Nitrogen Atmosphere Cooling

[Emperor Que]

Hail all, I'm starting a project that involves creating a 100% sealed Nitrogen Environment. Then use Peltiers to cool it. The bigest problem with any kind of cooling is the constant threat of condensation, By useing a 100% nitrogen environment with only 4parts per Million of water it serves as a perfect platform for cooling. The whole thing is hard to describe but I'll post more if ppl are interested. Plz tell me if I'm a quack for thinking about doing this. (But I've already started) he he

Added bonus with nitrogen environment is that it's never refreshed with ambient warm air. So the only heat is coming from the Cpu and Gpu.

Anyways just wanted to post and get any feed back possible.

[8-Ball]

Would it not be simpler to make a sealed container and throuw a couple of those dehumidifier sachets in.

8-ball

[MadDogMe]

Does the nitro enhance cooling or is it purely to eliminate condensation?, cause condensation proofing a TEC setup is'nt that much of a biggie, far less than sealing a PC case gas tight!. a few pieces of neoprene, some silicon sealant, a jar of Vasilene is all...

[bigben2k]

It's certainly a better idea than some others I've seen: this one actually looks like it could be beneficial.

Your main issues will be:
-keeping the environment sealed
-getting rid of the peripheral heat

Details would be nice...

[Emperor Que]

The thing is that I've done the whole silicon and neoprene trick and honestly it's messy and makes the mainboard a little goopy. Not that that's a big deal it's just I really want to try something different. The biggest potencial is that you can still use any kind of heat exchange you wanted on the CPU or the GPU, water cooling with peltiers ect. Which i'll most likey be trying once I have a seal atmosphere. If I can eliminate the ambient temp, it's one less enemy I have to battle. And true it's maybe overdoing it, but I'm not sure if this has ever been done before unless on an industrial platform. Now I doubt i'll be able to get the ambient down below 0c but like I said once I have a seal atmosphere I can install just about any kind of cooling system onto my CPU.

Have to run to work I post specifics when I get home later tonight.

[Skulemate]

Originally posted by Emperor Que
Added bonus with nitrogen environment is that it's never refreshed with ambient warm air. So the only heat is coming from the Cpu and Gpu.

There are lots of things in a computer case that make heat besides these two things. If you don't do something to remove the heat from the case the nitrogen will heat up, just as air will. If you watercool your CPU and GPU you can effectively remove these two heat sources, but what about the PSU, hard drives, memory, voltage regulators on the motherboard, etc? Perhaps you could add a chiller coil into the sealed chamber, and then circulate the air past this to keep the air both moving and cool.

[myv65]

I think it's a little optimistic to expect a perfect seal. You're still going to have quite a few electrical connections and these aren't designed to be air tight. Perhaps you would consider a "nitrogen pad".

We used these on hot oil systems where oxygen would cause accelerated degradation of the heating oil. Think ~5000 gallons of cooking oil heated to ~550°F and flowing around 500 gpm. The system could not be completely sealed due to the change in volume heating that much oil from room temp to 550°F. We used an expansion tank with relief valve (~5 psig) to handle the initial expansion and had a bottle of nitrogen connected at ~1 psig. You don't need much positive pressure, but any "gap" would have nitrogen flowing out rather than exchanging with ambient air.

A dessicant could be added as an additional safeguard.

[bigben2k]

Lookup "explosion proof" plug and outlet: if I remember correctly, they are airtight (but expensive). Good point Dave.

Here's a stupid question: how would you fill it up? I mean, how do you evacuate the existing humidity?

I like the dessicant idea, but how do you maintain it?

[Skulemate]

We use salts in the dessicator at school, and all you need to do is replace them when they become saturated. Some will even let you draw the accumulated salt solution away, leaving pure crystals behind.

[myv65]

Explosion proof. Yeah. Good luck finding explosion proof USB connectors, 15 pin D-sub, and 1/8" audio plugs. A better bet would be to build a sealed box around the ports and pot the lines passing through the box.

A dessicant is fine as once you get everything sealed (or go with net positive pressure) you should theoretically have no further intrusion of moisture. In those events where you break into the case, the dessicant can typically be recharged by a brief stint in the oven. It's only real job is to "clean up" the moisture left over when you close the seal.

