Anyone here ever bench test or repair an ATX power supply?

[zzyzzx]

I was given a crappy old Emachines 533i (Intel Celery processor) that I wanted to experiment with (i.e. Linux box, of something) but I can't get the thing to start. So I removed the PS, and tried jumpering the green lead on the MB connection and I do get the PS cooling fan to rotate sum, but it's very slow and the voltage output fluctuates. I tried this with and without a load and got the same results.

Never having done this on a known good power supply, I can't say 100% for sure what a good one would do. But I'm pretty sure it's bad.

Anyway, I take the thing apart and there is a charred section on the Power Supply board, indicating either a small diode or semiconductor (or something, I forget the exact packaging style name but it remonds me of a commn votaeg regulator) is bad, but it looks like this:




Anyway, are these things fixable? Anyone here sucessfully fix an ATX power supply? I've done it with a AT one before, but mostly because of blind luck. I don't have another one I can swap into theis old eMachines for disgnostic purposes, and I'm really more likely to remove the HD and toss the thing before I'd but a new PS for it.

[redwudz]

The one pictured appears to be a -5VDC 500MA voltage regulator. But it's representative of other 3 pin VRs. You could change it easily enough, but I'm thinking it blew from either a dead short or a out of spec input voltage. Either way, it's more likely not the cause, but the effect of a problem. E-machine PSs have a bad reputation for dying. I would be more concerned with what it may have taken with it when it died. (Worst case, MB, CPU, RAM, all drives. }

You really need to try a substitute known good PS to see if the motherboard or other parts are fried.

If you jumpered the the power-on lead to ground, you should have gotten operation. That's usually pin 14.



However, most ATX PSs won't operate properly without a load on some of the rails, most likely the +12VDC and the +5VDC. Without a load, the regulation may not be very good.

[minidv2dvd]

no. like a lot of electronics it's not worth the time spent diagnosing and repairing them. and if you did repair one voltage section, you'd go on to find that the others were weak anyway. e-machines probably bought it for a dollar wholesale in china.

you can get a new 300-400 watt unit for around $20 at newegg.com

[bendixG15]

You can jump start the unit by connecting the green and black as shown here ... http://www.overclock.net/faqs/96712-how-jump-start-power-supply-psu.html

[Nelson37]

Attempting to repair a power supply in general is pretty much a bad idea.

Doing this with an E-machines power supply is pure foolishness.

Spending time, effort, and energy on a $25.00 part that might very well endanger $300.00 worth of hardware is something I have a hard time getting a handle on as a general concept.

[dannyboy48888]

i concur , go to the store and get a new 300w for 30.00 and consider it a insruance policy at the same time. even if you order that part from digikey you are looking at s&h and a week wait. beside e-machines! burn it and move on.

[zzyzzx]

Spending time, effort, and energy on a $25.00 part that might very well endanger $300.00 worth of hardware is something I have a hard time getting a handle on as a general concept.

Wouldn't it be more like Spending time, effort, and energy on a $20.00 part that might very well endanger $0 worth of hardware...