Silicon fluid immersion?

I wasn't positing an inert gas cooling system. There was some talk earlier in the thread of using a gas-filled expansion column as a way to handle thermal expansion of the cooling liquid. An inert gas would be preferable, since it wouldn't promote oxidation in the coolant. My point was simply that small quantities of inert gases can be had from any wine shop for a reasonable price, without having to deal with a commercial gases supplier.
 
Yep.

Post # 43....

--------

Commercial gas suppliers are the only way to go !

Now , granted I have an account at Praxair , but it's one stop-shopping.

I can get helium,nitrogen,argon,etc.....Dry Ice ? no problem....Liquid nitrogen ? Sure... Got a dewar ?

hehe..

-----------

Yes ... Mark305TBI yah nailed it.

Motor oil is not a pure liquid.It's a mash of different hydrocarbons , not including the 'additives'.Here's what that means : It's a really good solvent for rubbers and plastics.

It certainly would work initially , but it would eat the crap outta stuff and would fail eventually.

:D
 
This is not a 'troll' << I just want to say that so as to avoid getting flamed, and to say that I am asking a question here along with my suggestion...lol.

For all the 'hassle' that an immersion system is likely to be, or that using complicated fluids inside loops, wouldnt an active phase-change system (compressor) system just be easier, cheaper, and give better results? I mean, to pump/push such thick fluids, the work that a pump would have to do to circulate would jump, so you are talking about adding larger pumps to the system which are going to add even more heat that you have to compensate for... and a radiator is a passive method of cooling, so its easier/more effective to minimize the heat going into the cooling system in the first place rather than taking on more heat and then adding surface area to the radiator system to compensate. You are sort of taking one step forward and two back in this regard. A pump that moves 300gph of water at a given head pressure may only be 15 watts, but to move something thicker like oil will cause a rapid jump in wattage... easily 50 watts or so, and then you are adding 50 watts of heat to the system.

It just seems to me that an actual phase change system would be more worth it since then you can go lower than room temps.

Speaking of, if anyone knows about alternative compressor-cooling substances which can be used to cool even lower than normal... Im all ears. Thats my kind of project... get a CPU down to -100 or something running 24/7.
 
undertheradar said:
This is not a 'troll' << I just want to say that so as to avoid getting flamed, and to say that I am asking a question here along with my suggestion...lol.

For all the 'hassle' that an immersion system is likely to be, or that using complicated fluids inside loops, wouldnt an active phase-change system (compressor) system just be easier, cheaper, and give better results?
Click to expand...

In my opinion, the best reason to build a system in an immersion environment, isn't for active cooling. It's for totally fool-proof condensation prevention. So I would use it in conjunction with a direct phase change, or water cooler, or pelitier, or some combination of the above, active cooling system.
 
Mark305TBI said:
In my opinion, the best reason to build a system in an immersion environment, isn't for active cooling. It's for totally fool-proof condensation prevention. So I would use it in conjunction with a direct phase change, or water cooler, or pelitier, or some combination of the above, active cooling system.
Click to expand...

What he said. The colder you get with immersion, the more issues you get with fluid thickness. However, that's really pennies compared to issues with condensation. One requires a stronger pump, and the other has the possibility of dripping water or frost onto your motherboard (gross oversimplification, of course).
 
Thanks for entertaining my question. Okay... so if you want to prevent condensation, wouldnt putting the components into a sealed box with some of that silicon gel work just as well? I mean, if you wanted, you could vacuum seal the box... or even pump it full of Argon then. I mean, if you are going to have waterblocks and phase change blocks on everything... not that I have ever found this to be a problem as long as you insulate your phase change piping and block. If you were still concerned, you could simply run the motherboard with its backside sticking straight up so that any condensation would drip away from the components. Not to mention condensation (rather pure compared to tap water) has very little chance of shorting things out since the conductivity of pure water is 5.5 · 10^-6 S/m (Copper is 59.6 × 10^6).
 
OK, if you want to go that way: then you would want to fill it with inert gas. I don't think that I'd want to see what a lot of computer parts would do when pressure is lowered down to nil. Not to mention at that point you've got a very delicate system, *and* you have no way to distribute heat (anything without a waterblock or the like will keep on heating as heatsinks just plain won't work in vacuum).

Keep in mind that this is "extreme" cooling; insulation at these levels is bound to be impractical at best, and short-term in general. If you're using a coolant significantly close to freezing point, you'd need one hell of an insulation job to keep the exterior of the insulation from eventually forming condensation.

Using something like Argon gas is an option, but it's significantly harder to make a good seal with a gas than a liquid (even a liquid oil). Though it's certainly a better option than vacuum, not the least of reasons is thermal conductivity. I do plan on trying a gas-contained system out at some point, but not until I get my already significantly over-complex 2000-watt peltier system up and running. And even then, I would never consider it a good idea; there are just far too many hassles involved, and I haven't even put much thought into the issues yet to know what all of them may be. Not to mention that while whole-system cooling isn't the strongest selling point, it's still a pretty damn good one, and even with argon gas you arn't going to have the same thermal conductivity by far.

And frankly: water is water. If you are going to have water condensation, you will eventually have a short. The water will pull out enough material from the surroundings to gain enough conductivity eventually, and in general it's just a plain bad idea to allow condensation at the electronics.

