Statistical Large-Signal Model Enabling Yield Optimization in High-Power Amplifier Design

Abstract

A statistical large-signal model is presented that allows for optimizing yield of high-power amplifier MMICs. The modeling technique is based on the transformation of process control data into modeling parameters of an empirical, compact large-signal device model, followed by a multi-variant statistical analysis, resulting in a full set of principal components for both the current and the charge model. The model component has been implemented into ADS (Agilent) and an automated software periodically updates the statistical model parameters.