Running head: ZERO UNPLANNED MAINTENANCE 1
Implementing Zero Unplanned Maintenance into Daily Work in a Factory
Elizabeth Shockley
LEAD 585 – Leading Quality Improvement Initiatives
March 25, 2012
Dr. Ronald Beach
Southwestern College Professional Studies
ZERO UNPLANNED MAINTENANCE 12
Table of Contents
Introduction 3
Current Process in DMOS6 Factory 3
Improvement and Potential Benefits 5
Malcolm Baldrige Criteria 6
Implementing ZUM 8
Challenges Faced 10
Conclusion 10
Glossary 12
References 13
Implementing Zero Unplanned Maintenance into Daily Work in a Factory
Texas Instruments has developed and initiated Zero Unscheduled Maintenance (ZUM) which is an “equipment-focused continuous improvement program that mandates fire prevention and provides the tools needed to achieve continuous and cumulative equipment downtime reduction” (Boone, 2009). Fire prevention is not literally fire-related, but fire-fighting is a term commonly used to describe the response to equipment failures in the most complex manufacturing environments, such as the semiconductor factory environment. Usually a part or component is identified as being faulty and is replaced in order to get the equipment running again and processing the product. Yet those intimately involved with the repair --- the repair technicians or equipment engineering team --- know that in time this failure may occur again if the root cause of the failure was not also addressed and eliminated. Implementing ZUM into a repair team’s daily routine in a high tech factory provides the tools necessary to drive improvements instead of reacting to failures. Here the discussion will be describing the team’s current process of responding to equipment failures, a proposed implementation of ZUM into our daily routine, the potential benefits of ZUM, how these improvements are linked to basic tenets of the Malcolm Baldrige National Quality Award, and discuss the challenges faced during this improvement.
Current Process in DMOS6 Factory
Our current ZUM process is somewhat limited. We have one tool owner or engineer (both terms used interchangeably) who is our assigned module ZUM “champion” and who sends out a compiled snapshot of our equipment Ao for the last four months prior to the current month to date percentages. This is communicated to our entire team --- repair technicians, engineers, and supervisors --- via email on a biweekly or less frequent basis. See Figure 1, ZUM Dashboard snapshot for our module, representing data from November 2011 through March 2012 (month to date). This is in the exact format which is sent out to our group by our module ZUM champion.
Module / Grouping / NOV / DEC / JAN / FEB / Model / Goal / MTDMetrology / AMAT SemFIB Defect Review / 98% / 95.90% / 96.80% / 94.40% / 92 / 94 / 92.60%
AMAT Semvision Defect Review / 89.50% / 92.10% / 95.10% / 94.40% / 84 / 90 / 88.50%
August AXI Edge Inpspection / 87.20% / 96.40% / 86.40% / 70.80% / 88 / 92 / 95.90%
KLA 8300 CD Measure / 93.30% / 94.10% / 90.50% / 92.40% / 89 / 92 / 90.20%
KLA Optiprobe Thickness Measure / 90.30% / 72.60% / 91.50% / 89.50% / 84 / 90 / 67.90%
KLA RS100 Resistivity Measure / 97.20% / 95.20% / 95.50% / 94.70% / 92 / 94 / 96.30%
Technos Trex631 TXRF Measure / 98.30% / 97.90% / 100% / 99.50% / 91 / 93 / 99.60%
AMAT Compass Pattern Inspect / 91.10% / 75% / 97% / 99% / 92 / 93 / 98.20%
AMAT Verity CD Measure / 85% / 94.10% / 94.70% / 93.90% / 93 / 95 / 86.10%
KLA ATIXUV Dark Field Inpsect / 99.30% / 97.60% / 98.70% / 91.60% / 88 / 92 / 96.40%
KLA AIT Pattern Inspection / 90% / 86.90% / 91% / 97.10% / 92 / 94 / 87.90%
KLA Archer Overlay / 92.50% / 97.50% / 97.20% / 74.80% / 93 / 95 / 98.10%
KLA Archer Aim Overlay / 94.70% / 97.40% / 97.20% / 90.80% / 94 / 95 / 93.20%
KLA Puma Dark Field Inspection / 93.20% / 96.20% / 98.50% / 92% / 86 / 92 / 98.10%
KLA SLF77 Reticle Inspection / 96.40% / 84.30% / 85.60% / 90.80% / 90 / 92 / 91.90%
Metron RMS6 Reticle Sort / 98.80% / 86.30% / 97.20% / 95.80% / 93 / 95 / 98.70%
Nikon 3100 Micro Inspection / 81.50% / 99.90% / 99.60% / 99.40% / 96 / 97 / 99.40%
Veeco VX330 Prolile / 88.60% / 77.40% / 91.90% / 98.90% / 88 / 92 / 63.40%
KLA F5XP BEOL Thickness Measure / 91.90% / 86.70% / 89.60% / 93.20% / 80 / 90 / 91.70%
KLA F5XP FEOL Thickness Measure / 73.80% / 98.30% / 93% / 96.80% / 88 / 90 / 86.70%
Asyst SMSA4300 Sort / 93.80% / 93.70% / 98% / 97% / 88 / 90 / 96.10%
Asyst Spartan Sort / 95.90% / 98.10% / 84.60% / 77.20% / 90 / 95 / 99%
KLA 2350 Pattern Inspection / 81.20% / 92.30% / 86.60% / 89.40% / 94 / 95 / 84.20%
Leica INS3300 Defect Review / 63.60% / 39.90% / 88% / 96.60% / 96 / 97 / 99%
KLA Es32 EBeam Inspection / 67.60% / 95.50% / 90.80% / 59% / 85 / 88 / 72.80%
KLA SP1 Particle Inspection / 95.40% / 93.10% / 95.80% / 96.40% / 86 / 90 / 95.50%
KLA SP2 Particle Inspection / 91.30% / 66.20% / 86% / 61.60% / 93 / 95 / 78.30%
KLA Viper Macro Inspection / 90.60% / 97.40% / 96.50% / 98.60% / 94 / 95 / 98.80%
Figure 1. Zero Unplanned Maintenance (ZUM) Dashboard, DMOS6 Texas Instruments, Inc.
This dashboard is available via our company’s intranet only and pulls data from the actual equipment available time or Ao (Personal communications, March 23, 2012). The email does not include any plans to address or identify the root cause nor does it schedule maintenance to address the failures. Most of our technicians glance at this but the snapshot does not contain the specifics needed to make it a useful tool. In addition, the amount of data it contains is a lot of information which makes it less appealing and less useful as a tool. The engineers are required to brief or present to our factory Leadership Team their equipment Ao every fifth Friday. This brief also includes what they are doing to address repeat failures. We do not have a daily ZUM activity or discussion to drive these improvements with our team of technicians. Our daily morning startup or pass down meeting is focused on reactions to equipment that is in need of repair. A small portion at the end of this thirty minute meeting is devoted to reviewing the scheduled maintenance coming up that day and who will be performing it. In attendance at these meetings are the tool owners, technicians, supervisor (me) and our module manager.
Improvement and Potential Benefits
The concept of ZUM is not a new one. It is called different things by different companies depending on the perspective of maintenance, such as Total Predictive Maintenance or Reliability Centered Maintenance (Boone, 2009). The benefits of ZUM include a structure by which “comprehensive analyses, closed-loop problem solving, and effective equipment knowledge management can be achieved” (Boone, 2009). Utilizing these tools within ZUM also helps to reduce disruptions to production in the factory.
An important benefit to ZUM is that by resolving repeat failures on our equipment, our team supports our factory’s 2012 Priorities (Personal communications, March 23, 2012) which includes improved product cycle time. ZUM improves equipment availability and that translates into more equipment available for more time (capacity) so that our product can be processed on this equipment. This impacts cycle time directly.
Another potential benefit to ZUM impacts our technician job satisfaction. These technicians received their specialty education in electronics repair and most took jobs as equipment engineering technicians because they enjoy troubleshooting a problem and then repairing the problem. By implementing ZUM in a proactive manner, technicians can take ownership of resolving the problems they are seeing on a recurring basis, become more involved with our engineering staff, and learn how to apply data seen in the pareto charts to a plan of action which will resolve these repeat failures.
