BQ40Z80: DSG/CHG pin series resistor limitations

Eyal Roesnthal

Part Number: BQ40Z80

Hello,

For my BMS design, I've used 10k ohm resistors for the CHG/DSG gate drive pins as mentioned in the BQ40Z80 datasheet. I've also used BSO033N03MS as my protection MOSFETs. I've only recently discovered that under extreme short circuit conditions (~1000 amps) the MOSFET seems to burn off before the BQ40Z80 can respond.

In my next iteration I'll probably use stronger MOSFETs, but for now I'm wondering if by lowering the 10k gate resistor I could maybe reduce the short circuit response time. In the EVM schematic TI used 1k and 4k resistors, and I can see in the pin equivalent circuit that there is an internal 2k resistor at the pin output as well.

So my question is how low of a gate resistor can I use in series to the CHG/DSG pin? I assume it's depended on the pin's max sink current but I'm not sure what is the max allowed from the datasheet.

I'm using the BMS mostly for 3S-6S Li-Ion applications.

over 5 years ago

0 Frank Boeh83 over 5 years ago

Intellectual 585 points

I've been implementing a few high current MOSFET switches (up to 2000A) in the past and noticed one general caveat: at reasonable switching speeds (e.g. 0 to two-digit microsecond numbers) it often doesn't matter much how fast you actually switch. In many cases it is mostly inductively stored energy that is dumped entirely into the MOSFET's silicon die (for example, a 1 meter loop carrying 2000A contains about 4 Joules of energy), which heats it up instantly without any possibility of the heat to spread into its surroundings. It literally boils down to the amount of silicon, and its corresponding thermal capacity, in the transistor.

The single pulse avalache energy rating of a MOSFET is a good indicator of the die size, but TVS diodes in parallel to it will also help as they are purposely built using physically large dies.

0 Eyal Roesnthal over 5 years ago in reply to Frank Boeh83

Prodigy 70 points

Thanks for the informative reply.

It's unfortunate though if I wouldn't be able to fix it via the switching constant. I have a fairly large quantity of manufactred BMS and it would be a while before I could work on a new revision. However, the short circuit test which caused the MOSFETs to burn was quite extreme and very unreasonable to happen in practical situations, so maybe it is still acceptable.

Still, I would like to try and see if changing the gate resistor would have any effect so my original query still remains.