In article ,
Jeff Liebermann wrote:

On Sat, 03 Jan 2009 18:53:59 -0800, Jeff Liebermann
wrote:

Thermal Conductivity Viscosity
W/mK cSt @20c
Water 0.6 0.9
AF (glycol) 0.24 2.0
Water+AF 0.8(?) 1.5 50%/50%
Silicon Oil 0.1 varies radically
Mineral Oil 0.138 34.5
Fluorinert FC-77 0.063 0.75

Air 0.025
Copper 370.
Diamond 1000.

Ok, I see why. Water has 1/5th of the thermal conductivity of mineral
oil. 50/50 water and antifreeze won't work. That raises the boiling
point but ruins the thermal conductivity. Pure ethylene glycol looks
tolerable. Other than the health and ecology issues, any reason that
100% antifreeze won't work?

Sorry, I goofed. Vicodin etc. It should be the higher the W/mK, the
better the thermal conductivity.

So why is Fluorinert favored for cooling when it has such a lousy
thermal conductivity?

High resistivity
High dielectric strength
Low maintenance
Low corrosion = very compatible with most materials
Leaks don't cause more damage
low viscosity
Wide useful temperature range

http://multimedia.3m.com/mws/mediawe...E5X46EVuQEcuZg
Vs6EVs6E666666--
Since it's usually circulated with a pump and involves direct
immersion, is it because of it's low viscosity and superior electrical
characteristics?

It would also appear that water has 5 times the thermal conductivity
than mineral oil. So, why use mineral oil?

DI water is very corrosive and requires stainless steel heat transfer
radiators among other components. It is difficult to keep uncontaminated
and you have to keep changing filters for example. Get a leak and it
usually causes additional damage.

Mineral oil makes a mess and is old technology.

I know of one ATE manufacture that used DI water for cooling in the
mainframe and test head. What a mistake that was. Every time they had a
leak in the test head expensive boards got damaged.

--
Telamon
Ventura, California