About WTS Prof. Clair Brown Faculty, Students and Affiliates Research Areas Online Research Reports Working Papers

THE COMPETITIVE SEMICONDUCTOR MANUFACTURING HUMAN RESOURCES PROJECT:

Second Interim Report
CSM-32
Clair Brown, Editor

7.3 Equipment Maintenance


This section, in concert with Section 7.4, probes more deeply into the specific responsibilities for each occupation. This section focuses on equipment maintenance responsibilities, and Section 7.4 examines SPC activities. Given that equipment that is chronically down, dirty, or out-of-alignment can prevent world class manufacturing performance, equipment maintenance activities should be a distinguishing factor in fab performance. However, initial correlations between a fab's level of equipment maintenance and the five performance metrics presented at the end of this section are primarily negative, although not significant. The correlations between the level of participation in equipment maintenance across occupations and the five performance metrics show mixed results across the occupations. The positive correlations for the operator job category support our hypothesis that involving line workers in problem-solving can heighten performance.

Figures 7-1 through 7-4 demonstrate, as expected, that technicians and equipment engineers shoulder the greatest level of responsibility when it comes to equipment maintenance and troubleshooting. The charts depict the weighted scores for each equipment maintenance activity summed across all fabs for each job category. (See Appendix 7-1 for complete descriptions of each activity.) The scores were weighted as follows: High=3, Some=1, None=0. Since 14 fabs responded to this question for all job categories, the maximum score per activity per job category is 42. Although operators do not play an extensive role in either long-term maintenance or modifications, they are in intimate contact with the equipment—recognizing and documenting abnormalities, cleaning and/or lubricating the equipment, and performing daily or weekly inspections.

To capture the variation across fabs in terms of their use of equipment maintenance activities, we grouped the activities in our survey according to their degree of difficulty as shown in Appendix 7-1.

Then, for three occupations—operator, technician, and equipment engineer—we derived scores for each fab by:

1. weighting each equipment activity according to Appendix 7-1 (High-Level=3, Medium-Level=2, Low-Level=1);

2. then, multiplying the scores from (1) by weights based on the fab's response (High=3, Some=1, None=0) for each activity;

3. and finally, summing together the double-weighted scores from (2) for the thirty equipment maintenance activities for each occupation for each fab. [Therefore, the maximum score a fab could achieve for one job category is 204: 15 (High score on the 5 Low-Level activities) + 72 (High score on the 12 Medium-Level activities) + 117 (High score on the 13 High-Level activities).] The resulting scores for operators, technicians, and equipment engineers at each fab can be found in Figures 7-5, 7-6 and 7-7, respectively.

There appears to be some substitution between the work activities of technicians and engineers. As shown by the striped bars in Figure 7-6, the fabs with the highest equipment maintenance scores for technicians are located in Asia, Europe, and the U.S. As shown by Figure 7-7, those five fabs do not rely as heavily on their equipment engineers—only one of the top five in the technician chart remains in the top five in the corresponding chart for equipment engineers.

Table 7-5 presents the correlations between the performance metrics and the three occupations. Consistent with our hypothesis that the involvement of line workers in equipment maintenance is important, operator involvement is positively related to line yield performance. Surprisingly the level of technician involvement is not significantly correlated with higher performance.

Table 7-5. Equipment Maintenance Across Occupations and Fab Performance

  Correlation with:

Performance Metric
Operator

Eq. Maint.

Technician

Eq. Maint.

Eq. Engineer

Eq. Maint.

Defect Density (dd_pcout) 0.26 0.034 -0.065
Stepper Throughput (wopd_out) 0.28 -0.23 0.22
Line Yield (lyd_pout) 0.61** -0.14 -0.28
Cycle Time (ctpl_out) 0.41 0.017 -0.16
Direct Labor Productivity (dlp_pout) 0.21 -0.089 0.032

**Statistically significant at the 5% level.

Finally, we calculated an "overall" equipment maintenance score for each fab by summing a fab's scores for three occupations: operator, technician, and equipment engineer. These "overall" scores for the fabs are found in Figure 7-8.

The correlations between the rankings of the fabs from Figure 7-8 and their rankings for the five performance metrics are presented in Table 7-6. Although a number of the correlations between the use of equipment maintenance and performance are negative, none of the correlations is statistically significant.

