After all that Sith, you didn't tell us if we had a good theoretical design...... also we included BOTH water cooling...and chemical moisture management. Sooooo to summ up again, we want to use a water cooling kit with resovior in the freezer portion(no need for a radiator i think but if so that be in freezer too). In the water cooling system I want to use antifreeze or similar solution, to achieve subzero temperatures running through the tubing. Next I have suggesed several times to use a product known as DampRid which is a dry chemical which absorbs 150% of its weight in moisture(its main ingediant might just be calcium sulfate). So with a steady amount of this stuff in the fridge case we should be able to compensate for most of the condensation, right? I think if any one had an extra fridge to "operate on" we could test this with pretty easily. As a heating element we could use something simple like a curling iron......or a light buld like everyone else mentions.

So if you feel its theoretically possible...... have we designed any thing of a worthy prototype????/

sith you can hopefully answer some questions i have. do you by any chance know the cooling capacity of the average freezer you know the small box ones inside the main fridge ? also can you see any other problems with the two section idea ?

I just wanted to add some technical explanation to some of the elements you discussed.

The water chilling idea is superb, but be careful with an antifreeze system. They use a glycol which won't just evaporate if there is a spill inside your system. Sub zero temperatures is a great idea, but again, the key element isn't really the moisture removal, its the assurance that the case is perfectly sealed. Moisture removal is more of a backup for leaks. Room temperature air hitting a sub zero cooling pipe will immediately condense large volumes of water before it even gets exposed to the moisture absorbent salts.

I'm not sure what the flowrates are for your water cooling system, but you may want to use a small coil of copper tubing in the freezer along with the reservoir (ie a radiator). Surface to volume ratio determines how much heat can be absorbed. Its really a function of the reservoir size that determines if this is needed. If the water in the tank cycles once every five minutes, well, we all know how much colder a warm can will get in a freezer in just five minutes. This system is going to be fighting more than just the heat from the computer, but also the natural operation of the freezer itself. Remember, the better you seal everything, the easier that's going to be. That goes for the freezer as well, so get out your caulking gun.

The use of an iron is a good idea, but be careful this doesnt cause air leaks from the cord. Are you trying to simulate a certain wattage? Try something more similar to a lightbulb. A desk lamp perhaps? The iron gets awfully hot and gets there quick. I would estimate the average iron can overpower a simple fridge because (like I said earlier) its easier to create heat than remove it. Either way, its an interesting feasibility study.

Good luck. One of the key problems we had with the nitrogen system I made was avoiding getting too cold! LQ N2 isn't just sub zero, its around 70 K. Thats minus 200 C! We ended up introducing mechanical stress on elements such as the chip, simply from the temperature differential. The cold side contracted, the hot side expanded, and actually broke a memory stick. There was no visible damage, but the warping was evident under SEM. So watch that you don't try to go too far in an uneven manner with your heat exchangers.

so assuming everything is insulated and sealed it should work?

Wow, that was an interesting point thanks sim. Hmmmm I'm not 100% sure but I believe we also agreed not to keep the freezer on full chill, but to keep a few degrees below room temperature. I suggested anti freeze due to the fact that it should be able to both transfer cold, and absorb heat more readily than water. Also I appologise but I meant to also add it was to be a water/anti freeze mixture. Anti freeze being a bit thicker may cause some uneeded stress on pump.

Also noting, not keeping the freezer soo cold, but definately in the cool range should help with condensation. You are correct though, I did my own experiment with the mositure absorbers. Their is another thread that speaks about building an ac (basically its a box full of ice with fans). The problem with the idea was humidity. I used two 12oz absorbers (each to cover 12sq ft) directly in the box with ice next to the out fan. The air comming from the ac unit seemed to be cold, and as far as I could tell .....dry??? Well I tried some things to tst for the moisture, lastly using another absorber infront of the fan... this remained dry. But thier was still some condensation in the box.

Lastly, I'm sorry to hear about your Nitrous system not working. Too cold, guessing you meant that it literally was too cold, or cooled too quickly. Either would casue warping I'm sure. And I'm sure you tried these Ideas but hey you never know. How about using less nitous at a time, or allowing the nitrous to mix with another gas as it hit your heat sink. OR making your heat sink out of another material that transfered cold more efficianctly..... or a THICKER heat sink. Make the system cool the heat sink and not the processor itself. A kind of indirect cooling.
I'm sure that if the problem was making it too cold, then lessing the efficancy of one aspect may be enough to counteract.

