Statistical modeling of GaN HEMT based on Monte Carlo simulations

Abstract

A statistical model of GaN high electron mobility transistor (HEMT)'s parameters is presented. This statistical modeling approach includes principal component analysis and factor analysis and multiple regression model. When applied to the database of extracted equivalent circuit parameters (ECPs) for GaN HEMT devices, it has been validated that this approach can generate accurate statistical model by comparing original and Monte Carlo simulated means, standard deviations, correlation matrix and S-parameters. In particular, the primary goal of this work is that the statistical model has been used to perform a ku-band power amplifier design and analysis for the first time. From the results, it has proven that the parameters are statistically indistinguishable from a measured ones.