This paper presents a novel implementation of the Sawyer-Tower system for characterizing the large-signal Coss loss of power semiconductor devices, enhanced to be cost-effective, compact, and fully automated. The system integrates a hybrid class D converter, a control and signal generation system, and programmable test equipment to measure Coss losses based on user inputs to a Python script. The resulting characterization values show strong alignment with previous literature, but with a system that demonstrates more than tenfold reduction in cost, time, and space. This system significantly enhances accessibility for rapid, large-scale characterization of Coss loss across various drain waveform parameters, including frequency, voltage amplitude, dV/dt, and waveform shape.
