[Ganf]

Any of you fellas with the appropriate job descriptions ever get the opportunity to snag a gallon or two of this stuff and experiment with it?

There's no question it'd work. The various types are used for everything from cooling military supercomputers to motor vehicles.

If you have, I'd like to know which formula you tried, if you chose it for specific reasons, and what you learned from the experience such as rates of evaporation, issues with pressure buildup if any, quirks, advantages and disadvantages over distilled, etc...

I'm thinking about doing some sleuthing to dig up some consumer-sized quantities of this stuff to play with in my next build, and I'm interested to hear if it's going to behave as I expect.

 

[u3b3rg33k]

How exactly do you want to use the stuff?

 

[Ganf]

Standard OC'd build. If I still lived up north I might have looked into dropping a radiator out the window since I'd never have to worry about the Novec freezing, but since it only hits 30 degrees once or twice a year here I'll have to settle with just amusing myself by comparing it's performance to distilled water.

 

[u3b3rg33k]

well if you just want to fill a loop with it, use the same mix you use in your car - water/antifreeze mix. comes with corrosion inhibitors, and a cool green/blue/yellow/orange color.


good down to -80F. if that freezes, you have a problem, and that problem is not enough overclock.

 

[Ganf]

You're headed in the wrong direction. Antifreeze has a lower thermal conductivity than water and a higher boiling point. The cooling formulas of Novec have a higher thermal conductivity and lower boiling point.

I could use alcohol and get similar results, but alcohol is flammable and I like my couch unburnt. My house, also. Alcohol would also evaporate through the seals quite a bit faster than Novec if I'm reading the specs right. I say a standard OC'd watercooled build, but I've had thoughts that an atypical high speed pump or a chiller may be necessary to prevent it from boiling off too quickly. Hard lines are going to be mandatory, and an adjustable pressure valve to avoid anything busting or burning up the pump if I do something stupid like walk away to go grocery shopping and accidentally leave IBT and FurMark running simultaneously for the duration.

This will be well shy of requiring anything like a condenser. That type of closed loop heat exchange is designed for near 0% coolant loss over an indefinite period. I don't expect a gallon of Novec to last me more than a couple months, and that's assuming I pick the optimal formula and have everything set up correctly.

I'll likely screw this up and vaporize it in a week.

But when I'm done playing I'll have a standard loop to fill up with water.

 

[u3b3rg33k]

Quote:

Originally Posted by Ganf

You're headed in the wrong direction. Antifreeze has a lower thermal conductivity than water and a higher boiling point. The cooling formulas of Novec have a higher thermal conductivity and lower boiling point.

What are you trying to do exactly?

Water is 1 calorie/gram °C
novek 7100DL is 0.28 calorie/gram °C
http://multimedia.3m.com/mws/media/104167O/3mtm-novectm-7100dl-engineered-fluid.pdf?fn=prodinfo_nvc7100dl.pdf
ethylene glycol in a 50-65% mix with water only has a 20-28% respective penalty for heat capacity, which still blows Novek out of the water. pun not originally intended, but kept.
i don't see thermal conductivity listed for 7100DL.

which version of the fluid are you thinking of using? if you can present a scenario in which Novek makes more sense, I'll be excited to see what you come up with. otherwise, I may remain rather skeptical.

 

[kevmatic]

Wait, wait, so you want to use the novec as the working fluid is some sort of strange, high-temp phase change system?

It'll put a fairly hard limit of 61C-65C or so on the temp of of the CPU block (depending on pressure), but water does way better than that!

Unless you plan on using it like a bong cooler? If that's the case, you'll likely run through a gallon of novec in a few minutes thanks to its low heat capacity.

Otherwise, there's no benefit to a lower boiling point. If you're trying to keep the stuff in liquid form, then the lower boiling point and heat capacity will work against you... You'll have to keep the flow up very high- 4x as much as with water.

 

[Ganf]

Quote:

Originally Posted by kevmatic

Wait, wait, so you want to use the novec as the working fluid is some sort of strange, high-temp phase change system?

It'll put a fairly hard limit of 61C-65C or so on the temp of of the CPU block (depending on pressure), but water does way better than that!

Unless you plan on using it like a bong cooler? If that's the case, you'll likely run through a gallon of novec in a few minutes thanks to its low heat capacity.

Otherwise, there's no benefit to a lower boiling point. If you're trying to keep the stuff in liquid form, then the lower boiling point and heat capacity will work against you... You'll have to keep the flow up very high- 4x as much as with water.