That may be an interesting idea, to simply set the ceramic cup filled with LQ N2 on top of the heatsink. Guess I might have to get things back out and play again. One of the neat things about LQ N2 was not having to worry about moisture as much. Any condensation on the ceramic cup simply froze to it. Spilling LQ N2 was no big deal either. We were using a fairly old system. Perhaps a more massive heatsink like on modern day machines would avoid the problems. TY for the idea

Also, if you're wondering, water normally has a higher heat capacity than the chemicals used in antifreeze. Hell, water has a higher heat capacity than most anything liquid phase. Its a result of the small molecular size and massive hydrogen bonding. Fun fact: water is one of the only substances where adding pressure can cause it to transform back into a liquid, rather than a solid. This is why water expands when it freezes as well ^^ It's how ice skates work! I'd avoid antifreeze if you arent going below zero. To kill the moisture, Id seal up the system, start an internal fan to encourage mixing, and let the salts do their job. Get the water out of the air before you start bringing the temp down. Also, one of the key elements of a labratory is environment control. Humidity shifts from day to day can dramatically effect a lot of experiments. This may kind of seem dumb, but it would almost be easier to build the system in the winter. Humidity is lower, you may have better luck trapping less moisture in the case in the first place, by building it outside (most homes add humidifiers) on a low humidity day. Even fluctuations as it rains/doesnt are significant. Watch the humidity outside and try to aim for a good day to build!

Hi,

I've also considered using a mini fridge as a case. How well would it work if you use a metal plate between the fridge cooling element and the CPU heatsink, as you'd be relying on conduction rather than convection. If you fill the unit with CO2 you should also get rid of any moisture issues.

Actually mounting the board so that a seal is maintained and the necessary external connections (Power, video USB etc) are available could be interesting though.

Anyone here think it's viable?

Thanks,

Zummi

the product that floppy linked to on the third page is basically a fridge compressor hooked up to a copper block that sits on the CPU

Interesting points raised. I really think a PC in a fridge (or a freezer) would work very well.
My objectives for this would be:
1. No dust in the PC or any of its internal components. Dust causes heat to accumulate and also makes the fans more noisy.
2. Run the PC at a low temperature to allow reliable overclocking.

The PC does not need to be modified; just put the whole case in the fridge. A remote power switch would be handy as would external removable media drives.

Out-of-scope:
No need to be able to serve cool drinks. Every time the box is opened, moisture and dust are introduced. However, a more expensive 'frost-free' fridge/freezer might allow this.

Would prefer a glass front door so the lights etc. show.

Provided the fridge is not opened more than say once a day, I doubt that condensation will be a problem. Unlikely to open it more than a few times a year anyway. The reasons why there won't be a condensation problem are:
1. The PC will always be the warmest thing in the fridge, so condensation will go there last.
2. The fridge will always be the coolest thing, so water-vapour will condense where it's safe, drip down to the drainage system (which most fridges have) and out to the evaporator (sitting on the hot compressor). That is why (unlesss the door-seal is damaged) the inside of a fridge is always dry. Definitely drier than the outside.
3. The PC fans will all be working so you would need a huge amount of moisture for it to settle.

The only time condensation would happen is if the (cold) PC is removed from the fridge. To avoid this, just switch off the fridge, switch off the PC ( so it doesn't overheat ) and allow it to warm-up to near room temperature; (air would be even drier now) then open the fridge door for a while... should be fine.

Would need to drill a hole ( at a safe location in the back of the fridge ) for the cables... re-seal with expanding foam, etc.

I doubt that a wine-cooler will be powerful enough for this... a freezer is probably best (because the whole system would be able to work harder and not give up) but with its thermostatic switch modified such that the unit runs a bit above freezing... placing even less load on the compressor etc.

For even cooler and more efficient operation create a fanless CPU-heat-pipe, similar in design to say the one at http://www.silentpcreview.com/article114-page1.html
but if possible connect it between the CPU and the chilling metal component inside the fridge / freezer (it could perhaps be soldered to the cold metal tubing in the freezer).

your gonna need to open the fridge to use the CD drives, unless you get some external ones, in the design i came up with it doesn't matter if its open all the time the PC will still be cooled because the compartment the PC is in is roughly at room temp, the only thing would be dust going into the fridge and settling on the PC components, an exhaust would be needed because it needs to move the heat created by the PSU and parts that can't be cooled by the watercooling system

food and drink can be put in the cold compartment because the only thing in there will be the radiator and resevoir so there should be plenty of toom for atleast a 6 pack of beer

a glass door would be a bonus but i haven't seen any fridge freezer combos with one, the condensation will be minimal because the only cold things will be the tubes and blocks of the watercooling system, all of which would be heavily insulated

there would only need to be the one exhaust fan everything else would be cooled by the watercooling system

port extensions onto the side of the fridge would be the best setup purely for making it easier to move the setup, making holes to plug things in is an option but adding a new peripheral would be difficult, also the holes wouldn't need to be sealed because a minimal amount of dust would go in and the exhaust fan would blow it back out

the cooling would not need to be modified because the thermostat would keep the temp constant and not too extreme, leaving the pump on constanlty could cause the temps to be too cold

not knowing the relative efficiencies of watercooling and heatpipes its hard to say which would be best, finding heatpipes long enough could be difficult and expensive, it would also mean that a suitable block would need to either be made or bought and then brazing the two together if possible without ruining the heatpipe, for this reason it would be far simpler to buy a watercooling kit and use that

obviously having everything called by a single watercooling loop may be asking a bit much of a single pump but spliting the circuit into two and having a pump for each but still using a single radiator and resevoir would overcome this problem, if you really wanted to get maximum cooling for overclocking you could have a radiator and resevoir on each loop