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

6.3 Percentage of Time Spent in Training After the First Year of Employment
Jumbi Edulbehram and Dan Rascher

The average percentage of time spent in training after the first year of employment is shown below:

Table 6-5. Average % Training Time after the First Year

Operator-OJT 16
Operator-Classroom 6
Technician-OJT 18
Technician-Classroom 6
Engineer-OJT 12
Engineer-Classroom 10

Even after the first year of employment, technicians and operators spent considerable amounts of time in OJT. Engineers spent a larger percentage of their time in classroom training than technicians and operators. However, engineers still spent slightly more time in OJT than in classroom training.

The breakdown of the average percentage of training time by the number of fabs in each category is shown below:


Over half of the fabs required that less than 5% of operators', technicians', and engineers' time be involved in classroom training after the first year of employment. Engineers spent less than 10% of their time in OJT after the first year in 9 of the fabs. Surprisingly, technicians and operators spent less than 10% of their time in OJT in 6 and 4 fabs respectively. In 3 fabs engineers spent at least 20% of their time in OJT which shows that there is more variation in engineer OJT training than for operators and technicians.

The training data shows large variations across fabs. Surprisingly, however, training practices are not statistically related to the performance measures for operators and engineers. For technicians, classroom training was correlated with stepper throughput. These findings are in contrast to our earlier findings based on interview data. See Figures 6-2, 6-3, and 6-4 in the appendix and the accompanying text for more information.

Operators

In the First Interim Report, it was found that the amount of on-the-job training from mentors was positively associated with performance; course-based or classroom training was negatively associated with performance. For the current set of data, neither the amount of operator on-the-job training nor the amount of classroom training is significantly associated with measures of performance. However, as will be shown below, the type of training and how it is delivered can make a difference.

Technicians

In the First Interim Report, it was found that none of the technician training was associated with performance. The new data show only a positive correlation between the amount of technician classroom training and stepper throughput.

Engineers

In the First Interim Report we had found that the training index (made up of the different types of training that engineers receive) was positively correlated with performance, especially with stepper throughput and direct labor productivity. But the new set of data failed to show any significant correlation between the amount of engineers' training and performance.

6.4 Topics and Types of Training: On-the-Job (OJT) or Classroom

Operators

The number of fabs that provide the particular types of training are shown below:

Table 6-6.

  # Fabs
Type of Skill No Training Just OJT Just Classroom Both
Basic skills (math, reading) 7 0 7 1
Basic science (chemistry/physics) 7 0 7 1
Statistical process control 0 1 5 9
Company-specific orientation 3 2 9 1
Machine operation 0 12 0 3
Machine maintenance 5 9 0 1
Teamwork and communication 0 3 9 3
Problem solving methods 1 1 11 2
Design of experiments 13 1 1 0
Safety procedures 0 3 1 11
Cleanroom procedures 1 2 4 8
  • Only about half the fabs provide their operators with training in basic skills and science.

  • Most fabs provide both OJT and classroom training for SPC.

  • Machine operation and maintenance training is predominantly provided through OJT.

  • Most fabs provide classroom training to operators on teamwork and problem solving.

  • Training in problem solving methods is provided mainly in the classroom.

  • Most fabs do not provide any training in the design of experiments.

  • Most fabs provide training in safety procedures and cleanroom procedures through both OJT and in the classroom.


Correlations between Different Types of Training and the Performance Metrics

In the first report it was found that for operators, OJT (especially for SPC and general computer skills) was positively associated with high performing fabs, while classroom training was negatively associated with fab performance.

This report shows that fabs which provide just OJT or just classroom training have negative correlations with the performance metrics. Specifically, cleanroom procedures taught solely on the job were significantly negatively associated with direct labor productivity. Design of experiments, teamwork and communication, and company-specific orientation were significantly negatively associated with defect density, stepper throughput, and cycle time, respectively, when taught solely in the classroom.

On the other hand, when both OJT and classroom training are provided, most types of training are positively correlated with the performance metrics. Problem solving methods and cleanroom procedures are significantly positively associated with stepper throughput and cycle time. Additionally, cleanroom procedures is significantly positively correlated with direct labor productivity.

Examples of Training

At F8, one of the top overall performers, training is of great importance. The fab has a separate training center, which consists of conference rooms, a computer lab and an auditorium. The fab has a separate training department, which reports to corporate services. The employees at the fab were proud of the fact that the fab had won two national training awards.

F5, an outstanding overall performer, tailors OJT to the particular needs of the operators. During the first year, about 5 of the operators are picked as leaders and provided special training. Each operator's training is tailored to individual needs, and the supervisor keeps track of needed and completed training. The supervisor also keeps track of each operator's ability to run certain machines; operators are ranked on ability in their particular equipment area.

Though the practice is rather rare, operators at some fabs like F11 do pursue outside education at local schools for an AA degree in electronics, and are reimbursed for it.