Table 7-6. Equipment Maintenance and Fab Performance

Performance Metric Correlation with Equipment Maintenance
Defect Density (dd_pcout) 0.0022
Stepper Throughput (wopd_out) -0.083
Line Yield (lyd_pout) -0.17
Cycle Time (ctpl_out) -0.033
Direct Labor Productivity (dlp_pout) -0.12


7.4 Statistical Process Control

The charts depicting employee involvement in SPC show a similar pattern as those for equipment maintenance (Figures 7-9 through 7-12): The tasks performed by the operators and technicians overlap to some degree with the engineers' tasks (e.g., creating X-bar, R charts), but in many areas, they are complementary (e.g., operators and technicians enter quality data about the process flow into the computer and the engineers use the data for problem identification). In Figures 7-9 through 7-12, the example of a "shared task," creating X-bar, R charts, is shaded gray and the examples of "complementary tasks" are striped. In constructing Figures 7-9 through 7-12, the same weighting scheme as described for Figures 7-1 through 7-4 was used. The maximum possible score for each SPC activity varies across the job categories according to the total number of fabs that responded. The maximum possible scores are as follows: 42, 45, 42, and 39 for the operation, technician, process engineer, and equipment engineer job categories, respectively. Of the three occupations, process engineers are clearly the most involved in advanced problem-solving.

In a similar exercise as described above for equipment maintenance activities, we grouped together the SPC activities in our survey to allow us to calculate the intensity of SPC-use at the fabs. The groupings are presented in Appendix 7-2. Then, for three occupations—operator, technician, and process engineer—we derived scores for each fab by the same double weighting scheme as described in Section 7.3: We weighted both the fabs' responses as to the level of participation of each occupation in the activity (High=3, Some=1, None=0), as well as weighting the activity for its degree of difficulty (High-Level=3, Medium-Level=2, Low-Level=1). We then summed together the weighted scores for the eighteen SPC activities to calculate the SPC score for each occupation for each fab. [The maximum score a fab could achieve for one job category is 114: 18 (High score on the 6 Low-Level activities) + 24 (High score on the 4 Medium-Level activities) + 72 (High score on the 8 High-Level activities).] The resulting scores for operators, technicians, and process engineers can be found in Figures 7-13, 7-14, and 7-15, respectively.

As found for equipment maintenance activity, some fabs use their line workers more intensively instead of relying exclusively on their engineering resources. Of the five fabs with the greatest degree of operator participation in SPC, two are found in Asia and three are found in the U.S. (Figure 7-13, striped bars). When examining the involvement of process engineers in Figure 7-15, one can see that those same five fabs (striped bars) no longer rank at the top which suggests that those fabs rely more heavily on their operators for SPC instead of relying exclusively on their engineers. The five top ranked fabs in Figure 7-15 are all located in Asia and produce memory products.

As for the relationship between involvement in SPC activities across occupations and fab performance, a high-level of involvement by process engineers is clearly correlated with higher fab performance (Table 7-7). The correlations between performance and operator involvement in SPC support our hypothesis to some degree that the participation of line workers in problem identification and solution boosts performance.

Table 7-7. SPC Across Occupations and Fab Performance

  Correlation with:

Performance Metric
Operator

SPC

Technician

SPC

Process Engineer

SPC

Defect Density (dd_pcout) 0.67** 0.45 0.66**
Stepper Throughput (wopd_out) -0.54* -0.30 0.54*
Line Yield (lyd_pout) 0.60** 0.40 0.53*
Cycle Time (ctpl_out) -0.041 -0.13 0.44
Direct Labor Productivity (dlp_pout) 0.18 0.0062 0.74***

*Statistically significant at the 10% level.

**Statistically significant at the 5% level.

***Statistically significant at the 1% level.

Finally, we calculated an "overall" SPC score for each fab by summing together the scores for three occupations: operator, technician, and process engineer. These "overall" SPC scores for the fabs are found in Figure 7-16. As presented in Table 7-8, the correlations between the fabs' rankings in Figure 7-16 and the five performance metrics support our hypothesis that a high degree of SPC activity in the fab is vital for success in terms of defect density and line yield performance.

Table 7-8. SPC and Fab Performance

Performance Metric Correlation with Fab SPC Ranking
Defect Density (dd_pcout) 0.68***
Stepper Throughput (wopd_out) -0.29
Line Yield (lyd_pout) 0.52*
Cycle Time (ctpl_out) -0.16
Direct Labor Productivity (dlp_pout) 0.18

*Statistically significant at the 10% level.

***Statistically significant at the 1% level.

Go to Figures 9-16

End of Chapter 7

Go to Chapter 8
Go to Appendix for this Chapter
Go to Table of Contents for this Chapter
Go to Table of Contents for the CSM-HR Interim Report

 IRLE HOME
 LIBRARY
 PROGRAMS
 
© 2005 Institute for Research on Labor and Employment. 
2521 Channing Way # 5555 
Berkeley, CA 94720-5555