Resolution of linear/nonlinear inconsistencies in charge-conservative FET models

Abstract

The presence of a nonlinear capacitor in a circuit model mathematically implies the existence of a parallel transcapacitor, if the value of its capacitance is a function of a remote voltage. As a consequence, simulating small-signal excursions with a nonlinear model produces data which is inconsistent with the simulations using the small-signal model, if the latter does not include the transcapacitance. However including the transcapacitance in the small-signal model is complicated by several effects. Furthermore, the transcapacitance seems to confound the convergence of the harmonic balance simulation algorithm. Herein, a new mathematical formulation has been concocted which renders the transcapacitance equal to zero at any prescribed quiescent point, and thus negligible for moderate large-signal excursions therefrom. As a result, the small and large signal models are consistent and harmonic balance convergence capabilities is improved.