FabTime Cycle Time Management Newsletter Abstracts

FabTime Cycle Time Management Newsletter Abstracts - Volume 6 (10 issues)


Operational Recommendations for Wafer Fab Cycle Time Improvement (Issue 6.10)

In our main article this month, we have gathered a collection of operational recommendations and metrics for driving cycle time improvement efforts. We have heard from several fabs recently that they are planning cycle time reduction initiatives for the New Year, and we thought that this collection would be helpful. We include recommendations related to variability reduction, tool utilization (capacity loss), and tool qualification. We also briefly define 9 metrics and data sources useful for short-term cycle time improvement projects.

This month we have a job change announcement and a press release regarding a new version of our software. Also, due to positive response from our subscribers, we are repeating our offer of a free one-hour talk on cycle time management for fabs in the U.S. Our FabTime software user tip of the month describes a short-cut for home page management. This month we have subscriber discussion about WIP States and cycle time estimation formulas.



Estimating and Using Operation Cycle Times (Issue 6.09)

In our main article this month, we discuss uses for both actual and planned operation-level cycle time data. Actual values can be used to flag operations for which the ratio of cycle time to process time (x-factor) is higher than expected. Actual values can be compared with planned values, to identify short-term problems. Actual operation cycle times are also helpful in calculating planned cycle times. Planned cycle time values are in turn useful for estimating lot completion dates and facilitating dispatching decisions. In this article, we also introduce a new sidebar feature: a brief exercise for FabTime software users. Our goal is to make the newsletter articles more hands-on and relevant to our customers, the ones who make this newsletter possible.

This month we have an announcement and call for papers for a conference to be held in France in the spring. We’ve also provided a brief review of the ISMI conference held in Austin last month. We are also repeating an offer to give a free one-hour talk on fundamentals of cycle time management for fabs in the U.S. This month’s FabTime user tip of the month is about clearing Internet Explorer’s temporary file folder to improve FabTime performance. Our subscriber discussion forum is brief this month, but contains an excellent extension to our WIP States proposal.



Cycle Time and Hot Lots Revisited (Issue 6.08)

In our main article this month we revisit the topic of hot lots. We talked about hot lots back in Issue 3.02, but thought that it was high time for a fresh look. This article is adapted from a section in our two-day cycle time management class, as well as from various discussions that we have had with our course and software customers. We discuss reasons for hot lots, the two primary types of hot lots, and the impact of hot lots on cycle time, and conclude with recommendations regarding hot lot management. We also include several references for further information on hot lots.

This month we are pleased to announce a free one-hour talk on cycle time management, which we are offering to deliver onsite for fabs in the U.S. We have several announcements related to conferences and journals. October seems to be a popular conference month! We also have a notice about two jobs that are available at a U.S. wafer fab. Our FabTime software tip of the month concerns independent sorting of chart and data table information. We have subscriber discussion related to two open topics: fundamental drivers of fab cycle time, and identifying the cause of declining moves in a wafer fab.



Setup Avoidance and Dispatching (Issue 6.07)

In our main article this month we discuss setup avoidance policies for lot dispatching. Where present, setups reduce the available capacity of tools, and tend to increase cycle time. This has led to frequent use of setup avoidance policies in fabs. However, running a pure setup avoidance policy can lead to long cycle times for low volume recipes. Various methods exist for forcing a setup to occur to prevent long queue times – several are discussed below. We also discuss integration of setup avoidance into a dispatch factor paradigm, and parallels between setup avoidance and batch size formation decisions.

This month we have a conference announcement about ISSM, to be held in San Jose next week. Our software user tip of the month is a primer on using FabTime’s Excel export functionality. We also have subscriber discussion related to identifying the cause of declining moves in a wafer fab and improving lot tracking in less automated fabs (both from Issue 6.06), and the fundamental drivers of fab cycle time (from Issue 6.05). A new subscriber discussion topic about formalizing methods for setting operation cycle time goals is also included.



Cycle Time and Holds (Issue 6.06)

Our main article this month is about the impact of holds on fab cycle time. This is not a topic that we’ve seen addressed in much depth in industry publications, despite being something that people who work in manufacturing deal with on a day to day basis. Holds negatively impact cycle time in two ways. First, the hold time itself is a direct addition onto cycle time. Second, holds increase variability in the fab, particularly when the time until a lot comes off hold is highly random. And as we know from our previous discussions, anything that increases variability in the fab is also increasing cycle time. We offer a few recommendations for managing holds, and we look forward to hearing your ideas.

This month we have an announcement about FabTime’s sponsorship of the upcoming ISMI Symposium on Manufacturing Effectiveness. We hope to see you there! Our FabTime Software User tip of the month describes how to quickly view the status of all tools in a particular production area. We have no subscriber discussion this month.



The Three Fundamental Drivers of Fab Cycle Time (Issue 6.05)

In our main article this month we have opted to go back to basics. The article discusses the three fundamental drivers of cycle time at the tool level: utilization, variability, and number of qualified tools per tool group. We introduce each of these factors, reviewing why and how they affect cycle time. Each discussion concludes with suggestions for mitigating the effect of the factor, and hence improving cycle times. While we have discussed each of these issues in previous newsletters, this article brings the topic together into one convenient format. We do have a one-hour presentation that is similar to the content in this article. If you would like someone from FabTime to visit your site to give this talk (perhaps to help you to kick-start a cycle time improvement project), please contact us.

This month’s FabTime user tip of the month is about setting up personal goals and displaying them on chart pages. We have subscriber discussion related to last month’s article on lot dispatch for wafer fabs, as well as on the practical application of WIP turns and the cause of declining moves. We also have a conference announcement and call for papers for the 2006 Advanced Semiconductor Manufacturing Conference. ( Return to newsletter subscription page to subscribe and receive the current issue free each month)



Lot Dispatch for Wafer Fabs (Issue 6.04)

In our main article this month, we discuss WIP management, with emphasis on lot dispatching. We begin by defining scheduling and dispatching. We then define several standard dispatch rules, and examine the common factors underlying these rules. Finally, we briefly review some of the fab-specific issues that lend particular complexity to dispatching: batching, setups, and time constraints between process steps. This is far from being a comprehensive description of all of the possible fab dispatch rules. Instead, our intention is to introduce a common vocabulary for dispatching, so that in future discussions we can explore more complex scenarios. We welcome your feedback.

We are lacking in community announcements for this issue, but we do have subscriber discussion related to last month’s article on overall WIP effectiveness, and a new question about troubleshooting declining moves in a wafer fab. Our FabTime software user tip of the month is about excluding extended holds for marketing purposes from the shipped lot cycle times reported by manufacturing. ( Return to newsletter subscription page to subscribe and receive the current issue free each month)



A WIP-Centered View of the Fab: Part 2: Overall WIP Effectiveness (Issue 6.03)

In our main article this month, we continue last month’s discussion of “A WIP-Centered View of the Fab”. Last month we proposed a set of six basic states for recording how lots spend their time in the fab. We also discussed several subtleties regarding measurement of these states, and possible extensions for capturing more information. In this issue, we expand upon this topic by proposing a higher level metric derived from the WIP States, Overall WIP Effectiveness. We welcome your comments and feedback, as we attempt to develop useful metrics to provide a WIP-centered view of the fab.

This month we have a conference announcement and call for papers for the second ISMI (International Sematech Manufacturing Initiative) symposium on manufacturing effectiveness. This was a great conference last year, and we recommend that you participate if you can. We also have a new FabTime software tip of the month describing methods of copying chart images to other applications. In this month’s subscriber discussion forum we have three responses to last month’s main article on WIP States, one response to a previous article about WIP Utilization %, and a response to another subscriber’s question about experiences in improving lot tracking for less-automated fabs.



A WIP-Centered View of the Fab: Part 1: WIP States (Issue 6.02)

In our main article this month we propose a set of basic states for recording how lots spend their time in the fab (we call them WIP States). These are analogous to the tool states defined in the SEM E10 standard, except that they apply to time spent by a lot, rather than time spent by a tool. We propose a set of six basic states, which encompass most of the time that a lot spends in the fab, and discuss several subtleties and possible extensions. Next month we will expand on this topic by proposing a higher level metric derived from the WIP States, Overall WIP Effectiveness. We welcome your comments and feedback, as we attempt to develop useful metrics to provide a WIP-centered view of the fab.

This month we have several community announcements, including notice of the formation of a new semiconductor manufacturing group called the Fab Owners Association (FOA). Our FabTime software user tip of the month is about viewing home page charts in slide show view. This issue also brings subscriber discussion related to WIP utilization %, product mix, shift change effects, and improved lot tracking for less automated fabs.



Product Mix and Cycle Time (Issue 6.01)

In our main article this month we discuss the interaction between product mix and cycle time in a wafer fab. Specifically, we identify a number of reasons why increasing product mix may drive up cycle times. Although product mix itself is not a knob that people in the fab can just turn down to improve cycle time, we believe that exploring the underlying issues in more detail will suggest opportunities for cycle time improvement. We welcome your feedback.

This month’s community announcements section consists of a call for papers that we thought some of you might find of interest. In our FabTime software tip of the month we describe the use of the Dynamic X-Factor chart for looking at shift change effects. This month we are in the rare situation of having no new subscriber discussion. We have listed the recent discussion topics, however, and invite your comments for future issues.