Microwave Power Sic Mesfet Large-signal Modeling Techniques

Silicon Carbide (SIC) is the third generation (wide band gap) of semiconductorelectronic materials. SiC metal semiconductor field effect transistors (MESFETs) areexcellent for high frequency, high power, high temperature and radiation resistantelectronics. Compared with physical-based models that most commonly studied thelast decade, measurement-based models are simple, fast, and accurate, and are morepopular at present. In this work, large signal models (LSMs) including empiricalmodels and black-box models, which are based on measurements, are studied withsmall-signal and DC I-V analytical characteristics.At first, a new small signal equivalent circuit model (SSECM) topology formicrowave power SiC MESFET and a method for broadband determination of theSSECM are proposed. The established SSECM is accurate, relative simple andbroadband (0-20 GHz), and is very useful for large signal characteristics analysis.Then, the SiC MESFETs large signal characteristics, which are based on GaAsFETs LSMs and SiC MESFETs DC I-V characteristics, are modeled by modified(Triquint's own model) TOM, Angelov and Curtice Cubic DC I-V models, andAngelov nonlinear capacitor models. The LSM has been implemented in ADS by usingSymbolic Defined Devices (SDD), and its validity has been verified by measured DCI-V curves, multi-bias S parameters and harmonic balance (HB) simulations.At last, the support vector machine (SVM) and its application in large-signalmodeling are introduced briefly. And a new black box large-signal modeling approachbased on support vector regression (SVR) has been proposed. Experimental results aregiven out to validate its good ability in predicting SiC MEFETs nonlinear performance.Accurate power device modeling is critical to achieve device's maximum powerand efficiency and can speed the development of SiC MESFETs devices and circuits,and their application in kinds of domain.