An Accurate Crosstalk Evaluation and Prediction Method for SiC MOSFET Considering Nonlinear Capacitance and Stray Parameters

Abstract

In high-frequency applications of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{SiC}$</tex> MOSFET, crosstalk restricts the switching speed, increases additional switching losses and reduces the system stability. This paper proposes an accurate crosstalk evaluation and prediction method for <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{SiC}$</tex> MOSFET, which considers nonlinear capacitance and stray parameters. The proposed method is a programmed prediction and evaluation process, including analysis of stray parameters, nonlinear capacitance analysis and measurement of nonlinear transfer capacitance. First, the influence of drain-to-gate stray capacitance <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$C_{\text{dg}^{\prime}}$</tex> caused by PCB layout and probes on crosstalk peak voltage is analyzed in details for the first time. Second, detailed analysis and explanation are made to reveal the disadvantages and inaccuracy of the traditional method which regards the reverse transfer capacitance as a constant. The simulation results show that the error of crosstalk peak voltage under different transfer capacitance values is very large. In addition, the influence of drain-source capacitance is also studied deeply. Third, this paper considers another challenging problem, the accurate measurement of nonlinear reverse transfer capacitance and propose a practical transfer capacitance big-signal measurement method for SiC MOSFETs based on dynamic transient, which can overcome the shortcomings of the frequency measurement method of small signal and shows highperformance and easy-operation. Finally, the performance of this proposed crosstalk evaluation and prediction method is verified and compared experimentally by a double-pulse test (DPT) platform.