Page 1 of 2 12 LastLast
Results 1 to 10 of 14

Thread: Old PC problem

  1. #1
    Join Date
    Nov 2015
    Location
    Singapore
    Posts
    17
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default Old PC problem

    Hi all,

    I am currently debugging an issue I am seeing on my old pc.

    psu : op650 (>7 years)
    mb : asus p7p55d-e i5-750 (yeah pretty old)
    gpu: r7 265 (brand new) / used to run hd6950 for several years

    6x hdd + 1ssd

    setup had been running stable for many years, but recently I am seeing shutdowns (no blue screen), so I swapped out the old hd6950 (which was running way too hot), but did not solve the problem.

    memtest : passed
    futurmark : passed (wall watt-a-meter shows ~250W)
    prime95 : passed

    if I stress both the cpu and gpu to 100% concurrently, the system will shutdown in less than 1 sec. Checked temps and they are within range (~70C for GPU and CPU). hwmointor shows 12v supply dropping to 11.34v (below ATX spec) so I am suspecting the PSU, but could be the motherboard.

    What do u folks think is the problem here ??

    Thanks in advance .

  2. #2
    quest for silence's Avatar
    quest for silence is offline Silencer & heathen relic of a human (™ by OW)
    Join Date
    Sep 2013
    Location
    Italy
    Posts
    3,353
    Thanks Thanks Given 
    1,037
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Actually it could be something related to the mobo VRM circuitry: did you experienced shutdowns or more properly reboots?

    Anyway, you may try to overvolt the CPU (within safety margins) in order to draw more than 250W at the wall under load, and then re-test with Prime95 SmallFFT, or better Linpack libraries (IntelBurnTest 2, LinX, Intel Math Kernel Library), in order to investigate a bit furtherly that possible VRM issue, ruling out GPU from the equation. Reducing cooling will also rise mobo/VRM overheating, if you dare to do that.

    In case, as a general rule I wouldn't use your running system to perform those tests but a fresh OS installation on any spare hard drive/ssd, and I would make such a test with the mobo outside the case.

    Last but not least, if you're running Windows, another safety measure could be running sysprep with OOBE/generalize/shutdown options before running some test with a spare hard drive: in the unfortunate occurrence you will kill that mobo, that will let you to restore your environment on any replacement system (you may google about that to gather some more information).
    Best, Luca

  3. #3
    Join Date
    Oct 2012
    Posts
    45
    Thanks Thanks Given 
    7
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Look in the event logs, there might be something there even if you don't get a blue screen.

  4. #4
    Join Date
    Nov 2015
    Location
    Singapore
    Posts
    17
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by quest for silence View Post
    Actually it could be something related to the mobo VRM circuitry: did you experienced shutdowns or more properly reboots?
    Yes I also suspected the motherboard, the problem is consistently a power down, so there isn't any crash dump to debug .

    Quote Originally Posted by quest for silence View Post
    Last but not least, if you're running Windows, another safety measure could be running sysprep with OOBE/generalize/shutdown options before running some test with a spare hard drive: in the unfortunate occurrence you will kill that mobo, that will let you to restore your environment on any replacement system (you may google about that to gather some more information).
    Good idea, I'll take an image of my current windows for insurance.
    Another observation is that when I reduce the cpu speed (from 2.6G to 1.6G) it becomes more stable.

    Since most of my hardware is old, I'll go slowly upgrade an item at a time. Planning to get a new PSU this week ....

    Thanks for your suggestions.

  5. #5
    Join Date
    Nov 2015
    Location
    Singapore
    Posts
    17
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by Frick View Post
    Look in the event logs, there might be something there even if you don't get a blue screen.
    Good suggestion, I will double check ....

    Thanks!!

  6. #6
    Join Date
    Nov 2015
    Location
    Singapore
    Posts
    17
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Update.

