An evolution algorithm for noise modeling of HEMT’s down to cryogenic temperatures

Abstract

The implementation of an Evolution Algorithm (EA) for the cryogenic noise modeling of microwave devices has been the object of the present work. Several approaches exist to treat the modeling issue as far as advanced microwave transistors are concerned. The aim of such procedures is mostly oriented to extract a linear equivalent circuit including noise performance that allows to reproduce the global device behavior in a small signal regime. In the present paper, the authors show how EA's can be employed to solve the noise modeling problem according to a blackbox approach. The application refers to the simulation of the Noise Parameters of High Electron Mobility Transistors (HEMT) in the 6-18 GHz frequency range and down to cryogenic temperatures (90K) compared with experimental data. The quality of results indicates that EA techniques represent a truly alternative way to determine the microwave noise performance of HEMT devices, thus furnishing a flexible tool to support CAD of high sensitivity -ultra high speed circuits.