EMI Reduction for Smart Power Switches by Iterative Tracking of a Gaussian-shape Switching Transition

This paper presents a digital control strategy to track a Gaussian reference switching profile at the output voltage of a Smart Power Switch by a feedforward gate current profile. The gate current profile is adapted from switching cycle to switching cycle by an Iterative Learning Control strategy to compensate for model inaccuracies, load variations and temperature dependencies. The presented strategy reduces the generation of conducted electromagnetic interferences in a power optimal way. The strategy is verified by a series of measurement results on a rapid prototyping test bench and by an EMC compliance test according to CISPR 25.