CO2 / liquid Air - cooling

Hi there,

Has anyone ever tried to use these small 12g CO2 capsules for instant cooling?

I wondered whether it is possible. I guess all it would need would be a valve tht can be controlled by arduino, a cartridge holder and an aluminium/copper-block with some holes in it.

cheers!

Those small cartridges are gas charged, not liquid. What you are thinking could be done is -no where- near as easy as it sounds.

For instance, lets say you were using a liquid refrigerant (which is what liquid CO2 can be considered), something more normal though (ie, smaller working pressure) like R-134 (like you would see in a vehicle A/C system).

Here's an explanation of how an A/C system in a car works:

http://www.familycar.com/ac1.htm

You would essentially need to replicate this, with your CPU cooling block replacing (poorly, unless designed right) the evaporator coil. If you notice in the diagram, there is a low and high side to the system. The low side is under fairly high pressure (approx 50 psi), while the high side is under even higher pressure (approx 200 psi) - note that the pressures will vary due to a variety of reasons; these are just very approximate values.

This isn't a system that can be easily built from scratch by someone in their garage; you certainly wouldn't use CO2 for the refrigerant, either, as it doesn't have the proper characteristics needed for such a use (there's a reason freon and later refrigerants were developed - and a reason ammonia gas was first used, though it is a nasty and poisonous refrigerant that, while it works very well and is still used in industrial refrigeration/cooling - it is not something you want to play with).

Any homemade system would incorporate a none-too-small compressor, plus all the attendant tubing (mostly copper or stainless steel line), fittings, heat exchangers, etc. You have probably seen a water-cooling setup for a PC; maybe you've even built one or installed one. An A/C system for a CPU, while more than possible (in fact, I think there have been one or two -expensive- kits available for extreme overclockers of PCs in the past), isn't something you easily homebrew; and what you end up with is going to be more bulky and potentially dangerous than any water cooling setup. Leaks will be more likely at the pressures being dealt with...

What you are saying can certainly be done; but unless you are an experienced A/C technician who wants to play around with such a system, or you know someone who is... Let's just say a little knowledge can be a dangerous thing in this situation. You're ultimately going to want someone you can go to with questions and help; you're going to want to do a ton of studying on the subject. You'll probably want to read about how both A/C units (houses and cars) are configured as well as refrigerators/freezers (the principles are the same, though). There is probably as much to learn as there is in electronics...

Once you get to the point where you can possibly think about building such a rig with confidence that it will work, you can expect to spend a lot of money doing it, unless you want to reconfigure an existing system (like a small window A/C unit or a small refrigerator); even there you are going to need equipment and other tools just to do the recharge after you break the system for the reconfiguring, not to mention the refrigerant (some refrigerants can only be purchased by licensed individuals, btw).

Good luck with your efforts!

:)

Or… instead of re-using the coolant like an AC system does, you can let it escape the system…

I don’t see why you couldn’t use a blast of CO2 from such a cartridge to cool a heatsink.
As opposed to the AC system crosh posted, you’d have to replace the cartridge when it is exhausted… I geuss you could detect an empty cartridge when a blast of CO2 doesn’t drop the temperature on a heatsink.
A direct blast from a cartridge on a heatsink will work, a very fine mesh for it to blow through will enhance its effects… (the ‘deeper’ the mesh (more allong the lines of tubing I geuss), the more heat will be exchanged).

edit: keep in mind that those cartridges are exhausted fairly rapidly =P
If it needs to be used on a regular basis, I’d look for alternatives.

Enviromentalists don't like refrigerant in the air :p. R12 was superceded by r134a due to enviroment. There's another coolant they are working or researching on in Europe if I'm not mistaken. I'd say any refrigerant that's enviromentally safe would be best... You could have a heatsink like in the liquid cooled computer systems.

mmm ammonia ;-)

Sorry for the misunderstanding, I actualy mean the system Imahilus talked about... something like an emergency- cooling system.

I visualised something like this: Heat-critical application with temperature sensitive components that could accidentally overheat. So a temperature sensor will be attached to the heatsink and if a critical temperature is detected, gas will be released and the heatsink shock-cooled. After that, shutdown of system.


