Development of a High Power Density Inverter

Rising in switching frequency directly reduces passive component requirements in power converters, improves achievable control bandwidth, and in theory, enables miniaturization. Soft switching is used to reduce frequency dependent loss, so that power converters can be scaled, with good efficiency, to very high switching frequency. With new circuit architecture, new optimal control method, and the new high bandgap power mosfets, it becomes possible to reduce the inverter size and achieve dramatic improvements in transient response. While the soft switching technique diminishes mosfet turn-on loss, it often rises other types of losses such as inductor core loss, and mosfet’s body-diode conduction loss. Unfortunately, these losses are also frequency dependent (meaning the losses increase with frequency). To reduce these losses, existing techniques focus on optimizing either inductor ripple current or mosfet turn-on delay. Since the two types of loss mechanisms are closely related, existing techniques only achieve sub-optimal results. Simulations and experiments in this thesis suggest that the true optimal operating point, where the loss is lowest, can be found by tuning in both directions (optimizing inductor ripple current and mosfet turn-on delay). In this thesis, zero voltage resonant transition of a half-bridge is reviewed with modifications to make the analysis correct in high frequency converters where the resonant transition occupies a large part of a switching cycle.