[jtroutma]

I came up with an idea....

You could keep a bottle of nitrogen next to the system at all times and keep pressure on the system from the bottle. Of course you would need to "charge" the system first by removing all the O2 + humidity by using a vacuum and replace the vacuum with nitrogen gas. Once the system is "charged" with nitrogen, you use the bottle to keep a pressure on the system. If any gas is released due to a slow leak, the bottle will keep a constant pressure on the system. When the bottle gets low (hopefully not too quickly), refill it and keep going.

Just imagine having a N2 tank attached to the side of your computer....you could claim that its Nitrous powered..

You could use a heat exchanger to cool the nitrogen gas sealed inside the case and cool the exchanger with an outside source. The water cooling system would HAVE to be completely sealed inside the case though or else any humidity would immediately collect on any pelts or cooler surfaces due to being a different molecule inside the almost pure nitrogen environment.

This could be done but would require A LOT more maintenance than a standard H2O cooled system......

BTW it would be best to have an external box with all your cable hookups.

[RoboTech]

If you're going to go to all the touble, you might as well use Helium - almost 6 times the thermal conductivity of nitrogen...

[Alchemy]

How is pure dry nitrogen better than dry air? Air is 80% nitrogen already.

Seal it with dessicants as everyone else says. Much smarter.

If you want to purge and fill with dry gas already, spray DFE in there with one of those "compressed air" cans.

Alchemy

Edit - I really like Skulemate's idea.

[merlinx29]

there are other components in a computer that heat up other than the cpu, video and chipset... unless you plan on cooling the nitrogen inside the case with some sort of heat exchanger sealing the case is not such a good idea and it seem like a big hassel having to get it air tight


good thinking though... never give up on the quest for perfect cooling

[Joe]

My question is why Nitrogen. Why not Aargon, or one of the many other noble gas's. Cost?

Why not CO2, its cheap.

[Emperor Que]

Answers: Jtroutma you got the idea.

I’m fabricating an Acrylic Cylinder about 16 inches in Diameter and about 16-18 inches in length. Inside is just the motherboard, Cpu, GPU and as Skulemate pointed out the voltage reg as heat sources. The Psu, hard drive and cd-rom are external of the cylinder. I’ve found Military spec Explosion proof Cannon plugs that’s I’ll use for the Atx power supply and 12v,5v system. Also included is atx switch and leds if I really want to.

There’s a product out their name Amp tape that’s used as a general sealant for the Telephone companies and DSL providers and provides many other uses. This tape is rubbery and is rapped around cables eg. my Video, USB 2.0, RJ45 and some others that will fit through a hole and with Silicon and other methods seal it, but the amp tape is doing most of the work in filling in those tough to get (round edges). The biggest problem so far is how to get my ATA cables outside or at least just one for my hard drive. (Hard drives don’t like the cold) I think I can do it by cutting a slot and use the amp tape combined with small acrylic plates to create a seal and some silicon.

Keyboard and mouse will be wireless, I’ve considered doing a lot more wireless things like USB and Network but the cost will be more insane than this project.

I’ll use a vacuum pump to suck out the O2 and after a short time turn on my nitro tank (Laying beside the cylinder) and purge the cylinder. Then turn the valve off to the vacuum pump, then turn off the pump and let the nitro build up pressure inside.
One of the key points is that I’m only putting 3 to 5 PSI inside nothing major. And with a Regulator I can control the amount of Nitro that goes inside. One thing to remember is that as long as I have a positive pressure inside the cylinder nothing can get inside (regular air).

Ok great now let’s just assume that this all works and I get a 99% seal nitrogen environment a little leak is ok just not a major one. Skulemate also pointed out that heat still works in a nitrogen environment absolutely correct, so I’ll be installing inside the cylinder a coil of copper tubing pumped with water. I’ll have 2X120mm fans circulating the nitro around inside. Here’s what I got for the cooling….