    With wall watt-a-meter at 280W, I see that voltage rails at:
    3.3 @ 3.12
    5 @ 4.68
    12 @ 11.24
    Screen is slightly unstable (I can visually see some jitter).

    When watt-a-meter reaches 290W, system power is cut.

    Checking dumps show the gpu driver (atikmdag and atikmpag).
    I am back to suspecting PSU as all the voltages are below -5% mark .

    Further testing with furmark and video driver consistently stops responding, but is able to recover ....

  7. #7
    quest for silence's Avatar
    quest for silence is offline Silencer & heathen relic of a human (™ by OW)
    Join Date
    Sep 2013
    Location
    Italy
    Posts
    3,353
    Thanks Thanks Given 
    1,037
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by eawtan View Post
    Checking dumps show the gpu driver (atikmdag and atikmpag).
    I am back to suspecting PSU as all the voltages are below -5% mark
    Unfortunately you can't trust voltage software readings.

    Do you have any other GPU handy? And (I don't recall atm) which PSU is your one?
    Best, Luca

  8. #8
    Join Date
    Nov 2015
    Location
    Singapore
    Posts
    17
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Took home a hx850i at a fantastic price. Everything works great now. Maxx'ed out at 380W (from wall watt-a-meter) and voltages remain stage.

    Software voltage does not correlate to corsair link voltage report, but I learnt that if the software voltage drops when load increases , there will be a point where the voltages will drop past the ATX specs (again subjective if the system will ever reach this point), but in my case it happened below the system max out wattage .

    I will be putting up a teardown of the op650 (bought on 2007) in the psu section and provide a link here. A peek inside did not reveal any secondary caps issue (teapo), the primary caps are hidden behind a heatsink so I will need to open the case. Stay tuned ...

    Thanks all for the various feedback, suggestions and comments.

  9. #9
    Join Date
    Oct 2006
    Location
    jonnyGURU forums, of course!
    Posts
    15,513
    Thanks Thanks Given 
    501
    Thanks Thanks Received 
    45
    Thanked in
    36 Posts

    Default

    Quote Originally Posted by eawtan View Post
    Software voltage does not correlate to corsair link voltage report
    Well.. there's a reason for that.

    Link actually gets its readings from the PSU.

    The BIOS and any 3rd party software gets it's voltage readings from the motherboard. To be more specific, an IC towards the middle of the motherboard, quite some distance away from the actual connection the PSU has to the motherboard. So... yeah.

  10. #10
    Join Date
    Nov 2015
    Location
    Singapore
    Posts
    17
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by jonnyGURU View Post
    Well.. there's a reason for that.

    Link actually gets its readings from the PSU.

    The BIOS and any 3rd party software gets it's voltage readings from the motherboard. To be more specific, an IC towards the middle of the motherboard, quite some distance away from the actual connection the PSU has to the motherboard. So... yeah.
    Hi Jonny,

    on the 12v rail, what should be a good/acceptable voltage drop from PSU to motherboard ?
    I am currently seeing ~0.2V drop when loaded at ~350W . Is this normal?

    Again the ADC on the (old) motherboard may be slightly off calibration and I don't have any precise measuring equipment at home...

    Thanks.

Similar Threads

  1. what's the problem?
    By salejel in forum General PC Hardware
    Replies: 2
    Last Post: 07-19-2011, 08:54 AM
  2. Problem here?
    By ice445 in forum PC Power Supply Discussion
    Replies: 5
    Last Post: 04-09-2011, 02:35 PM
  3. problem ?
    By wolf_288 in forum PC Power Supply Discussion
    Replies: 4
    Last Post: 02-14-2011, 07:26 PM
  4. PSU Problem????
    By kenny_k in forum PC Power Supply Discussion
    Replies: 5
    Last Post: 08-10-2007, 09:32 AM
  5. Do I have a PSU problem?
    By Halibass in forum PC Power Supply Discussion
    Replies: 5
    Last Post: 10-15-2006, 07:07 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •