MOSFET failure modes in the zero-voltage-switched pull-bridge switching mode power supply applications

As the demand for the telecom/server power is growing exponentially, the need for higher power density increases each year. Increasing power density relies on less component counts, smaller reactive component size, and/or better system efficiency. Higher switching frequency leads to the smaller reactive and filter component size. Better efficiency, which reduces the heat sink or paralleled devices, requires the reduction either in conduction losses or switching losses, or both. Therefore, phase-shifted zero-voltage-switching (ZVS) full bridge topologies are gaining popularity due to their extremely low switching losses in the power devices even at higher switching frequency. However the intrinsic body diode is required to conduct in order to create the ZVS turn-on condition for the power MOSFET. Due to the extremely low reverse voltage, the reverse recovery charges might not be swept out before turning off the MOSFET. Therefore, the body diode might be subjected to the dv/dt stress when it is not yet capable of blocking reverse voltage. Also not able to maintain the ZVS operation at low load will force the on-state MOSFET to turn off at hard-switching condition. Like in the hard-switched full bridge topology, the cdv/dt shoot-through current might produce a voltage spike at the gate of the on-state MOSFET on the same leg and cause devices failure. Several silicon technologies will be presented to resolve the above-mentioned failure modes in the ZVS topology. Fast reverse recovery time and better dv/dt ruggedness make this new MOSFET technology suitable for higher frequency ZVS full-bridge applications. Inherent with extremely high silicon density and low gate charge, these new MOSFETs ran reduce the component count with the same or better performance and will enable much higher power density for the next generation telecom/server SMPS designs.