#12, ok. Just wanted to make sure you weren't trying to convey something to me that I wasn't getting.

cmartinez: There was 20-40 amps flowing through a cable to the loose bolt. The looseness caused increased resistance, and thus heat. Lots of it. It was finger tight onto a lock washer so very little actual metal contact was occurring.

I've built 4 different variations of a positive gate driver. I'll have test results tomorrow. I'll post the design that is getting the best actual rise/fall times as measured, not simulated.

One point that I would add to this is that the battery voltage coming in is dirty. I've seen spiking up to 20V+ above the nominal voltage. So, I went with products that had higher maximum voltages. Anyway, once I see the results, I'll post the designs for comment.

I appreciate the feedback guys.

 

@mikekehrli

Here is a simple solution. Use the inboard comparitor or ADC to measure the supply voltage (left) and supply power to the appropriate NPN transistor to power the P-MOsFET with about 80 mA current to turn on your MOSFET very fast vs your existing design. Off speed will be a bit slower but still significantly faster than your existing design. Make sure to use 1 Watt Zener.

 

Thank you. I've designed, built and tested 4 designs. Attached is the winner. It had the best rise and fall times, and I've run this design on the bench at 50 amps all day long, and the heat sink and fan remove the heat just fine.

I used an adjustable negative voltage regulator to create a floating "ground" for my mosfet driver. I set the "ground" to be nominal 12 volts below batt+ (mosfet+ in my schematic). This allows a steeper off slope to my off voltage than if I had used 10V, same as my zener.

I really haven't seen many good p-channel mosfet driver circuits when googling. This seems a bit odd. Almost everything looked like a variation of my first design above. Very inefficient. The one I have now is at least 20 times faster than the original, which is vital for high amp applications.

I appreciate all the ideas I got from this forum. I invite your comments on the design:



Addendum: I'm running the unit on the bench now at 80A @ 26.5 volts, and it is not getting hot.