Addressing GaN Converter Challenges: False Turn-On Issues & Switching Loss Modelling

Abstract

Wide bandgap devices are the future of this dynamically changing technological world. Considering Gallium Nitride (GaN) and Silicon Carbide (SiC), GaN has exceptional characteristics that will likely allow it to proliferate greatly in the area of low-and-mid-power power electronic converters. One of the current challenges in this context to completely utilize GaN are the reliability issues, especially false turn-on events, which is a main focus of this thesis. False turn-on due to its momentary short circuit capability deteriorates the converter performance. Furthermore, simple and accurate modeling of GaN device losses is critical to help electronic designers optimize converter designs. This thesis focuses on three contributions to help reduce false turn-on events and improve GaN modeling efforts. First, this work uniquely investigates the optimal pulse-width-modulation (PWM) scheme to balance efficiency and false turn-on. The experimental results lead to a recommendation to use a larger negative bias than is currently recommended by device manufacturers. Secondly, the work proposes a new simplified switching loss model with high accuracy that can be used with different gate drive circuits (including negative gate bias voltages) to make it more useful for power electronics design engineers as a tool. And thirdly, since the main contributor to false turn-on events are the parasitic inductances in the switch and on the PCB, this work proposes a new parasitic inductance measurement methodology which can be implemented using only simple laboratory instruments.