[Dvorak]

I was considering building a phase change or liquid cooling system, and then I thought it'd be cool to build both in one! So instead of cooling the system with room temperature water, I could cool the water while it's being fed to the system.

Only problem with that is that water freezes at 0°C, so I figured I would use something else so as to get it cooler.

Thinking back to chemistry, the first thing I thought of was Propylene glycol (C3H8O2). As I recall it has a low freezing point and a very high boiling point.

Is there any reason I can't use this as a liquid coolant? thanks =)

[Burn]

Propylene glycol - Wikipedia, the free encyclopedia

under "Applications", it says it can be used as a coolant for liquid cooling systems. dunno what aspire has to say though...

[ComatoseClown]

i dont see any reason why not....although what is its heat transfer rate? .....also with temperatures that are lower than the due point...watch out for condensation

[c0rr0sive]

I use Feser One, not sure what the lowest temp it can go to is, but I do know, in a room that was at -15F, I got a good 3.78Ghz out of a 6000+ that hardly gets above 3.2Ghz stable with room temp air... And my processor was at a cool 30ish F...

[aspire]

This should definitely be moved to the cooling sub-forum.

Anyways, you will need some kind of non water coolant to prevent it from freezing. Propylene Glycol is often used in radiator fluid so you might want to start there and see how well that works.

You will need to choose your parts for the loop very carefully however, many parts, blocks, reservoirs, tubing, is not designed for temps as low as it sounds like your hoping for.

Condensation will be an issue and the cpu socket will need to be filled with dielectric grease and the area around the socket insulated as well.

Some additional information on how cold your looking to go, as well as yours planned specs for the phase unit would be helpful in giving you advice.

[Hefemeister]

Moved to cooling sub forum.
Ask and ye shall receive.

Thanks for the report.

[Dvorak]

alright, I looked around for a bit for some numbers and formulas but I'm not sure this is right (in fact I'm almost positive that it's wrong) so if anyone's good at chemistry, feel free to help out

Heat transfer coefficient - Wikipedia, the free encyclopedia

plugging in the correct numbers here should give an answer in the form of (Joules/second)/(m^2Kelvin) which would then tell me the amount of heat my solution should remove from the heatsink

Assuming the following numbers: (I use D for Delta)
(If someone could verify these it would be great!)
DQ = 1055 J (1 btu)
A = 0.007808 m2 (a 97.6 x 80mm heat sink)
DT = 333K (temp of heatsink) - 245K (temp of coolant) = 88K
Dt = 3600 seconds (btu is J/hour)

some quick math and I get 1055/(0.007808*88*3600) = ~0.4285 (J/s)/(m^2K)
so, using the heat sinks surface area and high end temperature above, I get ~0.4285 (J/s)/(0.007808*333) = ~0.1648 J/s per 1K per BTU

If you have 5000 BTUs of power, you can move remove 824 J/s of heat from the heat sink to the liquid

I've run outta time so I can't finish my calculations -.-

anyway I'm sure all the numbers are wrong so I'm looking forward to all the corrections ^_^