New cooling liquid?

[dima y]

http://abclocal.go.com/wpvi/news/041..._sapphire.html

i know there are currently fluids available, but still. I wonder how much a gallon of this will cost

[Cathar]

Some google results:

http://www.tyco.com/tyco/press_relea...l.asp?prid=718

[gruntledweasel]

linky

Low boiling point (<50 degrees celsius). Notably nicer viscosity than H2O. I don't know enough to hazard a guess at its thermal conductivity/capacity.

[Cathar]

Originally Posted by gruntledweasel

linky

Low boiling point (<50 degrees celsius). Notably nicer viscosity than H2O. I don't know enough to hazard a guess at its thermal conductivity/capacity.

Excellent.

That low viscosity is especially attractive.

Has about 40% of the thermal capacity of water by volume, but given its low viscosity, it should be possible to get flow rates up high enough (>4LPM or so) to offset the loss in thermal capacity.

The thermal conductivity is the key element here though. Really need to know that to figure out if it could be better than water.

[bigben2k]

Given the formulae, CF3CF2C(O)CF(CF3)2, I'd have thought it would be denser, but I'm no chemist... Wouldn't there be a similar substance that we could compare it to, for a very rough ballpark figure?

Still, if this stuff is new, it's not likely to be readily available.

[jlrii]

Link to the stuff used HERE

[AngryAlpaca]

bigben2k: If there were a substance similar to it, we would have had this thread loooong ago.

[jlrii]

Maybe we should think about critical heat flux...Might work with restrictive blocks and the right fluid...perhaps R11 or 123??? Not much temperature rise there ...but it might be made to work.

EDIT:Spelling, content

[bigben2k]

Originally Posted by AngryAlpaca

bigben2k: If there were a substance similar to it, we would have had this thread loooong ago.

Actually, I'm sure that there is already something similar. I've looked into various other cooling liquids, and they're out there, it's just hard to get some decent info on them, just like this one... and they're also usually very expensive, which is why we don't usually have these threads in the first place...

[dima y]

well if it boild at less then 50C if you can make a pump less luquid cooling system like a heatpipe

[pelle76]

Originally Posted by dima y

well if it boild at less then 50C if you can make a pump less luquid cooling system like a heatpipe

My first thought to... It would not work for an overclocker, but maybe for a silent PC solution?

If it boils at 50C then i guess it could maybe hold the CPU at 60C? And then some passive heatsinks could take the heat away ontop of the Case?

a closed loop with no moving parts.

[Cathar]

Originally Posted by pelle76

My first thought to... It would not work for an overclocker, but maybe for a silent PC solution?

If it boils at 50C then i guess it could maybe hold the CPU at 60C? And then some passive heatsinks could take the heat away ontop of the Case?

a closed loop with no moving parts.

Uh, you're just describing a heat-pipe here.

[pelle76]

Originally Posted by Cathar

Uh, you're just describing a heat-pipe here.

Yepp, but with the advantage that it should not be quite easy to bulid due to the fact that the system would not have to be pressurized.


EDIT: Or maybe I'm totally out of line here... I did not give it so much thought before posting it...

It was my first thought

[jlrii]

It wouldn't actually be a heat pipe as it is not passive. The fluid would still be pumped but a portion of it would be allowed to flash to a gas when absorbing heat in the CPU block. This aproach could take advantage of both the thermal heat transfer properties of the liquid and the latent heat of vaporization. Getting a little more extreme, a positive displacement pump could be used, allowing a higher pressure drop at the CPU and more heat absorbed from vaporization. If you took it to the limit and used a compressor you have a system that absorbs all of the heat at the block via vaporization. So with a normal pump and pressures that are not sky high, it would sort of work like a cross between liquid cooling and phase change. The power of the pump would directly effect heat transfer in this case, as the more vacuum created at the block the more liquid would evaporate there, creating, in effect, a very mild phase change setup that works at close to ambient pressures.

[bigben2k]

You guys ought to spend some time trying to get a sample of this stuff...

[jlrii]

A tidbit of pricing info

Q. What is the price of Novec 1230 fluid?
A. The price of Novec 1230 has been established for
purchase by approved OEMs. However, we
anticipate that the overall cost to the end user for a
Novec 1230 fluid extinguisher or system will be
within the range of other in-kind alternatives that
are already on the market when the overall value
associated with its features and benefits are
considered. This includes consideration of its
sustainability as a viable fire protection technology
that can be expected to be available long into the
future.

Not much there I know.

I dunno how viable it would be for what I was describing earlier as it has a fairly low heat of vaporization.

[jlrii]

This is what got me thinking about the subject.

[redleader]

Originally Posted by Cathar

Excellent.

That low viscosity is especially attractive.

Has about 40% of the thermal capacity of water by volume, but given its low viscosity, it should be possible to get flow rates up high enough (>4LPM or so) to offset the loss in thermal capacity.

The thermal conductivity is the key element here though. Really need to know that to figure out if it could be better than water.

Where does it say that?

