Small signal and large signal modeling of HBT's using neural networks

Abstract

This paper presents the small signal and large signal models for an AlGaAs and a SiGe heterojunction bipolar transistor, using neural network techniques. The main advantage of this technique is the wide range of frequencies over which the small signal model is valid and the great accuracy of the large signal characteristics. Both the models have been verified by comparing the simulated values with the measured ones of the HBTs for both the material systems.