yieldHUB can host many types of data. 

Customers mainly upload datalog data from their subcons, but we also support MES/ERP data, genealogy data and inline Fab data.

The most complete understanding of yield is reached when you have all of the above integrated into yieldHUB.

We realise that this is not common as fabless companies rarely have in-line Fab data at their disposal. IDMs can have all the above data sources but rarely are able to bring t hem all together like is possible in yieldHUB.

yieldHUB supports dozens of data formats, including STDF, Excel and Text formats. Many formats will need slight tweaking of our parsers. The key requirement for fast setup of your new yieldHUB site is consistency.

It’s more expensive to set up and maintain manual data entry formats.

Here are some more tips:

– The original datalogs are preferable, not derivative (or converted) data
– The format needs to be consistent to eliminate ongoing maintenance
– Where possible, include test limits with the data
– The datalogs need to contain program name and lot ID
– Make sure to update the program name when there is any change to test limits or any edit at all to the production program

yieldHUB already parses data from multiple foundries(fabs).

However some fabs have varieties of formats (i.e. they are not consistent company to company).

Often the limits in these fab datalogs are of no use to you their customer. yieldHUB allows you to generate useful limits for clones of these fab tests. Hence you can set alerts and do trend analysis that allow you close control of the fab performance.

Read more about WAT Analysis and PCM

If you can get a feed of genealogy data from, say, your MES system, then yieldHUB can use that so you can do genealogy correlations. The best case is wafer level linking. So a list of wafers from fab has a link to the final lot split it is associated with.

yieldHUB also supports multi die correlations. This is where more than one die is in a package. Often the dice will be from different fabs. We can specify a comma separated value (CSV) format that will be compatible with yieldHUB. Our customer simply runs a query on their MES system to generate the CSV file, often on a daily basis. This updates the yieldHUB database and is seamless to the user.

Manufacturing Execution Systems (MES) and Enterprise Resource Planning(ERP) systems are valuable data sources when you link them up with datalog streams.

With yieldHUB they can be the top layer of drill down analysis all the way to even individual die performance.

They also enrich the data for any production lot stored in yieldHUB. In just one example, the package type for a lot can be derived from the MES feed and this allows you to analyse parametric performance in yieldHUB by package type.

MES/ERP can also help with data integrity. We can use this information to check if all the data for a lot is in yieldHUB and help identify gaps caused for example by testers not uploading data to the data pipeline.

This is an example of a feed from an MES system which can be linked by yieldHUB with the lot number read from the datalog.

Final Test Lotid, Qty In, Qty Out, Yield, Package, Handler Id
L34287.2, 1000, 972, 97.2, PLCC, XT2332

There is such a variety of in-line fab data and the data can be event-based as well as time-based. The great thing is that this is the most powerful data for yield improvement.

Correlations can be done across multiple steps in fab once the data from each stage is in yieldHUB. There was a time when yieldHUB only catered for PCM/Wat, Wafer Sort and Final Test. yieldHUB now has an unlimited number of stages across which you can analyse and do correlations.

Can yieldHUB Help You?

Contact yieldHUB today! Our global sales and support team will be happy to help.