Forecasting Methods for MMIC RF Yield

R. Tsai, M. Nishimoto, and R. Lai
TRW, Inc., Redondo Beach, CA  90278

Accurate MMIC RF yield forecasting is a critical tool for proactively defusing yield problems and planning the efficient use of manufacturing resources.  TRW has investigated several methods for forecasting the RF yield of HEMT MMICs.  We have considered and investigated:  Monte Carlo statistical models, correlated statistical models, boundary device models, transforming maps, database models, and Semi-Physical device models.  We compare the advantages and disadvantages of these methods, and illustrate examples of yield prediction with the two best solutions that we have found.  Of the two, the semi-physical device modeling approach has provided the most elegant and versatile solution, but is more complicated to implement.




Return to TOC