Sub-zero degree Cooling

[HAL-10K]

Hi,

has anyone experiences with sub-zero degree cooling, when no air can flow in the computercase. If there is no water condension, is hardware (without HDD- and DVD-Drives) capable to operate in temperatures like -110°C ?

thanks,

Sebastian Till
aka HAL-10K

[Random Nonsense]

welcome to the forums!!!

i've heard of people using liquid nitrogen, but to be honest anything lower than -50 celsius is a bit pointless in my opinion.
I know with a peltier set up you can keep load temps sub zero, and if you chilled the water in the cooling system too it would get even colder! but negative 110 celsius is hard to achieve

[r0ckstarbob]

yes it will

the only thing on the mobo that has a problem with temps like that is actually the CMOS battery which tends to die the death of a rag doll funny enough. thing is i dunno anything that would go that cold. even N2 with fluroinert doesn't even hit those temps....

welcome to the forums

what did you have in mind???

[RangerJoe34]

yo
one of my friends pumped vodka through his old suburban radiator. he overclocked his 1.4 to 2.4 and the thing was running at -14 idle and -10 load. which is pretty damn cold if you ask me

[Thelemac]

Originally posted by RangerJoe34
yo
one of my friends pumped vodka through his old suburban radiator. he overclocked his 1.4 to 2.4 and the thing was running at -14 idle and -10 load. which is pretty damn cold if you ask me

What else did he use?

Actually, you shouldn't go below -50C cause you'll kill your cpu. It'll run...until it pulls apart. Then you'll be sol.

[Oni]

Originally posted by RangerJoe34
yo
one of my friends pumped vodka through his old suburban radiator. he overclocked his 1.4 to 2.4 and the thing was running at -14 idle and -10 load. which is pretty damn cold if you ask me

Thats EXCELENT! When I get some liquid cooling, I'm using VODKA! The people @ the Liquor Store are going to think I'm NUTS buying 3 bottles of Absolut and NOT INTENDING ON DRINKING ANY OF IT!! :D

[fuzzba11]

Um, just to let you know, I saw an article on this somewhere about a month ago and it turns out alcohol is far worse than water for cooling purposes. I have serious reason to not believe that 1.4 at 2.4 story....

[Warlord2]

even at -14 2.4 I just do think is posible

even people with vaporchill cant get 2.4 out of a 1.4 atleast for for over 10 secs.......

I would have to see some hard evadance to support a 2.4ghz story

[RangerJoe34]

yea, i didnt believe him either. until i saw it, played games on it and some other ****, its the fastest ****ing thing ever, and stable. he leaves it on most of the time, except when he sleeps. i have no idea how he did it, the setup is baddass though. and yes, vodka does cool quite well, especially if you leave it in the freezer for a few days before you use it in the cooling kit. all you have to do it have something to keep the vodka at REALLY low temps

[-=UR=- Ranger]

Why don´t you invite your friend to post some screenshots and some pics of his setup here. I think we would be all very interested...At the moment it´s hard for me to beleive that story:rolleyes:

[dozier768]

vodka is 40/60 alcohol water soulutin, weve seen the test that say alcohol isn't a good coolant except for possible evaporative purposes, and it is not possible to get below ambient with any water cooling without a peltier or something except with evaporitive wich in anycase wouldnt be -14 c or f

[-=UR=- Ranger]

If I´m right, RangerJoe says, that his friend puts the alcohol in the freezer for some time....But even at -14 or what he´s getting 2.4 Ghz seems not to be possible for me.

[r0ckstarbob]

well, you're right, the alcohol has pretty bad heat transfer properties... at least compared to water wihich is some of the best stuff out there for heat transfer. it's ALOT better though, then say, antifreeze, which is little more then liquid styrofoam and has absolutely wretched properties.

heres the thing about alcohol though.

although it has less then half the heat transfer potential of water, it has almost 3 TIMES the viscocity of water. that means you can pump three times the amount of alcohol (roughly) that you can water. the faster that liquid moves, the better the heat transfer ratio. that means you can take a hit in the pound for pound heat transfer potential of water and make up for it in velocity. a quantity instead of quality sort of thing, and bully your way over the mark. use as much water as you can and find a balance of water/alcohol so it doesn't freeze and you're going to be good to go. this is something that i've been EXTENSIVELY researching for about 2 or 3 weeks now...

does your friend have an email or can you send him mine? am working on a project with similar temps and would like to ask him a question or two.

better yet, lets invite him over for drinks. tell him to bring his vodka er i mean computer with him!!!