No phase change, just replacing the water. No pressure beyond maybe 10 or 15 lbs. Haven't decided on that, it'll depend on the pump. I know the standard issues like a D5 Vario will burn out in minutes with any amount of pressure in the system. Some of that flow rate will be accounted for by the lower viscosity. The thermal capacity isn't the important number, it's the speed of the exchange, or the thermal conductivity if you prefer. Water doesn't conduct 100% of the heat available in any given water block for the volume traveling through the block, nothing near that, in fact. That is why the Delta of your water is significantly lower than the Delta of your CPU or GPU, and in many setups is often close to ambient.

Standard watercooling pumps already move water through the blocks too fast for them to pick up the entire 1 calorie per gram, but disregarding that the higher viscosity and higher thermal conductivity should guarantee an interesting result. The point, however, is not to guarantee lower cooling temperatures. It's to find out the practicality and efficiency of using these industrial coolants in an everyday application.

However, If I wanted to dummyproof the entire thing and guarantee that I would get lower temperatures I would get a sturdy 120v horticulture pump and use the 7600 fluid, which has a higher thermal conductivity AND boiling point than water, lower viscosity than 50/50 Ethylene Glycol and dielectric properties comparable to neoprene rubber. Then I could drop a lump of dirt in the resevoir, mess up the TIM on every block, push all of the components to their thermal limits, refuse to fix a minimum of 4 catastrophic leaks and get the same experience as if I was using water in a non-derpy loop.

But that's not the point. The point is that if I use one of the fluids with a low boiling point and put a mechanical pressure release valve in the loop, I can then take a small, cheap MP3 player, butcher the play switch on it, place it so that the mechanical valve presses it, and hear any of these sound clips every time my PC blows off some steam.

Fun is the word of the day. I like messing around with stuff like this when I get the chance, it doesn't necessarily have to be the most efficient or the most effective method, it just has to appeal to my sense of novelty.

Edit: Link to the tech data sheet for all Novec and Flourinert variants, including some that are used in missile warheads. Just being able to say that your liquid cooling loop is built to milspec is enough to warrant the trouble.

 

[lin2dev]

I would use 649 because of the low boiling point, you would have to be careful since your loop just got turned into a pressure vessel, but other than that, it would transfer a lot more heat than water because of the phase change.

added advantages would be the ability to use significantly less fluid

3M Novec 649
Heat of Vaporization: 88 kJ/kg
Specific Heat: 1.103 kJ/kg-K
Density: 1.6 kg/L

heating from 20C to 80C
132 kJ/kg

the transition from liquid to gas uses the same energy as raising the liquid by 80 degrees

Water
Specific Heat 4.184 kJ/kg-Kr
Density 1kg/L

heating from 20C to 60C
167 kJ/kg

note that these are not volumetric, multiply by 1/Density for volumetric quantities

 

[Ganf]

Quote:

Originally Posted by lin2dev

I would use 649 because of the low boiling point, you would have to be careful since your loop just got turned into a pressure vessel, but other than that, it would transfer a lot more heat than water because of the phase change.

I would rather any phase change be incidental and find a way to keep the coolant stable whether through loop organization, flow rate or maybe I'll just throw a glass beaker in between the components and the pressure valve to give it a chance to condense naturally. I'm not going to be wholly scientific in my comparisons for sure, but I will at least try to compare apples to apples n' such. Who knows where this will lead, but I am going to start digging for someone to get a gallon or two from before my interest wanes.

 

[lin2dev]

Quote:

Originally Posted by Ganf

I would rather any phase change be incidental and find a way to keep the coolant stable whether through loop organization, flow rate or maybe I'll just throw a glass beaker in between the components and the pressure valve to give it a chance to condense naturally. I'm not going to be wholly scientific in my comparisons for sure, but I will at least try to compare apples to apples n' such. Who knows where this will lead, but I am going to start digging for someone to get a gallon or two from before my interest wanes.

If you would like to avoid phase change, then there are no advantages over water and quite a few disadvantages:

• higher viscosity
• lower heat transfer coefficient
• lower specific heat
• higher density

So it pretty much defeats the purpose of high molar mass liquids

 

[u3b3rg33k]

Quote:

Originally Posted by Ganf

No phase change, just replacing the water. No pressure beyond maybe 10 or 15 lbs.

