GPU1 overheating, Please help

Discussion in 'Extreme and Water Cooling' started by Slayer_T, Aug 28, 2017.

  1. Huggy_Bear64

    Huggy_Bear64 Member

    Joined:
    Apr 24, 2010
    Messages:
    3,096
    Location:
    Melbourne, VIC, 3127
    Is the fan connector just the RPM reporting wire, or does it have all 3 wires for some reason?
     
  2. OP
    OP
    Slayer_T

    Slayer_T Member

    Joined:
    Jul 31, 2013
    Messages:
    63
    Its got all three
     
  3. Huggy_Bear64

    Huggy_Bear64 Member

    Joined:
    Apr 24, 2010
    Messages:
    3,096
    Location:
    Melbourne, VIC, 3127
    https://www.ekwb.com/shop/ek-d5-pwm-motor-12v-dc-pwm-pump-motor
    Is this the pump you've got?

    If if is, the fan connector isn't used for power anyway, it's just RPM reporting and PWM signal, so either plug it into your motherboard for that, or you can just unplug it and the pump will run at 100% all the time.
     
  4. OP
    OP
    Slayer_T

    Slayer_T Member

    Joined:
    Jul 31, 2013
    Messages:
    63
  5. Huggy_Bear64

    Huggy_Bear64 Member

    Joined:
    Apr 24, 2010
    Messages:
    3,096
    Location:
    Melbourne, VIC, 3127
  6. OP
    OP
    Slayer_T

    Slayer_T Member

    Joined:
    Jul 31, 2013
    Messages:
    63
    oh right...it looks different
     
  7. Huggy_Bear64

    Huggy_Bear64 Member

    Joined:
    Apr 24, 2010
    Messages:
    3,096
    Location:
    Melbourne, VIC, 3127
    So yeah, I don't think you're having pump problems.
    Try turning your pump up to 100% (or unplug it from the motherboard) then move your case around a bunch (lay it on its side, back, upside down, etc.) and see how much air comes out. If that doesn't fix it, you might have to pull the GPU block apart and have a look inside.
     
  8. OP
    OP
    Slayer_T

    Slayer_T Member

    Joined:
    Jul 31, 2013
    Messages:
    63
    I'm starting to believe that too...
     
  9. OP
    OP
    Slayer_T

    Slayer_T Member

    Joined:
    Jul 31, 2013
    Messages:
    63
    Last edited: Sep 2, 2017
  10. radeon_freak

    radeon_freak Member

    Joined:
    Sep 19, 2007
    Messages:
    3,155
    Location:
    Sydney
    Mate you're at the point not where you have no choice but to tear down the loop. It's obvious even without the thermal camera that one gpu isn't getting cooled sufficiently.

    It's either a blockage or a bad mount which I doubt. A bad mount usually results in temps spiking rapidly as there is poor contact with the block. But since you are pulling the loop down there is no harm in checking the mount whilst you're at it.

    Also I'd be going back to series as there is really no reason to run parallel in such a simple loop and a D5. Temp wise you will only be looking at a few degree's difference between GPU's.
     
  11. OP
    OP
    Slayer_T

    Slayer_T Member

    Joined:
    Jul 31, 2013
    Messages:
    63
    I have reconfigured my loop and ran the thermal scanner again and it looks a lot healthier. See Photo / Video

    https://www.youtube.com/watch?v=M4BEjktc3zE

    https://www.youtube.com/watch?v=MLIGG4p7QPM
     
  12. kEv_

    kEv_ Member

    Joined:
    May 5, 2012
    Messages:
    536
    Location:
    Adelaide
    so its ok now? good work:thumbup:
    what are your temps like?

    my 290x's in series have ~1 degree variance (iirc, that rig is in pieces)

    looks pretty clear to me in that first thermal vid that the flow:
    but you know, i could be wrong...
     
  13. Flat-6

    Flat-6 Member

    Joined:
    May 1, 2014
    Messages:
    422
    Location:
    Adelaide
    All the thermal scanner is doing is showing you what you already know - ie the first GPU/block has an issue.

    By re-configuring it as a serial connection, all you are doing is masking the problem.

    Pretty much confirms now that you have some blockage issue in the first GPU block.
     

Share This Page

Advertisement: