Take the water out of watercooling ...

[Olly]

There's an article in the latest edition of Custom PC which reviews a new Armari PC which uses 3M Fluorinert instead of water in it's cooling loop.

More interestingly, it suggests that Armari are planning to sell Fluorinert itself in the near future (admittedly at £54 for 500ml !)

It raises some interesting possibilities ... :D

Olly

[adwhitworth]

BAH. They used a Koolance system though :rolleyes: :angry: -_-

Besides, where's the fun in having a sealed system. The diy jobs are the best!

[Edd]

Wow, I was thinking of getting some Fluorinert after reading that article, but if it's that expensive I'll pass!

[Starbuck3733T]

the point?

was it b/c flourinert is INERT?

or were they chilling the coolant?

[Pug]

I need to read the article before passing too much comment but I'm with Star on this.
Fluorinert's strength is its non-conductivity, not its heat transfer properties. It's generally used for submersive cooling because of this.

Did they quote its thermal efficiency in relation to water anywhere? Or are they just saying it's safer to use if you have a leaky system?

[furious trout]

It was the safety angle, they're pushing - no comment on it thermal qualities so one can only assume it's ok ( i suppose if it was worse they wouldn't be singing it's praises ) , having said that they're a funny bunch journalists so who knows :p

[Turbosatan]

Says the coolant can be chilled further than water (-180 degrees freezing point) it doesnt carry any of the organisms lime etc that water does so will stop harmful buildups and they sprayed it all over the inside of a machine and it worked fine . i think even at that price with the right colling unit it would be worthwhile.

HS

[fillip]

welcome to the forums turbosatan.

i've not seen the article either - i'm now trying a google for info. but i still dont think that you could justify the price of it right now. Not when you consider you'll need prob a litre for a w/c setup if not more on the larger setups.

Ok, so it doesnt have organic properties that cause harmful build ups like lime - but thats why we distilled water and coolant solutions/anti freeze which goes for a few quid a bottle and assists in lowering system temps.

As for system safety, well anyone who is capable of putting a half decent setup together can ensure for the most part it will never leak - even if it does, providing it isnt too severe, then the system should be ok if turned off immediately and left to dry out (this ofcourse assumes that deionised water was used)

To reach those sub zero temps you're gonna needs some kind of chiller/*thingy from a fridge - name escapes me now* to get the fluid to that temp.

So we're looking at a vapochill system with a litre of Fluorinert or some other high quality components... my pocket money doesnt cover that :(

[Starbuck3733T]

Ahh, but what is the thermal coefficient of flourinert?

[Grand Monkey]

Good thread this, i was gonna post something similar. One of my mates is being tempted to water cool and was being put off by the thoughts it might leak, then he read about this flourinert, and how they managed to start a machine even though the soaked it with it. Now i said from talking to you guys, i hadn't heard any of you losing components to leaking. But he still was determined this was the route he was going, as it gave him that extra piece of mind. Perhaps what we may be seeing is a further drive of water cooling becoming mainstream. I mean look how many manufactures are now selling water cooling kits this year (coolermaster for one!). If they can now turn round and say," if you use this with your system there is no risk of you destroying your compenents if it leaks", i think this may persuade people to go towards it.

[Da_Rude_Baboon]

This has come up loads of times on Bit-Tech. Fluorinert itself is used in super computers to cool them, or more precisely super computers sit in a chamber FILLED with Flourinert. :o

I dont know what the thermal properties of it are but i beleive it does provide cooling at least equal to water. Theres loads of different types of fluorinert as seen HERE (http://products3.3m.com/catalog/us/en001/manufacturing_industry/specialty_materials/node_W1JMJGXGVWbe/root_GST1T4S9TCgv/vroot_FG8FTD9L7Wge/gvel_RN2BNC4D33gl/theme_us_specialtymaterials_3_0/command_AbcPageHandler/output_html) but with out knowing the particular type i cant give you any more info. As previously mentioned its the fact its inert which makes it attractive but then i though that was the point of using De-Ionised water? <_<

The main problem with fluorinert is that its rather toxic and needs to be disposed of properly, i.e. you cant pour it down the sink. I think as far as things go its VERY damaging to the enviroment, something which CustomPC neglected to mention.

[Grand Monkey]

custom pc mentions it as 3M Fluorinert PF5080, the website you linked to rude baboon, only says this about it
3M™ Performance Fluid PF-5080&nbsp;
PF-5080 is an ozone depletion substance replacement for heat transfer applications.&nbsp;

The Armari website says this
The Armari T800-LQX uses state of the art InertX F9090 liquid cooling.
Not a whole lot of info.