You realize most watercooling gear isn't rated for 10-15 psi, right? that's hardly "no pressure".

This should be dropped here as well:
http://koolance.com/cooling101-heat-transfer

by the way, mercury has AMAZING thermal conductivity. sure, it's toxic, but the novec fluids are bad for the environment, too (they cause global warming).

diamonds and mercury, man! get with the program!

 

[Ganf]

Quote:

Originally Posted by u3b3rg33k

You realize most watercooling gear isn't rated for 10-15 psi, right? that's hardly "no pressure".

This should be dropped here as well:
http://koolance.com/cooling101-heat-transfer

by the way, mercury has AMAZING thermal conductivity. sure, it's toxic, but the novec fluids are bad for the environment, too (they cause global warming).

diamonds and mercury, man! get with the program!

Mercury vapor makes me nervous, or I might have already tried it.

A gallon of Novec has less global warming potential than a cow that's been fed a 50lb bag of grain, so I'm okay with this.

Edit: And yes, some of the flourinert formulas are pretty awful and just used for their very high stability and ability to maintain purity, I understand. FC-43 is almost a direct replacement for tool oil, and not water by any means.

Pretty much all of your radiators are tested to 40+ PSI before being boxed up, so as long as you aren't foolish enough to use soft tubing and a flush mount pump with no traditional bearings on the shaft, you'll be okay up to 25 PSI. Even if the blocks leak that's easily solved with a tube of automotive silicone, but that won't happen unless they were milled wrong or given the wrong O-rings to begin with.

Lin2dev. I suggest you look things over again. Every Novec formula outperforms it's consumer counterpart in spades, and all of that tech data is in single phase heat exchange tests. Try to keep in mind that you're not arguing with me, you're arguing with the engineers who decide to use these liquids over something simple like distilled water with no gas traps and a high purity filtration system, the cost of which to build is less than what it costs to fill up a server rack with Novec and more easily sustained.

It's nice to look at the tech sheets and argue, but at the end of the day Novec is being bought and used by people whose professional careers are built around knowing the advantages and disadvantages of these products and how to use them in applications that are far beyond anything you will find in a PC. I didn't start the thread to argue the pros and cons of Novec over anything else, to sing it's praises or to convince anyone that this is the coolant of the future. I asked if anyone else has tried it, because regardless of whether it's going to be better or worse than a bottle of boiled water I am going to try it.

If in doubt, please refer to the OP.

 

[HITTI]

Quote:

Originally Posted by Ganf

Any of you fellas with the appropriate job descriptions ever get the opportunity to snag a gallon or two of this stuff ?

Quote:

Originally Posted by B NEGATIVE

Yes. I also have couple of liters for sale....

 

[u3b3rg33k]

Quote:

Originally Posted by Ganf

No phase change, just replacing the water. No pressure beyond maybe 10 or 15 lbs. Haven't decided on that, it'll depend on the pump. I know the standard issues like a D5 Vario will burn out in minutes with any amount of pressure in the system. Some of that flow rate will be accounted for by the lower viscosity. The thermal capacity isn't the important number, it's the speed of the exchange, or the thermal conductivity if you prefer. Water doesn't conduct 100% of the heat available in any given water block for the volume traveling through the block, nothing near that, in fact. That is why the Delta of your water is significantly lower than the Delta of your CPU or GPU, and in many setups is often close to ambient.

Standard watercooling pumps already move water through the blocks too fast for them to pick up the entire 1 calorie per gram, but disregarding that the higher viscosity and higher thermal conductivity should guarantee an interesting result. The point, however, is not to guarantee lower cooling temperatures. It's to find out the practicality and efficiency of using these industrial coolants in an everyday application.

However, If I wanted to dummyproof the entire thing and guarantee that I would get lower temperatures I would get a sturdy 120v horticulture pump and use the 7600 fluid, which has a higher thermal conductivity AND boiling point than water, lower viscosity than 50/50 Ethylene Glycol and dielectric properties comparable to neoprene rubber. Then I could drop a lump of dirt in the resevoir, mess up the TIM on every block, push all of the components to their thermal limits, refuse to fix a minimum of 4 catastrophic leaks and get the same experience as if I was using water in a non-derpy loop.

But that's not the point. The point is that if I use one of the fluids with a low boiling point and put a mechanical pressure release valve in the loop, I can then take a small, cheap MP3 player, butcher the play switch on it, place it so that the mechanical valve presses it, and hear any of these sound clips every time my PC blows off some steam.