At F18, one of the top performers, each employee receives an individual "training roadmap." Operators receive refresher training if they fail their periodic re-qualification tests. Operators attend two statistical process control training sessions. The first is training on the 7 Quality Control Tools: Pareto diagrams, histograms, cause and effect diagrams, control charts, scatter diagrams, graphs, and checksheets. The second is training on basic SPC, including topics on variation, probability, histograms, normal distribution, mean, range and standard deviation, X-bar R and P charts, and interpreting control charts. Operators also receive Quality Circle Team (QCT) training if they are members of QCTs.

Additionally, operators are cross-trained for 5 jobs on average; cross training is stressed by the manufacturing supervisors who decide and plan cross-training needs. Some operators felt that while cross training improves the flexibility of the work force, it reduces productivity on any given machine. One operator said that they feel less confident doing more diverse jobs, but it clearly helps to keep the work moving and to cover for operators who are out. Cross-training was seen as being key to reducing cycle time.

The same fab periodically administers a written test, which covers applied reading, math skills, and problem solving. Operators have to retake and pass this test periodically. The applied reading sections test for comprehension of main ideas and factual details. The math skills sections test proficiency in arithmetic computation, arithmetic reasoning, statistics, measurement, and graph comprehension. The problem solving sections test logical ordering, reasoning, problem solving, and analytical judgment. Those who fail the test must take a remedial curriculum or quit. They get 4 chances to pass.

Technicians

The number of fabs that provide the particular types of training are shown below:

Table 6-7.

  # Fabs
Type of Skill No Training Just OJT Just Classroom Both
Basic skills (math, reading) 7 1 5 0
Basic science (chemistry/physics) 3 0 9 1
Statistical process control 1 2 4 6
Company-specific orientation 2 2 7 2
Machine operation 0 6 0 7
Machine maintenance 0 4 1 8
Teamwork and communication 1 0 7 5
Problem solving methods 0 3 5 5
Design of experiments 6 0 5 2
Safety procedures 0 1 2 10
Cleanroom procedures 0 1 4 8
  • Most fabs provide classroom training in basic science, while over one third provide it in basic skills.

  • For SPC, only one fab in the sample does not provide any training to its technicians, while others provide it mostly either in the classroom, or combine it with OJT.

  • Training for machine operation and maintenance, ostensibly the most important type of training for technicians, is provided through both OJT and classroom at most fabs, though some fabs provide it only through OJT.

  • Most fabs provide training in teamwork and problem solving, either only in the classroom, or combined with OJT.

  • Only some fabs provide training in the design of experiments, either only in the classroom, or combined with OJT.

  • One half of the fabs provide training in the design of experiments.

  • Most fabs provide training in safety procedures and cleanroom procedures through both OJT and in the classroom.


Correlations between Different Types of Training and the Performance Metrics

The data show that the following kinds of training are positively correlated with line yield regardless of the ways in which the training is provided: basic skills, company-specific orientation, machine operation and maintenance, problem solving methods, and safety and cleanroom procedures. Except for basic skills and company-specific orientation, these types of training, along with SPC and teamwork/communication, are also positively correlated with direct labor productivity.

Classroom training in basic skills and problem solving methods is positively correlated with line yield and defect density respectively. The use of both classroom and on-the-job training together is correlated with high performance for many of the skills being taught. Both machine operation and maintenance training are positively associated with defect density. Training in SPC, teamwork/communication, and safety and cleanroom procedures is positively correlated with direct labor productivity. These types of training, along with problem solving methods and design of experiments are also positively correlated with stepper throughput.

No training of any type is significantly associated with cycle time for technicians. In addition, on-the-job training of any type is uncorrelated with any of the performance metrics. Also, training in basic science is uncorrelated with any of the performance metrics.

Examples of Training

At F3, an outstanding performer, 'machine-keepers' (not formally called technicians) are always receiving some sort of training. Heavy emphasis is placed on mentoring-type of training within groups and departments. That is, the head of a group takes a class and then teaches his direct reports. It is felt to be extremely important to learn from one's supervisor. When asked what percentage of time was spent in training, one machine keeper explained that this was a difficult question because he could not separate his work and training activities. He considered it to be a part of his 'work' to learn from everything he does.

At F18, one of the top performers, technicians have to take the same written test operators, as described above. Each technician receives an individual "training roadmap." Technicians receive OJT for most areas related to machine operation and maintenance. They may also receive training on specific equipment by the equipment vendors. Like the operators, all technicians attend the 3 SPC training sessions. Technicians also receive Quality Circle Team (QCT) training if they are members of QCTs.

At F17, a good overall performer, technicians receive 2 hours of training per month. Technicians receive quality control and process training related to their work area. Technicians can also attend advanced engineering courses. These are offered every two weeks and last from 1 to 2 hours. The classes are conducted by senior engineers. Classes in management information systems are also available.

Engineers

The number of fabs that provide the particular types of training are shown below:

Table 6-8.

