Converting a computer PSU to a bench power supply

Question

)I have read a number of articles dealing with the conversion process and think I can manage to get there if I am careful and take it step by step.

However my starting point is a box of PSUs that I picked up all over the place (from my old PCs, stuff discarded at work (when we moved to new premises) and so on). I don't know what condition these units are in or if they even work.

So my question is, before I go through all the hassle of drilling, soldering etc to make the conversion, how can I know for sure that the PSU candidate is in good working condition? I have read that these units have two power on leads that need to be shorted and that they should have a load across the +5v lead and ground, so if I do both of these things and then check the voltages in the other cables (as per colours) with my multimeter would that be enough?

Even so, are there any obvious indications that a PSU is starting to fail (apart maybe from a noisy fan)?

Thanks.

Answer 1

I'd start by shorting the green and black wire on your 20/24 volt connector. If the fan turns on 'it works'. If it doesn't, throw it out. Nice thing there is all you need is a paperclip and a powersource.

If the fan doesn't turn on, don't even bother, its not worth salvaging. Its a good idea to have a load for long term use, or to wire up a switch between them, but for spot testing, without a PSU tester (which would be a nice investment here, and not too expensive), its the way to do an immediate spot check. You can then check voltages. Unfortunately most PSU failures tend to be sudden, I'd probably consider opening them up and spot checking the chips and capasitors inside, then closing them back up. If any PSU seems dodgy or dangerous (see an example below), throw it out.

I'd recommend not opening up, drilling and otherwise directly modifying a cheap PSU. They have quirks - I've had a PSU that had a heatsink that was live at 240V separated from the case by a blob of silicone putty. Its quite frankly not worth the risk. I didn't die from a 240V shock, but it hurt quite a bit. I had a small burn on my finger and a sore arm for three days. I'd recommend NOT using any PSU that has a blob of silicone between the case and a heatsink as a result!

I wouldn't bother changing the fans (which are often simple 2 wire units), if you have sufficient spares.

It makes more sense to either work from the wires coming out of the PSU or better yet, get a 24 pin molex female (perhaps salvaged from a dead motherboard) and do all the connections to that. As a bonus, if your PSU dies, you simply unplug the PSU from the female connector, and attach another PSU

(for reference, I confirmed the colours of the wires you need here, and the need for load resistors here I usually used the old PSUs to power something so I never needed the latter but this may be of interest for someone building a benchtop power supply)

Answer 2

yes. Potentially you can also switch the multimeter to AC voltages to check how clean the DC power is, which is not an ocilliscope and would be completly useless information if you did not have other stuff to compare to , and a load placed on the voltages. Seeing on a cheap multimeter what the DV voltage shows when switched to AC is a cheap trick that can provide minor clues to how well the cleanup curcuits are working.

Open them up prior to plugging in, and clean out the dust, check for charred components , loose heatsinking, useless thermal compounds, bulging capacitors both top and bottom. Inspect the bottom of the boards (quickly) for any charring or odd substances. Any time you open them up, you have to watch out for charges on high voltage caps, and double insure that everything is unplugged and all each time, for all that electrical safety.

I would say just immediatly replace the fan of any one you pick to use, but that is probably because I like to change them often to "better" fan items like ball bearing style, or for my prefered thermal vs RPM desire at the time. If they do not have one, adding a fan with its own thermal control could provide it thermal (and sound) control, that might improve it.

There is not much that I can think to add to what you already know. Some PSUs will turn back off again if no (even minor) load is on the 5V line. A wire is set to low (grounded to the black) to turn them on. Old ones even properly working will have aged capacitors, and could have reduced Clean output.
Many PSUs (cheaper) will vary the different voltage lines slightly when the load on the Other line changes, so it still is not a lab instrument, and you may notice those quirks when loading only one voltage down a lot.

The last thing to mention would be that there are now loads and testers that would make dealing with a lot of these slightly easier. That purchacing loads and testers that can only do so much anyway, is a lot cheaper than they were. Cheaper both ways of course :-) but cheap enough to have one hand to do what they can.

Answer 3

IMO this belongs in electronics.se

That having been said, I am going to disagree with everyone here. Yes, the pass/fail tests that have been described here are good pass/fail tests. However, in your situation you have a great opportunity for strategic thinking. Why not use it to plan out your workbench?

First, read some basic electronics text. A great place to start is Horowitz & Hill, The Art of Electronics, but there are many options on Amazon.

Then, come up with tests for the basic components you will likely find (diodes, caps, transformers/coils, etc). Hint; if it looks burnt, probably no point testing it.

Finally, go online (or to your local hardware store) and find a nice bits bin with enough compartments for all the components. Having done that, proceed to tear apart all the supplies except the one you think will best serve your needs, and then follow whatever instructions you like to build a bench supply. Now you have a nice parts collection, improved solder skills, and a working supply. Productive, no?