Wafer Data Fab

Empowering design and production engineers to focus on analysis and not wafer data management.

Engineers spend a significant percent of their time managing data, including updating Excel/Spotfire models, creating files, and running reports. 70-85% of an engineer's time can easily be consumed by creating scripts to parse log files and executing the various data management tasks.

Semiconductor design engineers can easily process 50 to 100gb of log files for a lot of 25 wafers. Often they limit the outputs of their test programs as they do not have the tools to process the data. Having the tools to rapidly analyze big data sets of semi-structured log files can have a material impact on the new product development life cycle.

Product and test engineers need to rapidly parse and analyze test log files to identify and improve yield problems. Quick identifications of yield problems and yield enhancement have a direct impact on supply and profitability.

Traditional industry software solutions require significant investment in infrastructure and software. Typically this investment puts these tools out of the reach of semiconductor startups still in design as well as small and mid-size companies in production.

T2VSoft's Wafer Data Fab provides an affordable solution to this problem.

Our solution is a combination of customization to your specific needs and a powerful framework built on Microsoft Azure that is fully managed by us.

Our solution consists of :
  • Initial customization to your specific log file formats and on-going changes to adapt to your engineering needs. Since our loaders are customized programs we can process ANY format,
  • Powerful framework built on Microsoft Azure. We leverage Microsoft’s implementation of the Hadoop ecosystem to process gigabytes of log files in minutes from multiple sites globally,
  • Summary data defined by you. We rapidly combine data from multiple sources to produce summaries such as test, wafer and lot summaries for yield. Allowing engineers to quickly focus in areas requiring further detailed analysis,
  • Cloud storage for terabytes of raw log data,
  • Rapid queries against terabytes of raw log data producing structured output to be imported into popular analytics tools,
  • Easy to use, secure, cloud based portal for distributed engineering groups to collaborate and share data,
  • Geo-redundant infrastructure on Microsoft Azure. Fully managed by us,
  • Fixed implementation cost and monthly operating costs
  • Initial implementation in 4-8 weeks with monthly enhancements based on your evolving needs. We understand that requirements change and the time to adapt to changes is critical,
  • A dedicated team that is an extension to your engineering but with clear exit strategies protecting your investment should you ever want to take the application in-house.

Nantero process flow