Fast(est) shoot-through protection

[kathmandu]

Don't see a difference. Usual "desaturation" circuits are covering both cases.

For a switching Mosfet (one of the h-bridge's switches), I have to continuously monitor a Vds in 0-48VDC range while for this steady conduction mode Mosfet the normal operation Vds is in 0-500mV range.

There are hall sensors with < 10 µs (even down to 1 or 2 µs) response time.

Do you think that using an analog comparator (op amp) to detect this low voltage peak value (500mV) won't be reliable/fast enough?

My actual LEM sensor has a response time of 7us but I thought it wasn't fast enough. Anyway, I might use a comparator to check it's output, too (beside that Vds monitoring).

If there's no significant downside, I'd like to start designing/building this electronic fuse.

 

[dick_freebird]

A Hall sensor that can sacrifice accuracy, can be fast.
Ones meant for high accuracy and noise rejection will
be slow (op amp in the loop).

My concern with the active approaches is, a failed FET
is perhaps not going to listen to your commands. I'd
suggest you walk through a poor-boy FMEA based on
your top few threat scenarios, what fails, and what
would prevent that from pushing through to the next
level (like, if HSS FET fails shorted, a commanded HS
gate = L might have no effect; then you need some
independent interrupter in the VIN leg).

 

[kathmandu]

That's my final option, indeed: an independent interrupter (electronic fuse).

After all, all the shoot-through/short-circuit conditions are only valid as long as there's a DC link voltage applied to that circuit.

Another advantage of an external e-fuse: it's much easier/reliable to control an independent low-side switch (moreover, in a steade state conduction mode) than four swithing Mosfets (high-side/low-side) with interlock circuits.