Helium results!!! Window air con chill box build.

Hi all,

I'm new here and want to share my project that I've just started with you.

The initial project objective was to provide better than ambient water cooling temps by building a chilled sealed and thermally insulated box in which my pc components would be housed and use air cooling. The air in the sealed enclosure will be cooled by the evaporator of a window type air conditioning unit.

My hope is that the substantially reduced air temp in the enclosure will boost the efficiency of the air cooling components and so once the enclosure is built then my future pc builds will be relatively simple air cooled builds not requiring the complexity of water cooling and yet delivering better temperatures.

This is an experimental build and I'm yet to see how efficient this will be.

Since starting the project I've realized how close the air cons evaporator will be to the major components. So I also want to try and thermally bridge them using many metal strips or wires connecting the fins of the air cooling components to the fins of the air con evaporator. Having done some rough thermal conductivity calculations I feel sure that this will provide extremely low temps so long as I can marry them together adequately.

Again this is also experimental and I'll have to see how effective either method actually turns out.

So with that said my ac unit arrived today so lets start logging my progress.

So here she is, in her pristine glory before I get my screwdriver out, all 12000btu of R407c goodness.



Uploaded with ImageShack.us

The front panel came off real easily as did the outer case which basically slid off as a complete sleeve leaving all the internals on a base tray.



Uploaded with ImageShack.us

The first problem was the electrical control board, I had to disconnect everything from it and push the wiring through a hole in an internal metal shroud.....just tagged all the cables and photographed them so I could remember where they went



Uploaded with ImageShack.us

Still soon got the electrics out and reconnected them to the side of the unit so I wouldn't forget anything.....when you're 50 that happens a lot..lol.



Uploaded with ImageShack.us

The next thing were those 2 metal shrouds either side of the evap fan and the polystyrene moldings..... these were quite difficult as everything interlocks together and you have to be gentle working around the evap so as not to damage it or the pipework.

I had to resort to bending a long flange down the one side of the metal shroud nearest to the evap so I could slide it out. I was then just able to get my hand in and remove the fan from the motor shaft.
The polystyrene molding had to be broken out bit by bit as everything was tied in.

Finally I was able to remove the larger metal shroud which also supports the fan motor. I removed this so I could cut the sides off and so make a large gap between the motor and the evap once I've cut the shaft on the motor.



Uploaded with ImageShack.us

It's actually working out quite nicely, I've got a 17cm space between the evap and the ac fan motor, that gives me plenty of room to go through the wall of my chill box and place a home made shroud and and fans behind the evap blowing into the chill box.


So another day done...an some more progress...I've cut the spare end of the fan motor shaft off...whilst doing it I covered the end of the motor to stop metal fillings getting into the bearings with a piece of paper...simple but hopefully effective.



Uploaded with ImageShack.us

Then I cut off the projecting end of the base tray so I can get my chill box as close as possible and sitting squarely......was tough going even with power tools.

Also mounted it on a piece of mdf on casters (thanks Drew)...makes it much more maneuverable.



Uploaded with ImageShack.us

And from a different angle so you can see the nice space I've got between the evap and motor.



Uploaded with ImageShack.us

So all seems to be going good so far, I've tested her out after all my pulling about and everything seems okay, evap was developing a nice frost anyways...

Next stage is to drop her into my chill box.

On that note I'm a little concerned about that insulating material on the pipes as they approach the evaporator. I'm going to have to remove it so I can seal around it as the pipes enter the chill box......just not sure if that is going to cause a problem....just don't know why its there really?
Edited by technogiant - 7/7/12 at 7:34am

Thanks for the welcome and the advice Drew.

I've made pretty good progress today in that I've completed the chill box and done some preliminary tests.

I already had a chill box of sorts built for the project but as my ideas on the build have changed I had to chop it about a bit, but luckily it actually adapted quite well. That's another thing I've learned with this build, don't go too far ahead of yourself as your idea's will change and evolve.

The pictures are self explanatory so here we go.



Uploaded with ImageShack.us



Uploaded with ImageShack.us



Uploaded with ImageShack.us

Actually that makes a nice little shelf, I was wondering where I was going to put my hard drives....sorted!



Uploaded with ImageShack.us



Uploaded with ImageShack.us

It's all gone together quite nicely. I tested the fans and they seem to give a good circulation through the evap (3x 120mm scythe gentle typhoon 5400rpm 150cfm each).

I've also tested the fans and air con working together briefly......WOW....I've unleashed a beast.

I haven't got my temp probe yet but used a simple bulb/alcohol thermometer.

To be honest I'm a little scared of using it.....so only briefly tested it out.
with the lid off holding the thermometer in front of the evap the temp quickly plunged to -5deg C.

So I stuck the lid on.... within 15 sec or so the temp was down to -20 dec C or there about.....tbh I was in a bit of a panic to read it properly as my fans were frosting up and I was in a rush to shut down the beast.......I'm going to have to get some control on this me thinks.

I wanted to do some load testing but there is water from condensation on my fans so I think I should wait till that dries off before I try running them again.
Don't want to blow them or my psu particularly.
Edited by technogiant - 6/3/12 at 9:34am

Thanks, I'm hoping I wont get too much of a problem when its under load but was kind of hoping the condensation would be restricted to just the evap/rad. Will have to sort this before my pc components go anywhere near it. Atm I'll be using my current rig which is in dire need of up grading Q6600 @3.5GHz and 2Xgtx460 2GB models in sli @820MHz. But once I've finished experimenting I'm due an upgrade...not sure what yet though.


They're 120mm Scythe Gentle Typhoon 5400rpm 150cfm models...should give me enough air flow
I'll keep the post updated Decali....can't do much tomorrow but my next job will be insulating the pipe work back to the compressor as Drew suggested.....then I want to test again with some load and see if I still get condensation problems.

One more pic so you can see how it will finally look.



Uploaded with ImageShack.us

Looks a bit of a mess with the silicone everywhere...but I've been chopping things around as my ideas changed. (PS that's not my hand, I don't wear nail varnish...anymore )

As you can see the cooling fins of the graphics cards and cpus cooler are real close to the evap rad fins...thats why I want to try and directly thermally bridge them rather than using air or water as an intermediate medium for heat transfer.....we'll have to see how that one goes though.
Edited by technogiant - 6/3/12 at 1:40pm

Thanks Malo...It's early days but going well so far


When I decided to do a chilled build "mineral oil submersion" was my first thought because of condensation. But it has a lot of problems, firstly the oil attacks many plastics and pc components, so you end up having to seal components against the oil instead of condensation, so that is kind of self defeating.

Secondly if you are considering sub ambient cooling then due to its increasing viscosity with lower temps its not very useable below 10 deg C.

Thirdly its specific heat capacity is about 1.7 compared to 4.2 for water, so there are problems when you consider how you are going to cool the components, you can't pass it through a water block as it doesn't cool as efficiently as water and the block being designed for water would not have sufficient surface area.
If you use air cooling components, how much movement of oil are you going to get between those cooling fins spaced to allow air not thick oil to flow through them?

So all things considered.....meh.

My hope by doing it this way, certainly if I stick to the air cooled route rather than the "thermal bridge" route, then the pc components will always be hotter that the surrounding air and so should not attract condensation.
Of course I say "should not" as this is experimental.....if condensation is a problem then I will have to offset the thermistor so I can control the temperature in the range of my choosing not that set by the air con mrf.
Edited by technogiant - 6/3/12 at 4:49pm