Extream cooling solution theory

Discussion in 'Extreme Cooling Solutions' started by Mewohkie, Sep 28, 2012.

  1. Mewohkie

    Mewohkie [H]Lite

    Messages:
    88
    Joined:
    Feb 14, 2005
    So I have been thinking long and hard. I would love to try and combine several cooling solutions to try and create a working sub ambient cooling system.

    The freezer method with the radiator in the freezer.
    Why this does not work: Freezers are not meant to keep constant sources cold. The unit would burn out before long.

    How this could work: Running a series of radiators before the water gets to the freezer to make sure the water before entering is already ambient temperature. Using an antifreeze coolant in the lines, I could even submerge the radiator in the freezer in a bucket of frozen water.

    Why this does not work: Condensation will build up on components inside the PC one the coolant reaches the computer again.

    How this could work:
    Submerge the components in mineral oil. The thermal properties of the mineral oil should be enough to prevent condensation from forming on components.

    Let me know what your thoughts are if this could be doable or not.
     
  2. Jorona

    Jorona Font Snowflake

    Messages:
    3,236
    Joined:
    Nov 6, 2011
    Sounds dumb to me. Frozen water actually acts as a insulator in some situations. I think you're looking at minimal gains for a mess fo work and a mess of a setup.
     
  3. Tsumi

    Tsumi [H]ardForum Junkie

    Messages:
    12,687
    Joined:
    Mar 18, 2010
    Bong chillers and proper insulation of the case would be far more effective and easier.

    Even if the water was at ambient temperatures, the freezer is still not designed to work that way. You are still going to burn the unit out.
     
  4. Mewohkie

    Mewohkie [H]Lite

    Messages:
    88
    Joined:
    Feb 14, 2005
    Having 2 heat exchangers in a loop may work.
     
  5. Tsumi

    Tsumi [H]ardForum Junkie

    Messages:
    12,687
    Joined:
    Mar 18, 2010
    You basically have three options for sub-ambient cooling:

    Bong chiller.
    TEC plates.
    Refrigeration unit designed for continuous operation.
     
  6. thesecond

    thesecond [H]ard|Gawd

    Messages:
    1,051
    Joined:
    Jan 15, 2010
    Uhm. Running radiators outside of the freezer will just put more heatload into the water (by "removing" the coldness that the freezer put into it). At the rate the fluid should be flowing, the fluid temperature should be more or less equalized across the entire system.
     
  7. Archmage

    Archmage 2[H]4U

    Messages:
    3,019
    Joined:
    Jul 13, 2000
    thesecond is correct.

    Evaporative coolers, sometimes designed into the shape of a "bong" = 5-10C below ambient, but the loop is open and thus requires significant maintenance. I'd use a bit more silver in a loop like this, and I'd change the water more frequently. This is easily the most power/cost efficient method to get sub-ambient temps, and it can actually be combined with other methods (cooling of hot-side of TEC loop, or water-cooled refrigeration condensers). - This may still require some insulation, but you could probably design a unit to stay above the dew point.

    TEC: Are inefficient in general (with respect to power), but not at low voltage. However at low voltage, they cannot produce a large temperature delta (up to about 25C between the 2 sides before you lose a hell of a lot of efficiency). The most effective way to use these (unless you're cooling...say GPU memory or a mobo chipset- which can probably use TECs directly (sandwiched between two tightly clamped pieces of copper - the top one being the waterblock)) - is to chill the water with multiple of the biggest wattage TECs in parallel. This is expensive, and will require radiators on the hot-loop.

    Edit: I just realized that since Sandy-Bridge, CPU power has decreased a bit. Perhaps the power-density is still high, but the overall heat load has decreased. Direct-TEC cooling is likely possible again: http://www.xtremesystems.org/forums...-Bridge-(No-Insulation-Required)&daysprune=-1 - these guys claim to have made an ~ambient temp solution. I bet one could go with a little more power, and even cool the hot-side with an evaporative cooler (probably a lot of evaporation though) for an additional decrease in temps.

    Phase Change / Refrigeration: I'd probably chill the water just like with the TEC. You can go colder with this setup and still maintain some decent level of efficiency (but I'd still say it isn't worthwhile at this point).

    Insulation: Use it if you need it. There are a bunch of guides using various materials: kneaded "eraser" and similar compounds like dragonskin or liquid electrical tape + maybe Dow Industrial conformal coatings, neoprene of course... dielectric grease is probably still used...

    I don't think it's worth the hassle anymore. I'd build an evaporative cooler if I wanted to experiment with cooling again - minimal insulation at worst. I'd probably actually use some colloidal silver solution in the loop in addition to the silver kill coils. I'd be annoyed at having to top it off constantly...so I'd build in a big reservoir.

    Now it's actually possible to control the case environment such that no condensation occurs (essentially chill the air, seal it) - and it has been proposed and even completed in one instance, but again...so much work...
     
    Last edited: Oct 12, 2012
    honegod likes this.
  8. Risant

    Risant Limp Gawd

    Messages:
    161
    Joined:
    Jun 19, 2004
    I have actually been thinking of doing something along these lines. Kind of a hybrid between a watercooled setup and a oil submerged setup. I want to set up a watercooled setup with a Hailea water chiller and then insulate the setup from the air using Mineral oil to prevent condensation. I know it sounds like a huge pain in the ass and I would probably regret it as soon as I got started, but it just sounds like a fun challenge to me. Also I would like to try to mod the water chiller to allow chilling below the 4C that they are designed to run. Of course that would also require running some other fluid than water to prevent freezing, but I should be able to go to just above the freezing temp of mineral oil (-30C). I know I sound crazy, but I like challenges like this.
     
  9. schmuckley

    schmuckley [H]ard|Gawd

    Messages:
    1,237
    Joined:
    Oct 21, 2011
    tall bong :D
    My res water is usually around 11c lower than ambient.
    people take windows AC units and stick the err..outside finny piece in their reservoirs for a chiller effect.
     
  10. Lord of Shadows

    Lord of Shadows 2[H]4U

    Messages:
    3,312
    Joined:
    Mar 9, 2004
    I have an old peltier and AMD64 system that I was afraid to get below ambient for obvious water condensation regions that I'm not afraid to completely destroy now that it is outdated. (once I have enough free time from work that is)

    But at work we have temp chambers, and I noticed one thing that these temp chambers have that I have never seen mentioned before anywhere. They have a hose that constantly pushes dry air into the chamber to help prevent moisture and condensation. I figure the warm air in your case is probably the driest air you can get and a cheap pump sorta like those fish tank air supplies could move it to where you fear condensation will get you around your motherboard and super chiller
     
  11. Ice Czar

    Ice Czar Inscrutable

    Messages:
    27,424
    Joined:
    Jul 8, 2001
    LOL
    deja vu

    or a twist on the same theory

    (and the rest of yall...don't bother you can't have that :p)