Hi there,

I've put together a circuit to drive a P-Channel MOSFET from an Arduino Uno. My first mistake is that I got the PCB upside down and all my surface mount components face the Uno, making troubleshooting difficult. My second mistake I cannot identify, but it involved smoke that wasn't related to a BBQ or fireplace.

I've inserted the circuit below (the section I was testing and smoking).
The top connector, J7, had a 12V sealed lead acid battery connected.
The lower connector, J8, had my benchtop power supply connected.
Line 3 coming in from the left is an Auduino digital pin. I set it low to turn on the FET, high to turn off the FET.
Capacitor C9 on the solar line was not installed, as the capacitors I have are all 16V (no good for a 12V nominal solar panel).

When I increased the voltage on my benchtop power supply to 17V (current limit at 2.2A), a lot of smoke started coming out.

Once I'd isolated all power and confirmed no active flame, it became apparent that the Si8261BAD-C-IS isolated gate driver was the victim (U39 in image). The driver side has the case rupture, the signal side looks ok still (+1 to isolation).

I can't for the life of me work out why the gate driver fried. Datasheet says that VDD can go up to 30V. The FQP7P06 shouldn't take much if any gate current (they don't do they?), so the low output current of the gate driver shouldn't matter surely?



Please don't judge my soldering, I'm way out of practice.
The bottom connector in the photo (2 pins, connector on other side) is teh battery. The top one is the solar input.
The trace from solar going elsewhere is simply a potential divider into the ADC.
The battery trace going elsewhere is simply driving an 8V linear regulator and a potential divider into the ADC. The battery trace also drives another FET outside the image for load control, but I haven't tested this yet.


If anyone can see anything obvious, I'd really appreciate the assistance.

Cheers,
Andrew

 

Capacitor C9 on the solar line was not installed

The minimum value specified is 1uF. Without that it is possible the IC was oscillating at high frequency, resulting in over-heating and its demise.
Your schematic shows the IC's NC pin connected to ground. That could be a no-no if that IC pin has an internal connection but must not be connected externally.
The pic seems to show pins 2 and 3 shorted by the pcb copper?

 

The manufacturer recommends connecting NC pin to ground (plane).

 

That's fine then.

 

The manufacturer recommends connecting NC pin to ground (plane).

Which ground plane? The input side or the output side? I'd suspect that using the output side would break the
minimum clearance distance...

 

Apologies for lack of reply, I placed extra parts on order and moved onto other things.

I loaded a new gate driver this morning and added the 1uF capacitor as recommended and now all appears to be operating fine. Thank you Alec for your clear and concise response!

I never even considered a bypass capacitor, trap for rookies I guess.


To answer the other concerns in replies:
Definitely no shorting between pins, the lighting may have interfered there.
The input and output ground planes are joined by a 0ohm resistor (if you can still call it a resistor). I just kept the 12V ground section separate on this design after a lot of problems on project attempt #1.
I'm not too worried about isolation clearances and creepage distances, the only isolation I really need is to be sure that the nominal 12V solar and load switching doesn't find it's way back to my control pins. Clearances for 1500V compliance aren't required and only being a home project...