Testing ATX PSU with scope

Discussion in 'General Electronics Chat' started by DerStrom8, Dec 29, 2013.

  1. DerStrom8

    Thread Starter Well-Known Member

    Feb 20, 2011
    2,428
    1,328
    Hi all!

    So some of you may remember several months ago (maybe a year already?) that I was having issues with my desktop PC shutting down randomly with no warning for no apparent reason. After replacing parts one-by-one, running software checks, etc, I eventually decided it was probably an issue with the PSU. Well I just recently started having problems with my desktop again so I finally broke down and bought a new one this afternoon. However, I'd like to test my old one to see if I can figure out what caused the problem. My question is this: How does one go about testing an ATX power supply using an oscilloscope? Unfortunately I don't have one of those fancy digital ones, all I've got are my scopes from back in the late '70s--analog CRT, one digital storage, the other not. What would I be looking for on the output? I expect I'd be looking for ripple and noise (perhaps something has gone past the filter caps, which I just replaced recently), but I'm not sure what levels of noise/ripple are normal, and what is unexpected. Has anyone here done this sort of testing, and might you have some tips/recommendations? Any thoughts are appreciated!

    Regards,
    Matt
     
  2. Dr.killjoy

    Well-Known Member

    Apr 28, 2013
    1,190
    156
    Lol your in my neck of the woods now .....The psu are tested all the time and are given ratings on there performance .... Check out reviews of psu online and they will show you how they test them .... Also is this a oem or aftermarket psu???? They do make testers for psu for pretty cheap..

    Check out
    Anand Tech power supply methodalogy
     
  3. DerStrom8

    Thread Starter Well-Known Member

    Feb 20, 2011
    2,428
    1,328
    It's a CoolerMaster 650W psu, about 5 years old. Never found much information on testing those particular units.
     
  4. Dr.killjoy

    Well-Known Member

    Apr 28, 2013
    1,190
    156
    Check out Overclockers.com and scoll down to the psu thread and find the psu reveiw thread or I will post it up later for you...
     
  5. DerStrom8

    Thread Starter Well-Known Member

    Feb 20, 2011
    2,428
    1,328
    I'm thinking Overclockers.com is one of the sites I looked at and couldn't find any information about testing PSUs with scopes, especially ones that are 40 years old. A link would definitely be appreciated though!
     
  6. Dr.killjoy

    Well-Known Member

    Apr 28, 2013
    1,190
    156
    Ok
    Did some leg and found this .. Hope it helps out
    Cooler Master GX 650W Power Supply Review


    http://www.hardocp.com/article/2010/11/22/cooler_master_gx_650w_power_supply_review/1#.UsGVirQwLsw


    AnandTech Power Supply Test Methodology


    http://www.anandtech.com/show/2259
     
  7. DerStrom8

    Thread Starter Well-Known Member

    Feb 20, 2011
    2,428
    1,328
    Hmm, so according to those links, there's really no chance of me testing the supplies with just a scope then? I had a feeling I'd need programmable loads, but I was really hoping I could do some very simple tests without them.

    Well thanks for the links anyway. It's been very helpful!
    Regards,
    Matt
     
  8. paulktreg

    Distinguished Member

    Jun 2, 2008
    612
    120
    You could check the noise levels whilst the power supply is in use. If I remember correctly it's 120mV max on the 12V rails and 50mV max on the rest. Load the power up as much as you can with a CPU and GPU stress test program (OCCT, Furmark, etc) but watch your temperatures! You could have a look at the DC rails on the scope and see if you can catch any dips or spikes?
     
  9. DerStrom8

    Thread Starter Well-Known Member

    Feb 20, 2011
    2,428
    1,328
    The problem with testing it in-computer and under regular load is that the computer doesn't stay running long enough. It shuts down after only a few seconds to a minute, usually before I even have a chance to log in (sometimes it shuts down during its boot). I would really need to test it on a bench, since it's impossible to do it when it's installed in the system.
     
  10. paulktreg

    Distinguished Member

    Jun 2, 2008
    612
    120
    First off try loading it up with an HDD or two and short the green wire to any black on the 20/24 pin connector. Does it turn on and stay on?
     
  11. DerStrom8

    Thread Starter Well-Known Member

    Feb 20, 2011
    2,428
    1,328
    I haven't tested it with hard drives attached yet, I'll try that later today. I have turned it on (by shorting the green to ground) with nothing attached and it stayed on, but under no load I expect I wouldn't see issues anyway.

    I'll let you know what I find.
     
  12. GopherT

    AAC Fanatic!

    Nov 23, 2012
    6,072
    3,845
    Based on behavior, I would look for a cold solder joint or a hairline crack in the pcb and heating upon startup is causing an open circuit.

    Cheers
     
  13. DerStrom8

    Thread Starter Well-Known Member

    Feb 20, 2011
    2,428
    1,328
    Good thinking, I'll have to check for that as well.
     
  14. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    It is not really much use using test gear and techniques designed for soak testing known good power supplies (or any other gear for that matter) to trace an intermittent fault.

    Paul's hairline crack or cold solder is a good one.

    So is a shorted turn within the HF transformer, a failing capacitor, resistor ...etc..etc.

    Another possibility is the return good signal from the motherboard. I know you are providing this by shorting two wires, but if there is a poor circuit on the psu board then this could shut down the psu. Momentary failure of the control signals can induce the symptoms you describe.

    Do you understand the control signal sequence for an ATX power supply or would you like me to post it?
     
  15. DerStrom8

    Thread Starter Well-Known Member

    Feb 20, 2011
    2,428
    1,328
    Unfortunately I really know next to nothing about ATX power supplies. I replaced the two filter caps about 6 months ago and the problem stopped until just a few weeks ago, then it started again. I have a feeling something might be causing the caps to go bad quickly, though like I said I really know next to nothing about PSUs.
     
  16. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
Loading...