Factory Explorer® v2.5 Technical Alert #24

Capacity Analysis and Scrap
Inside Rework Loops


Summary: For models that specify scrap inside rework loops, if the scrap percentage is high, or the rework percentage is high, Factory Explorer®'s capacity analysis line yield estimates will be biased on the high side. If product volume is specified in terms of throughput (e.g. Daily Going Rate), the calculated start rate will then be biased on the low side, leading to lower-than-actual product flows at the beginning of the process. Simulation results are affected, but only to the extent that start rates for the simulation will also be biased on the low side.

Details: One of the first steps in Factory Explorer®'s capacity analysis calculations is to predict the flow rate of lots between all steps in the process flow. To simplify this problem, it breaks the flow into rework and non-rework lots and solves each problem separately. The code makes one notable approximation during this process: it assumes that rework child lots do not experience significant yield loss (scrap) within rework loops. For models that specify scrap inside rework loops, if the scrap percentage is high, or the rework percentage is high, Factory Explorer® will not capture this effect. The net result will be that Factory Explorer®'s estimate for line yield will be biased high (it is not capturing the scrapped rework lots). If product volume is specified in terms of throughput (not starts), the calculated start rate will then be biased low (start rate = throughput rate / line yield, so if line yield is biased high, then start rate will be biased low). This start rate bias leads to lower-than-actual product flows at the beginning of the process. Thus, tools that are used primarily at the beginning of the process flow will have lower-than-actual predicted loadings. The capacity analysis impact for tools used primarily at the end of the process flow will likely be negligible.

Simulation results are affected, to the extent that start rates for the simulation will also be biased on the low side. Since the simulation does account for scrap inside rework loops, simulation throughput numbers will also be biased on the low side.

Workaround: The first step is to identify whether or not this problem affects your model. To do so, check to see if your model specifies scrap inside rework loops (the rework loop consists of all steps that a rework child lot will visit). If so, check the rework and scrap percentages. If they are both small (say on the order of 1-3%, then the effect is likely negligible). For example, if the rework% is 3%, and the scrap% is 1%, then the net effect is around 0.03%. But in some cases (development lines in particular) the rework and scrap rates may be much higher. For example, if the rework% is 50%, and the scrap% is 10%, the net effect is closer to 5%. If there are many rework loops with this behavior, then the overall effect can be quite large. In this case, it is best to move the scrap outside of the rework loop. To do so, you'll need to revise the scrap percentage to reflect the fact that it is no longer inside the rework loop. Please contact technical support for the details on how to perform the necessary calculations.

Techalert tafx24-9906


SITE MAP 1SITE MAP 2SITE MAP3

6200 Stoneridge Mall Road • 3rd Floor • Pleasanton, CA 94588 • Tel 925-399-6246 • Fax 925-396-6174
©1995-2015 Wright Williams & Kelly, Inc. • All Rights Reserved • Email: [email protected]