DocumentCode :
3066425
Title :
An effective optimization for microwave power device modeling [GaAs MESFET]
Author :
Gao, YiFan
Author_Institution :
Chang´´an Univ., Xi´´an, China
fYear :
2004
fDate :
24-27 Aug. 2004
Firstpage :
290
Lastpage :
292
Abstract :
A new systematic optimization strategy for extracting the model parameters of a small-signal equivalent circuit of a microwave power GaAs MESFET is proposed. On the basis of sensitivity analysis of error functions of the model elements, a rearrangement of the optimization sequence of sub-error functions referring to the model element sensitivities from large to small, and a transformation of optimization directions of error functions along with optimal directions, the global-error function is fast convergent under given starting values. In addition, the logarithmic damping algorithm is also proposed for stabilizing the final results by controlling the updates of insensitive model elements. The proposed approach is applied to extract the parameters of a microwave power GaAs MESFET chip with gate-width 2100 μm under different biases. The results show that the agreement is good between the calculated values and the MESFET measurement data.
Keywords :
III-V semiconductors; convergence of numerical methods; equivalent circuits; gallium arsenide; microwave power transistors; optimisation; power MESFET; semiconductor device models; sensitivity analysis; 2100 micron; GaAs; error function convergence; error function sensitivity analysis; logarithmic damping algorithm; microwave power MESFET; microwave power device modeling; model parameter extraction; modeling optimization; small-signal equivalent circuit; Damping; Equivalent circuits; Frequency; Gallium arsenide; MESFET circuits; Microwave devices; Power system modeling; Reactive power; Scattering parameters; Sensitivity analysis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Radio Science Conference, 2004. Proceedings. 2004 Asia-Pacific
Print_ISBN :
0-7803-8404-0
Type :
conf
DOI :
10.1109/APRASC.2004.1422462
Filename :
1422462
Link To Document :
بازگشت