Calculation of turn-off loss in ZVS

gjw

Hi team,

I trying to calculate the turn-off loss of the MOSFET in a Four-Switch Buck-Boost（FSBB）converter when it working in ZVS control. According to the test circuit in the data sheet, when measuring the turn-off loss, the measured current includes the current flowing into the output capacitor, so the measured Eoff given in the data sheet includes the actual Eoff of the MOSFET and the output capacitance energy Eoss. However, the energy of the output capacitor will be returned to the main circuit in ZVS, unlike the hard switch, which will be dissipated in the MOSFET, so the calculation of the shutdown loss under ZVS should be (Eoff（given in datasheet）-Eoss).Is this calculated correctly? if is correct，When I calculating the turnoff loss in the case of high voltage and low current, there is a case where Eoss is greater than Eoff, resulting in a negative value in my calculation, which is obviously incorrect, please tell me how to achieve the correct calculation.Taking the 800V10A scenario as an example, the Eoss obtained according to the calculation method in Figure 16 or Figure 18 is greater than 50uJ, but the Eoff in the datasheet(Figure 22) is obviously less than 50uJ.

I also used the PLECS model for simulation verification, but I found that the PLECS model uses a lookup table interpolation method for loss calculation, Eoff does not subtract Eoss as given in the datasheet, and the value is missing in the range of 0-8.8A, which I think may cause inaccuracies.

I would like to ask if the correct method of calculating the turn-off loss under ZVS conditions and whether the data sheet and PLCES model take Eoss into account, and if so, should I just use Eoff instead of Eoff-Eoss in the data sheet.

Thanks

GJW

Forum_Moderator

Thank you for your post, it has been approved and we will respond as soon as possible.

YuequanHu

Thanks for using our Power Applications Forum to raise your questions.

Yes, you are correct. In ZVS applications, the actual turn-off loss should not include  Eoss. Please note that the energy curves in the datasheet are curve-fitted based on the actual test data, which can result in a negative value when the difference of Eoff and Eoss is very small. What we can say is that the turn-off loss is negligible for your case.

Please keep in mind, Eoff in the datasheet highly depends on the gate resistance you are using. If you use RG = 10Ω for turn off, Eoff will be much higher. Please see below plots. Unfortunately we don’t have the curves for 800V/10A with RG = 10Ω for turn off in the datasheet for C3M0032120K.