:D

thanks

RSB

langauge@speakeasy.net

[dimmreaper]

Boy I've heard some woppers in my time, but this tops 'em all! Nobody in there right mind would waste pefectly drinkable Vodka in a liquidcooling system . . . .

[r0ckstarbob]

HA!

i don't hear anyone claiming overclockers to be the most balanced folks around!!!

*chuckle*

must go faster mustgo faster mustgofaster mustgofastermustgofastermustgofastermustgofastermu stgofastermusgofstrmsgfsrmgfrmfrmfmfmmmmmmmmuuuuuu uuusssssstttttttggggggggoooooooooffffffffaaaaaaaaa assssssssssssstttttttttttteeeeeeeeeerrrrrrrrrrrr:: ::::::::::::::::::!!!POP!!!

@#(%#)(#$))!@!&!!@!!!!!!!!! EEEEEIIIIIIGGGGGGHHHH!!!!!!!!!!!!!!!

tee hee

oh the things we won't do to make Quake run faster *shakes head*

btw dimreaper, i see you're currently on Borg Cube 119... did you have a chance to check out that link to that project i'm working on? funny enough the nickname for it is the Borg Project...

[dimmreaper]

Originally posted by r0ckstarbob
btw dimreaper, i see you're currently on Borg Cube 119... did you have a chance to check out that link to that project i'm working on? funny enough the nickname for it is the Borg Project... Link? What Link?

[r0ckstarbob]

oops, my bad

http://216.254.0.2/~language/interface/computer.html

check it

[dimmreaper]

Looks like you have a pretty well thought out plan RSBob, but with one problem.

Don't put the HDAs(harddrives) in the "core". HDA components are machined to precises tolerances. Metal expands and contracts with temperature. If the HDAs get too cold critical components like the spindle bearings, and the actuator arm, will no longer be within spec. This could possible result in seized plators, or even a head crash . . . .

[r0ckstarbob]

yah thats actually a problem that i'm working on overcomming as we speak. didn't want to put the HDD in the ICU because i wanted em to have the protection of the core. The ICU won't exactly flop around mind you but still... it's a problem thats definately being looked at. thus far the solution i think i'm going to run with involves using some real small 2 or 3 watt surface heaters and a *little* bit of styrofoam insulation to keep them at decent temps and within operational parameters.

i dunno. theres lots of little tweaks to the design that myself and several others are working out right now. thus far i'm getting pretty steady feedback from abit, amd, ibm, maxtor and western digital on this topic so am certain things will work out eventually. those are just the concept sketches, theres alot that didn't get thrown in there but will when i'm done.

feel free to ask questions if you like. it definately helps me troubleshoot the design.

[dimmreaper]

Originally posted by r0ckstarbob
yah thats actually a problem that i'm working on overcomming as we speak. didn't want to put the HDD in the ICU because i wanted em to have the protection of the core. The ICU won't exactly flop around mind you but still... it's a problem thats definately being looked at. thus far the solution i think i'm going to run with involves using some real small 2 or 3 watt surface heaters and a *little* bit of styrofoam insulation to keep them at decent temps and within operational parameters.

i dunno. theres lots of little tweaks to the design that myself and several others are working out right now. thus far i'm getting pretty steady feedback from abit, amd, ibm, maxtor and western digital on this topic so am certain things will work out eventually. those are just the concept sketches, theres alot that didn't get thrown in there but will when i'm done.

feel free to ask questions if you like. it definately helps me troubleshoot the design.
You shouldn't need any heatercoils for the HDA, remember that they generate quite a bit of heat on there own, just a bit of foam should do, provided you run the system 24/7.

