FabTime Cycle Time Management Newsletter Abstracts

FabTime Cycle Time Management Newsletter Abstracts - Volume 11 (5 issues)

Time Constraints and Reverse Dispatch in Fabs (Issue 11.05)

In this final newsletter issue of the year, we have community announcements about Future Fab International’s new email publication, and the Fab Owners Association upcoming member meetings. Our FabTime user tip of the month is about adding complex logic within standard filters, using the new in-filter WHERE clause. We have subscriber discussion related to early lot delivery times, OEE, and the number of columns in the PDF newsletter.

In our main article this month, we discuss the application of “reverse dispatch” in the presence of time constraints between process steps. Where there are time windows between steps, lots are usually held at the upstream step, and only released when needed by the downstream step. This poses an implementation issue for dispatch systems, which usually look at one tool at a time. What FabTime has been working on is a system called “reverse dispatch”, by which the person running the downstream tool can look back at all of the WIP queued at the upstream tool, and then make the best dispatching decision for both steps.

Early Delivery Times in Wafer Fabs (Issue 11.04)

In this issue, we have a community announcement about a new position for Professor Scott Mason. Our FabTime User Tip of the Month is about using the Quick Jump feature to navigate within FabTime. We have several subscriber discussion topics, ranging from a couple of changes requested in the newsletter format to responses to the topics raised in the last issue (OEE and capacity planning). We welcome your feedback.

Our main article this month sprang from a subscriber discussion comment about the importance of tracking lots that are early. In most fabs that we’ve observed, lots that are late garner a lot of attention, while lots that are early exit without fanfare. It turns out, however, that early lots can be a symptom of an underlying problem, such as too wide a distribution of cycle times, or inaccurate planning models. In this article, we discuss both the underlying fab behavior that contributes to early lots (including a brief mention of hot lots), and list a few potential consequences of having lots be significantly early.

Computational Issues in Estimating OEE (Issue 11.03)

In this issue, we have three community announcements: the results of the raffle for a complete set of past FabTime newsletters; a notice about the upcoming Fab Owners Association meeting; and a conference announcement for the ISMI's Manufacturing Week. Our FabTime user tip of the month is about setting goals for Pareto charts. We also have one subscriber discussion question about data structures for capacity models.

In our main article this month we discuss some of the computation issues that go along with using OEE (overall equipment effectiveness) as a metric. At its simplest, OEE measures the actual good units produced on a tool relative to the most units that could have been produced on that tool if everything went perfectly (no scrap, no rework, no downtime, no wasted time). Much of the value of OEE lies not so much in the actual OEE values, but in analyzing the underlying components, and understanding why a particular tool is not performing as expected. We last wrote about OEE in the newsletter in detail back 2002. In this article, we revisit the basics of OEE, with emphasis on calculating OEE values from fab performance data. We also specifically address the issues of calculating OEE for batch tools and estimating planned OEE.

100th Newsletter Issue Celebration (Issue 11.02)

This is our 100th issue of the newsletter. Issue number 1.01 was sent out on April 14th, 2000, to 33 subscribers from 17 different companies and universities (plus several consultants). Some of those initial companies don’t even exist anymore, while others have merged and split and changed almost beyond recognition. It’s hard to keep track. But we’re happy to still number some of those initial subscribers among our current 2731 newsletter recipients. And, given the year that the semiconductor industry went through last year, we’re happy to still be here. For our main article this month, we’ll be revisiting the 10 years of newsletter issues, in a bit of a 100th issue celebration.

In this issue, we also have a call for papers for the 6th International MASM Conference and links to several recent news articles that we thought that our subscribers would find of interest. Our FabTime user tip of the month is about using customized text in email alerts.

Product Mix and Cycle Time Revisited (Issue 11.01)

In this issue, we have two community announcements, and one response to a previously introduced subscriber discussion topic. Our FabTime user tip of the month is about using the new Forecast Arrivals Charts to predict future arrivals to a particular tool or step.

In our main article this month we return to a topic first discussed five years ago, the effect of product mix on fab cycle time. Our return to this topic was triggered by a question raised in discussion with a friend: Is it inevitable that cycle time increases as you add technology mix to an existing fab (because you have more dedication, smaller tool groups, tools at higher utilization, more setups, smaller batches, etc.)? In light of this question, we have revised and expanded our previous thoughts on product mix and cycle time, and added some new suggestions for mitigating the negative effects.