Even better would be to have two valves:

CO2 ===||=== pressure chamber ===||=== heatsink

So if I put the CO2 capsule upside down, I could fill the pressure chamber with CO2 and then on command expell only the small amount residing in the pressure chamber, correct? And if I can control the first valve delicately enaugh it should be possible to control the amount of CO2 used for cooling burst. I wouldn't want to empty the whole cartridge at once.

If you could recycle the system somehow you'd have an endless supply. Open up a valve to dispence in sealed containment. Once full, release/open recycle...

I have had a play with these cannisters.

I can tell you that they are a B*tch to try to puncture. You are going to need a nice powerful spikey thing to puncture it. I can confirm that they are good at emergency coldness as once they have been punctured they cool down to quite a bit lower than 0 degrees C. They do try to go liquid when you puncture them fast enough so that would suggest a temperature of less than -78 degrees C at that point. CO2 doesn't actually liquify at normal atmospheric pressure so it'll only go liquid while there is a high enough pressure in the cannister...

You can purchase cold spray cannisters... Maybe you want one of those and a servo or similar to push the spray can head.

If you could recycle the system somehow you'd have an endless supply. Open up a valve to dispence in sealed containment. Once full, release/open recycle...

Umm, well you would have to lose energy somewhere... You can't make things endless...

Mowcius

Heat-critical application with temperature sensitive components that could accidentally overheat.

If this has a possibility of occurrence, you try to design your system to minimize the chance of this happening - first step being to use a larger heat-sink. Beyond that, you start looking into thermal cutoff switches and other current sensing devices to shut off current flow and/or shut down the system, and if all else fails, you go with fuses or fusible links.

Finally, if you’re shutting down the system anyhow and you want to “shock cool” everything (which will probably cause further damage - who knows?), why -not- empty the cartridge; they aren’t that expensive.

I can understand what you are attempting to do, and maybe you are trying to save a $600.00 CPU (for all I know) - but I think maybe you are trying to design a system more complex than is needed, while trying to save money saving a part that is (or should be) more expensive than the system saving it…

:slight_smile:

I can understand what you are attempting to do, and maybe you are trying to save a $600.00 CPU (for all I know) - but I think maybe you are trying to design a system more complex than is needed, while trying to save money saving a part that is (or should be) more expensive than the system saving it…

Oh come on… Extreme cold temperatures are cool!
Maybe he’s just trying to cool an arduino. I don’t care and I think he should still do it!

Oh look. I have an arduino cooled by carbon dioxide. It can reach temperatures as low as -50 and then I can get it to as much as 4MHz more than its normal clock speed!

;D

Mowcius

Hmm the CPU was a good guess (I actualy want to make a homebuilt aruino enhanced watercooling system für my next computer... but thats still in the future)

Yes the coolness factor having something that is liquid carbondioxide cooled is actually one of the major reasons I'm thinking of building this :-)

Damage due to shock-cooling... didn't considere this, has anyone good information on this?

Also I could make an arduino-controlled CO2-marchmallow gun :-D

The pressure vessel thing is probably a dead end =P Rule of thumb: the more compressed, the colder it is. Thermal dynamics is fun stuff, dealing with all sorts of energy types.. (heat is a 'waste' energy in nearly all energy conversions) but likely far too complex to apply to a project idea like this =P

I also like the premise of this idea, even if it isn't all that suitable =P As for damage from the 'cold snap', if it is on a heatsink.. the effects will be a lot more gradual on the component it is attached to.

Hmm ok than its a one-valve design…

Any idea where one could get his hands on an electrically controllable valve??
(And I wouldn’t like to use a servo ← I’m past the skill level to use a servo for everything ;-))

cheers!

Or... instead of re-using the coolant like an AC system does, you can let it escape the system..

I don't see why you couldn't use a blast of CO2 from such a cartridge to cool a heatsink.

High school chemistry should give you an estimate of how much heat you can eliminate. Assume something like 12g of liquid CO2 at 20C becomes 12g of gaseous CO2 at 20C, so all of the heat absorbtion comes of the "heat of vaporization." Or do the whole PV=nRT thing assuming a perfect gas at cartridge pressure (800psi) vs atmospheric. Both are simplifications, and don't take into account efficiency get getting the heat out of whatever is hot, but I suspect that the numbers will be sufficiently depressing to eliminate the idea.

