The key to practical and sustainable yield improvement is actionable data that is timely, clean, relevant, intuitively presented and comprehensive.
Timely: You can see the latest data when you login to yieldHUB. You don’t have to download data, you can click to analyse straight away, even for high volume data.
Relevant: The data you are looking at is relevant to yield improvement. For example, the highest failing tests are presented first. You can quickly trend the tests vs all other lots to see the trend over time.
Intuitively Presented: Charts and tables are available which clarify any issues. For example, if you look at final test data, immediately you can see the data from fab, which is very likely what you would be looking for next anyway to solve a yield problem.
Comprehensive: You have a large array of analysis tools and capabilities at your disposal to work with the data in question. For example, you can do commonality, ANOVA, genealogy correlation, characterisation, sensitivity analysis and many more. You can select the data and see what the yield would look like if you changed limits on certain tests.
The result of using the power of yieldHUB is that yields will improve over time. Engineers will become adept at using the power of yieldHUB to get a few extra % of yield out of their products. They may even come up with a new type of insight that could be added to yieldHUB to further speed up yield improvement. This is how yieldHUB evolves and everyone wins!
If you only can analyse small amounts of data at a time, which was traditionally the case, you will not be able to easily see important trends over time of bins and tests. These are very important facilitators for long term yield improvement. Connecting such trends with fab parameter variation is really important and has been hard to do in the past being manual. With yieldHUB, the links are automated and the reports showing the effect of fab parameters on yield in wafer sort and final test is a key capability and critical for true long term yield improvements.