Tell: Cap Charging with ATX PS

Discussion in 'General Electronics Chat' started by rektide, Jul 25, 2011.

  1. rektide

    Thread Starter New Member

    Jul 24, 2011
    2
    0
    Greetings, thanks for the forum space & thanks for reading,

    I'm looking to build a discharge tab welder (largely based off a very long thread at Endless Sphere). I'm dealing with the MOSFET discharge circuitry now, but I have a question I don't know how to go about researching with regards to powering the capacitor bank that this discharge side will interface with.

    I plan on using a pair of very powerful ATX power supplies I've ganged together to make a 24V power supply to charge the capacitor bank. The power supplies, Cooler Master RS800s, have a host of very fancy overload protection acronyms in their spec sheet. I have no idea what to expect when I put them in front of a decent sized capacitor bank though. I take it that an ATX power supply is more likely to shut itself down than it is to try to provide it's maximum constant current? I have audio amplifiers that will turn into rather serious current sources if placed in front of a "short", and will keep at it basically indefinitely; should I assume otherwise for an ATX power supply?

    I'd like to get a feel for what to expect now. For now, I'm going the "build first, test latter" route. There are two charge and discharge capacitors used to charge and manually discharge the capacitor bank via PWM, so I can start with a trickle charge and build up. Hopefully the power supplies don't mind very very brief shorts, and I can keep ramping up until something starts going badly.

    I suppose I could investigate current limiting resistors, but a hundred amp charge capacity would mean a .24 ohm resistor, and i^2 r power dissipation on that resistor would mean I need a 2.4kW resistor. :/

    Texaspyro mentioned his cap bank has a 0.3 milli Ohm ESR, although mine will be higher (not shooting for 3 in parallel). We're both using Rockford Fosgate RFC1's, which are 1F 16V.
     
  2. rektide

    Thread Starter New Member

    Jul 24, 2011
    2
    0
    Ah, one of the example units on endless sphere, to give an idea:

    http://endless-sphere.com/forums/viewtopic.php?f=2&t=2633&start=582

    Why did he use P-channel gates for charging? every other example I've seen uses the same N-channels for charging/discharging the caps as for the discharging welding? Any particular reason to go P-channel here?
     
  3. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    P channel devices are more expensive, harder to find, have higher on-resistance and higher gate charge. The reason they were probably used despite these limitations is because it gives more safety. A P-channel can easily be used on the high side of a power supply rail. The cap bank can't discharge without positive onto anything. Shorting negative to the case would do nothing. But an N-channel solution, unless a bit more complex, would probably have a ground switch. That way the positive would always be connected and the negative would be floating, not pretty if the positive touches anything earthed like the PSU's case!

    A power supply will likely shut down if you attach a large cap bank to it as the overcurrent protection will probably trip. One solution is a *very* high power NTC thermistor which will limit charging current, or switching the capacitors in slowly, so the load is ramped up gently.
     
  4. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    These resistors are usually quite big and it takes some times for them to heat up. It depends also on how often you charge the caps. If you use a resistor the current will be high if voltage is applied to a discharged cap, sure, but the current decreases rapidly while the cap is charging.

    We had some 300W 0.1 resistors at my company, we regularily passed a lot more peak currents through them than their rating would allow and there wasn't any problem. When they were used as constant load, a fan was blowing the generated heat away.
     
Loading...