Umm, well you would have to lose energy somewhere... You can't make things endless...

ya you're right... What about liquid nitrogen (dry ice)? I mean this stuff is cheap if you can keep wayyy cooled in some kind of container. If it goes away could just refill...

If you go for the marchmallow gun, please consider firing a few rounds in my direction :-)

Hmm the CPU was a good guess (I actualy want to make a homebuilt aruino enhanced watercooling system für my next computer... but thats still in the future)

Yes the coolness factor having something that is liquid carbon dioxide cooled is actually one of the major reasons I'm thinking of building this

How much $$$ are you willing to spend just for a "coolness factor" which may or may not enhance the function of your gizmo? I've seen a "boiling" heat sink setup that used a liquid that boils at a temperature lower than water but is liquid at room temp. It had clear plastic sides and lighting so one could see the boiling bubbles streaming off of the heat sink attached to the cpu. Very similar to the "bubbling" christmas tree lights.

Any idea where one could get his hands on an electrically controllable valve??

You can't just use -any- solenoid valve; it has to be one that can withstand the pressure. I have seen small solenoid valves used in the coolant loop of a Cornelius ICEE machine (we have one at my work - my boss and I are constantly fixing that beast - picked it up cheap on Craigslist, tho). I doubt they are cheap, though I've never priced them (nothing has been cheap on that machine to fix/replace - heck, syrup is like $30.00 a box).

If you are really serious about doing this, you first need to do a ton of research, and perhaps find someone who knows what they are doing so they can help you. At the pressures you are talking about, one slip and you could give yourself a skin embolism (ie - forcing compressed gas thru your skin - painful, and potentially deadly).

Once you have that knowledge and help, then you'll need to source the parts; start by going to an industrial gas supplier - they can help point you to a supplier for parts (note: you need to find a place that sells to the public, unless you have a business tax ID or whatever is similar in your country). Expect to spend some money building such a system; gas handling parts aren't cheap.

Could you potentially cobble something together from parts at a big box home improvement store? Possibly, but I wouldn't want to be near it while it ran (the parts you can get at such places likely aren't rated for the pressures/temperatures involved).

I am not trying to discourage you or anyone else from attempting this; I just want people to be aware of the realities of technologies - both what kind of knowledge and experience would be needed to be successful, plus the amount of resources that will be needed to pull it off. If you were attempting to build an Arduino controlled steam engine, I would be just as concerned, even though the basic tech is just as old and older - simply because working with steam, boilers, etc is -dangerous- if you don't have the knowledge or experience.

:)

What about liquid nitrogen (dry ice)?

First off, liquid nitrogen and dry ice are two radically different forms of two different gases. Are you not aware of this?

I mean this stuff is cheap if you can keep wayyy cooled in some kind of container. If it goes away could just refill...

Yes, dry ice is fairly cheap, but I doubt there would be any need to get anything electronic that cold, unless you were working on some form of cryogenic research. Cray computers originally used liquid cooling, which was cooled by ordinary A/C chiller machines; we're not talking freezing temperatures here.

So - ordinary ice would suffice. For such a system, I would actually just pump a chilled (ice bath, for instance) coolant mixture through an ordinary water-block cooling system. It would still be overkill for all but the most demanding tasks. I honestly can't see the need for doing this and going to this much trouble unless you have a real need for it. Just to do it to say you've done it...seems a waste of time and resources for little gain.

On the other hand, if you were building a super-overclocked multi-cored cluster machine, each stocked to the gills with NVidia stream processors, all to do a parallel processed neural networking simulation for stock market prediction (or something similar) - where you needed to eek out every last bit of performance; then such a cooling system monster would be well worth the investment - hopefully!

:)

all to do a parallel processed neural networking simulation for stock market prediction

Didn't you get the memo? The markets already melted down, so need to cool it down. ;)

Hmm thanks for all the input.

I guess I'll drop the cooling idea..

So the marshmallow launcher isn't totally out of the question, I found an instructible on how to employ a bicylce CO2 pump which uses the same cartridge. But the thing was poorly built, mainly using duct tape.

Maybe I'll pick this later up but for now it just seems too much work with too little intersting things to accomplish.

Thanks again for all the helpful information!

cheers!