How to get -60c?

[Player0]

Well, I aim high :) Or is that low... anyway, I was having a discussion with Punisher earlier today about this topic. My new system isn't even finished yet and already I'm looking in to something else.

I've taken peltiers to the edge of the universe so far. In fact, my poor circuit breakers can barely keep up with the power requirements of nearly 900w of peltier cooling. But, the results are quite amazing. -10c water temperatures. With a 226w direct mounted on the CPU giving me a delta T of at least 20-30c, Im guessing a loaded CPU temperature of under -30c. I'm replacing the pumps and working on a new mounting system, but I think its doable with what im running now and results ive seen with other 2262/chiller systems. It however takes no less than 15amps to run just the cooling system on my box. Thats 15a 120v mind you.

Okay, so its crazy, and certainly not economical. What it does do nicely is fit all nice and snug in my little cube case, something which cant be said about most high power chiller systems. But now Im thinking, if I could just get that water to -30c...

I've seen compressor systems chill stuff this cold before. Certainly with a peltier, the CPU could get really damn cold. Will the CPU even work that cold? Who cares lol. Its what i want to find out. Of course, I still want a self contained and portable (although granted, heavy) comptuer system. No tubes sticking out or nothing. So space is limited.

Prometia did it right, all nice and small and self contained. Not sure if it could cope with the extra heat of a peltier but...im sure a larger version could. The question is, do you chill fluid down to -30c, or do you direct mount the evaporator right to the peltier for the most potent punch?

Im thinking the refridgeration unit has to cope with at least 1500w of heat to handle the 400w peltier/cpu output. Anyway, Im in no rush to do anything quick. My next system will probably be compression based... maybe even a prometia. But, Im going to finish up my current pelt system to see just how it *really* performs. I just need my new 600GPH pumps to arrive.

[PiLsY]

Dual evaps should do it m8. Get a pair of promteia style systems and use them to cool the water, then run a 220watt pelt on the cpu. A single prometia can cool 250watts to 0c assuming 100% efficiency. A pair could then cope with 500watts again assuming 100%. Its more like 80% real world, so that leaves 400 watts. 180watts for your cpu and 220 for the pelt. Should see some VERY low temps with that.

Another way of doing it would be something I brielfy toyed with. Instead of air cooling the vapor phase radiator why not chilled water cool it in an enclosure....

Im pretty damn sure if you rigged up your current chilled setup to pump water through a small res (150 x 150 x 80) and stuck the promteia radiator in it youd see some awesome results with a pelt on the cpu chilled by the evap. Id be willing to bet around -20c hotside on the pelt giving at least -45c coldside.

Anyway, food for thought - hope you come up with something :).

PiLsY.

[dmitriyaz]

i know its not much help,
but i'd recommend going to a public library near you and pick up a book called "basics of refrigiration", or something of that sort.
read it and understand it, then you'll know precisely what to do.
i doubt there is an "easy" way there.:slap:

[Player0]

Actually, the 226w peltier produces 360w of heat by itself. Maybe another 100-120w for the Athlon. Maybe more. So, at least 500w input temperature we're looking at.

Punisher mentioned a dually system as well. But then it just takes up too much space.

As for cooling the phase-systems 'radiator' with chilled water, that sounds like a good idea. My knowledge of refridgeration is weak at best, but I was under the impression that there is simply a limit to the cooling power of the R404/r290/etc gas itself. By cooling the radiator portion, im not sure if your increasing the wattage capabilites of the system, but you are lowering the systems temperature. Of course, there comes a point (-40c or so) that the system just wont get much colder? So, i dont know, but from the little i have read so far, it doesnt seem like that would work so good. it would work, but only to a point i guess?

Looking at my case, i dont think it would be too hard fitting two small danafloss compressors in here. Some tweaking could be done. I worry about the things cycling on and off all the time, since if they do this in phase, it could blow the breakers.

Dont worry, id definately go buy every book I can find before buidling a dangerous system like this :)

[MrIcee]

Player0:)

Go check out the phase change gallery starting on page 2. Theres a pic of my phase change"refrigeration" chiller, which I first fed a CPUFX Z-4...then later a Swiftech MCW462UHT w/226 watt pelt.(which is now clamped to my 9700 Radeon:D ) It worked very well..but to be honest..the Prometeia by itself is colder overall....and I've done extensive testing of all systems:)

Randi:D

[bowman1964]

-60 I am there now ..unloaded of course but -44 loaded.new system i am building i am trying to get -50c loaded temps and i should make it.this system is being built inside of a pc tower.with a chiller unit also for video card.going to be a first if it works.
check out the progress in these fourms.

http://www.phase-change.com/phpBB2/viewtopic.php?t=18

http://forums.procooling.com/vbb/showthread.php?s=&threadid=4327

[Player0]

Thats good lucking stuff bowman. Id like to have half your skill :)

[|PuNiSh3R|]

It takes a lot more then half skill to do anything in refrigeration..

[PiLsY]

Player0,

If my understanding of phase change is right then you are quite correct on the actual temps. However, when you cool the radiator lower it increases the thermal capacity of the gas, therefore allowing it to more efficiently draw heat away. While your overall evap temp may therefore remain the same, cpu temp will drop closer to evap temp. IE the difference in temps between idle and load will become closer.

