A Simplified Method for SiC Mosfet Spice Models Creation

With advances in electric vehicle (EV) technology, the use of SiC MOSFETs for power converters has increased. To facilitate converter design, accurate and robust SPICE models for MOSFET devices are essential. Although existing solutions, including the modified Enz-Krummenacher Vittoz (EKV) model, exist, they often lack detailed specifications of model parameters. This study provides detailed insights into the parameters of the Wolfspeed C2M0040120D model, highlighting the dominance of the threshold voltage, drift resistance and transconductance fitting curves. By adjusting these curves, the manufacturer's model accuracy in the on-state was improved from 69% to 50%, with a maximum mismatch with the datasheet for the first EKV model variant (EKV model v1). To further improve the results, the paper proposed an optimization algorithm that defines the best set of fitting parameters based on the total error cost function to find a second EKV model variant (EKV model v2). The new model showed a better match, reducing the maximum static mismatch to 22%. The double-pulse test conducted on the EKV model v2 shows improved matching in switching, which was verified by experimental results and faster simulation.