FabTime Cycle Time Management Newsletter Abstracts

FabTime Cycle Time Management Newsletter Abstracts - Volume 10 (9 issues)

Improving Factory Cycle Time through Improvements at Non-Bottleneck Tools (Issue 10.09)

Our FabTime user tip of the month is about reporting open lot cycle times. We have one community announcement, about ISS Europe 2010. We also have a new subscriber discussion topic (well, something that we haven’t discussed in several years): operator productivity metrics.

In our main article this month, we return to something that we think is important for cycle time improvement efforts, but that we haven’t discussed in detail since Volume 1 of the newsletter: cycle time improvement at non-bottleneck tools. It’s well-known that in order to increase overall capacity in a fab, it’s necessary to focus on the bottleneck (or bottlenecks, in most cases). However, when seeking to improve cycle time, it’s possible to make improvements at tools that aren’t capacity bottlenecks, and see improvement in overall cycle times. In this article, we explore the impact of improvement at non-bottleneck tools in a reentrant environment, and then offer concrete suggestions for deciding where to begin, and taking action.

The Hawthorne Effect Revisited (Issue 10.08)

In this issue, we have only one brief community announcement, about FabTime’s sponsorship of the coming Fab Owners Association golf tournament. Our software user tip of the month is about our new custom chart feature. We have subscriber discussion concerning dispatch execution / dispatch compliance, and target percentages of hot lots in the fab.

In our main article this month, we revisit and refresh our very first newsletter topic. The Hawthorne Effect, based on studies that took place at the Western Electric plant in Hawthorne, Illinois, suggests that worker productivity improves as a result of workers having their performance monitored, and then working harder. Although the Hawthorne Effect was formulated in the early 1920’s, it remains of interest today. There continues to be debate over whether or not the Hawthorne Effect is “true”. That is, people question whether in fact the productivity improvements recorded could be rightly attributed to the study at all. FabTime’s feeling is that the general conclusion, that people will work harder if management pays attention to their outcomes, is intuitive and valid. For those who believe this, too, we’ve included some recommendations for leveraging the Hawthorne Effect in wafer fabs.

Using Short-Term Indicators to Improve Long-Term Performance (Issue 10.07)

In this issue, we have announcements about registration for two upcoming industry conferences, a new member of the Fab Owners Association, and a new issue of an industry publication dedicated to cost modeling. Our software user tip of the month is about filtering the WIP line on the tool state charts in FabTime (and other modifications to the tool charts). Our subscriber discussion forum this month has two responses to last month’s main article about forecasting lot completion dates. We expect readers to find these responses quite useful.

In our main article this month, we discuss a central management issue in running wafer fabs, the constant need to translate short-term signals into actions to drive long-term goals. Of course the translation of longer-term goals into shorter-term actions is a task that people undertake every day, in many areas of their lives. This task is particularly complex in wafer fabs, however, because of the high volume of data available, and the high degree of complexity and variability. In this article, we discuss some of the real-time signals that indicate problems in fabs, as well as some intermediate goals that are used as early warnings regarding longer-term goals. Our conclusion is that while access to good data is essential in translating from short-term signals to long-term goals, the human element remains necessary, too.

Forecasting Lot Completion Dates (Issue 10.06)

In this issue, we have two announcements, one about a book focused on improving factory performance, and the other about registration for the upcoming ISMI Manufacturing Week in Austin, TX. Our software user tip of the month is about using the new "export all data" capability in FabTime. We also have one subscriber discussion topic, about bringing a fab back up to full speed after a slowdown. We hope that this question will be relevant for many of you in the coming months.

In our main article, we discuss forecasting of lot completion dates. We believe that projecting shipment dates for individual lots is likely to become increasingly necessary for fabs. In this article, we offer a general method for predicting lot shipment dates using the sum of planned cycle times by step. We review several implementation details, particularly in regards to computing the step-level cycle times, and varying x-factors to account for changes in lot priority. We also briefly touch upon estimating earliness or lateness for in-progress lots, by comparing actual cycle time to expected cycle time to this point.

Problems that Stem from Broken Assumptions (Issue 10.05)

In this issue, we have two brief community announcements, and a response from a subscriber to two previously introduced discussion topics (dispatch precision and tool state reporting). Our FabTime software user tip of the month is about using the alert functionality in FabTime to send alerts to other people from your team.

