PDA

View Full Version : A problem you all wish you had :)



Serra
03-10-2005, 11:05 AM
Okay, here's my problem: I have (and will continue to have for some time) access to unlimited quantities of free liquid nitrogen, but don't know how to make use of it! Even as you read this I have a large tank of LN2 just sitting outside my room not being used to cool anything. I know it's just plain sick and wrong, but thats the situation. I know if anyone can give me at least a link to a decent (and simple - I'm n00b to all but air cooling) LN2 cooling guide, it's someone who comes here. Thanks!

DuPu
03-10-2005, 11:16 AM
I guess u can find everythink what u need on the internet, google can be u new best friend :D
Just take u time, and u know when u are ready..

afireinside
03-10-2005, 12:14 PM
Just look around here and make a tube like other peoples. How are you getting free ln2 in large quanities? LN2 is everyones dream...

[XC] moddolicous
03-10-2005, 06:43 PM
If u want, u could pm chilly for one of his cpu tubes, and when u do do it, take pics and make sure u have good insolation.

Serra
03-10-2005, 07:01 PM
Just look around here and make a tube like other peoples. How are you getting free ln2 in large quanities? LN2 is everyones dream...

Well, I still need some help in someone sending me some good link (the links I've found - while alright, aren't quite "step A: <blah> step B: <blah>" enough for me. In the end I suppose I'll end up just designing my own (I am an engineering student, I *should* be able to do this), but I do have two main questions:
- What sort of system to use for the pump?
- Do you just use standard water-cooling type hookups for the CPU/RAM/GPU? That's what seems logical to me, but I've never tried before to know...

Anyway, to respond to your question about where I get the LN2: my dad has a freezer at a local cattle AI facility, and they give us free tanks that they bleed off their main feed. Most of these facilities end up dumping this stuff (which they don't enjoy because they pretty well have to fill up containers, walk them outside, and dump them themselves) or giving it away about once a week when they replenish/do maintenance (which they prefer, because it's one less bucket they have to drag outside). It might be an avenue worth looking into if you're interested.

afireinside
03-10-2005, 08:47 PM
You dont pump it. You pour it into a tube connected to a plate bolted to the CPU.

Serra
03-11-2005, 01:00 AM
So... a LN2 cooler essentially just consists of an insulated holding tank, a tube, and the use of gravity to ensure constant contact with the tanks resevoir? If so, that's a lot more simple than I had this working out as in my head. I suppose that makes me ask a few more questions:
1. Why is a metal plate used on the processor rather than, say (to take the lazy mans way out) a Koolance CPU Liquid Cooler adapter? I understand that they have two nozzels because they're for cooling that does use pumps, but I could just block off one of those pumps, couldn't I? Or, versely, split my LN2 tube to attach to both of them to get a more even cooling (not that it really matters). I suppose I'd have to insulate it a little more, but that wouldn't be any trouble....
2. What sort of rate does an average CPU go through this stuff at? If it helps, I'm running a 2.4GHz P4 (though I assume I'd have it OC'd above the 3.xGHz I'm at currently). I ask because I'd rather not have to tap the giant tank I have here, it'd be much more convenient to be able to make and use a smaller, more space-friendly tank.

blinky
03-11-2005, 02:31 AM
i have access to free LN2 as well ;)

i gotta make a container sometime :(

afireinside
03-11-2005, 06:12 AM
A waterblock wont work... Hows the ln2 get in it?

charlie
03-11-2005, 07:04 AM
You stick an 8" copper tube with 2.5" diameter on top of a .5" copper plate, rectangle somewhere around 3" x 4"....
Braze the copper tube on top of the copper plate. Drill holes in the plate corresponding with the motherboard mounting hole. Attach plate to mobo. Insulate board and tube to guard against condensation.

That's the easy part. Making it work, not killing the CPU and or the mobo is the HARD part. THAT's where you need to do some reading.

C

afireinside
03-11-2005, 01:09 PM
How would you kill the CPU? I'd be worried about the caps on the board freezing...

bh2k
03-11-2005, 07:25 PM
If you short a cap or a mosfet. Also if you somehow sort cpu pins, but that's tough to do.

Serra
03-11-2005, 08:21 PM
Okay, I think I've been typing like I've been taking some kind of drugs. Having no experience with anything beyond water cooling (and even then, not direct experience), I was thinking the system would work about the same... using tubing and whatnot as opposed to just placing metal tubes full of the liquid on top of the processor. But that makes me wonder: why can't you just use tubing and whatnot? I understand that direct pressure is the *best* way to do it, but I for one would be perfectly happy with a slightly less than perfect way (which would also look significantly better). Still, because no one does do it this way, I'm led to believe there is some good reason behind it... but why is that? I'd be looking to use a CPU liquid cooling cap like the one in this picture:
http://www.memoryexpress.com/productfiles/images/5547.jpg
and basically tube LN2 in and out of it... or, versely, just into it using a splitter off the source tube (insulating the tube well, of course). Not cycling the LN2 would work under the theory that the LN2 at the end nearest the processor is still more than cold enough for my needs. I've also attached a jpg I made at work of the basic setup. Oh, and concerning possibly breaking the mobo: I have an ASUS P4P800-VM board lying around somewhere that I'm never going to use to test on (if you know the specs, you know why).

lalPOOO
03-11-2005, 08:51 PM
ln2 evaporates at something like -200. You can't use any kind of tubing other then brazed copper really. I suppose if you made a copper block, and then used a really small tube feeding the block ln2 you could run it for extended periods of time, but I don't really think it'd be worth it. You're better off just getting a ln2 tube and using your vast quantities for benching.

Serra
03-11-2005, 09:09 PM
I realize that the LN2 evaporates at pretty low temps, but I'm curious as to whether it *would* evaporate, given that it is connected to a much larger supply to keep it cool. That's also where running it in a loop might be good. Physics dictates a current would necessarily be generated in the tubing, with one tube taking it away and one drawing it in.

lalPOOO
03-11-2005, 09:11 PM
I don't know what kind of equipment you have at your disposal, but I doubt you can run a closed loop with ln2.

Jort
03-12-2005, 03:31 AM
:p:

you need a huge installation to bring the evaporated lnĀ² to liquid again.
just use a container like some people are trying to say here :)

its not possible so use is 24/24, temps are WAY to low.

http://www.reviewers.be/images/post/pic5.JPG

just read the sticky's here :)

Serra
03-12-2005, 08:04 PM
Because it would run on a little, I've typed up some thoughts on looped LN2 cooling and attached them as a word document(which I maintain should be possible unless that annoying "pressure" thing wants to get too involved). Anyway, if you have a few minutes take a look and let me know your thoughts.

Also: I know that LN2 must boil off.. I'm not trying to stop that at all.

wdrzal
03-14-2005, 01:08 AM
your gonna get hurt! give that stuff back!!!