With the high switching speeds of GaN HEMTs, there is an increased risk of dv/dt induced shoot-through events in bridge configurations. An accurate method to quantify the false turn-on energy loss is essential to understand the severity of it and GaN HEMT dv/dt immunity. The capturing and evaluation of a shoot-through event is challenging due to the limited accuracy in instantaneous electrical measurements, and the inaccessibility of the gate node at die level. First, we revisit the concepts of dv/dt immunity and its relation to device Cgs and Cgd. We show that the maximum value of the induced vGS spike occurs at a dv/dt value much lower than the peak dv/dt value, providing important insights. Then, we propose a simple and robust experimental approach, adapting a no-load circuit, to quantify dv/dt induced shoot-through energy loss. The technique does not require any high-frequency voltage or current measurements on the device, and relies on average current and voltage measurements external to the device. It is also experimentally demonstrated that e-GaN HEMTs with a monolithically-integrated Miller clamp offer significant shoot-through immunity as opposed to conventional solutions.
