Qiang Li, Chair
Tuesday, April 11, 2017
Fred C. Lee
The difficulty of modeling a resonant converter is the fact that, unlike a pulse-width-modulation (PWM) converter, an average concept doesn’t work. The only successful equivalent circuit model, proposed by E. Yang, is based on fundamental approximation and harmonic balance theory. However, it is conceptually hard to understand because of an unidentified "equilibrium state." On the other hand, electric machine theory has already tackled AC signals for decades. By a rotating coordinate transformation, or dq-transformation, AC signals become DC signals. As a result, an equilibrium state is clearly defined and the modeling process is straightforward. Therefore, this paper present a new modeling methodology for resonant converter: rotating coordinate modeling. As a result, the modeling of resonant converters can be as easy as PWM converters. The simulation result shows the proposed model well predict both large-signal and small-signal behaviors of a series resonant converter.
Fred C. Lee
These days, constant on-time current mode (COTCM) control schemes are widely used in the industry VR controllers for its light load efficiency, higher BW design with simpler compensation requirement. The issue of this ripple based current mode control is that when the inductor current ripple becomes small because of ripple cancellation effect for multiphase operation, control becomes very noise sensitive and create jittering at the output. Authors have proposed a new non ripple based 'Inverse Charge constant on- time (IQCOT)' control which can operate seamlessly at ripple cancellation point in multiphase operation. This new control improves the transient response of constant on-time control dramatically also. In this paper a high frequency model for IQCOT control has been derived using describing function. Then an auto-tuning method for Q-value control is also proposed. Simulation results are also presented to verify the derived model.
Niloofar Rashidi Mehrabadi
This paper predicts the behavior of a high-switching frequency SiC-based modular power converter based on low-power validation experiments. Switching model of the modular power converter is developed and low-power validation experimental results are provided. Different sources of uncertainties causing the deviation of the modeling and simulation results from the real system behavior are identified, characterized, and quantified via verification and validation analysis using the switching model and low-power validation experiments. A regression-based model is then used to extrapolate the estimated model form uncertainty (MFU) to the full-power condition where there is no experimental data available.