[a_yeap]

RockStarBob: That whole thing looks a monstrosity...an amazing work of mangled art...Its beautiful...I want one dammit...Have you got a physical version instead of an on-paper drawings and stuff...

It'd be astounding is it works...

Just some questions...what temp do you think it can maintain?...and how much would it cost if it retailed?...

and what do the 'evaporation coils' do?...what is there to evaporate if you've got the scrubbers sucking all the moisture...

an eye-popping setup

[HAL-10K]

Originally posted by Random Nonsense
welcome to the forums!!!

i've heard of people using liquid nitrogen, but to be honest anything lower than -50 celsius is a bit pointless in my opinion.

Why do you think so?!

... but negative 110 celsius is hard to achieve
Well you need an powerfull compressor that produce enough liquid nitrogen per hour.


Originally posted by r0ckstarbob
the only thing on the mobo that has a problem with temps like that is actually the CMOS battery
OK, no problem... i could simply provide the mobo with an extern powersuplie

even N2 with fluroinert doesn't even hit those temps....
Ohhh.... it does!! Keep all those scientific fridges in mind, that can achieve even lower temperatures.

Originally posted by Thelemac
Actually, you shouldn't go below -50C cause you'll kill your cpu. It'll run...until it pulls apart. Then you'll be sol.
Why should temperatures below -50°C kill the CPU? Do you have any experiences in cooling the CPU this down?


thanks,

Sebastian Till
aka HAL-10K

[r0ckstarbob]

Originally posted by a_yeap
RockStarBob: That whole thing looks a monstrosity...an amazing work of mangled art...Its beautiful...I want one dammit...Have you got a physical version instead of an on-paper drawings and stuff...

It'd be astounding is it works...

Just some questions...what temp do you think it can maintain?...and how much would it cost if it retailed?...

and what do the 'evaporation coils' do?...what is there to evaporate if you've got the scrubbers sucking all the moisture...

an eye-popping setup

thanks! no, this is still in the design stage and we should be able to break out an actual prototype within the next 6 to 9 months. theres ALOT of designing and planning being done on it still. i just threw a rough concept page together to illustrate the premise of whats happening.

we're not entirely certain what temps we're able to maintain, but we know we can get the core down to -35F to -40F/C and can get the liquid coolant down to -25F to -30F. what temperature we'll be able to operationally maintain is hard to estimate due to several variables though we're almost certain to be able to smoke both the Vapochill and the Kryotec with this design in overall performance guaranteed. both those products cool directly with freon which is a gas. we're using liquid which has much higher heat transfer properties even though it might be a little bit warmer then the freon itself. this design also emerses the entire board in a sub zero environment instead of just the CPU and as you can see, we're using a minimum of 2 and probably use 3 seperate pumps to spot cool the components that need it the most... CPU, AGP, Northbridge/Mosfets.

the evaporator coils are the refrigeration coils. i won't go into the refrigeration process to explain it, but freon only exists as a liquid when its run through the condenser and the condenser coils. when it's vented back into the refrigerator itself through the evaporator coils, it becomes a gas that is evaporating and sucking up heat as it does it. the evaporator coils port back into the condenser which condenses it and the process starts all over again. the scrubber is used to remove the liquid in the air and drop the dew point prior to turning the refrigerator on to prevent condensation or ice from forming in the case itself. it doesn't actually cool, it just removes the moisture in the air so condensation CAN'T form.

we're not certain if this will ever hit the retail market and for now we're not concentrating on it. it is a distant thought though and remains an intregal part of the design process in that at some point, we're hoping that we will be able to get it to a retail stage. we won't have a very solid idea of what the pricing schematic will be until we actually build it, but the hope is to put it on par with the vapochill retail between $600-$800 for the case. definately higher end we're thinking, targeting the devoted overclocker and the server markets.

will post the finalized concept drawings within the week for public troubleshooting. there should be alot more detail at that point.