FWIW, that tech data sheet you listed compares 3m's fluid to PAOs and other oils - no where do they make the claim that it will outperform water.

if in fact you did mess up the TIM on all your blocks, it wouldn't matter what fluid you used, because you wouldn't be able to get the heat out of the chips, so no, it would NOT still give you the same performance as a non-derptastic water loop.

as to your remarks regarding water's performance in cooling systems, I can't tell if you're joking around, serious, or just curious about what happens when you use random fluids with watercooling gear.

if you are serious, you should be able to use the specs and calculate roughly what would happen, and then verify/disprove your hypothesis with lab(project) testing. The laws of thermodynamics don't change when you pick different fluids, for the most part just the coefficients on your variables change.

 

[Ganf]

Quote:

Originally Posted by u3b3rg33k

as to your remarks regarding water's performance in cooling systems, I can't tell if you're joking around, serious, or just curious about what happens when you use random fluids with watercooling gear.

A bit of both. If I was intending to see whether these fluids could replace or beat water I'd be building the loop around their use, and not around the use of water, which is what the Novec will be replaced with after I've had my fun.

I'm more interested in the solvent/heat exchange hybrids for playing with in this manner, and you've got to remember that fluid dynamics are a key factor also, not just thermodynamics. It's also a bit of curiosity after having plenty of hands-on experience with ethyl-ketones and wondering how the two compare.

But anyways, I got my answer. There is someone who has used it, so unless any of you would like to counter-argue with his hands-on experience we've "discussed" all we can discuss.

 

[u3b3rg33k]

This may amuse you:
http://www.egr.msu.edu/mueller/NMReferences/KiliscarslanMueller2004CMES_COPsR718inComparisonWithRefrigerants.pdf

likely explains why they use water vapor in heat pipes and vapor chambers (instead of some other refrigerant), environmental and cost benefits notwithstanding (water wins here hands down):
Quote:

CONCLUSIONS
Water as a refrigerant (R718) is compared with current refrigerants including R717, R290, R134a, R12, R22, and R152a by using a created computer code for calculations of a simple vapor compression refrigeration cycle
...
The computed results show that the use of water as a refrigerant can result in a higher coefficient of power (COP) than if the other refrigerants are used. From the presented results, it can be concluded that for evaporator temperatures above 35 ̊C the highest COP can always be obtained with R718.

in short: water is awesome.

 

[Jollyriffic]

novec 7000 boils at 34c
this is more for a submerged build than putting it in anything like a rad.
using this you shouldnt run a heatsink.

once the novec boils you need to convert the gas back to liquid, so you need a heatsink above the liquid attached to a tec to drop the temps to convert back to a liquid.
also the entire build would need to be liquid and gas sealed.
liquid is easy to seal, gas not as easy.

i spoke with phil over at 3m back on 7/31/12 to see if i was missing any info about doing this.
his email [email protected]

no idea if he still works there.

novec 7000 is around $20 a gallon, if i remember correctly.

here are a few videos of this in action.
https://www.youtube.com/user/petuma1 <--- phils youtube

my idea back then was to have a wall hung computer that looks like art but is actually functional.
never did get around to it, but wish i had.

 

[superstition222]

Quote:

Originally Posted by Jollyriffic

novec 7000 boils at 34c
this is more for a submerged build than putting it in anything like a rad.

There is a range.

7000 = 34C
7100 = 61C
7200 and 8200 = 76C
7300 = 98C
7500 = 128C
7700 = 167C

(fluorinert range is 50C - 174C)

recommended temperature range

7000 = -123 - 25C
7100 = -110 - 50C
7200 = -115 - 60C
7300 = -30 - 85C
7500 = -70C - 120C
7600 = -60C - 125C

Toxicity range. (All data is set by the manufacturer. Time weighted averages.)

7000 = 250 ppm
7100 = 750 ppm
7200 = 200 ppm
7300 = 100 ppm
7500 = 100 ppm
7700 = >100 ppm (NOAEL 6 hr/day for 5 days, highest dose tested)

Global warming potential rating.

7000 = 420
7100 = 297
7200 = 59
7300 = 200
7500 = 90-100
7600 = 700
7700 = 420

Thermal conductivity.

7000 = .075
7100 = .069
7200 = .068
7300 = .063
7500 = .065
7600 = .071
7700 = .069

(fluorinert range is .057 - .066)