High Frequency Modeling Of EMI Filter-Investigation Of Parasitic Effects

Power electronic equipments generate electromagnetic interference (EMI) noise due to the switching dv/dt and di/dt of power devices, and EMI filter is an effective measure to suppress the EMI noise. However, the real high-frequency (HF) performance of an EMI filter is usually worse than the ideal expectation result in respect that both the components parasitic parameters such as the lead inductances of capacitors and the winding capacitances of inductors and the stray EM couplings between filter components are detriments to the HF performance of EMI filter. For this reason, it is necessary to model the parasitic effects of EMI filter in order to evaluate and predict its performance more accurately through 150kHz to 30MHz.Common mode (CM) choke is an important component in EMI filter, and it is usually employed to suppress CM EMI noise. Considering that the parasitic characteristic of CM choke influences its HF performance significantly, the dissertation presents a procedure to build CM model and differential mode (DM) model for CM chokes from impedance measurement, and CM circuits of subdivided frequency ranges and Foster network are both used to characterize the change of the parameters of CM model with frequency. It also presents a simple method to determine the leakage inductance of CM choke by introducing equivalent magnetic permeability and average leakage flux path length, and it gives a method to calculate the winding capacitance by equating the energy stored in stray capacitance and in winding capacitance.Near-field couplings of capacitor and CM choke are analyzed. The near-field couplings of capacitor are mainly due to the magnetic couplings between the partial loop of the capacitor lead and the circuit loops nearby in the filter, and the near-field couplings of CM choke mostly behave as leakage DM flux coupling with circuit loops in the filter. It introduces how to build up the circuit model considering the parasitic effects and how to extract the coupling parameters by virtual prototyping in FEA tool Ansoft. According to the prediction model, EMI filters with several different layouts are investigated and the influences of couplings related with different layouts on filter performance are analyzed. A simple prediction model forπshaped CM EMI filter is proposed, in which the coupling between the input loop and the output loop of the filter is considered as a key coupling parameter while other couplings are ignored for simplification.In final, the dissertation analyzes the generation and propagation mechanism of EMI noise in power converters. It is found that the transient voltage across the power devices, for example, vds for MOSFET, is the common voltage source for both DM and CM noise, and the primary difference between DM noise and CM noise lies in that they are with different propagation paths. All of the circuit parameters have influences on CM and DM noise level, while for convenience, and in terms of contributions to CM and DM noise of the parameters in converter, it is acceptable to analyze DM noise without considering the influences of stray capacitances distributed between converter nodes to earth, and the influence of Boost inductor on CM noise can be ignored in Boost PFC converter. Based on the generation and propagation mechanism of CM noise, a simple method to determine the CM noise source impedance for power converter is presented, and the simplified CM equivalent circuit model for power converters is deduced.