From a reservoir of water, water is pumped into a coil of copper tubing attached to a cold plate with has some peltiers on it, the tube comes out of the Cold coil to the coil located inside the cylinder cooling the nitrogen environment. The hot side of the pelt is cooled with tradition air-cooling and heat sinks. The tubing from the cold coil is insulated as it travels to the coil inside the cylinder just to make sure I don’t lose anything in the process. That deals with bringing the ambient air temp down to maybe +5C Not sure if I can get into the negatives but I’m guessing NO, now compare that to the +15C of the ambient air inside of a house, it’s a little but a lot.

Now I’m still deciding weather to use water cooling on my CPU, GPU, If anything I’ll go with water cooling with out support from Peltiers, that is until I can get enough money to buy good enough Pelts and a power supply that is rated for the big pelts. I may even go with traditional Air cooling for my Cpu, GPU and volt reg. Not sure at this point still need to see what kind of results I’ll get with just cooling the ambient environment. That means at least one more water Reservoir. I was also thinking that if I had my Water reservoirs in a bucket (Crude) or something I could throw in some Ice to cool the water down, and Insulate the Reservoir.

Now to the Question of Why Nitrogen?
A: Non corrosive
B: Only 4parts per million of water atoms (Regular Nitrogen, I’m not using a Spec Gas/research grade, UH+ or UHP)
C: Nitrogen won’t hold Positively Charged particles at least not a lot. Co2 tends to hold onto dust and hair and small positively charges objects.
D: It’s not Explosive in it’s natural state
E: It’s not toxic and you don’t need to put silly stickers on your car when you transport Nitrogen
F: Helium is ok, but It’s expensive due to the fact that is has to come from Decaying Hydrogen pockets from Oil wells or under the ground. Also if there was a major leak I’m not sure how many ppl would like to have silly voices at a lan party.
G: Nitro is cheap and there’s a lot of it in the Air we breathe
H: Nitrogen is easier to cool down and heat up

I think the using the Salts is a good idea, I’ll keep that in mind.

Any ways that’s the Huge spiel so far, hard to get into specifics but ya.

Hope that answers some question it was hard to get into detail before on the other post due to time. Plz give me more feed back, I love it. It's great to talk to ppl that know what I'm talking about.

I'll try to post some more Dimentional data in the next few days specifics about size and stuff.

[Alchemy]

Originally posted by Emperor Que
I’ll use a vacuum pump to suck out the O2 and after a short time turn on my nitro tank (Laying beside the cylinder) and purge the cylinder.

It's not oxygen, it's air. It's 80% nitrogen, or 70% by weight.

Also:

A: So is air

B: Dry nitrogen is dry, as is dry air.

C: Whether or not you're right about the ionization of nitrogen, it shouldn't matter because dust particles cannot collect in the case if it is sealed.

D: Neither is air

G: Dry nitro might be cheaper than air, so you might be right here. That may be reason enough for your plan.

H: Thermal and thermodynamic properties of air and nitrogen should be very similar

I mean, it's not a bad idea, but if you think nitrogen is some wonder-gas that will work any better than sealing the case with air, you may be disappointed.

Alchemy

[Skulemate]

Originally posted by Alchemy
G: Dry nitro might be cheaper than air, so you might be right here. That may be reason enough for your plan.

I'm not sure I follow... is there some cost for air that I don't know about?

Sounds like you have quite a plan Emperor Que. I do feel that watercooling the CPU and GPU is definitely the way to go, especially if you're using pelts. Just this simple thing will remove the vast majority of the heat created inside your cylinder and dump it outside. It would make cooling the inside of the chamber that much easier.

I was wondering if you could build an air chiller for the cylinder using a couple of heatsinks and a pelt (I bet it would be a fairly low power pelt at that). Have one heatsink in the chamber that's cooled by the pelt, and circulate air through it. The other sink cools the hot side of the pelt. I am envisioning some sort of setup where the two HSFs are on opposite sides of the cylinder walls, so you'd have to deal with another hole, but it may be worth thinking about (then again, maybe it's just a dumb idea). Alternatively, if you used a waterblock to cool the warm side it could simply be added into your coolant loop, and be contained completely within the loop.

