HEY! "Liquid" question for u pro's

[Ottoman]

ok I've been contemplating water cooling for awhile...

but I've been wondering, why does everyone use water?

u guys go to extremes in other situations, why use lousy water?
it's not exactly the best thermal absorbing and dissipating liquid...

why not use alcohol? or some other liquid...

and if it absorbs and dissipates heat better,
wouldn't that mean u could use a smaller pump, a smalelr resevoir and a smaller radiator, effectively making it smaller quieter and cheaper?


and one other question, with the peltiers, why not use it to cool the liquid in the resevoir? instead of the heatsink?
and using alcohol, which has a much lower freezing point, and it doesn't conduct electricty, so if u had a leak, your less likely to have a short circuit?

I'm no pro, nor a computer engineer,
but I'm just tossing a few ideas around
how about some feedback

 

[William]

alcohol actually abosorbs heat worse than water. It is used to cool people down because it evaporates so quickly, but it is really bad at absorbing heat. There are better liquids, but water is by far the easiest to use as far as a liquid goes. Water is not dinky, it is really really good.

 

[Bad Maniac]

Well, there just aint anything better to use than water. (or well, there is, but its liquid unobtainium) Basically Water has the best Heat transfer/ammount of fuss. I know for a fact that there are highly corrosive and toxic oils that are better, but they are simply not useable for the "normal person". So water it is.

 

[Ottoman]

ok.. that showed how little I know..

what is the heat transfer constant measured in liquids?

is it C/P? C/W?
something like that,
I wanna do some research..

 

[cjtune]

Ottoman (Jul 10, 2001 05:28 p.m.):
ok.. that showed how little I know..

what is the heat transfer constant measured in liquids?

is it C/P? C/W?
something like that,
I wanna do some research..

Liquids, like all other matter, have conductive heat transfer coefficients (symbol: k, unit: W/(m.k)). Invert the coeeficient around and you get...thermal resistance! (T=RQ where T=temp difference, Q=heat transferred, R=1/(kA/x)=thermal resistance) This simplifies heat transfer analysis if you have some basic idea about electrical circuits. Yes, C/W is thermal resistance. But generally, C/W ratings of HSFs and watercooler rigs also include many other factors and (an important) one of which is due to convection. This is the mode of heat transfer where moving fluid carries away heat and not just by virtue of touching (conduction). You cannot determine an over C/W of a system just by knowing a fluid's k so you'll either have to do some experiment or be damned good in mathematics (or a combination of both). Convection heat transfer coeffients (or resistance) depends on the flow rate of the fluid, the surface area in contact, and the way the fluid flows around the surfaces. Obviously faster flow rates are better, more fins or pins lead to more surface area, but at a danger of posing a obstruction to the fluid flow so these factors hang in balance in a good HSF or waterblock design.