Youd need to use r404a to get down to -60c - I believe its capable of around -70c, but youd need a small nuclear reactor to manage it :p. Using a sub zero cooled radiator along with an r404a filled prometia youd see idle temps of around -55 - 60c. Load temps would most likely be of the order of -45 to -55c. Without the super cooled radiator youd be looking at about -30 to -40c, but the same evap temps give or take a degree.

PiLsY.

[Player0]

Well, it sounds like what I want to do isnt so easy hehe. I should probably start on a easier project.

But, from what im hearing, you guys recommend direct die evap cooling on the CPU with a hefty compressor system, and NOT having a peltier on the CPU for the added oomph.

[|PuNiSh3R|]

Haha.. prometia with r404.. You need a good 1/2 or 3/4HP compressor.. and then you need to put the condenser in your yard.. about 10 feet down.. Make it a big ass one too.. like.. get several AC unit condensers use them all in series.. yum..

[|PuNiSh3R|]

Actually.. why not just get a 1HP? I mean.. all the damn freon you'll be able to stuff in there.. hot damn.. that would own..

[aenigma]

:confused:
Cascade Prometeia?haha no I dont think so....

Just do what I do, build a dual stage system.I dont have a thermometer that goes low enough to test temps but it goes well below -51c(thermometers limit)...
Damn cheap acu-rite thermometers...
Oh and I use propane....

[Player0]

What size compressors do you use m8?

you think a system like that could fit by itself an a large mid-tower case? (Antex 630 for example)

[|PuNiSh3R|]

Rofl.. no it wouldn't he (aenigma) uses a 3/4HP and a 1HP compressor

[aenigma]

Originally posted by Player0
What size compressors do you use m8?

you think a system like that could fit by itself an a large mid-tower case? (Antex 630 for example)

No...
Very few systems can fit in an avg. size case.(good systems)
But you could make it somewhat small and make a case for it....
My compressors are pretty tall too....
Bust way would be to build a system in some kind of aluminum box and put your pc on top of it....It would look pretty nice that way.

My system wouldnt even fit in a computer case, the condenser is about 4 times as wide as the case....Probably longer too...
But I dont go for looks either, so I am sure something could be worked out if you get creative.

[bowman1964]

yes it is hard to build one inside a case that will cool.now i have one running now inside a old pc case and is compact.-60 unloaded temps.but i am almost half way there building a complete alumiuim case and cooling system all in one pc tower.and i am going to get some better load temps i hope ,with the new system.now i run -44 loaded with a xp2100@2150 1.85volts.just a matter of time before i have the new one up and running.
you can keep up with my progress in the forums.

[|PuNiSh3R|]

BOwman isn't sparing any expensive either..

[bowman1964]

well we all have to have a hobby i quess.i dont do drugs and have a high stress job Managing a department.i have 2 kidds that i take care of by my self .so i need some form of relaxation.???????????????????why am i working on computers?

and yes i am trying to build this one a little different.speaking of spending money i am buying a simpson LED temp controller to handle the display and computer startup.reads down to -200f i am told.is programable and complete with switching relays.so yes this system will go first class.i all ready have gotten the new motherboard AT7-MAX2 and a stick of corsair pc3200cl2 512mg waiting on xp2600 price to drop in a couple of weeks.

[Player0]

Got a link to that Simpson thing? Sounds nice.

I got that board, that ram, and that CPU (well on backorder).

Be interesting to see what you can do with it :)

[|PuNiSh3R|]

Bowman.. you are going to make me sick.. lol.. That temp thinger musta cost atleast $150.00.. ?

[aenigma]

I want that temp probe!!!! :)

[sysfailur]

Another method to reach mad low temperatures would be to do what the show "The Screensavers" did. They took HEF and doused a system with it, and then they pumped it through copper piping which was went into a seperate tank which was full of LN2. The LN2 cooled the HEF, and then the HEF was pumped onto the processor core. They didn't use a heatsink which I thought was a bad idea. I would say still use a heatsink to spread the heat out rather than run without a heatsink. If you actually sealed off the LN2 and HEF so it wouldn't evaporate, then you'd be set! They didn't do a good job at all of sealing it off or they could've run the solution for a much longer time.

[aenigma]

The reason LN2 cools is because it boils, you seal it off and it gains pressure and it cant boils anymore, which means temps keep climbing and so does pressure...no more cooling.....

LN2 cooling is stupid, unless you actually use a compressor.But just temporary ln2 cooling is weak.

[|PuNiSh3R|]

i'm a big meany

[dmitriyaz]

hey come on now guys,
this is not [H], remember? :slapass:

[|PuNiSh3R|]

yea.. I know.. on [h] there are some seriously smart people!!!!

[sysfailur]

OMG you're still a fa**ot on this forum as well. Well, geuss I'll just have to ignore your lame ass like I did on [H].

[JBELL]

Originally posted by sysfailur
OMG you're still a fa**ot on this forum as well. Well, geuss I'll just have to ignore your lame ass like I did on [H].

I won't tolerate flaming from EITHER side.. please do not continue discussion in this manner.

[JBELL]

Originally posted by |PuNiSh3R|
yea.. I know.. on [h] there are some seriously smart people!!!!

especially from you

no flamming will be tolerated.

[bowman1964]

well this is how you do it.picture is a thousand words