[EDIT] One other thing to consider, unless I've missed it: how are you going to prevent heat loss from entering the cylinder from the outside environment? Are you going to insulate it? If you don't I'd be surprised if you reach an internal temperature of say 5°C.

[MadDogMe]

MkII:

Does the nitro enhance cooling or is it purely to eliminate condensation. I thought it was only cold when liquid and evapourating?...



_______________________________________________

PS, George Bush has kindly offered to have my PC launched into orbit to avoid my peltier condensation issues, he refused my satilite uplink request so I'm now in the market for a (multi plug) extension lead(s) :? ...

[8-Ball]

The N2 is purely as a gas which can be bought "dry". And for the other reasons listed, not for cooling. However, this can be cooled without condensation problems.

8-ball

[bigben2k]

I'm surprised that gmat hasn't made his negative pressure speech yet! In any case, positive pressure is what is required here: Only a minor pressure should do, and it would be easier to keep the system filled, if there's a leak.

So which dry gas is cheaper to get? Is it nitrogen? What about CO2?

Joe: I think that Argon (or other noble gas) isn't necessary, because he's doesn't need to use a chemically inert gas here: I don't think there's any expected reaction. I like the CO2 idea.

As for cooling the insides, it's only needed to compensate for the heat generated by the mobo. As pHaestus pointed out, a waterblock will still dissipate up to 20% of its heat in the surrounding air, so if we're talking about a CPU making 100W, the cooling solution would need to pick up 20 Watts. To that, add a few watts for the ram, and then there are those MOSFETs, chipset, etc...

Seeing that the heatload isn't insignificant, I would opt to use a heatercore inside the tube, for a more efficient heat transfer, instead of a copper tube. This is where a BIX would be nice, because it's small.

This is getting interesting. I never thought I'd see a heatercore used "in reverse", for a PC!

[MadDogMe]

Seems totaly pointless to me :shrug: ...

There is a lot of heat generated by lots of various components, these will all build up until you have a high (comparitivly) equalibrium, I'd guess between 30~40DegC, my mobo reaches 30DegC out of case!...

All that work just to prevent condensation?, a maintanance nightmare! , why not just thermostaticly control the TEC's so they don't dip below the dew point?...

Here's an interesting linky on mobo heat: mobo thermal imaging...

[Emperor Que]

The only reason for using Nitrogen is cause there's little moisture in it and well just cause. I mean I could probably use Co2 but at this point Nitrogen seems just as good. From my paintball Experience Co2 is hard to find pure with out going into to a research grade or spec gas. But I mean it’s totally possible to use Co2 but Nitrogen sounds cooler He he. Sorry not much of an answer there.

In terms of maintenance well, with a sealed environment no dust enters other than me opening the case up every month or so, and the vacuum pump will help keep the dust levels down. So basically there is not much maintenance.

After thinking more about it I might go with a Phase change cooling system and use R22, but I’ll have to do more research about it. The phase change cooling would replace the Coil of copper tubing with Peltier’s cooling the water. Any links or suggestion about that would be great, just reading the other forums seem to give good amount of detail about how it’s done so I’ll need to read for a while.

Now Skulemate was right about probably losing some cooling through the fact that my Cylinder is Clear acrylic ect. So there will be a loss there but I’m thinking that if I use Phase changes Refrigerator style I think I’ll get better results.

Oh and I’m surely expecting a ton of condensation on the outside of the cylinder, I may insulated the outside but for display reasons I’ll keep the insulation off. But it’s something I’ll keep in mind.

Now Bigben2K have I missed something about using the 3 to 5 psi pressure to maintain a positive pressure to keep regular air out? If So plz explain.

[myv65]

Three to five psi is way more than needed to keep ambient air out of the system. Heck, a few inches of water would be more than enough pressure.

As for condensation, plastics tend to be fairly good insulators. If the inside of your box is within 10 or 15 degrees of the dewpoint, I'd be surprised if the outside surface would get cold enough for condensation.

[MadDogMe]

I was thinking more in the line of 'resetting the c~mos' or 'reseating the block' incase kinda maintenance (I finaly know how to spell maintenance!, w00t!)...