submersion cooling

well, i've read about this and have wanted to do it for a while, and i once i get a job this summer (can't get one during school) i will have enough money to make it happen. Of course i can't afford flourinert, but i have heard that 100% mineral oil will work as well (albeit not as well and will leave residue) i was going to put it in an acylic container that i will fabricate (aluminum or steel outside to hold it in place, im not sure what to use to not let the oil out) but i want to know is if the container will have to be open to let air in and out or do i have to spend money to have a pump and radiator for the fluid to exit and enter the main submersion chamber and cool the liquid off? will i need to put heatsinks on anything (not having them would look so tizight? will cables and cards need to be insulated (will it affect data transfer/ power)? and how would i keep the oil from splashing out (obviously cables have to be able to get to the mobo. i will be using a cheap ecs kyvta3 mobo and probably an 2000+, and an old mx440 vid card, how will heat be and what to i need to cool off the mixture and what can be expected as an oc (i know this isn't exactly an ocing board but i don't want to spend money on somthing that's gonna possibly break)?

-me

thanks, that's a lot of questions, but i don't want to risk anything and im not very educated on electronics and the like (like i said about the interactions between cable and card and the mobo)

is this even realistic (i hope so i just don't know enough details to doso)

please help me

it'll cost a butt load of money. The liquid you need is like 200 dollars a gallon. Go phasechange for cheaper and better results.

ChiefScout said:

it'll cost a butt load of money. The liquid you need is like 200 dollars a gallon. Go phasechange for cheaper and better results.

Click to expand...

he already said he will be going with mineral oil not flourinert. (it will still be kinda expensive to get that much pure mineral oil though)

you will need a pump and radiator. air want transfer the heat fast enough. and you'll want to putt heatsinks on everything.

you'll probably want some kind of lid over the majority of the top(room for wires to leave of course) so that contaminants want find their way into the system.

good luck, i'll be doingthe same if i ever get any money.

: ) :

It wont be that bad, all you need is to pump the mineral oil outside the case, through a heat exchanger, and then split to the heatsink on the CPU and the GPU (possibly the NB as well) This way there is turbulence over the heatsinks and a transfer of heat.

The cool thing about doing it this way is that you can put the heat exchanger in a bath of ice water or something like that and get super cold temps without having to worry about condensation much

Is oil the only alternativ?

I understand that water is out of the question
But what about an alcohol like methanol or maybe ethanol?

I know that both are toxic (well, methanol is really toxic, ethanol might actually be a little fun )
But besides that, are there any other problems with those?

Finken said:

But besides that, are there any other problems with those?

Click to expand...

BOOM!!!!

: ) :

haha

yeah I guess

I was going to do the submersion route with mineral oil. I did all research and had/have most of the components necessary....BUT..... there is one hurdle that I could not overcome.
The caps that are used on mobos have a small rubber seal at their base...the oil will eventually (6 months +/-) deteriorate it and then you find that weird things start to happen and finally you will be totally out of luck (no boot).
I was going to seal them with epoxy....the slight expansion of the gases within caused problems.
So...the caps cannot be sealed, the caps cannot be subjected to the oil for a sustained period of time without them failing, the project won't work with mineral oil(on an on going basis). I'm afraid that the 200 buck a gallon flurinate is the only submersion game in town.
Sorry.
wj

i was reading somewhere, not sure it's been awhile, from a guy that had done submersion to his computer that after about 6 months his computer started giving him problems. looking closer he found a cap. on the bottom of his tank.

aparantly the problem with them is that they use a dry material (for insulation or something) and that fills with mineral oil and expands and pushs itself free from the motherboard. he checked the current(or whatever) it was rated for and found that it still performed at the required level so he reattached it to the motherboard and it worked perfectly. he put a small spacing in between the cap and the board so it couldn't happen again.

this was from his first hand experience so there appears to be a possible way around this problem when/if it accures.

: ) :

he extended the caps wires so that when the cap-case expanded it wouldn't pull itself free from the solder on the mobo. He continued to chase other caps that 'fell' off. If you are willing to remove, and affix different type of caps (maybe tantilum-epoxied dipped of the same original values) then sure it will work. I am not one that is going to remove all caps, find those much higher cost replacements, measure their values before and after soldering in order to do this conversion. But maybe someone is willing. That original project fell by the wayside btw. Read through all XXX pages and he was there every day or so for months then finally the cap thing happened and within a few weeks he was MIA.
Sorry, it is not a reasonable mod IMO.
wj

what was that web site? been along time since i was there and cant remember it.

i wouldn't mind doing the mod. may be pointless, but fun and worth the bragging rights.

: ) :

read till your eyes bleed:

http://www.ocforums.com/showthread.php?s=&threadid=216968

wj

I read about someone who did submersion cooling for his powersupply to get rid of all the fans. he sealed the caps by carefully spraying them with PCB sealent. Difficult to seal the bottom of them though. Maybe a better idea, you can seal the bottom of the caps to the motherboard with silicone sealent. That wouldnt be affected by the oil (I think) and will still allow expansion with heat.

We need this stuff.
Demonstrated here.

Notice they submerge a laptop and a plasma TV in it, turned on none the less, nothing happens to it and they take it out and it is dry again.

we have talked about that a few times in the watercooling section. most everyone agrees that it can't transfer enough heat and will evaporate to fast to be effective.

be cool to see some temporary setups with it though.

: ) :

and to put the final nail in the oil coffin,

http://forums.2cpu.com/showthread.php?threadid=32859&perpage=30&highlight=mineral&pagenumber=13


this posting is from the guy that started that 13 page thread....read it all if you want or read the first page and the last.
wj

thanks for the link wymjym, that's the one i was refering to earlier.

haven't read it in awhile, glad to see someone revived it

: ) :