In our main article this month, we discuss problems that can stem from broken assumptions. Whenever you implement a series of steps, whether this is in software code, a spreadsheet, or an operational process in a fab, you make assumptions along the way. Often these assumptions seem so obvious that you don't even document them, let alone plan for them to be broken. But of course sometimes they do break. When that happens, the root cause is often difficult to identify. We decided to open up a dialog with our newsletter subscribers on this issue of broken underlying assumptions.

Responses to Four Recent Discussion Topics (Issue 10.04)

In this issue, we have two conference announcements, one about the IMEC Technology Forum scheduled for Brussels in June, and another about the AEC/APC Symposium scheduled for Michigan in September. Our FabTime user tip of the month of the month is about using new average WIP Trend and Pareto charts in FabTime (and using the same averaging functionality in the WIP Turns charts).

This month we have rolled the subscriber discussion section into the main article. We have four interesting and detailed discussions ongoing with subscribers related to: dispatch precision (a dispatch compliance metric); equipment uptime reporting (the main topic of the last issue); granularity of tool state reporting and modification of transactional data; and calculation of degree of lateness for in-process lots. In light of the substantive nature of these discussions (and with many thanks to the subscribers who have contributed), we’ve decided instead of a new main article to simply highlight these four topics.

Equipment Availability versus Equipment Uptime and Manufacturing Time (Issue 10.03)

In this issue, we have two announcements, one concerning social networks, and the other a call for papers. Our FabTime user tip of the month of the month is about identifying cumulative cycle time contributors, across the lifetime of lots. In this month’s subscriber discussion forum, we have two responses to topics raised last month (dispatch compliance metrics and correlation in wafer fab data), as well as a new subscriber question about tracking of late lots.

Our main article this month is a relatively brief discussion of equipment state and availability-related definitions. We review the SEMI E10 definitions for equipment states, and discuss our intention to transition from using the term “Availability”, which is not defined in terms of the E10 tool states, to using separate terms relevant for maintenance personnel vs. manufacturing personnel. For maintenance effectiveness tracking, we will use the metric Equipment Uptime (Productive + Standby + Engineering), reflecting the time that the tool is available for either production or engineering use. For manufacturing personnel, however, we will continue to report Manufacturing Time (Productive + Standby), which is the time that the tool is available for manufacturing use. It is, of course, the utilization of this Manufacturing Time that drives cycle time performance.

Correlation in Wafer Fab Data (Issue 10.02)

In this issue, we have two community announcements (one about an industry survey and another about an industry-specific networking site). Our FabTime software tip of the month is about identifying current top cycle time contributors in a fab. We have no subscriber-submitted discussion, but we have introduced a new topic (dispatch compliance reporting).

In our main article this month, we discuss potential charts to explore data correlation in wafer fabs. We begin with a general discussion on correlation vs. causation, and then propose several potential data pairing that we think would be useful in increasing our understanding of fab behavior. These range from the obvious example of looking at tool group cycle time vs. utilization to less obvious examples, such as overall fab cycle time vs. number of current single path operations. We hope that this article will stimulate discussion among our subscribers on data relationship in the fab.

Setting WIP Goals in Wafer Fabs (Issue 10.01)

In this issue, we have two community announcements, one about a special issue of Future Fab magazine, and the other a call for papers for the next MASM conference. Our software user tip of the month is about analyzing MTBF and MTTR data in FabTime. We have no subscriber discussion this month, but we have listed some recent topics, and welcome your feedback for future issues.

In our main article this month, we return to a topic addressed in Volume 9, Number 9, controlling WIP in the fab. In that previous article, we discussed the management of WIP bubbles. In this article, we discuss setting goals for WIP in the fab as a whole, and by area, and the tracking of the absolute delta from WIP goals as a measure of variability. We also discuss the importance of ensuring that WIP goals are consistent with other fab goals, and illustrate this with a detailed example. While WIP levels are probably declining right now in many fabs, we reiterate the point from last month that a downturn is a good time to focus on fundamentals. Understanding and tracking your WIP levels in more detail is a good place to start.