  # Fabs
Type of Skill No Training Just OJT Just Classroom Both
Basic skills (math, reading) 6 2 5 2
Basic science (chemistry/physics) 4 3 5 3
Statistical process control 0 0 7 8
Company-specific orientation 4 3 6 2
Machine operation 0 9 0 6
Machine maintenance 0 5 1 9
Teamwork and communication 0 1 8 6
Leadership and supervision 1 1 7 6
Problem solving methods 0 1 8 6
Design of experiments 0 1 4 10
Safety procedures 0 3 3 9
Cleanroom procedures 2 1 3 9
  • Over one half of the fabs provide their engineers with training in basic skills and science, mainly in the classroom, with some OJT.

  • All fabs provide SPC training, either in the classroom or combined with OJT.

  • Training in machine operation and maintenance is provided either solely through OJT, or combined with classroom.

  • Training in teamwork/communication, leadership/supervision and problem solving methods is provided mostly in the classroom, or combined with OJT.

  • Most fabs provide both OJT and classroom training in the design of experiments.

  • Most fabs provide training in safety procedures and cleanroom procedures through both OJT and in the classroom.


Correlations between Different Types of Training and the Performance Metrics

The data show that just OJT and training of any type (either OJT or classroom, but not both) is completely uncorrelated with the performance metrics. Just training in the classroom is mostly negatively correlated with the performance metrics. Specifically, training in basic skills, SPC, and company-specific orientation is negatively associated with defect density, stepper throughput, and cycle time, respectively. SPC training in the classroom is also negatively correlated to direct labor productivity.

Like technicians, using both OJT and classroom training is positively correlated with high performance. SPC and cleanroom procedure training is positively associated with direct labor productivity. These types of training along with teamwork/ communication, leadership/supervision, and problem solving methods, are positively correlated with stepper throughput.

Examples of training

At F5, one of the outstanding performers, a relatively senior (second or third year) employee is assigned to a new hire as a mentor. The mentor teaches the new employee to compile a yearly work plan and goals, and to make updates/milestone checks every three months.

At F3, another outstanding performer, engineers constantly receive some sort of training. There is an in-house curriculum for engineers, but heavy emphasis is placed on mentoring-type of training. This curriculum develops around the needs of the job. Engineers receive additional training in quality control. Their training subjects include 15 hours in quality assurance, 13.5 hours in experiment planning methods, 3.5 hours of introductory reliability engineering, and 3.5 hours in semiconductor reliability. It is estimated that an engineer or manufacturing manager (group leader) spends over half his/her time in training & mentoring activities.

F18, a top performer, provides each engineer with an individual "training roadmap." Continuing engineering training is primarily OJT. Engineers attend training on Taguchi methods, including topics on Taguchi's loss function, orthogonal arrays, linear graphs, parameter and tolerance design, and signal to noise ratios. Engineers also receive Quality Circle Team (QCT) training if they are members of QCTs. Most engineers take advantage of an educational assistance program, through which they can earn advanced degrees (MS or MBA).

At F8, another top performer, new equipment engineers receive 1 week of "common training" concerning safety, fab rules, etc. After that comes 2 weeks of 'line training' focusing on the specific tools and troubleshooting procedures, which is followed by 3 months of mentorship by a senior engineer. In addition, many equipment engineers receive vendor training on particular pieces of equipment. Most process engineers take a 2-3 month course in semiconductor physics during the evening from a local university. Engineers also get reimbursed for courses taken towards advanced degrees from the local university. In addition, there are formal in-house training programs for engineers. These training programs are tailored for the specific areas: etch process training amounts to 4 hours per month on an on-going basis, implant equipment training is 20 hours per quarter, etc.

At F20, a moderate overall performer, training is less important. New equipment engineers get training from vendors on new equipment (3 to 4 weeks). Other kinds of training are less common and amount to about one week of training in five years. Engineers spend less than 3% of their time training others. Engineers train the supervisors, who then train the operators.


6.5 Summary of Preliminary Training Findings

The average number of days of initial training is similar across job categories. Operators, technicians, and engineers averaged 24, 21, and 26 days of initial training, respectively. The number of days of initial training is positively correlated with defect density and direct labor productivity for all three job categories. Additionally, there is a positive association between initial training days and line yield for engineers and operators. For operators, cycle time is positively correlated with initial training days.

About 40% of each worker's time is spent in on-the-job training during the first year of employment across all three job categories, with an additional 5-8% of their time spent in the classroom. No significant correlations were found between this training and performance.

After the first year, approximately 15% of each workers time is spent in on-the-job training, with an additional 6-10% spent in the classroom. Again, no significant correlations were found between this amount of training and the performance metrics. The results show that the type of training and how it is delivered is important, rather than the time in training. It appears that using OJT training alone isn't useful with respect to increasing performance. For operators and engineers, using only classroom training hinders the goal of high performance. Technicians respond positively to solely classroom training. As can be seen in Figures 6-1 – 6-4, there is a positive correlation between the use of both types of training and the performance metrics across all job categories.

End of Chapter 6

Go to Chapter 7
Go to Appendix for Chapter 6
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