Tubular Case Cooling
 

I've been thinking about building my own case for a while because it seems no case is perfect for what I want. When I was planning out the frame I saw some square tube aluminum. This sparked my memory of a show on the Discovery Channle where they ran oil through the frame of a motor cycle for coolant.

Before I go any firther I'd like to get input on the idea of running water through a welded square tube frame as either a passive radiator / large resevoir.

It's a valid idea, the question is, how well would it perform?

If you look at the current set of cooling parts, you'll notice that the heatercore is by far the most widely used. The reason is that it's cheap, and effective: the water is in turbulence, and there's a maximum of surface area for the air to pick up the heat.

Also, you have to consider that aluminium mixed with copper in a cooling loop, will accelerate corrosion significantly, so you'd definitely have to use an additive.

Ben's point regarding aluminum's corrosive propertys is quite valid.

This of course brings us now to the thought of making use of stiff plumbers copper tubeing for the frame. Round tubes would be harder is some other ways to use regarding attachment of the case sides ect. But copper would also be a better dissipater of heat from the water.

By biggest question is regarding just how much cooling you could hope to gain as the inside case temps are going to have a very pronounced impact on cooling in this tubuler frame.

but a interesting idea none the less.

Agreed.

By putting the heat load on the case, it transfers the heat dissipation surface outside of the PC, which carries the ventilation issue outside the box.

Strictly as a passive (no ventilation) system, and relying on convection from the case to cool the CPU, is something that may work, but certainly won't outperform a good old radiator (aka heatercore).

Using copper tubing which, to my knowledge is only popularly available in the round form, would solve the corrosion issue. I suppose that a "little banging" might create an effectively flat surface, but attaching it to the steel or aluminium case will create an additional thermal barrier, which will limit the effectiveness. Welding seems like a natural choice, but it's extremely difficult (read: near impossible) to weld copper to Alu. Welding to steel would still required that the weld creates an effective contact with the steel, i.e. minimal gaps. Thermal epoxy could be cost prohibitive.

Many challenges, but still possible.

Still interested?

If you examine the home brew radiators, there is not any more tubing in those than there would be in a case frame made of copper tubing. Then just use plexi for the rest. No need for a thermal bond. You wouldnt do this for extreme performance, you'd do it for passive cooling and "neato" factor.

It would work well for a resevoir, but that depends one the ammounts of tubing and its size. It would be really cool, but it will take lots of hard work to weld the frame, cap the ends of the tubes, and put in holes to allow the water to go through.

I did a little number crunching and if you used about 17 feet of tubing in the making of your frame. This would be square with internal dimensions of .5" by .5". You would get about 0.8357 L of volume.

So how would the heat escape? Or did you have an external frame in mind? Or were you thinking about an open top and bottom, to allow for natural convection to create an airflow?


Game_ender: Welcome!

17 feet, and 0.8357 L? Sounds about right.;)

external frame, of course.

Here's a new article by a guy who thinks like I do:
http://www.overclockers.com/tips1093/

I would have opted to either build the tubing into the concrete floor (rarely an option, unless you're having a house built ;) ), or enhance it by cementing it to the floor.

Not sure whether you are mocking me or not, but hear goes the break down:

17' (length of tubing) * 12" (number of inches in a foot) = 204" (length of tubing inches.

204 * .5 * .5 = 51 in3 (inches cubed) thats the volume of the tubing

Now using this converter 51 in3 = 0.8357 l (Using 'L' instead of 'l' was my only mistake).

Oh no, i trusted that you were right.

but since you bring it up, in round tubing, the ID (inside diameter)is more close to 9/16", aka 0.5625", which, when round, comes out to .25"^2. coincidentaly, it's the same as a square 0.5" ID tube.

It's all good!

Well, the good news is that I can tell you for sure that this can work.

The bad news is that I can say this because it's been done.

http://zzz.com.ru/119.html

One of the most impressive water cooling setups I've ever seen.

Check the message board at the bottom for his temps, some of the numbers sound a little to good to be true, but some very cool stuff none the less.

Oh that's no surprise. There's a fellow, in Australia, if I recall correctly, that built a case completely out of copper tubing, and forewent the panels alltogether. He had a much larger base though.

hmmm...a good idea, it really sound similar to this...
http://www.overclockers.com/tips1085/
I am kind suprised on the performace that he got by using this simple setup...by no means is it designed for high overclocking, but if you are looking for quiet, what you are planning sounds like a good option. Funny, I remember talking to a guy about something like this, but he was planning on attaching copper pipes to the side of a all alum case...therefore basically turning the whole case into a radiator. I do think that alum tubing could be used...with some caution. spend the extra $$$ & get them anodized...it will look better & help some fo the corrision concerns. Now if you decided just to use the tubing just for a resevior & have a radiator/fan in there too...sounds like a big res to me & defintaly could not hurt. You would be able to engineer some trick filler cap in the whole thing too. Fun project.

My plan was originally an 18x18x8 inch case. The longest flow plan I could make ran all the long lengths and bypassed 3 of the short ones. Basically that means 150inches3 of H20 or 0.6 gallons. Not much but better than nothing. I was planning on having a radiatior in the loop as well the frame would just be an extra.

The loop would be pump, radiator, cpu, split between nb and gpu, frame, pump...

The idea being the hotest water is dumped into the frame where it has the opportunity to lose some of the heat the the outside. Then its pumped into the rad where it loses the rest and the coldest possible water is sent through the hardware.

I thought that the corosion would be mainly halted by an additive and if needed I could always have the case anodized. I'm pretty sure the anodization should protect the aluminum against any galvination.

Thanks for all the input.

actually, i was drawing osme inspiriation from Rild over at ZZZ (sorry, didnt give credit).
IMO, some of the numbers he claims are retarded though:

Anodizing the exterior of the aluminium will not be of any help, in reducing the accelerated corrosion (a.k.a galvanic effect). You would have to anodize the interior of the aluminium tube. Also, you would have to make sure that all the joints do not leave any raw aluminium exposed to the water, because the accelerated corrosion would then be concentrated at those joints!:eek:

Aye...I think this is the biggest mistake when people mix copper & alum. They make a nice pretty alum top with some pipe threads in it. You then screw a barb into the tapered thread, which by its very nature is going to want to scrape a bit of the anodizing away...leaving you with very high potential for problems. Another problem with anodizing threads is the process tends to build up material in the root of the thread. Typically on the parts I have seen anodized the threads are masked from the anodizing process becuase of this...it builds up enough material to make it so a bolt won't fit. The other problem with anodizing the interior of the square tubing is it would have to be done after all the welding was done, making it damn tough to actually inspect how good of coverage was done. I have not seen square copper tubing, but I would not be suprised if it was out there. I have seen square brass tubing tho...might be ideal for intergrating a resevior into the frame for a case. Probabally would cost a few $$$ tho.