|
What is the recommended sample size of the reference data? |
|
 |
 or high confidence in the results, it is recommended that you take 1000 wafers or more to generate "good" and "bad" distributions. The application presents the "goodness" of the distributions by checking for its ability to separate "good" data from "bad" data. It is possible to use less data depending on the characteristics of the specific ET parameter.
|