Psu Testing Not Always Gonna Be Correct With Dvm

[sam67]

Testing a psu with a multimeter is not a sure way to determine whether or not it's faulty..I had an intermittent fault where my pc would only boot every now and again ..checked all voltage rails and they were correct when having a load i.e dvd writer,hdd and mobo all connected yet my psu was faulty !

So my question is ,how can we determine that a psu is 100% working when testing with digital v meter when we have intermittent faults ? is a dvm that accurate ..I say the only true way is to replace with a known good psu .

Would like to hear your thoughts..

sam67

[370forlife]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Oscilloscope to check ripple, also need a loading device.

[Toasty]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Matters not the quality or type of meter or test instrument. Intermittents are the bane of every tech in every industry from the Maytag repairman to NASA.

You can load it and measure ripple and voltages and everything checks great, and it still drops out on the 3rd day after you reinstall it... usually exactly at the second before you click "Save" on that proposal due in 1 hour.

[seanc]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Mmhmm.

I have an Akasa AK-P400FG PSU, looks well built inside.

I removed it from a customers PC, it seemed to be dead, 5vsb wasn't on.
Got it home, tested it with my cheap PSU tester, powered on and looked fine.

Re-capped Teapo for Panasonic, tested with PSU tester - fine, shorted it on and measured with DMM, all looks good.

Hook it up to any mobo, 5vsb comes up, then it shuts down :S

[sam67]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Mmhmm.

I have an Akasa AK-P400FG PSU, looks well built inside.

I removed it from a customers PC, it seemed to be dead, 5vsb wasn't on.
Got it home, tested it with my cheap PSU tester, powered on and looked fine.

Re-capped Teapo for Panasonic, tested with PSU tester - fine, shorted it on and measured with DMM, all looks good.

Hook it up to any mobo, 5vsb comes up, then it shuts down :S

Well thats 2 of us that have tested psu's that seemed ok then not work ? yet something must trigger this to cause the problem (i know mine is a different problem)..just had mine to bits and noticed a frayed wire so re-soldered it put psu on pc ..still same fans spin but nothing..switch it off unplug the swine then plug it back in and hey presto all working.

Surf the internet running perfect but once powered off it might take a couple of tries and then it comes on again..Weird I know i could bin the psu but i want to find this problem as it has got me thinking what it could be.Not sure though where to go as everything inside looks ok .

[seanc]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Actually, I read that some Antec power supplies have issues when hooked up to motherboard using an nVidia Nforce 4 chipset - which is the chipset on the original test board I was using. I'm wondering if it's the same issue.

I hooked mine up to an old P4 board yesterday and it powered on. Didn't have long to test, I had to tidy things away.

I'll do more testing over the weekend.

[sam67]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

My psu came with the pc i got 5yr ago and has never let me down till i opened the psu to see what caps were in it and removed some of the brown glue. This psu was never a great design to start with it's a bestec but like i said it was running fine ..At least my mobo ain't fried as i tested it with my lite-on out of my other pc ,so i know it's defo psu now.

[seanc]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Well, this Akasa PSU has been running for most of the afternoon, doing Memtest86.
The machine die with a blank screen, but still had power, about an hour after runnning - it's working fine at the moment though, I could have just knocked something.

[sam67]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Just been bench testing this psu that has an intermittent fault...

+12 v rail is 11.87
+ 3.3 v rail is 3.32
+ 5 v rail is steady 5v
5vSB is steady 5v
-12 v rail is -12.3 v
power ok grey wire 4.92v

All done with dvm while idling ..

Then ran prime 95 torture test

+ 12v was + 11.77 v
+ 5v was steady 5v
+ 3.3 v was 3.2 v
5vSB was steady 5v
power ok grey wire 4.92v


Now this pc worked 3 time in a row without any problems when testing it on idle.. ran torture test to see what the voltages would be and then powered off ,restarted pc and problem is there fans spin but monitor won't show logo just sits with orange led on..

This happens like this even when not running torture test .

Could this be due to a faulty start up resister or cap that once warmed up causes an intermittent fault ? (most caps replaced)

Think i may just strip the complete psu and test all resisters and diodes..Not bothered about the psu but it does make me wonder why this problem occurs and all voltages are fine ? would like to find the cause ...I know when it's going to do it cause usually the pc seems to slow for a split second when it works but does not do this when it wont boot and display the logo on the monitor.

I don't like things like this to beat me and also this is a good learning point for me Just using this thread instead of my other one because this does show a dvm is not a sure fire way to say psu is good ..I know this for a fact as the problem is not there when using another psu.

Weird how all voltages are there even when it does not show on the screen.

[seanc]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

I assume you've checked for bad solder joints, might just be worthwhile re-soldering some bits that look questionable anyway.

[sam67]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

I assume you've checked for bad solder joints, might just be worthwhile re-soldering some bits that look questionable anyway.

Hi seanc , yes i am gonna do that tomorrow when i strip psu . .. Just curious that's all .. At least it helps me with my trouble shooting techniques i suppose it's better this way as i have nothing else at the mo to check as i have recapped the 4 monitors i have here lol.

[seanc]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

I have a mobo (failed cap) and and PSU (preemptively) to recap tomorrow

[Toasty]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Over sensitive PG signal line. Tripping PSU into protection/shutdown mode on start up. Substitute PSU may not be as sensitive and runs fine.

Be nice to know what make/model PSU this is...?

[sam67]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Over sensitive PG signal line. Tripping PSU into protection/shutdown mode on start up. Substitute PSU may not be as sensitive and runs fine.

Be nice to know what make/model PSU this is...?

Hi Toasty,

The psu is a bestec 300w 12 Z

[Toasty]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Recapped?

[sam67]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Recapped?

Yep but did not do the 22uf 50v cap and the 0.47uf 5ov ,0.1uf63v

and the 10uf 50v and 1uf 50v..

was thinking about soldering these out to test ESR when i start stripping psu..Did recap all the other caps with pan FM and FC though..

[seanc]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Replace even the smaller ones! they're important and they're really cheap! I'm not surprised you're having problems with it then.

[everell]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

I have noticed a similar problem on the Bestec ATX-250 12Z, ATX-300 12E, and ATX-300 12z: all have a input pi filter for the 5vsb located next to the heat sink, and this capacitor is the first to bulge. Also, any glue that has turned black indicates lots of heat, and is usually found in the area between the two heat sinks. So I recently fixed a ATX-250 12Z by moving this capacitor to the other side of the heat sink. Have to get a very small drill bit and drill two new holes. Plenty of room in the location shown. Just follow the circuit traces for the 5vsb and ground and you will see that this modification is very easy. Attached photos show old location and new location.

[sam67]

Re: Psu Testing Not Always Gonna Be Correct With Dvm

Replace even the smaller ones! they're important and they're really cheap! I'm not surprised you're having problems with it then.

Hi Seanc ,

The psu was working fine without the recap,i only had problems after i opened up psu to find what caps it had and there values ...I reckon i disturbed some think when separating some of the caps that were glued together..I am gonna replace the smaller ones anyway and test them with my ESR meter.. Just like to add none of my caps had bulged and the ones i did remove checked ok with ESR meter.

everell,

Thanks for the pics ,there was quite a bit of decayed glue that i have removed which i know can cause conductive issues.. The only reason i did not replace the smaller caps when i recapped was that there were no issues at all with psu when i decided to change caps (only changed due to running psu for over 5 yrs without any issues at all).

Will let you guys know once i sort it ..Removed transformers already and heat sinks to do a proper job