In addition, application of ZUM provides a method by which our team can capture lessons learned and further identify which parts are needed on hand to be ready for the most frequent repairs of our equipment. This information can be archived and made available via our team’s share point for easy access by all technicians and engineers on our team. This improves our knowledge management and provides potential for more detailed accounts of what methods or repairs are done for which failures. Capturing this not only improves knowledge management but also could save a technician time who may encounter a problem for the first time. If this problem and associated repair notes are documented on our share point then this could potentially save hours of troubleshooting.
Malcolm Baldrige Criteria
ZUM implementation in our module would reflect and support three of the Malcolm Baldrige National Quality Award (MBNQA) criteria categories of Leadership, Measurement, Analysis & Knowledge Management, and Process Management (Summers, 2009). Our company’s quality policy is the best place to start when establishing how ZUM implementation aligns with these criteria:
“We will achieve business excellence by:
· Encouraging and expecting the creative involvement of every TIer.
· Listening to our customers and meeting their needs.
· Continuously improving our processes, products and services” (Personal communications, March 23, 2012).
Our factory Leadership Team has communicated the ZUM vision for our factory. It is a process that will be applied to improve cycle time and equipment availability (Ao). They have communicated the initiative via our department quarterly meetings and to the engineering population via our monthly engineering meetings. They have further tied how this process of identifying repeat failures on equipment specifically impacts product cycle time, equipment availability, and quality. They have solicited the involvement of everyone in the organization in order to make this a successful effort with measurable results.
Summers notes that “Leaders use measures of performance to communicate what activities are important. Good measures of performance are designed based on what is valued by the organization and its customers” (2009, p. 250). Our factory uses a balanced scorecard to measure our progress towards our priorities and objectives. Much of the data that is reflected on this scorecard is restricted when specific numbers are included. A sanitized scorecard for our factory includes categories of New Products (including prototype cycle times), Execution (includes cost per pattern wafer, defect density, cycle time, and ZUM), Quality (includes excursions, customer returns), Safety (lost time, recordable incidents, safety training compliance). The measurement of ZUM is part of this balanced scorecard and therefore is a measurable entity that is tracked. Finally, ZUM does align with the MBNQA criteria of Process Management by establishing this as a process whereby we can gather the needed information that will be used in our improvement efforts. ZUM impacts cycle time and equipment availability. This is where our efforts are directed towards improvement, and the data collected specifically provides the information needed to address repeat failures.
Implementing ZUM
ZUM is definitely a culture change within our module. It will not completely turn around in a short period of time, but over a period of time it will continue to become more of our routine if implemented properly. According to Boone, “There are three well-known key elements for creating change: 1) make it available; 2) make it simple; and 3) make it mandatory” (2009, para. 11). Making it simple will be very important for our technicians to embrace and want to accept. It should also be non-disruptive, which would be satisfied by making it part of our daily routine.
Another equipment engineering supervisor who manages a team of technicians in another module in our factory slowly got buy in towards ZUM from his technicians and engineers by keeping it simple yet mandatory. Our team can do this by reviewing ZUM data once a week during our team’s morning pass down meetings, utilizing the quality tool known as the pareto chart (Summers, 2009). The pareto chart is easy to understand because it sorts the data by failure code and is a graphic representation of what is happening with our equipment. Our team can implement ZUM into its daily routine by having one tool owner or engineer (on a rotating schedule) review their equipment available time (Ao), show the pareto chart, as seen in Figure 2 of the top three failures, and then open up the discussion for everyone to brainstorm the probable root causes of those top three failures. The total time spent on this is would be no more than 20 minutes. At the end of the 20 minutes, the designated recorder will send the captured brainstorming ideas to the engineer via email so that the engineer can formulate a plan of action to schedule maintenance addressing the root cause(s). Ideally, if all parts and resources are available the scheduled maintenance could occur that same day or the very next day. By pursuing the plan in a timely fashion nearest the brainstorming session, the ideas are fresh with the technicians and the motivation is also more likely to still be there. The following weeks would have other tool owners presenting and then leading the idea sessions, with follow up reports given by those who have resolved root causes and have data to show the elimination of the previous chronic failure prior to ZUM. This would tie our activity to measurable results.
Figure 2. Pareto Chart Indicating Number of Failures by Type and Equipment ID.