Title :
Global dynamic FET model for GaN transistors: DynaFET model validation and comparison to locally tuned models
Author :
Jianjun Xu ; Halder, Sebastian ; Kharabi, Faramarz ; McMacken, John ; Gering, Joseph ; Root, David E.
Author_Institution :
Agilent Technol., Inc., Santa Rosa, CA, USA
Abstract :
Extensive results are presented validating a recently enhanced large signal FET model, DynaFET, applied to an advanced 6×75um periphery, 0.5um gate-length GaN HFET transistor, manufactured by RFMD. Excellent results are achieved for DC (including leakage current), S-parameters versus frequency and temperature, harmonic and intermodulation distortion, as well as load-pull figures of merit, over a very wide range of bias conditions, complex loads, powers, and frequencies. The DynaFET model features detailed dynamic trapping and de-trapping mechanisms for gate and drain lag, and dynamic self-heating. The multi-variate constitutive relations are represented by artificial neural networks (ANNs) trained from large-signal waveform data obtained using an active source injection NVNA-based characterization procedure. The new model is compared with two empirical models, each independently extracted from pulsed I-V data from different quiescent bias points and tuned for specific applications. The DynaFET model is demonstrated to have a much better fit than each of the individually tuned models, even at the tuned conditions. The new global model is sufficiently accurate over the entire operating range of the device that no tuning of model parameters is required for different operating points or amplifier classes of operation. The model runs in transient, harmonic balance, envelope, and all other simulation modes.
Keywords :
III-V semiconductors; circuit analysis computing; circuit tuning; gallium compounds; harmonic distortion; high electron mobility transistors; intermodulation distortion; neural nets; nonlinear network analysis; semiconductor device models; wide band gap semiconductors; ANN; GaN; HFET transistor; NVNA-based characterization procedure; RFMD; S-parameters; active source injection; amplifier classes; artificial neural networks; detrapping mechanisms; dynaFET model validation; dynamic FET model; dynamic self-heating; harmonic distortion; intermodulation distortion; large signal FET model; leakage current; pulsed I-V data; quiescent bias points; size 0.5 mum; size 6 mum; size 75 mum; Harmonic analysis; Load modeling; Semiconductor device modeling; Transistors; Voltage measurement; Compact models; GaN; Microwave FETs; Neural networks; Nonlinear Vector Network Analyzer; Semiconductor device modeling; power transistors;
Conference_Titel :
Microwave Measurement Conference (ARFTG), 2014 83rd ARFTG
Conference_Location :
Tampa, FL
DOI :
10.1109/ARFTG.2014.6899526