So you wanna short a big battery...

[tomv]

Yeay! It looks like I have a small contribution for the failed components thread. I'll get the pic soon. But now I'm really puzzled about what happened :|

So... I thought, what could be more fun than shorting a big fully charged battery The battery is 12Ah, 12V SLA B&B SLA. Spec says it's 18 mOhm. Capable of 180Amps for 5 sec....

So I hook it up to 2 Ohm 20W resistor, IRL3103 FET, 12 awg wires. My Arduino is pulsing the FET and logging the voltage.
- First test 200ms on time for FET. No noticable drop in voltage, resistor gets warm. Current should be arround 6A
- Second test. Lower resistance to 0.5 ohm. 50ms on time for FET. Resistor gets warm. Voltage drop 1.5V. Estimated current 20A or so.. First stupid mistake. I didn't check if FET was getting hot for 20A test. I have no idea now.

- Third test. Short out resistor with 1ft long, 30awg jumper wire. I had uneasy feeling about this. But this is the prize. That's what I wanted to see!

That's where it all went wrong. I lower pulse length to 20ms, I hook it all up, Arduino sends a pulse, smoke, sparks flying! From the FET!! I was watching the jumper wire and expecting that to melt first. So post mortem is:

- I got the voltage drop recorded. It went down 3.3V, estimated current 60A or so.
- FETs fail by shorting out! Current kept on flowing until i unplugged it. Version 2 will have slow blow fuse..

Why did the FET cook? Data sheet says 12mOhm, 64A continous, 220A pulse.
- Is it because gate voltage was 5V not 10V? But this is logic level FET and is supposed to be fully on at 5V. It looks that way from the chart in the spec.
- Could it be because of gate capacitance? I have no idea how long it took to turn on.
- Because I had no heat sink?

What do I do make this work? What's the right way to drive FET gate to make sure it turns on quickly and fully?

IRF1324S just came in the mail too... I'm willing to risk this precious jewel too if that's what it takes... 0.8 mOhms, 430A continous, 1640A pulsed

 

[lazarus2405]

- Because I had no heat sink?

I'd wager good money on that. Screw some aluminium to it...

As for the 1' of 30awg, that sounds like instant insulation burn. You'd have about the same resistance with 10 feet of 20AWG, but I'm sure that would melt the insulation at >50a too. Still, it wouldn't glow.... Unless you're trying to make a toaster, too.

 

[fechter]

With 5v on the gate, the FET may not be fully on, in which case the resistance might be way higher than the rated on resistance, causing the FET to overdissipate. You should drive the gate with 10-15v to make sure it is fully on.

 

[tomv]

I'm sort of making a toaster . Battery + switch could be

- Resistive soldering iron. This is in between spot welder and regular soldering iron. It needs less power than welder, but would spot heat quickly. So soldering on large metal heat sensitive things (batteries) becomes possible. Commercial ones exist but are rare for some reason. I'm thinking this could be nice to solder small SMD parts with solder paste.

- Who knows, maybe even spot welder will be possible. I'm thinking of doing a hybrid with capacitors. Combine battery (200Amps 20ms pulse - approx 50 Ws energy) to charge capacitors. PWM capacitors into the welding joint at say 5% duty cycle. It output impedance is 1mOhm most of the heat would go into the joint. At current of 8000 Amps . Capacitors could be small and cheap and whole thing could be hand-held. In theory

 

[tomv]

Before doing v2 of this I decided to simulate it just in case... Both of your guesses are correct, and my guess that gate takes too long to turn on is wrong.

- Gate dissipates 270W when fully on with Vgs = 5V. At Vgs = 10V it would be 210W
- The really relevant stuff is thermal resistance. Rth without heatsink = 62C/W, with 2 C/W (+resistance of heatsink).
So steady state power that this can handle at 25C room temp, 125C case temp is 2W without heatsink, 50W with good heatsink. That pretty much explains it.
- Changing R1 (Rgate) does have effect on switching time, but even with 500R it's in microsecond range.

 

[deardancer3]

.. I thought, what could be more fun than shorting a big fully charged battery The battery is 12Ah, 12V SLA B&B SLA. Spec says it's 18 mOhm. Capable of 180Amps for 5 sec....
[/b]

Dont know what to say here, Maybe the people I hang out with are just a little too boring. Dont know if Tomv is a better class of person or, one to be avoided.

Party on Garth.

dick