Leakage spikes on pulse transformer gate drive?

[cupoftea]

Hi
The attached LTspice sim shows pulse transformer gate drive of an LLC. We will instead use a pair of UCC27533’s in hi and lo side (as per attached pdf)

The leakage spikes on the output of the (hi side) gate driver look perilously high in the sim…..so the diodes were added and are shown in the lo side….the internal FETs of the UCC27533 incorporate these diodes but it seems a bit perilous to depend on them? Would prefer a more robust BJT based driver with the protection diodes, since BJTs are far more rugged than FETs.
Would you agree?

UCC27533 FET driver datasheet
https://www.ti.com/lit/ds/symlink/u...https%3A%2F%2Fwww.ti.com%2Fproduct%2FUCC27533

(LTspice sim and pdf schem attached)

 

[mtwieg]

The spikes in the simulation seem unrealistic for a driver based on MOS devices. Like you say, the FETs' body diodes ought to clamp the output voltage. But the model is a completely different part from the UCC27533, so why rely on it?

BJTs are far more rugged than FETs.
Would you agree?

No...?

 

[FvM]

It's a good example how you come to wrong conclusions by using a behavioural simulation model beyond its dedication. Quite obviously, the LTC1693 model uses ideal switches for the output stage instead of real MOSFETs. During deadtime the output can swing below and above the rails.