[Pug]

Welcome to WizD Turbosatan. :)

Fluorinert does have its uses, I think we'd all agree with that.
The question is, unless you are chilling or planning to submerse things, why spend that extra £50?
As fillip pointed out, we tend to use distilled/deionised water to prevent the limescale/organic buildup which can manifest itself when using tap water - that's one of the well-known basics of liquid cooling.

I've read the article now. -180C is pretty unrealistic even with a chiller and that's not just because of the liquid in the loop. There are the other components to consider too.
I've taken the older Cuplex Evo blocks to subzero temps but I wouldn't expect to get below -40C without problems and I doubt very much that the Koolance block would get as far as the Evo, let alone the tubing they use in that particular setup... :blink:

Also, even before you reach zero degrees, you have the dew point to cross - consider the following picture; that's 8/5.5mm tubing sleeved inside 12/10mm to provide a degree of insulation -

http://www.wizarddesigns.co.uk/images/chillDew.jpg

That water didn't come from the coolant, y'know... :lol:

I think the main benefit of using it in this particular system is to cover Armari's asses warranty-wise, tbh.
As it's inert, they won't encounter galvanic corrosion issues long term (we use additives to combat this anyway) and as it's non-conductive, even a leak wouldn't compromise them too much in terms of replacement parts.
I doused my 9800Pro good style yesterday (hadn't pushed a pipe fully home and was rushing things a little) but as the power was off, no problem. Dried it off, fired up and away I went.


Moral? Don't connect power till you're certain your loop is tight.
The majority of leaks I've ever had were during the initial fill. I've only ever experienced leaks after leak-testing if I abuse or stress the pipes in some way (and there's no sense in doing that anyway :rolleyes: ).

[Nexxo]

I've been looking into this. There's a company (TMC Industries, Inc.) which reclaims and purifies to "as new" a range of chemicals. They may be able to sell recycled Fluorinert (of the same purity as new stuff) at a much cheaper price.

The particular type you'd want is Fluorinert FC-74, by the way.

Now I'm not sure about it's toxicity. That would suggest reactance, and isn't it supposed to be inert? Animals have breathed the stuff, after all! Somehow I don't think Armari is going to sell it in their PCs if it is highly toxic.

It's tempting to have a short-circuit, corrosion and maintenance-free system, and £54,-- for a 500ml bottle (the content of an average WC rig) seems reasonable for that. However I think it can be bought cheaper. I'm exploring it... watch this space.

[mrplow]

Originally posted by Pug@Mar 31 2004, 12:47 PM
http://www.wizarddesigns.co.uk/images/chillDew.jpg
Is that all condensation?

:o !

[Da_Rude_Baboon]

Originally posted by Nexxo@Apr 5 2004, 08:08 PM
Now I'm not sure about it's toxicity. That would suggest reactance, and isn't it supposed to be inert? Animals have breathed the stuff, after all! Somehow I don't think Armari is going to sell it in their PCs if it is highly toxic.


Fair comment Nexxo as i'm just working off of memory from a link i read a while ago and there seems to be a whole range of the stuff. Also toxic to the enviroment and toxic to us are two different things. After all there is currently the new legislation about disposing of old PCs due to their enviromental impact but i've yet to meet anyone poisend by their hardware. :p

As for the animals breathing it is that the pink stuff they used to submerge mice in? i remember when that was on tomorrows world and all the science shows saying how it could be used for deep sea divers etc. Movie buffs might also remember it from the film The Abyss too. ;)

If memory serves me correctly, which it ussually doesnt, didnt they discover it kills you? Which is why you dont see it around anymore.

:EDIT:

The breathable liquid is made by 3M and is called FX-80 or the rather catchy perfluorobutyltetrahydrofuran.

A rather tenuious link HERE (http://www.totse.com/en/technology/science_technology/bpool.html) but ignore the part about Alien abductions lol.

:EDIT: [/off topic]

[Nexxo]

Your memory is not that bad, Rude. ;) Indeed there is a version of Fluorinert that is considered an environmental hazard (Perfluoroctane Sulfonamide) but that appears to be a different substance.

And Fluorinert did indeed kill the rats in the end... Thing is it allowed you to breathe, but it needed to have the correct vapour pressure and to be highly purified; long term exposure led to the bends (ironically) and metabolic poisoning.

Now the confusion around this stuff is that it has so many names and functions:

3M™ Performance Fluid PF-5080*
PF-5080 is an ozone depletion substance replacement for heat transfer applications.*

The Armari T800-LQX uses state of the art InertX F9090 liquid cooling.

:blink:

Well, as I said the Cray computer used Fluorinert FC-74. This was later replaced by FC-77, which is mostly perfluoroctane. Fluorinert is 3M's brand name for the stuff; another company (Ausimont) calls it Galden.

PF-5080 is also... wait for it... perfluoroctane. Same thing. :D

As TCM Industries explains:

"Chemically its exactly the same* C8 F18, chemical name Perfluorooctane.
The FC fluids by 3M have a tighter spec than the PF fluids, this is why we
sell them under the chemical name. All our reclaimed materials are 99% pure
or better unless noted."

Wait, back up! Who's TCM Industries?

Well, that's the thing. They are a company that reclaim and purify various compounds, one of which is PF-5080 a.k.a. Perflouroctane a.k.a. Fluorinert FC-77:

Good Morning Robin,

Yes, we do sell PF 5080. However we sell it as reclaimed Perfluoroocatne.
The pricing on this material is $16.00 per lb. And we do have a $100.00 min
purchase.

Best regards,

Robert Ansell
Eastern Sales Manager
TMC Industries, Inc.
1423 Mill Lane
Waconia,Mn. 55387
Phone: 800-255-5789
Fax: 952-442-1160
E-mail:rob@tmcindustries.com
Web site: www.tmcindustries.com

Let's recalculate: $16,-- for about 500ml (the content of an average WC rig) is about £8,82. A nice drop from £54,-- I must say. OK, there's a $100,-- minimum spend (still not much more than £60,--) but for that you get about 3 litres. :D

Anyone interested? :p

[Da_Rude_Baboon]

lol put me down a for a litre and i'll use the money i save to buy some mice... :o

[Grand Monkey]

nexxo resident chemical genius :D

[Da_Rude_Baboon]

Yeah i can handle fluorinert but i cant even pronounce the others!

[Starbuck3733T]

as much as i hate to endorse them one bit, Koolance has some interesting tables on specific heat and thermal conductivity for Flourinert (albeit FC-43-doh!) here:

http://www.koolance.com/technical/cooling101/002.html

[Nexxo]

Useful link, that. It also shows that Fluorinert is not nearly as good as water in absorbing heat... ah, well... back to distilled water and Zerex racing. :D

EDIT: Or is it? mclean007 on Bit-Tech.net comments:

Well, it's specific heat capacity is less (about half that of water IIRC) so yes, 1kg of water will absorb twice as much energy in heating up by a given temperature as fluorinert will. However, the correlary of this is that fluorinert will cool down by a larger amount when you remove a given quantity of energy from it (in your rad) so it isn't all bad. Additionally, spec heat capacity is measured in W/kg.K not W/m^3.K so, since (again IIRC) fluorinert is considerably heavier than the same volume of water, the same flow rate will result in a greater mass of fluorinert passing through your blocks. Therefore the fact that its spec heat capacity is less than that of water is misleading.

Argh... decisions, decisions... I still may give Fluorinert a try! :D

[Da_Rude_Baboon]

:wacko:

[Pug]

Nice research Nexxo. Maybe we'll see a resurgence in direct-die and targetted spray cooling if costs come down enough.

I had a link to a possibly cheaper alternative still the other day before my rig went down. If I can access the links I had (or find them again manually), I'll post 'em.

[Starbuck3733T]

Originally posted by Da_Rude_Baboon@Apr 12 2004, 08:05 AM
:wacko:
What he said!

[Da_Rude_Baboon]

Whats the viscosity (sp?) of flourinert like? If its heavier then water will it not need a heftier pump to move it around and maintain an equivalent flow rate?

[Nexxo]

The viscosity is the same as that of water. It is indeed heavier than water (mass, or weight, is unrelated to viscosity, or thickness, for those who are trying to follow this conversation), which means that initially the pump has more inertia to overcome. This may mean that flow takes perhaps a second (if that) longer to establish itself. After that, it has acquired momentum and the pump should not need to work any harder than with water.

[Da_Rude_Baboon]

Thanks Nexxo ;)

It does sound appealing and at a reasonable price i wouldnt mind trying it out. :)

[Starbuck3733T]

Originally posted by Nexxo@Apr 13 2004, 06:52 PM
The viscosity is the same as that of water. It is indeed heavier than water (mass, or weight, is unrelated to viscosity, or thickness, for those who are trying to follow this conversation), which means that initially the pump has more inertia to overcome. This may mean that flow takes perhaps a second (if that) longer to establish itself. After that, it has acquired momentum and the pump should not need to work any harder than with water.
In a system with a large reservior, such as mine, the coolant decelerates as it enters the res. Wouldn't the mass differential, in my case, mean that the pump would again have to accelerate the flourinert? I know it already does this with the water, but since the flourinert is heavier, wouldn't it mean an overall decrease in flow?

Now, in systems with a bleeder line and no reservior, I can follow what you're saying and I agree with it, since the fluid never slows down before hitting the inlet of the pump again.

[fillip]

Just out of interest, has anybody ever tried out FluidXP?

It's sold over at integrity pc for anyone who has never heard of it. It's cheaper than poor flurinert but allegedly has better heat transfer and dissipation properties than water.

Just wondered if anyone had ome more info. on it since IntegrityPC dont appear to carry any technical specs for it.

[gerbil47]

just got off the phone from tcm industries.
Their minimum order of pf5080 is 4.5litres which is $224 ,and shipping would be $40. At current exchange rates that works out at £32.83 per litre.
Any one fancy going Quarters in it?

[Starbuck3733T]

Welcome to the boards gerbil47! :cool:

I'm going to pass.

[Nexxo]

I might be interested...

[Da_Rude_Baboon]

I just bought a house so i cant i'm afraid. :mellow:

[Psykotik]

I was thinking about this a while ago, and decided (in my case) it isnt worth it, as my rad alone takes a litre and a bit! :o

And now you have to go and post a linky to a cheaper place! I swear this PC is trying to take over my life... :lol:

[iXXeon]

I just came across this thread, with read with interest some of the comments regarding perflurocarbon based liquid cooling. As I've been using this stuff for almost four years, I thought I could add some useful informed comments.

Just to give you an idea of cooling performance...
I'm writing this post on a liquid cooled PC-DL with dual overclocked 1MB cache Xeons (@3.6GHz), a 9800XT and northbridge all running with liquid blocks. The system is cooled by a Koolance EXOS and it's using InertX PF5080 perflurocarbon coolant.(www.InertX.com (http://www.InertX.com))
The machine currently holds the no.2 fastest PCMark2004 score on the futuremark orb site.

The Koolance EXOS never needs to be out of mode 1 (low noise mode) and the temp. now (i.e. under light load), is reading 35 deg.C. The point is perflurocarbon coolant works very well in place of water, even on a consummer friendly small bore pipe (6mm) liquid cooled system. If I fire up Seti@home on one cpu and run FarCry at the same time on the other, the Koolance will hit 42 deg.C tops. MBM is registering the CPU's temperature from the mobo about 4-6 degrees hotter than the EXOS disply. The EXOS temp probe is stuck to one edge of the underside of CPU 0's liquid block, and is designed to give you an idea of the coolant temperature. Using a Calex digital temperature prope to measure the real coolant temp. in the EXOS reservior, and hey presto it also reads 35 deg.C.
(MBM currently shows CPU0 @ 41deg.C & CPU1 @ 39deg.C)

Not bad when you consider the thermal load.!

There seems to be a lot incorrect information flying around on the web regarding the cooling performance of perflurocarbon coolants. Of course it's mainly by spread by people who have never actually used the stuff. I'm only trying to set the record straight here, as the advantages of using this coolant are numerious.

Re: it's use in Cray supercomputers.
Most Cray supercomputers actually use conduction-to-convection cooling.
A large aluminium coldplate in is direct contact with heat producing semiconductors (just like a PC water block) with Fluorinert FC-74 flowing through it. In recent times only the T90 series from around the mid to late 1990's used direct immersion cooling with FC-77 coolant, and Cray only built a handfull of those. Today there are very few T90's still in service, and the remaining T3E's and I believe the new X1 all use coldplate/Fluorinert cooling.

I'f anyone is interested in using this stuff for PC liquid cooling, I'll be happy to share what i know.

and for the Myth spreaders out there...
It's not toxic, unless you burn it. (usually above 200 deg.C)
and it doesn't destroy the ozone layer.

iXXeon

[fillip]

Thanks for the input, it's always nice to hear from ppl who have real world experience rather than 'number crunchers' who base their comments on pure theory alone.

However, i notice that you were using the Koolance system, which obviously requires a huge chunk of cooling equipment sitting outside your main box. If we were to take flurinert and stick it in a fully integrated system (i.e. no external cooling components) would it actually show more in terms of cooling performance than distilled water and antifreeze combo?

Obviously the benefits of using a non-conductive compound are clear, especially when crammed in a PC case, but most modern w/c equipment is of high enough quality to prevent any leaks at all if it is assembled in the correct manner and the necessary precautions taken.

So then, why spend so much more on the flurinert than on water? any experience of cooling that dual setup you have there with just plane jane water and antifreeze??

[Pug]

Hi iXXeon. Welcome to WizD. :cool:
Nice first post. Appreciate the input.


Obviously, we'd all be keen to share whatever experience you've had with any aspects of liquid cooling (inert or otherwise). I hope you find this place receptive enough to your findings to stick around & share them all. :)

I've personally trashed more hardware (ok, it was just an MS Natural Elite) by spilling coffee on it (or shorting metal parts) than coolant leaks but if there's no impact on cooling performance (or better still, an improvement) then I'd be keen to give it a go.

As with anything, economy of scale could bring the price of this stuff down a lot lower if there's a greater demand for it, so pioneers like yourself do the community a great service by helping to demystify things through actual real world findings and reassure people that it might be researching further (or not, as the case may be).

Thanks again. ^_^


As a side note though - it's difficult (and inadvisable) to compare readings between different types of temp sensor. I found this myself when taking a block to sub-zero levels and noticing a huge difference between what the BIOS was reading and the output from my Fluke 52 (see this post (http://forums.bit-tech.net/showpost.php?p=395384&postcount=10) on Bit)

[Psykotik]

Must resist experimental direct die cooling with Fluorinert on spare Barton.....*struggles pathetically with will*

[fillip]

Must resist experimental direct die cooling with Fluorinert on spare Barton.....*struggles pathetically with will*
I'm sure i read somewhere that the problem with direct die cooling was that after a while the liquid, no matter how inert, saturated the silicon wafers for the whole chip and it basically broke down. I dunno if anyone can confirm/deny this, but when i first heard about flurinert i had the exact same thought about direct die cooling with it.

[Psykotik]

Was just reading a thread on the [H]ard forums that mentioned something along those lines.

Bum clouds :(

[ACCS]

Most Cray supercomputers actually use conduction-to-convection cooling. A large aluminium coldplate in is direct contact with heat producing semiconductors (just like a PC water block) with Fluorinert FC-74 flowing through it. In recent times only the T90 series from around the mid to late 1990's used direct immersion cooling with FC-77 coolant, and Cray only built a handfull of those. Today there are very few T90's still in service, and the remaining T3E's and I believe the new X1 all use coldplate/Fluorinert cooling.The X1 sprays 'nert onto the MCM (Multi-Chip Module) that houses the CPUs, and uses a phase-change as the primary cooling method. Yes. Phase-change 'nert. It seems to work well at cooling the CPUs and the RAMBUS memory.

[Pug]

Hi ACCS. Recognise the name from Ars, I think.
Welcome to WizD. :cool:

I believe HP were investigating a similar technique to what you describe once using inkjet technology, where they sprayed a fine mist of coolant on the chip package (where it evaporated on contact)... can't remember reading how they were to collect & recirculate it but I can only imagine they were working for a lossless system with it.
Be nice to see what came of it.

[Psymon]

I've been toying with the idea of Submersion cooling of a PC, and have contacted serveral vendors reguarding Fluids. My two compeditors were

Midel 7173 avalable at www.Midel.com
Seemed like it would work better then any other solutions, HOWEVER, it was not tested or recommended for printed circuitboards, and processors.

So Crap, I scrapped that idea, besides a rep from Midel contacted me and explained the cost, and shipping would be $150 USD, so $400-500 wasnt doable for me..

I looked into Flourinert,
THIS was the IDEAL substance for computer circuitry, it would work effective as a coolant, however ecomonics again, the price, well over 400 dollars. For an experiment, this just wasnt doable for me.

Even Recycled Fourinert would be 75% retail cost, but still too much.

I've Scrapped the idea for a Petier Cooled Fuild Submerged computer, I thought it would be a fun project, however projects that I like to do, don't hurt my wallet this much.

In the end, I think my Project will be instead, A Pelt cooled Processor, with a water cooled Pelt, then a waterbloock cooling the fluid that cools the pelt, and air cooling the pelt, that cools the fluid. The idea... The COOLER i keep the HOT side of my PROCESSOR PELT, the COOLER the COLD SIDE gets. The fluid doesn't need to get subzero, just cooler, So with a radiator, and a pelt to cool it.. I can get away with a smaller radiator, and a much more efficent system.
A word on condensation. I know this is the biggest fear of Pelt users. However with a well made Unsulation, and a quick coat over the circuitry with a Spray Polyeurathane (Mask off areas that need METAL contacts.. lol like ur SIMS hehe) and nonconductive grease under, and around the socket, I think it will be ok.

[Starbuck3733T]

Welcome aboard psymon!

Say hello in the Hellos and introductions thread. (http://http://www.wizdforums.co.uk/showthread.php?t=51)

I'm ultra paranoid, if I ever go refrigerated I think I'll have sensors for condensate located outside my insulation.

[toby]

this flourowhatsit does sound very interesting, because i am prone to a bodge from time to time and if i bodged a watercooling setup i would want to know that the pc was gonna be ok.

[fillip]

Nexxo, did you ever buy that 3 litres worth of Fluorinert from TCM?

[Psykotik]

I bought 4.5 litres from them tother week, should be here soon.
Gotta wait for me other stuff to come (never ending wait for C-Systems pumps at the mo) before I can get to use the wonderous fluid :(

[fillip]

let me know how it goes, how much did u fork out for 4.5 litres then?

[Psykotik]

About £180 -£200, cant remeber off top of me head! I sorta looked and clicked while cringing :lol:

[fillip]

About £180 -£200, cant remeber off top of me head! I sorta looked and clicked while cringing :lol:
:h34r:
crap!

and you actually need 4.5 litres of it?

[Psykotik]

You've seen the rad right? That bugger takes about a litre and a bit, and then there's the rest of the primary loop to fill!

That's without the secondary loop!

[Nexxo]

@ fillip: nope, in the end I'm getting mine from http://www.inertx.com/. They're a UK-based company that sells PF5080 florinert in 1/2kg (280ml) and 1kg (560ml) bottles, quoting me £70,44 for a 1kg bottle, inc. vat and carriage.

My system (pretty average size: three blocks, a pump and a BIX on 3/8"ID tubing) takes 550ml, so 1kg is all I need.

[Cowd Tyke]

Hi Nexxo,

I have been thinking along the same lines.
Just using a normal water cooling set-up with a couple of litres of Inertx-5080. Seems some of the loss of specific heat is balanced by the increased density - but the gain in lack of algae and furring up of surfaces in contact with water should prove some long term benefits.

Is your system built? Any problems? What pump are you using to move the 5080 around?
How are your temperatures?

Cowd Tyke

[Loki]

Hi, I just stumbled accross you forum searching for Fluorinert liquid FC-77.

Ive been thinking along the lines of a completly imersed system in a "fish tank" and I didnt really think to check if others where trying the same thing. Im getting to the point of actually buying the bits I need and starting my project.

The main part of my plan is to do away with the hard disk and go for a static ram option to allow increased performance. Help in finding one in the UK that wont cost the earth would be greatly recived.

I would be interested to know what you lot think will be the best external liquid cooler/pump system.

If anyone is interested in the project durring construction and final tests let me know and I will post a site or something.

[cleverbob]

I have 110 gallons of fluorinert FC-74. I de-installed a Cray C90 and got to keep everything. Any offers, just post. :eek:

[Nexxo]

cleverbob, where do you live? US? I'm looking for, say, a litre.

Cowd Tyke: My system (in the Chop Shop: Metaversa Heavy Angel) is not yet finished, but it's using a Davies Craig EBP 12V pump to move things along. It's supposed to be quiet and long-lasting.

[Psykotik]

I had to cancel my order :(

I actually measured the amount that went into my primary loop, and nearly cried as its over 4.5 litres :mad: :(

[Pug]

I have 110 gallons of fluorinert FC-74. I de-installed a Cray C90 and got to keep everything. Any offers, just post. :eek:
Bloody hell! Got any pics for posterity? Would love to see 'em.

