I guess it quit typical situation involving inductance in the drain path resulting in transient voltage spike on the drain during turn-off.

Thank you very much Nikolay Pavlov for answering this question. Yes, stray inductance will have impact on turn-off loss, but what strange is the therotical result shows Eoff>Eon (basic data for therotical calculation are from manufacture's datasheet). What's more, I didn't add any Lloop to LTSPICE simulation, the result also shows Eoff>Eon.

Yes, the turn off losses of mosfet is greater than turn on losses, because back emf is produce only in turn off condition. The voltage spike due to back emf is multiplied with drain current to increase the Eoff greater than Eon. To reduce the Eoff lower than the Eon, try to add snubber circuit in parallel with drain and source

Thank you very much Afaq Hussain for answering this question. Back emf is an important issue in motor application. In this case, hard switching condition is a double pulse testing circuit (clamped inductive circuit) and the switching device is a advanced packaged GaN HEMT where the stray inductance is small and so does to Lloop.

Usually a MOSFET has less turning off loss than turning on loss unless you operate the switch at very low turning off speed. May suggest you to check again the gate resistance you used and meanwhile you may check once again the loss caculation approach both for turning on and turnin off. FYI.

Thank you very much for answering this question Alpha J. Zhang. The gate resistance is 10 Ohms (may be a possible reason) and the loss calculation method refers to a commonly used method i.e. Eon (10% Vds - 10% Id) and Eoff (10% Id - 10% Vds) in this application.

it looks the loss calculation problem. the turning on loss is more comlicated which includes the volt-current overlapping loss and as well as reverse recovery caused volt-current overlapping loss (uaually much bigger than previous) and as well as Coss discharging loss (which is quite big at higher switching operation).

Hi Minghai Dong,

For hard switching condition, the high voltage surges occurred during the turn-off behavior, due to the presence of stray inductance (Inductive load).

In that case yes, the turn-off loss may be greater than turn-on loss in case of power MOSFET.

For another power device like IGBT also, (IGBT takes longer time to turn-off as compared to MOSFET) the one can see the same behavior.

Best wishes.

Yes, Alpha J. Zhang Eoss is very hard to estimated and would enlarge Eoff, but the strange thing is therotical results also shows Eoff>Eon (Coss is obtained from manufacture's datasheet).

Mahesh Vaidya Thank you very much for discussion. The stray inductance is a critical factor in loss estimation. In this application, the power device is a GaN HEMT where stray inductance is only 0.1 nH, smaller than that of conventional power devices.