I've got this high side P-Channel MOSFET driver circuit that Sgt. Wookie hooked me up with on another thread: http://forum.allaboutcircuits.com/showthread.php?t=52037

Tonight the 10uF, 25V cap I had on the breadboard for C2 exploded pretty good. Smoke, noise, even the cats ran out of the room .

The problem is I'm not sure why it went. The Panel V+ from the schematic was hooked to my bench supply which was set at 20V. The cap was rated for 25V. I put the cap there to quiet down some ringing (which was going up in the mid 30 volt range before the cap). However after the cap I never saw the voltage spike much at all.

So, do I just need to get a 50V cap for that C2? I was thinking of adding a 20 ohm resistor between panel V+ and the collector of Q1 to limit the current when the switching the gate on. When I had the ringing... where does the 35V come from? Is it like an boost converter just using the inductance that exists in the circuit?

Anyone have any experience with this kind of thing and how to spec C2?

 

You may need to place a diode between C2's positive side and the solar panel in.

I didn't think about it before, but when the MOSFET turns on, it's going to tend to discharge C2 pretty quickly; then when the MOSFET turns off, the solar panel will charge it back up. That means lots of current charging/discharging C2, so it's heating up inside.

With a diode added between C2's positive plate and the solar in, the solar cell can charge the cap, but the MOSFET won't discharge it, as it would be blocked by the new diode.

Use a low Vf diode, like a 1N5818 Schottky.

 

I tried that, and it does keep the cap from heating up (I have a big 160V 100uF in there now). However, the ringing is back. I just measured 45V peak at the gate. I thought the cap kept the ringing at bay by supplying current when the MOSFET turned on. The voltage wouldn't collapse right at that moment and then bounce back and produce the ringing.

 

The big cap has too high of an ESR.

Add a smaller cap or two in parallel with it; keeping the leads short.

 

Is this schematic what you meant regarding the diode and the C2? I can see how that reduces the current passing through C2 put I don't see how the cap could ever discharge.

I put a 5 Ohm resistor for R4 and the cap w/out D1 and got these pictures. To me the look like pretty clean switching, but with only 2 amps, the MOSFET heats up pretty good. Could I have damaged it and now it doesn't switch cleanly?

This is really hard. I'm only switching at 5Khz and 2 amps and I have thermal issues everywhere. Even D2 is heating up. D2 is an 80SQO45N . The max instantaneous Vf is .5V and should be around .3. So thats 150mOhm - 250mOhm resistance. I guess that's nearing a watt of power dissipation at 2 amps so maybe that heat is to be expected?

At any rate I can't seem to make MOSFET switch w/out cooking. I have a 1" bolt attached through the hole of the TO-220 tab, which isn't much of a heat sink but if it heats up in less than a minute I think something's wrong. If I run it wide open (100% duty cycle) never switching off, then it runs cool. No problem. So it leads me to think that either it's not switching cleanly or not switching on fully. The Vgs is 12V and the on/off switch looks pretty vertical, the thing is spending 99% of it's time either on or off. It should not be heating up.

 

You were supposed to keep the transistors' collector connected to the cap high side.
See the attached.

I don't know why you're trying to switch it at a 5kHz rate. You don't need to switch it that fast.
If you had an inductor between the MOSFET and the battery, then switching fast would make sense.