Yield optimization of analog ICs using two-step analytic modeling methods

Innovative methods for statistical design optimization have been applied to the development of analog IC blocks. The most important feature of these methods is the derivation of an analytic function representing the yield surface in the design parameter space. Using this analytic model it is possible to optimize the yield accurately and efficiently. All the required operations are implemented in an integrated and fully automated CAD system. A comparison between simulated and measured data for several wafer lots demonstrates the validity of the approach.>