Dell ATX PWS Stupidity

[ugjay]

I'm a PC technician and have used several ATX power supply*testers on at least two DOA Dells,*so here's my quandary. The PWS tested good on both occasions(most recently with a Thermaltake Dr Power II), problem is, They were both bad. Last week I even ordered a new motherboard(I now have to return)*because I was fooled into believing the tester. Turns out it was the PWS! A new(old ) Dell PWS fixed the problem.

And Here's the kicker...I think The motherboard knew the PWS was bad and didn't/Wouldn't/couldn't*power it on(Dr Power II powered it on just fine.).

Why did Dr Power II Lie to me?!? Whats going on? Do I need to build a test jig with loads on each rail to test a ATX PWS properly?

[stj]

Re: Dell ATX PWS Stupidity

your tester checks voltage, but does not load down the psu to make it actually work hard.

if you want to make a small rig with 2-3 50w-100w lamps on it to connect to the 3.3v, 5v and 12v it will be a more realistic test.

btw, a psu that works with no real load but fails in the real world is probably suffering bad caps unable to smooth the larger power output.

[momaka]

Re: Dell ATX PWS Stupidity

Cheap power supply "testers" are nothing but a big joke (and as you saw in your case, a waste of time and money). And that's because they don't put any load on the PSU. Moreover, even if you did find a way to load down the PSU while still connected to the tester, that would not show you if the PSU is producing excessive noise or not. To find that out, you need a scope.

I find that it's cheaper and also often easier to just find an old Pentium 4 PC and use that for testing power supplies. Pentium 4 Prescott is preferred, as it has slightly higher power draw. Not to mention they are super cheap. You can get a motherboard + CPU online for under $15 shipped on eBay. If the PSU is outputting too much noise, then the PC may not start too. Of course it's not a fool-proof method, but it beats a simple load test without a scope.

That said, this is why it never hurts (or does it? ) to open the PSU's top cover and check for bad caps. In fact, I do that *before* I even try to test any suspect PSU. You'll be surprised how often the problem is bad caps - especially in OEM machines like Dell and HP. Speaking of which, the bad Dell PSU you were dealing with very likely had bad caps if it started with the load "tester" but not in the PC.

[Quaddro]

Re: Dell ATX PWS Stupidity

Yeah..
I never buy stuff like that.
Simple tester with led indication, multimeter and halogens 12v lamp is more than enough, better than throw dollars for fancy tester.

[PeteS in CA]

Re: Dell ATX PWS Stupidity

If you wanted to have fun creating your own tester, take an ATX PSU case with fan, wire up a piece of .1" spacing Vector board with load resistors (10%-20% load per output) and do window detectors for each output (those you care about) that drive an LED for each output. Power it with a wallwart. TL431s are good for the window detectors since they incorporate a voltage reference and an op-amp that can be used as a comparator (with an output that only goes to about 2V above the return rail).

[ugjay]

Re: Dell ATX PWS Stupidity

Thanks, I hadn't thought of a Thermaltake product as cheap but now I know it's a totally unreliable way to test a PSU. I guess next I'll buy several cheap (LED or analog? IDK) voltage readouts on ebay and grab a bunch of headlights from the junk yard and build a nifty test fixture for the future.

[PeteS in CA]

Re: Dell ATX PWS Stupidity

Think of those as having limited capability. It's useful as long as you understand its limitations.