Foundation Cooling

Foundation with Copper

Overclockers are always trying to come up with new, colder, and quieter ways to keep their PCs cool. [gigs] was so dedicated to this, he decided to lay 6 meters of copper pipe to use as a radiator in his new house’s foundation. As of now, the foundation is laid (copper pipes and all), and the forum posts come complete with finished slab pics, though there is no house to speak of yet.

[via Slashdot]

67 thoughts on “Foundation Cooling

  1. No worries mate, i just wish we could start the replies freshly with the right assumptions :-)

    I thought to read it here: “(like say 40000 watts and 1000 yards of tubing) can easily pumped by a single 150watt closed system pump and its pretty quit to.” To nit-pick: the original poster was building a house, you might be suggesting (1000yards?) a tall building; you did not say you HEAT your HOUSE with 40KW, but you did suggest 40 KW capacity. I agree with the closed vs open loop principle, and wanted to narrow the power requirements a bit. I’ve seen PC cooler pumps rated at 40W, my floor heating uses 7 Watt.

    I merely try to point out that a few of contributors may have started from the wrong assumptions. To correct a few: Concrete doesn’t insulate very well. It has a huge heat capacity. Floors with loops are warmer (as you tend to heat them and the insulation should be underneath) than a ground loop. Copper and concrete collide. A closed circuit needs a tiny (not even a small) pump. Joe _the gamer_ Average’s computer uses 400W. A passive house could heat 40 square meters with that. A well insulated house can simply overheat. A groundloop delivers 10-15°C water.

  2. ok first off
    its nothing new, it wont work, your pc will blow up and mice will start living in your eyelids if you make this

    the long copper tube and pump alone would dicipate enough for the average pc, concrete not even needed
    the only BAD part about this hack is the fact that it isnt silent, it IS passivly cooled but it still has noisy pumps

  3. @frogz
    NOISY PUMPS? What world do you live in? Oh,I see, the world where mice live in eyelids. Here (on the old continent) pumps are dead silent. By the way, if a pump is not ‘active’, how do you define ‘active’?

  4. “Thermal Conductivity – k – (W/mK)
    Concrete, light 0.42
    Water 0.58”

    Yes but water is a liquid. Most heat is moved around by convection not conduction. Also someone above mentioned the large thermal mass of concrete. True, but that only helps at the start; in the steady-state thermal mass is irrelevant.

    However, there’s so little heat involved here it probably makes little difference. I’m sure you could have just put the pipe in air and it would have worked just as well.

  5. I especially like the condensation will pool water comment; it goes on to say that this setup is ideal for heating, not cooling…

    Trouble is, this is a heater!

    In a heating setup you take hot water and put it in the concrete to warm the concrete (which happens to cool the water.) How does this differ?

    As far as pumps go, who cares how much mass of water you have, so long as the flow rate across the sink on the chip is near the intended one. Any pump sized to achieve that is correct whether you have 6 ounces of water or 6 gallons doesn’t matter. that just means the extra water will have more time to transfer into the slab as it makes the trip back to the pump inlet. The more water the slower the trip.

  6. This gives me an idea. Many of us had a distribution system where all our wires went back to a single block in our basement, or someplace else. Why not a centralized computer cooling system, for PC’s. or if you have one of those old homes with a central vacume system (vacume pipes in the wall), then it could be adapted as a centralized cooling system. No matter what room your PC is in.

  7. Now how about using a thermoelectric heatpump to make sub zero cooling for the pc and still only put about 300w into the ground depending on what size heatpump you use?

    As for the comments about the size of the pump it needs to be able to lift atleast 1m of water providing the computer is on the ground floor or about 3m if its on the second floor, this isnt easy with a small pump as computer cooling pumps are normally made for high flow not high lift.

  8. @tom
    These heatpumps are very inefficient, taking that you talk about Peltier cooling. You put in 3W at least for every W pumped; a 150W system would need (and produce) 600W (450W for the pumping, 150W from the cpu). And, as mentioned many times before, as long as you have a CLOSED LOOP system, the pump can be tiny. Oxygen is a bad thing for systems, so you don’t want to replenish the oxygen in your coolant by having an open system.

  9. This is not a good idea. Imagine a hot summer, humid, and the foundation cooling system working. Water condenses on the floor, making it slippery. The water also gets into any carpet/rugs/floor covering and produces mold. I wouldn’t do this for interior cooling at all.

  10. Obviously, the prototype testing has shown that it’s likely to work, but I want to try to smooth out some of the ripples in the thermal mass and conductivity discussions.

    Assuming his system runs continuously, he needs to dissipate ~150 W of power (heat) continuously. The thermal mass of the water, copper, and concrete play into the equation only when the system is turned on. That’s why you see the rise in temperature in the testing chart, as the mass heats up. Once it plateaus, it reaches steady state and the thermal mass doesn’t matter any longer. The thermal conductivity, flow rate, and the geometry of the pipe and slab are then the only things that are important. The heat transfer from the water to the pipe to the concrete to the far field can be modeled as a voltage across a resistor network. The voltage is analogous to the temperature difference between the water and the far field slab temperature. The convective heat transfer from the water to the pipe wall and the conductive heat transfer through the pipe wall into the concrete, and through the concrete are all modeled as resistors. This can be done to various levels of complication, but it’s a simple way to look at it.

  11. I have seen quite a few projects like this one being done with Pex tubing. Many of the others have used concrete as well so you shouldn’t have a problem. This one looks like a lot of hard work and using pex could have made it much, much easier.

  12. I find it amazing how many people have posted “That has been used for years to heat homes” … and then in the very next sentence, “But it will never work, Concrete is too good of an insulator”.

    You can’t have it both ways, people. Does this method conduct heat in to a room (and away from the floor) or not? If it can heat the house, it can radiate energy away from the pipes and cool the computer.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.