[talcum]

Could you use this to efficiently cool the hot side of a TEC and maintain a 20 - 30 C cpu temp? For a Prescott it might be worth it. Or would bubbles form in the cooling block and ruin the heat transfer when the stuff vaporized? Anyone cooled a TEC with a phase change system?

Originally Posted by pelle76

My first thought to... It would not work for an overclocker, but maybe for a silent PC solution?

If it boils at 50C then i guess it could maybe hold the CPU at 60C? And then some passive heatsinks could take the heat away ontop of the Case?

a closed loop with no moving parts.

:shrug:

[jlrii]

Originally Posted by talcum

Could you use this to efficiently cool the hot side of a TEC and maintain a 20 - 30 C cpu temp? For a Prescott it might be worth it. Or would bubbles form in the cooling block and ruin the heat transfer when the stuff vaporized? Anyone cooled a TEC with a phase change system?


:shrug:

Heat is absorbed from vaporization so the bubbles may be a good thing. I am unsure how well it would work but they are experimenting with allowing bubbles to form from vaporization HERE . Granted the experiments deal with much higher temps than we see but changing the fluid used would drop vaporization into our range and possibly keep pressures low enough to avoid more expensive fittings, tubing, and other pieces-parts.

The fluid that began the thread would work but not too well as it only absorbs about 40 BTUs per pound as it vaporizes while R123 absorbs about 357 BTUs per pound. The pump used may become an issue with alternate coolants not only becuse of material compatibility issues, but also because the bubbles that form will be passing through the pump. Much potential for cavitation noise and bearing damage there I suppose.

Lots of stuff online about using this setup to cool anyting from electronics to nuclear reactors. I found a couple of real interesting articles that get down to the nitty gritty details on using this process with micro channels in th 1-2 mm range. It should work...how well it would work for this aplication and how expensive it would be are questionable.

In short, if the lessened heat carying capacity of the coolant used as compared to water is more than offset by the heat absorbed from vaporization there are gains to be made there. The systen may be viewed as a hybrid of liquid cooling and phase change. The power of the pump has the potential to increase cooling more so than with a system using water as the coolant as, with a restriction imposed before the block, the fluid cooling will be under a slight vacuum, lowering the temperature at which the coolant evaporates. The rad will be under slight pressure lowering the temperature at which the bubbles condense.

[flashover52]

I know this is digressing from the topic a little bit but I figured I might share this information in case anyone's interested based on personal experience working with the product. As mentioned above, Novec 1230 is made by 3M for use as a fire protection extinguishing agent - namely, a halon alternative. It is one of many chemical compounds that have excellent environmental properties when compared with other clean agents (ie. halon alternatives). As for price, this is negotiated with thier OEM's but you could easily guesstimate anywhere between $12-20 per pound depending upon quantity. At 13 lbs/gallon, it won't be cheap for the common home or even power computer user.

With that said, there are a number of other chemical compounds that would be appropriate for this application that might be a bit more reasonable. I have some personal contacts that work on the Research & Development side at 3M on products used for refridgeration/cooling and would be willing to help anyone interested on a particular application and is something that my firm is starting research on. In fact, the Crayes supercomputer is operated in a 3M di-electric cooling fluid....

As for experience with pumps, we have just completed two years of research pumping Novec 1230 and I can attest that flows beyond 50 gpm are possible provided the right choices in piping and pump selection are performed to compensate for the slight vapor pressure. With a SG=1.6, it's a 'bit' heavier than water but with dielectric strength up near 60 KV, contact with electronics is not an issue. Similar 'engineering' is necessary for the other non-fire protection products but pumping shouldn't be a problem for smaller flows as long as you can prevent cavitation.

-David
http://www.capsfire.com

[Colt357tw]

Ehh, guys, are you thinking thermo syphon?

http://www.hpl.hp.com/research/paper...ermosyphon.pdf

[flashover52]

Originally Posted by Colt357tw

Ehh, guys, are you thinking thermo syphon?

http://www.hpl.hp.com/research/paper...ermosyphon.pdf

The fluid referenced in the article above PF5060 is yet another of 3M's performance fluids. As a former 3Mer, this was also another product that was used for cooling applications and fire protection. However, the environmental properties on PF5060 (or the other PF's) aren't nearly as good as the hyrdofluoroethers.

If anyone is interested in working together on a joint research project, I would be more than happy to help furnish some pump hardware and the appropriate fluid.

-dms

[RedPhoenix]

This?

Its sold by Dtek. I think its what your talking about.

[killernoodle]

I doubt you can submerge a computer in that stuff. It may not be conductive over a long period of time. I'd like to see OEM interest in immersion cooling outside of CRAY

EDIT: Just found a neat link on ebay for anyone willing to try immersion cooling: http://cgi.ebay.com/ws/eBayISAPI.dll...sPageName=WDVW
This stuff is commonly used in transformers as an electrical insulator and to keep them cool. Even if it doesnt have the thermal capacity or conductivity of water, it is still better than air.

Worth a try, I cant try it because I'm going to school soon :shrug: