Why am I killing NFET on low side of motor drive?

Glenn Monroe · ‎2020-03-09

I am using an STS10N3LH5 on the low side of a half bridge circuit. I have a schottky diode across the motor. For lighting applications we PWM the ground and everything is fine. However, during motor applications we PWM the high side and will sometimes blow the low-side FET. The circuit looks like the following:

The two flyback diodes are so we can try to separate parts during prototyping. Both are capable of the current we are driving the motor with.

There are two possible scenarios. During normal operation the ground is left connected while the high side PWMs. This should pin the low side to ground and sink any current generated by the inductance when the high side is off. The other scenario is when we turn everything off. Currently we turn high and low sides off simultaneously. This leaves the motor generating current until it slows down. However, the schottky has a Vf of about 500mV and the motor has a non-zero resistance in its windings. I would expect this to dissipate any residual current without issue. However, we keep blowing up the low-side FET.

Any ideas would be greatly appreciated. I have looked using the scope and do not see anything that appears high enough voltage to damage the FET.

Thanks

Ozone · ‎2020-03-09

FETs can be killed my slew rate (dV/dt) as well.

In difference to HV FETs, components below 100V Vdss do usually not specify a maximal value.

Perhaps this appnote helps: http://www.aosmd.com/res/application_notes/mosfets/dvdt_app_note.pdf

BTW, I hope you ar aware the STS10N3LH5 is marked "NRND".

Glenn Monroe · ‎2020-03-11

I was not aware, but thank you. We will change it out in a future rev.

I am only switching about 14V with this FET right now, information you did not already have. According to that article it shouldn't be a problem.

Thanks

uomo motore · ‎2020-03-23

Have you found a solution to your problem?

You can try adding a delay of around 800-1000nsec between turning off the high side PWM and turning off the low side switch? (High side PWM off>dead time>Low side switch off)

Also turn on low side switch of next step sequence first, implement similar delay before first high side PWM generation. (Low side switch on>dead time>High side PWM on)