Modular-Terminal-Behavioral (MTB) Model for Characterizing Switching Module Conducted EMI Generation in Converter Systems

This paper proposes a new frequency-domain modular-terminal-behavioral (MTB) modeling approach for characterizing conducted electromagnetic interference (EMI) noise sources in a converter. It models the EMI emission of a switching device module under specific switching conditions using a three-terminal equivalent Norton network, which can be extracted from a standard test. Experiments show that the MTB model can be used to accurately predict the EMI noise in a converter for the entire conducted EMI frequency range. It is more general and more accurate than the existing behavioral models, especially in the high frequency range, and it is more convenient than the physics-based model. The study also verifies the importance of EMI source impedance modeling and interactions between common mode and differential mode noises.