MadSci Network: Engineering

Re: Is there a liquid that is non-conductive and has a low specific heat?

Date: Wed Sep 13 13:07:14 2006
Posted By: michael pierce, Post-doc/Fellow, Materials Science Division, Argonne National Lab
Area of science: Engineering
ID: 1151555559.Eg

This is a very fun idea.  It's not going to be an easy project, but could
be a great deal of fun with a reasonable chance of success.  

The old Cray supercomputers used such a cooling system and there are in
fact commercial products that do exactly this.  One is Fluorinert made by
the 3M company.  The fluid will have a high dielectric and high resistivity
as required for the immersion of electronics.  However, these fluids
require special care in there use, maintenance and disposal.  I would not
recommend using them without proper training and knowledge.

So is it possible to do without bringing a potentially hazardous chemical
into your home?  Maybe.

It is possible that a mineral or vegetable oil would have the desired
properties and also prove to  be easier to work with.  High voltage
transformer oil may also work, but again you'll need to be sure that the
exact flavour you have is benign to work with.  Remember that despite your
best intentions, inevitably some will leak or spill(potentially a huge

I think the biggest problem you'll encounter with regards to the substance
is finding it in the correct purity.  While the oil itself may be ideal,
most easily available products will have other junk mixed in that will have
adverse effects.

The biggest problem in practice that I see is that you'll have a large vat
of oil to deal with.  Changing parts and maintenance will be difficult and
potentially very, very messy.  Aside from leaks, that's another reason to
use a safe liquid.  Anytime you need to work on the system it will be much,
much easier if you do not have to be concerned with the health and/or
environmental risks.  It will also be quite heavy and potentially difficult
to move around. My suggestion here would be to buy a fish tank.  

On the bright side... I imagine that the ambient heat exchange between the
oil and the outside environment should be enough that you can use simple
conductive/convection cooling.  ie, just stick the hottest part at the
bottom of the tank.  The oil heats, rises, cools, and descends.  If you
really need it, you can rig an external circulation system to either pump
the oil through a cooling system or to pump coolant through the oil. The
simple solution here would be to buy an old refridgerator.  You may also
wish to have a thermocouple to measure the temperature at the CPU where it
is most important to be cool.

All of this being said... I am aware of people trying to do this.  There
are rumors of success, though I've never actually seen it work in person. 
Have a look at

I have no idea how successful these guys were (I tend not to believe things
I read on the internet), or if their end product would be stable enough for
a long term system, but it looks fun.

Michael S. Pierce
Materials Science Division
Argonne National Laboratory

Current Queue | Current Queue for Engineering | Engineering archives

Try the links in the MadSci Library for more information on Engineering.

MadSci Home | Information | Search | Random Knowledge Generator | MadSci Archives | Mad Library | MAD Labs | MAD FAQs | Ask a ? | Join Us! | Help Support MadSci

MadSci Network,
© 1995-2006. All rights reserved.