DRV8350: Low voltage on the VGLS regulator

Hello Ti,

I have a custom board with the DRV8350H, developed for a few years. Now I wanted to use it to drive an e-foil which uses an inrunner brushless motor.
It works well at 60 amperes but when I increased the current to 80 amperes the gate driver got damaged (50V supply voltage). It also worked with 25V at 100A.

On another board I’m using the DRV8301 with 50V and 100A load, this works without any problems where the gate traces are much longer.

It's the hardware version of the DRV8350 chip. The resistances of the gates from the DRV looks ok, I compared it with a new device. The Mosfet is also good.

When enabling the device, it gets hot and what I can measure is that the VGLS is only about 3V and VCP is about 23V,
the supply voltage while I was measuring this was 20V. DVDD is at 5V, should be ok. VCP both sides measured to ground is about 1.33V and 22.3V. There is no voltage on all GLx AND GHx ouptus.

Do you have any idea why I damaged the VGLS regulator? I wanted to develop a new pcb and the DRV8350 would be nice for this, but it would be also nice to know how to prevent this on the new board.

The IDrive it set to lowest (50mA/100mA). This is the Mosfet I’m using, mti85w100gc-smd. (https://www.mouser.at/datasheet/2/240/MTI85W100GC-1549448.pdf)

Trise = 18nC/50mA = 360ns, Tfall = 18nC/100mA = 180ns. Fpwm = 20kHz.   

I might try putting some extra caps on the gates to further slow down the switching. But I’m not sure if this could help because the gate output resistances look good of the damaged device.

The Layout is maybe not the best, it’s only two layers and as I developed it I hadn’t much experience, but because of the package of the Mosfet I think it could work with higher currents.

Before increasing the current to 80A I soldered some TVS diodes(SMBJ14A-13-F) to the gates, this is the layout:



Here are the schematics, DRV, Mosfet, Snubber, Bulk:

, maybe to much caps

Picture of added diodes an bottom view:


Thanks in advance,
Daniel