liquid nitrogen cooling a possibility?

Robert G. Brown rgb at phy.duke.edu
Tue Jun 12 14:55:20 PDT 2001


On Tue, 12 Jun 2001, William Park wrote:

> On Tue, Jun 12, 2001 at 04:11:03PM -0400, Velocet wrote:
> > > The trick is just to cool the CPU and leave everything else on the
> > > motherboard untouched.  So, why not put dry-ice (CO2) on top of CPU?
> > > You would need something to hold it in place on top of CPU, and the
> > > dry-ice dimension can extend vertically.  I'm thinking long rectangular
> > > or round stick, freely held vertically from CPU, so that as dry-ice
> > > evaporates it will drop down.
> >
> > CO2 is liquid at about 6ATM and above (IIRC, remembering where
> > the triple point is from my university chem...)
> >
> > I dont think you want to mess with high pressure cooling systems...
>
> Dry-ice goes from solid to gas directly!  So, all you get is mist around
> your motherboard.  Hmmm, water vapour condensation is one concern.

Guys, I'm going nuts here.  Let's assume that your CPU runs at a nice
convenient 30W.  Let's also assume that the latent heat of the CO_2 on
top of it is oh, 300 KJ/Kg (correct within a factor of 2 depending on
the temperature.  The CPU thus releases 30 Joules per second, and
300,000 Joules are required to vaporize the CO_2.  We divide and
discover -- Amazing!  The Kg of CO_2 will keep our CPU a nice frosty
-78.4C for 10,000 whole seconds, or (dividing by 3600 and being sloppy)
three whole hours before the active release of heat turns it back into a
greenhouse gas.

Now, two minutes on the web reveals that the market price of dry ice is
around $0.60/pound (because it is, as I mentioned before, increasingly
expensive to cool things way below ambient temperature).  Let's call
this a mere dollar a Kg, since you'll be buying in bulk.  This means
that you can keep that CPU frosty cold for a mere $8/day, provided that
you work out some way to load a vertical hopper that will hold 8 kg of
dry ice in a vertical tower over the hapless CPU, or work out some other
way of refilling a smaller tower (your very own dry ice machine dropping
its precious and massive solids down only a CPU cooled WAY into the
brittle zone, for example).  And this is WAY CHEAPER than a liquid
nitrogen cooler, which I don't even want to THINK about.  Why do you
think the liquid gas (>>any<< gas) cooled supercomputers cost a
gazillion dollars each and required a small army of technicians to care
for?  Because piping a potentially deadly liquid that changes the
physical properties of things like metals from strong and flexible to
brittle is EXPENSIVE.

This whole discussion is a total waste of time.  Only a complete idiot
with far more money than sense is going to try to actually build a CO_2
or liquid nitrogen cooled PC for any reason other than the fun of it.
If that complete idiot actually builds the cooler, chances are excellent
that even if it functions as designed it will break the shit out of his
system and not TOO shabby that it will literally burn him a time or two
while doing so, quite possibly badly (unless he or she is an expert in
cryogenic technology).

Now can we just STOP?  This has nothing whatsoever to do with
beowulfery.  The next time somebody asks about exotic cooling
technologies, the appropriate answer from the list should be "No, no,
for the love of God, Montressor..."

   rgb

-- 
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 mailing list