exceeding 3,140 or less than 2,860
cannot be reworked; for these wafers, MM must
expense the variable cost of $500, which it has already
incurred.
Supplementary Material
I am grateful to Dr. Skip Weed, formerly of Motorola University, for suggesting the semiconductor environment as a
backdrop of the case and providing a short description of
the production process. The actual problem situation has
been modied for pedagogical reasons.