[Beowulf] Innovative liquid cooling
Robert G. Brown
rgb at phy.duke.edu
Thu Mar 7 12:55:32 PST 2013
On Thu, 7 Mar 2013, Lux, Jim (337C) wrote:
> Following RGB's suggestions, here is what I have implemented in my garage..
Simple proof that any good idea is already implemented by somebody,
somewhere. Especially one this good...:-)
> Cooled air goes into storage area for wine at roughly 12C, passing through to another box roughly at same temperature, but with controlled humidity, for curing of sausages, thence exhausts into the general vicinity of the server, fiber drop, network switches, and GPS disciplined time reference.
> This was evolved from a former, less effective system using the small 4 cubic foot "bar refrigerator" used to cool beer/ale/whathaveyou along with a set of metal tubes bonded to the "freezer" compartment of the bar refrigerator that carried coolant pumped around through tubing in the wine cooling area.
> Why change from liquid cooling to air cooling?
> 1) pumping losses.. the coolant pump actually puts more heat into the liquid than the fan puts into the air.
> 2) corrosion.. even with anti-corrosion additives (chromates and such as found in commercial antifreeze), the tubing on the freezer plate corroded away. I tried both copper and aluminum in various forms, and they ALL fail eventually. I do not like having antifreeze on the outside of the beer bottles.
> 3) design of liquid to liquid or liquid to air heat exchangers is a black art with which I am not skilled. In fact, even using commercially available exchangers (various and sundry heater cores and radiators), it is VERY difficult to get predictable performance from the system.
> 4) Too many failure points. For instance, if a critical amount of flow doesn't flow through the freezer heat exchanger, the coolant freezes and the pump then pumps against a blocked tube, adding heat to the system. (obviously, this was a time when I did not have antifreeze mixed in).
> 5) you still need a fan of some sort to transfer the cold from the chilled coolant to the wine bottles.
> I suppose one could go out and get a surplus lab chiller, etc. invest in a few dozen feet of stainless steel tubing and fiittings, etc.
> When one can buy a perfectly good tiny air conditioner at the end of summer for around $50, and cobble together some hose and cardboard with duct tape, it's just a lot easier. The smallest A/C is around 6000-8000 BTU/hr, which is an enormous amount of cooling for what is basically a large, warm refrigerator.
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
Robert G. Brown http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567 Fax: 919-660-2525 email:rgb at phy.duke.edu
More information about the Beowulf