Accurate small-signal and large-signal equivalent-circuit models of microwave active devices (e.g., diodes, FETs and HBTs) are very useful for device performance analysis (e.g., noise, gain) in designing microwave circuits and characterizing the device technological process. Nonlinear models of microwave devices are commonly in terms of state functions such as closed-form equations, Volterra series, or look-up tables. These quantities are classically determined via a small signal detour based on multi-bias S-parameters and DC measurements. The linear equivalent-circuit model is generally more accurate than a linearized nonlinear model for predicting the S-parameters.

Chapter Contents:

• 7.1 Overview of ANN Modeling Technique
• 7.2 ANN-Based Linear Modeling
• 7.3 ANN-Based Nonlinear Modeling
• 7.4 ANN-Based Noise Modeling
• 7.5 ANN Integration and Differential Technique
• 7.6 Summary

Inspec keywords: Volterra series; neural net architecture; microwave circuits; field effect transistors; S-parameters; neural chips; equivalent circuits

Other keywords: Volterra series; nonlinear models; FET; microwave active devices; diodes; HBT; large-signal equivalent-circuit models; closed-form equations; microwave circuits; artificial neural network modeling; linear equivalent-circuit model; small signal detour; noise; small-signal equivalent-circuit models; multibias S-parameters; device performance analysis; linearized nonlinear model; state functions; look-up tables; device technological process; microwave devices; gain; DC measurements

Subjects: Neural nets (theory); Insulated gate field effect transistors; Nonlinear network analysis and design; General circuit analysis and synthesis methods; Microwave circuits and devices; Other field effect devices; Neural nets (circuit implementations); Neural net devices