"PWM Spectral Shaping for Automotive SiC Controllers Based on a Random\u2013Periodic Coupled Spectral-Diffusion Kernel Model"

"High-switching-frequency automotive SiC controllers tend to generate pronounced narrowband harmonic spectral peaks, which may trigger electromagnetic interference (EMI) and motor noise, vibration, and harshness (NVH) tonal whine. This paper proposes a random\u2013periodic coupled spread-spectrum PWM (RPC-PWM) strategy and establishes a spectral-diffusion kernel model. By coupling random frequency dithering with periodic frequency sweeping, the proposed method enables controlled diffusion of harmonic energy. The model quantitatively characterizes the relationship between the dominant-peak attenuation and the 3-dB bandwidth, thereby providing a guideline for parameter design. Simulation and experimental results demonstrate that, under the same average switching frequency, RPC-PWM significantly reduces the power spectral density (PSD) peak at the fundamental switching frequency (by 20.1%) while achieving controllable spectral broadening (an increase of 107.1%). NVH measurements further show that the narrowband tonal components at the 24th\/48th\/96th orders in the order waterfall plot are substantially weakened and redistributed into a broadband pattern, validating the engineering feasibility of RPC-PWM for harmonic suppression and NVH mitigation."