1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

A bad, but effective WC experiment

Discussion in 'Project Logs' started by xvi, Apr 4, 2013.

  1. xvi

    xvi

    Joined:
    Nov 10, 2006
    Messages:
    2,027 (0.70/day)
    Thanks Received:
    1,221
    Location:
    Washington, US
    A bad, but effective WC experiment (Edit: ICE!)

    Specs:
    ASUS 580X AM2 something or other
    Phenom II X2 550 Black Edition
    CoolerMaster AQUAGATE Mini R80
    CoolerMaster Dual Storm 80mm
    Regular silicon thermal paste

    Story:
    Found the cooler at a local electronics recycler quite a while ago. It leaked even just sitting around, so I cleaned out what I could and it seems to have fixed it.. As I've been thinking about some more serious water cooling gear, I wanted to see what all the fuss was about first. I had no compatible mount and no reservoir, so it's just sitting on the CPU and I used a Ziplock container to hold extra water.

    [​IMG]

    The only way it would fit meant running the block/pump's outlet hose over the memory. It was moved to the back to compensate.

    [​IMG]

    First boot in to BIOS shows that temps are holding.

    [​IMG]

    Idling in BIOS, the temps leveled off about here. Temperature difference between CPU and "radiator" water was pretty good.

    [​IMG]

    [​IMG]

    Under full load for 10 minutes and temps are actually looking a lot better than I thought a 80mm radiator and crummy block could do.

    [​IMG]

    And finally, about an hour of crunching with the radiator fan at full speed.

    [​IMG]

    Conclusion:
    The good:
    • Roughly the same\better temps as stock cooler
    • Quieter than stock cooler
    • Watercooling on the cheap if you can find one

    The bad:
    • Heatpipe towers are the same price, cool better, and don't risk leaking.

    p.s. I know my water will evaporate if left out in the open like that. I'm putting the stock cooler back on tonight. Don't worry.:laugh:

    Will update rather soon with temps for the stock cooler.

    Update: I LIED!
    Further testing had to happen. What kind of testing? Well..

    ICE!
    [​IMG]

    Thanks to my handy-dandy infrared thermometer, the ice appeared to be -6c.
    [​IMG]

    Add ice to ziplock reservoir, watch temps drop. All recorded temps (unless otherwise noted) are under full load from BOINC.

    CPU temp measured through lmsensors.
    CPU-in measured from the side of the reservoir.
    CPU-out measured from base of radiator near intake.
    Rad out measured from base of radiator near output.
    Delta is difference between CPU temp and ambient temperature (the goal being sub-ambient).

    Column 1
    0 cpu temp ambient CPU-in CPU-out rad out delta
    1 no ice 41 21.5 19.5
    2 ice 22 21.5 0.5
    3 more ice 21 21 0
    4 ice2 7 10.8 0
    5 ice2 19 6.4 13 -19
    6 idle 18 19 9.8 15.3 -1
    7 load-fanoff 19 19 5.5 11.4 0
    8 18 19 5.5 10.9 -1
    9 18 19 5.9 10.4 -1
    10 moreice 17 19 4 7.4 8.4 -2
    11 17 19.5 4 6.7 7.9 -2.5
    12 17 19.5 3.5 6.4 7.4 -2.5
    13 17 19.5 4.1 6.5 7 -2.5
    14 18 19.5 4.6 6.8 7.2 -1.5
    15 18 19.5 6.8 8.7 8.8 -1.5
    16 19 19.5 6.8 9.1 10 -0.5
    17 20 19.5 8 10.1 10.1 0.5
    18 21 19.5 8.5 10.8 11.3 1.5
    19 ice gone 23 19.5 9.5 12.6 12.8 3.5
    20 24 19.5 11.2 14.4 14.2 4.5
    21 26 19.5 12.8 16.2 15.8 6.5


    I had the fan on for the first bit of the experiment. I eventually realized that I had sub-ambient water going in to the radiator and it was actually being heated up! Turned the fan off and results increased.

    Final thoughts:
    It took all the pre-made ice I had (about three containers full) to keep this thing cool for less than 30 minutes. There's no way this could be sustained without an army of icemakers.
     
    Last edited: Apr 20, 2013
    Crunching for Team TPU
  2. Kast

    Joined:
    Jan 4, 2011
    Messages:
    172 (0.12/day)
    Thanks Received:
    134
    Is that a HD6850 ?

    Edit:Now I can see your specs. Looks exactly like a 6850 that I sold.
     
  3. xvi

    xvi

    Joined:
    Nov 10, 2006
    Messages:
    2,027 (0.70/day)
    Thanks Received:
    1,221
    Location:
    Washington, US
    It's a HD 4850. Sig rig has the 5870. I see what you mean though. They're pretty close.
     
    Crunching for Team TPU
  4. xvi

    xvi

    Joined:
    Nov 10, 2006
    Messages:
    2,027 (0.70/day)
    Thanks Received:
    1,221
    Location:
    Washington, US
    Update. ICE! Ice as in frozen H20. Check the original post for some (just barely) sub-ambient CPU temps.
     
    Crunching for Team TPU

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page