Welcome to both yourself & Loki. Hope you can stick around. :)

Loki - you might want to look at a dehumidifier/chiller project to make the best use of things (assuming you're going for ultimate performance), unless you're going for the silent option, of course.
Let us know your main aims for the project and we'll try and advise accordingly. :)

Andy - 4.5 litres?! that's about three or four times the contents of all three of my main rigs put together! :p

[Starbuck3733T]

I'd take a gallon or two. Removes worries of leaks, that's for sure.

[cleverbob]

Good day! I am in the Detroit area. My e-mail is robertcleveringa@comcast.net.
I have a gentleman from Germany interested in buying the whole system, I could save some extra coolant. I do have pics if you contact me I can send them.

[Cowd Tyke]

hi fillip

i agree that water with additives is a proven method and cheaper

fluorinert (or what http://www.armari.co.uk/components.asp?SysID=168
are selling as InertX-5080 http://www.inertx.com/products/inert.html) has a Specific Heat Capacity around a 28% of that of water :unsure:

now the specific heat of a substance is the quantity of heat needed to raise the temperature of a unit mass of the substance 1 deg. C
inertx-5080 is 1.78 times as heavy as water (so volume for volume its about half as good as water at absorbing heat) :)
it has a slightly better thermal conductivity to water so may be better at shedding heat into the radiator(?) :)

over time the lack of inorganic and organic build-ups should mean the system efficiency doesn't drop (and doesn't need a periodic cleaning) :)

ignoring the property that it can be supercooled and leaks arent damaging (costly maybe) ...

...still gives a challenge of getting a traditional watercooling set-up working

has to be high flow as fluorinert is not as good volume for volume at absorbing heat so need to keep the volumes passing through the coolers higher than with water

needing a stronger pump to overcome the inertia of the heavier liquid

up and running for a reasonable cost - does that mean more efficient tubing routes and smallest radiator that works???

and challenges are fun :D

Cowd Tyke

[Nexxo]

BTW Armari Ltd. is now selling PF90, PF5070 and PF5080. Prices are cheaper by the litre than at Inertx.com (PF5080 is £58,-- litre instead of about £78,--).

[Fibbles]

Has anyone made the plunge and is now running with either of these fluids with good to great results?

[Psykotik]

I cancelled my order completely, and am gonna run water instead, as I cant be arsed to wait until I have enough in the bank to fund approximately £300 needed to fill ONE of my loops :eek:

[Nexxo]

Well, I ordered some of the stuff (PF5080). AS it happens, Amari gets their stuff through Inertx so the price turns out to be the same too! D'oh!

Some interesting facts in the accompanying docs:

PF5080 has a higher density (1.78 x that of water) so expansion rate is higher to: 5% in volume over a 30 degree C temperature range. A reservoir is a must. When filling for the first (and hopefully only) time, you have to avoid filling it above 90%, let the PC idle your PC for 15 mins. twice and top up the coolant slowly as you go along, to allow for expansion. Finally you have to release the reservoir cap once to allow excess air (compressed by expansion)to escape.

The good: you don't need as much as you thought. And it stores more heat.
The bad: you have to have a reservoir.
The ugly: your system collapses as the ncrease in weight tears the blocks, res and rad from their mountings.


PF5080 has a lower viscosity than water (80% of water viscosity). This means it moves more freely through the system and attains higher flow rates. It also means that it will find leaks easier...

The good: increased flow rates.
The bad: increased leak rates.
The ugly: watch £70,-- worth of coolant spilling out of your PC.


PF5080 has lower surface tension. Water has 72 dynes/cm; PF5080 has 12-13 dynes/cm. As such it makes much better contact with water block and radiator surfaces. It also manages to get through leaks much easier... PF5080 evaporates without leaving an appreciable residue. If there is a slow leak, good luck in finding it. Combine that with its higher density and expansion rate, and you want to make sure that your circuit is really leakproof.

The good: lower surface tension is better contact is better heat transfer.
The bad: lower surface tension is increased risk of leaks.
The ugly: watch your coolant level slowly decrease over a period of months... and go crazy trying to find that damn slow leak!

So this stuff is not for wimps. But hey, this is "home of the players". B) Let's play!

[Starbuck3733T]

I think i'll stick with water. :h34r: :h34r:

[Nexxo]

Wuss! :p

Really, it is not so bad. THe stuff comes in an ordinary plastic bottle with plastic screwcap (like the one in the supermarket, that you get your fresh orange juice in), it has been shipped via courier and it has not leaked. Any decent WC circuit should be able to hold it.

