whay does the cap. in my PWM get so hot

[Helz_McFugly]

the cap on my CCPWM gets pretty hot after about 15 minutes. any idea why? its set to limit 13 volts and the duty lets it run at about 20 amps. the freq. is just set in the middle because I have no idea how to set it. I know if I have a bad connection because i get a high pitch sound and I can change the sound with the freq adjustment. Im getting a good meter in the mail today that may cheack freq.

[Philldpapill]

The cap is in there to absorb the inductive "kick" when the PWM turns off. I looked the part number up on Mouser, and it's an "ok" capacitor, but under rated. The reason that it gets so hot, is that when the PWM turns off, large currents slam into it, then when the PWM turns back on, large currents gush out of it and into your cell. Repeat this process thousands of times per second, and you've got a hot capacitor.

Remember, everything has SOME resistance... even the plates in a capacitor, so the power dissipated as heat is a function of the current in and out of the capacitor(squared) times the internal resistance of the cap. My "Hulk" is going to have a good bit of capacitors. In fact, it's got half the board devoted to them so that it runs VERY cool.

[Helz_McFugly]

really looking forward to it. my CCPWM wont got up in volts. it tops out at 12.7 but it will go up to 30 amps. if I go brute force, the cell will pull 14 volts 30 amps. any idea on that one?

[Helz_McFugly]

Hey Shane, youre the one that built this PWM :p Its the ZFF CCPWM design with 2 mosfets. everything is 8 guage. You made this PWM for an E-lyzer that was 9x9 and drew 45 to 60 amps.

[Helz_McFugly]

and the e-lyzer is now 9x6 7 plates.

[Helz_McFugly]

cool, just so happens I may enough 4 gauge wire to do my whole system because thats what I used on my first wet e-lyzer. I didn't think I needed it because it wasn't getting hot so I went down to 8 when I rebuilt it because it was cheaper and easier to work with. I still doesn't get hot with the 8, warm at the e-lyzer connects but thats it. I just need some 4 gauge end connectors. Thanks man.

[Philldpapill]

Errr, this isn't MY design, so take what I have with a grain of salt... BUT, 8AWG should be MORE than good enough for this application. 10AWG is good for 30A with approx 1mOhm of resistance per foot, and only dissipates less than a watt per foot(barely noticeable let alone an issue).

Again, I may be totally wrong because I don't know the exact details, but one of the reasons the capacitor is in there, is to counteract the inductance of the wiring. Replacing the wiring with a larger gauge could actually make the problem worse as a larger wire gives rise to slightly higher inductance values... I think the problem may lie elsewhere.

Just a thought... But take Shanes advice over mine and try his idea first. I'm just speculating with theory and my own experience in high-power design.

[Helz_McFugly]

Biggy Boy told me ZFF says how to fix this. theres a wire that runs from the negitive terminal to the right terminal of each mosfet, its about 5 inches long, it acts as a resister I think. If i shorten it an inch or so it will let more current through, no need to rewire my system. I just need to shorten the two wires that go from the negitive terminal to the right terminal on each mosfet.

[Philldpapill]

Do you have a decent multimeter, Helz? You can get one that measures frequency/duty cycle for about $20 off ebay. If the duty cycle maxes out at like 50%, then that's definately it. Anyway, a good meter would be a nice thing to have when messing with this stuf.

[Helz_McFugly]

I just got this one in today http://grizzly.com/products/H0473
its got everything on it. have a look. I think it will work. it gives a good reading on amps with the clamp. I love it.

[Helz_McFugly]

ive only used the amp and volt meter on it so far. whats the difference in the 300 and the 1000 setting for reading dc amps? Ive been useing the 1000 BTW

[Philldpapill]

Precision and resolution. The 1000A probably isn't as accurate for lower currents(<100A) than the 300A range.

The other thing, is that when using a square wave PWM, some meters can be misleading. If they don't measure "true RMS" values, then chances are, you'll have alot of measurement jitter and flat out errors because the meter will take samples (seemingly) randomly as the current/voltage switches on and off. Sometimes it will get a measurement when the current is off, then another measurement when it is full ON, and it may screw with the readout.

[Tundra4x4]

make sure there isn't any current jumping across anywhere aka...make sure there are all good connections. I have had this problem and a good infrared temperture gauge works well to find the hottest point and this is usually the point of bad connection.

[Helz_McFugly]

yea I did have one bad connection at the neg connect at the e-lyzer and it gave off a loud high pitch noise that I could change with the freq adjustment on the PWM. I soldered the connection and it went away as did the heat that was there wich wasnt that hot, now everything runs great. the freq is great for finding the bad connections.