RSB

[r0ckstarbob]

the CMOS battery physically exists on the motherboard itself and cannot be removed. it maintains things like bios settings and so forth and is a completely independant entity from the powersupply. if you insulate it that should be enough and solve your problem there adequately.

more later....

[phungilax]

what would happen to a CPU at such low temperatures. could you get it so cold that it seizes to work?

[HAL-10K]

Originally posted by r0ckstarbob
the CMOS battery physically exists on the motherboard itself and cannot be removed. it maintains things like bios settings and so forth and is a completely independant entity from the powersupply. if you insulate it that should be enough and solve your problem there adequately.

more later....

I could simply remove the battery (yes, thats possible) and connect an extern battery via cabels to the batterie connectors. Insulation should be difficult (maybe impossible) , espacially if the system runs for a longer time.


Sebastian Till
aka HAL-10K

[dimmreaper]

Originally posted by HAL-10K
Why should temperatures below -50°C kill the CPU? Do you have any experiences in cooling the CPU this down?Remember that CPUs are a silicon waffer. They are essentialy plates of glass like silicon with the circuits in between. Silicon isn't too good at conducting heat you see. So the top "plate" might be -100C, and the bottum one might be -20C. Now think way back to those high school(or earlier) science lessons. What happens to matter with temperature change? That's right! It expands and contracts! Now if the top "plate" contracts more than the bottum "plate", what happens?

They seperate! POP!

[HAL-10K]

Originally posted by phungilax
what would happen to a CPU at such low temperatures. could you get it so cold that it seizes to work?

Well at the latest if the silicon loose his semiconductor properties. I don't know exactly at which temperatures this would happen. But for shure you need an REALLY cold sourrounding, because an working CPU heats every transistor of the CPU quit fast.

I have to ask an aquaintance here in Dresden who works at AMD, for sure he knows it.

PS.: Sorry for my bad english!

Sebastian Till
aka HAL-10K

[HAL-10K]

Originally posted by dimmreaper
Remember that CPUs are a silicon waffer. They are essentialy plates of glass like silicon with the circuits in between. Silicon isn't too good at conducting heat you see. So the top "plate" might be -100C, and the bottum one might be -20C. Now think way back to those high school(or earlier) science lessons. What happens to matter with temperature change? That's right! It expands and contracts! Now if the top "plate" contracts more than the bottum "plate", what happens?

They seperate! POP!


As far as i know, there are always the same temperature differences at an heating device. Equal which temparature sourrounds the device. Keep in mind that the low temperature cools the hotter areas of the CPU down. If the air arround the CPU is 20°C cold, the die might be 60°C hot. If the air is -20°C cold, the die is 20°C cold.


Sebastian Till
aka HAL-10K

[dimmreaper]

Originally posted by HAL-10K



As far as i know, there are always the same temperature differences at an heating device. Equal which temparature sourrounds the device. Keep in mind that the low temperature cools the hotter areas of the CPU down. If the air arround the CPU is 20°C cold, the die might be 60°C hot. If the air is -20°C cold, the die is 20°C cold.


Sebastian Till
aka HAL-10K OK don't listen to me. You will find out on your own what happens when the CPU gets below 50C. It might not die instantly, infact it probably won't but it will die a premature death.

I forgot to mention that the silicons crystaline structure is formed at or near room temperature, so the contraction can cause tiny cracks even if there is no substantial waffer temperature differential.

[r0ckstarbob]

i think he's just troubleshooting, not being difficult.

at -100C though... it doesn't matter what... better not have any moving parts like fans or whathave you going. instant snappage.

and if you wanna go through the trouble of rewiring your cmos battery, you'll probably be okay. i would watch out for the chip contracting at that temp though. it'll probably be fine running, but it's going to have a hard time in storage when it's not putting any heat out to beat the temperature difference. it'll be fully exposed to the weather so to speak. you'll probably be fine until you turn your computer off (providing you do) but when you wake up and turn it back on the next morning... watch out.

not that i'm one to talk, but for the life of me i can't forsee much performance increase at that temp vs say -30 or -40 but i don't know anyone who's done it either, so...

rock on crouton

let us know what happens