Blocks (and rad) actually hold very little liquid, so PF5080 is only going to make a few grammes difference. If that is enough to rip your circuit out, then perhaps it was not mounted all that well to begin with...

Considering people have no qualms installing TECs in their system, which in my opinion is far trickier, PF5080 really should not be such a problem. In any case, I'll let you know how I get on!

[Knipex]

Key question.

How did it effect your temps ????

[Nexxo]

I haven't tried yet. But the guy who did (read iXXeon's post on page 4 of this thread) says it's awesomely effective.

[Starbuck3733T]

Well I could give it a crack ;)

[RADCOM]

I am building my second liquid cooling system and I reckon InertX or PF5070 is a good idea. I once fried a motherboard with a drip of condensation from a compressed air can :(
I've found a UK company which sells 5KG for approx £150 (http://www.acota.co.uk/shop.sfxp) it is also available at thewatercoolingshop.com for £29.99/KG. I think this is a fair price and it does have a 5yr shelf life. I will let you know how I get on with cooling my system ( dual 3.06Ghz Xeon ) piccies and temps etc

[Tommygun]

Hi guys
I have used PP3 Inert cooling in my system for quite some times now. Not because i wont use water the main reason being accidents can and do happen as some of you guys will have found out.

With any liquid cooled system using water even with antifreeze/ additives will still cause desosits over time. PP3 or PFX-90 will not. Yes it is pretty expensive for what it is but the benefits outway the cost in my opinion.

If i am not mistaken the coolant sold through the watercooling shop is Fluid XP
Which claims to be non conductive. I have tried it and well lets say pouring it directly onto a power supply is the best test i could think of:duh: needless to the say power supply did not like it to much. I tried to fire the psu back up after leaving it to dry out nothing happened @ 29.99 i'd say leave well alone
stick with either PP3 or PFX-90.If you do not want to use water

I will post a small video file of 1 test i tried using PP3 if i can find file

The specification on PP 3

Test Spec Result
Distillation Range ( 5-95% ) C 92-103 100.6-101.7

Pour Point C Below -70 Poutable -70C

Density (20 C ) 1.85 0.02 1.853

Resistivity ohm-cm min 10/13 9.9x10/14


PP3 and PFX-90 are both inert the only problem with either coolant you cannot add dye because of the chemical makeup of these coolants the effect you get is like adding water to oil .

If you plan on buying PFX-90 from armari speak to dan at amari first or read Hexus forums.

These guys also sell PP3 same chemical as PFX F2 chemical A UK based company web addy www.fluoros.co.uk person you would need to talk to if intrested is Hellen Mcnamee

[Nexxo]

Oh, I forgot to come back to this thread...

I run my dual Opteron 250 and Radeon 9800 on a single BIX using PF-5080 for a good few months now. CPU temps are 43C and 41C respectively. Great stuff: system looks as clear and clean as the first day I filled it.

[fillip]

Yeah, FluidXP is NOT the same as Flurinert, even though the places selling it would have you believe it is.

I wouldn't say that you will definately get a build up of deposits from a watercooling loop, it depends on several variables and it is quite possible to have a clean lopp that'll last well over a year with minimal maintenance - e.g topping up etc.

As you say, Flurinert is far more costly than plain old distilled water and anti-algae/corrosion additives but it's far far easier to get hold of it quickly and in great quantity than it is to get hold of Flurinert.

I appreciate the relative merits of both options but unless you're extremely concerned about the loop leaking and you want that extra piece of mind as well as being able to afford it I don't see how the cost of Flurienrt is justifiable. Nexxo may well disagree but putting anything other than a clear liquid in Metaversa would have been sacrilege.

[Nexxo]

I appreciate the relative merits of both options but unless you're extremely concerned about the loop leaking and you want that extra piece of mind as well as being able to afford it I don't see how the cost of Flurienrt is justifiable.
I see it as cheap insurance considering the price if the hardware it protects --but to be honest generally no watercooling system will leak all over your hardware unless you do something really stupid (like not testing it). For the rest it's the low maintenance factor --I am just lazy. :D

Nexxo may well disagree but putting anything other than a clear liquid in Metaversa would have been sacrilege.
Absolutely. :D

[Tommygun]

I see it as cheap insurance considering the price if the hardware it protects --but to be honest generally no watercooling system will leak all over your hardware unless you do something really stupid (like not testing it). For the rest it's the low maintenance factor --I am just lazy. :D


Absolutely. :D

:D You hit the nail right on the head with that I am just so lazy lol that makes 2