I'm using a repurposed ATX power supply for my hobby projects since it's got 3.3/±5/±12 outputs, all of which are really convenient. But one thing I didn't really think about until I slipped my probes across the pins of an opamp, since I've always dealt with commercial/proper lab power supplies in my school labs, was that an ATX power supply will gladly deliver lots of current if that 12V line (or any other) is shorted to ground. The poor LM318 didn't stand a chance. My meter and supply survived, but in the interest of not killing anything in the future, myself included, I was wondering what the best option was for short/overcurrent protection?

I was thinking of sticking some high wattage resistors at the output of the supply before connecting them to the rails of my project breadboard (I use a separate breadboard with terminal blocks for power, which the ATX supply connects to). The problem is if I'm drawing a lot of current (LEDs etc) this will sag the voltage on the line. And, for example, if I use a 1W 200ohm resistor on the 12V line, it limits my current to only 60mA - if I want more, I need some really beefy wattage resistors. I can probably work around the voltage sagging by using a voltage regulator (eg at 10V), but this all doesn't seem like the best way to go about doing things.

I'd appreciate some input from someone more experienced than I am.

  • \$\begingroup\$ Something's wrong with the supply. The normal operation of ATX power supplies is to immediately shut down in case of a short circuit. \$\endgroup\$ – AndrejaKo Aug 12 '12 at 14:11
  • \$\begingroup\$ It did trip, but not before a nice spark. I'm trying to avoid all of that in the first place. \$\endgroup\$ – BB ON Aug 12 '12 at 14:14
  • 1
    \$\begingroup\$ So... how much current do you need? \$\endgroup\$ – W5VO Aug 12 '12 at 15:01
  • \$\begingroup\$ I'd like 1A on each rail so I never have to worry about hitting a ceiling, but something on the order of 500mA would also work. \$\endgroup\$ – BB ON Aug 12 '12 at 15:29

Use a fuse. You can buy e.g. PTC resettable fuses that will limit the high current and will automatically reset after some time.

  • \$\begingroup\$ Wow, I've never even heard of PTCs before. They sound like something I could use, but since they're thermistors, wouldn't they be too slow? The ones I can find say they take at least a second to trip. \$\endgroup\$ – BB ON Aug 12 '12 at 15:49
  • \$\begingroup\$ Depends on the current level. I can see 20ms trip time for 200mA fuse and 300ms for 1A. Sacrifice one LM318 to test if these times will be enough. If you will need faster times then I think you should look for some active over-current protection e.g. FPF2701MX. \$\endgroup\$ – Szymon Bęczkowski Aug 12 '12 at 16:32
  • \$\begingroup\$ Here is the links for PTC Resetabale Fuses sparkfun.com/products/8357 \$\endgroup\$ – masterleous Aug 12 '12 at 16:36
  • \$\begingroup\$ Or a faster acting PTC which fits nicely with the answer below....PTC NANOSMDC050F \$\endgroup\$ – rdivilbiss Aug 13 '12 at 2:39

The 'big spark' is hard to limit if the output of the ATX supply has lots of capacitance. There isn't a protection circuit fast enough to save you from that sort of instantaneous energy.

You may want to consider using the 12V rail to feed some buck converters, generating your own 5V and 3.3V rails. The bucks will have their own easily adjustable shutdown thresholds and provide you some measure of 'protection' from the stiffness of the 12V rail.

For the 12, you may want to consider a LDO regulator instead of power resistors. It will drop less power plus give you some measure of overload protection that resistors would not.

  • \$\begingroup\$ Having opened a few, even cheap ones have a 1000uF cap on the output of their 12V line. \$\endgroup\$ – Fizz Aug 30 '15 at 20:06

Use a low voltage MOV across the supply rails to protect the LM318. The resettable fuse (or PTC) goes between the supply and the MOV, to protect everything from the excessive current when the MOV turns on.

The MOVs are good parts but the specifications are a little hard to understand. They vary with temperature and have a tolerance. Use a higher voltage in order to prevent the MOV from turning on at too low a voltage.


I suggest checking this guide (in Italian) that allows you to not only limit current to arbitrary values, but adjust also the voltage, all using the original chip provided in the ATX PSU. A 350W PSU becomes a 0.15-20V 0.1-16A lab supply (of course, you don't get 20V and 16A at the same time...).

You will lose all positive power outputs except for one (the chip can regulate only one at time), but that one will be very accurate. Given the cost of used PSUs and the cost of this little mod, simply mod three of them.

The mod benefits from a small daughter board (easy to build) to avoid too much mess inside the PSU.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.