How to Test a Power Supply for Faults?

[jrgray93]

I'm wondering if there is a way for me to actually test a power supply for problems even if it powers up and appears functional.

I have a PC that came in for repair at work. Our PCs run our software to provide interfaces and controls for precision machining. Our customer said he experienced data loss on the hard drive twice, and between incidents, the PC stopped communicating with our drives. No matter what I tried, I could not get the PC to communicate with our drives or our office network. It only registered "limited or no connectivity," and wouldn't load web pages or detect network devices.

After thoroughly examining the PC and trying different setups and repairing the drivers, I can safely say this is a hardware issue. The motherboard would appear to be damaged. I'm also assuming the cause of the problem is the power supply because of the data loss incidents on an otherwise healthy hard drive (I checked into that as well).

So basically, I'm wondering if there is a way I can test these power supplies for problems. I should have the necessary equipment to do so.

 

[4745454b]

About all we can do is plug a volt meter into it. Will give us the voltage, but not the ripple.

I like to tell the story my Uncle told me about a machine that came across work where the hdds were failing at a much higher rate then normal. He said he told the tech to volt meter check the PSU, and they found the 5V rail was hitting 7V. Replaced the PSU, everything fine after that. It's certainly possible, but that's all I can think of unless you want to get fancy testing equipment. I'd try it at idle and as much of a load as you can. Check both the 12 and 5V rails.

 

[ram1009]

To test any power supply properly requires monitoring all rails for voltage, current and ripple simultaneously while under full load and at elevated temperature for an extended period of time. The results need to be recorded on a strip chart for analysis.

 

[jrgray93]

About all we can do is plug a volt meter into it. Will give us the voltage, but not the ripple.

I like to tell the story my Uncle told me about a machine that came across work where the hdds were failing at a much higher rate then normal. He said he told the tech to volt meter check the PSU, and they found the 5V rail was hitting 7V. Replaced the PSU, everything fine after that. It's certainly possible, but that's all I can think of unless you want to get fancy testing equipment. I'd try it at idle and as much of a load as you can. Check both the 12 and 5V rails.

Come to think of it, one of our guys has a test stand somewhere around here that runs cycles of some sort on power supplies and allows us to check the rails. I'll have to ask him about it. Thanks!

 

[ko888]

Make sure the power supply gets connected to an oscilloscope so that the ripple on each rail can be determined.

High ripple (i.e. outside of ATX specs) will eventually destroy the electrolytic capacitors on your devices.

 

[jrgray93]

Thanks for all the help, guys. I'll be testing it out within the week. It's really more of a learning experience than anything that NEEDS to be done. I just like to know what the root of the problem is.