An Improved Parameter Extraction Approach for GaN HEMT Small-Signal Modelling

Abstract

In this paper, a new parameter extraction approach is proposed for the small-signal modeling of GaN high-electron-mobility transistor (HEMT) devices. Based on the classical 17-element model, the inverse matrix is established for conversion and more accurate results are obtained using the backward difference technique. In addition, the frequency segmentation technique is also applied to better capture the nonlinear-characteristics of inductors and capacitors. The proposed approach is then adopted for the small-signal modelling of a $0.4\mu\mathrm{m}$ l GaN HEMT transistor with $2\times 50\ \mu \mathrm{m}$ gate-width. Comparing with the conventional parameter extraction approach, it shows closer agreement with measurement results.