On a side note: No need to thank for "entertaining". There are enough bat-shit-insane suggestions in here over time that they need a little bit of thought. ;)
 
Well, Im just thinking that there has to be some better way than 'immersion in a goopy fluid'. I get it, but you are voiding warranties and making those parts useless for anything else, and impossible to work with in any other way really. I would also be concerned about how many of your PC components will react over time to being immersed in mineral oil (or other). It might be considered 'inert' by most, but its not. It also serves as a medium for degradation and corrosion that a gas does not... so one of its main advantages can also be its undoing. There have not been any long term tests done on these effects, but a good example would be that over time, the oil will seep into caps and other 'gaps' in the circuits, and this will change the electrical characteristics of that component. I am trying to make my own version of a water cooled PSU, and sealing the PSU in a box with mineral oil being pumped through it (like koolance) was the best option so far, but as I was developing the idea, the interaction of mineral oil over time was a valid concern that many of my fellow EE buddies mentioned. These things may not conduct electricity (but they do conduct heat), but they are not inert. I cant imagine Si fluid to be much better in that regard. So a gas still seems like a better idea to me... want to remove a mobo or swap out the GPU... just open the box (later on close it and pump more argon in). Or just use silica gel moisture packs to keep the sealed box dry. If this isnt 'cool' enough, well think of this... if you were to seal a mini-fridge (which is already pretty much sealed) and then remove all internal humidity, you could use the fridge to cool the air inside to freezing levels. I know a fridge wont handle the whole thing on its own, but you would still route the phase change lines to a seperate cooler, so the little amount of heat that is left would be easy for a fridge to deal with. If you want, strap an even better compressor onto that thing.

Now you have a system where the CPU and GPU, etc can be kept below 0, and the rest of the board and components can be kept at chilly temps as well, but without fear of condensation, and to remove a component, just open the door. After you open the door every time, just shut it, clamp it shut, use a pressure pump to suck out the internals as much as possible, and then fill it back up with Argon. Or, heck, just close the door and keep a big sack of that silica gel in there.

Preventing condensation isnt all that hard if thats the main goal. Oh, and heatsinks DO work in a vacuum, just not through convection/conduction. But I hear your concern... a vacuum is a poorer cooling medium. BUT, putting another inert gas in there would be easy.
 
Voided Warranties? I would hazard a guess that most of the people here have already voided a few warranties over less ambitious projects as this. *chuckles*

The issues with fluid immersion are pretty well known, if not terribly well tested or documented. Mineral oil will of course cause damage over time, but there are a few things that can lessen it (sealant over components to keep out the fluid, not to mention maintaining a constant low temperature can help as well).

I'd personally love to see a gas-based cooling chamber. It would be, if nothing else, nifty. And yes, I'm aware that heatsinks would work, but to a significantly lesser extent. Also, when you're talking about running components in a vacuum... well, you're getting into a realm of danger where the casing/tank could rupture under the stress; I think that it would involve a few design issues to figure out.

Personally, I just wish that fluorinert wasn't so damned expensive.
 
No, not a vacuum, but a 'controlled gas' environment. A mini-fridge or freezer with pass-through bulkheads for the phase change blocks that go on the CPU, GPU, and perhaps some hoses to pass through to the NB/SB if needed would work. You could use regular air as long as you passed it through a dryer, or packed a bunch of silica gel in the fridge/freezer. Or, you could just pump some liquid argon in there after every time you open/close it. This would keep everything at freezing, and the CPU/GPU at well below, yet with no chance of condensation. Best of all... no warranties are voided, and even if it isnt so much warranties that worry you... the other concerns (long term damage, having a goopy component, etc) are dealt with. If you were careful, you wouldnt even have to vacuum pump the fridge/freezer compartment first... you could introduce the argon as a liquid at the bottom and simply allow the room air to exit the top through a bleeder valve. The heavy & cool argon would push all the nitrogen/oxygen up and out.
 
I'm actually using a 'controlled gas environment' on my thinktank revision 2 project , but not for condensation prevention.I'm using a positive pressure N2 system to provide an oxygen free environment.(the oxygen free environment is a result of my getting rid of the 'goopyness' of submersion.I'm using pure alcohol instead)

This brings me to my comment : Argon sounds good , but it's quite expensive.

I'm using nitrogen for that reason.A type 'k' tank of compressed n2 is about 180$ whereas a type K tank of Argon is about 400$......

:D
 
BrainEater said:
I'm actually using a 'controlled gas environment' on my thinktank revision 2 project , but not for condensation prevention.I'm using a positive pressure N2 system to provide an oxygen free environment.(the oxygen free environment is a result of my getting rid of the 'goopyness' of submersion.I'm using pure alcohol instead)

This brings me to my comment : Argon sounds good , but it's quite expensive.

I'm using nitrogen for that reason.A type 'k' tank of compressed n2 is about 180$ whereas a type K tank of Argon is about 400$......

:D
Click to expand...
Use CO2, it stores much denser [it is actually a liquid in the tank] also, it is cheaper than nitrogen, and a 40lb tank will hold about the same amount of gas as the K size tank from what I've read...
 
I've already got a nitrogen regulator , so that's why I'm using n2.

Technically , I could just use the co2 from the dry ice .....mebbie next time..

:D
 
best [486];1034140330 said:
Use CO2, it stores much denser [it is actually a liquid in the tank] also, it is cheaper than nitrogen, and a 40lb tank will hold about the same amount of gas as the K size tank from what I've read...
Click to expand...

CO2 bleeds through almost any plastic though. Trying to seal a fridge would be next to impossible since the CO2 would go through the gasket like a siv.
 
Hi - I'm the Technical Manager at DSI Ventures, Inc. We make a variety of electrical insulating fluids, including several that are used commercially to cool power electronics circuit boards and mini and micro-computer motherboards via total immersion. Totally inert, way less expensive than fluorinated fluids and biodegradable, too. If you're interested, drop me a line at dsundin (at) dsiventures dot com.
 
You must log in or register to reply here.
Back
Top