It's Time For Me To Go Extreme Cooling

[The Smith]

oct 4, 2008

 

Hi everyone,

 

With the recent aquisition of a C2D E7300, I thought I could try it out on dry ice cooling, as it could allow me to reach higher clocks, obviously. Right now I'm at 4.2Ghz, 1.5V, on air. I don't know where I'll stop, but then after I'll switch to dry ice to go further.

 

So I looked at some pots I could buy on the Internet. I saw many models, but God, they are really expensive! We're talking here of >$100, +shipping. I felt that it was not worth spending so much money on something I would use rarely.

 

Therefore, I decided to make one myself! This thread will be my worklog, so I'll update it when there will be something new.

 

At first, I thought to make it in copper, but then buying a big chunk of it would make the pot once again rather expensive. Then, Hardnrg suggested I could use my smithing skills to make one, but a steel pot is far from being the ideal, although it would have been original! Finally, I decided to go see what kind of metal I could get for cheap. I then found a piece of aluminum. Unfortunately it's not pure aluminum, it's an alloy. So the heat conductivity is decreased, but at the price I got it it was definitely worth it.

 

So here is the first picture of my worklog. It is the aluminum piece I bought for a mere 14$ CAD in a scrap yard. It is 25cm high, 7.3cm in diameter, and it weighs 7.2lbs. I don't know what this piece was before, maybe some sort of column.

 

So this is where I am now. I'll update the post in a few days. Next step is to draw a plan of the pot, so I can get it machined. A high-school companion will do it for me for cheap, so it will fit in my "Cheap pot" theme.

 

 

oct 8, 2008

 

Here are the first drafts for my pot, made with an academic version of Solid Edge.

 

THIS IS ONLY THE BASE. The smaller diameter part will tightly fit in a steel cylinder. I didn't include any measurements yet, but on the drawing, the aluminum piece is 7cm high and the holes between fins are 4cm deep and 0.5cm large.

 

Now the only thing is that I'm not sure if the guy will be able to carve these fins in the base... I'll get in touch with him.

 

 

 

oct 10, 2008

 

Here is another draft. I used the transparent view so you can see the fins.

 

This one is shorter and the fins are shallower. It will be easier to craft.

 

 

 

oct 13, 2008

 

Here is what the base will look like:

 

The base and the fins are thinner.

 

The machining is started. we will finish it tomorrow. The fins and the angle at the base is not done yet. Unfortunately I have no pics/videos of the making yet. Also, we will finally make the tube in aluminum, as it needs to have the same coefficient of dilatation as the base.

 

 

oct 20, 2008

 

Ok here is a big update. The pot is almost finished and ready to use. Look at that:

 

 

Obviously I will need to lap the bottom:

 

One bad thing is that when we inserted the base into the cylinder, the exterior fin has moved, so it leaves a spacing between the cylinder and it:

 

The second bad thing is that there is also a small spacing at the bottom on the exterior, on one half the circumference. I made some tests using water and unfortunately, it goes trough it. Slowly, but surely. IDK if this idiom exists in english though...

So I will need to weld it, but I'm not sure if I will be able myself. I have a semi-automatic welding machine with inert gas and MIG wire to do it, but every time I tried soldering aluminum I wasn't able to make a great job. Do you think there could be another way to weld it or to prevent liquid from passing through the spacing??

 

What if I melt the base and the cylinder together using my small oxy-acethylene torch?

 

What if I only put some epoxy in the spacing? BTW I can't remove the base now, it's too late. It has been inserted using a 30 Ton hydraulic press.

 

Finally, I also need to make the screw system to hold in on the socket. I already began the top wood piece:

I will make a backplate in a stainless steel 1/8" plate. And both pieces will fit socket 775 and 1366.

 

In conclusion, I think I asked too much of the guy... The round fins were very tough to do. He had to make a tool especially for that, which broke, so the fins are not as deep as I would have wanted. Drilling holes would have been much easier... But well, at least it didn't cost me an arm and a leg. (Another french idiom ) I gave him 40 box. It means I will have paid a total of 54$ for that pot. I'm already looking forward to make my next one.

 

 

oct 26, 2008

 

I will finally test my pot tomorrow. Here are some other pictures:

 

To attach it to the motherboard, I also needed some screws. However, there wasn't screws small and long enough to suit my needs. So I made my own screws from bigger ones I bought:

 

I also needed a back plate to hold the insulation on the back of the motherboard. Here it is, made of a stainless 1/8" plate. Later I will be able to drill other holes for the LGA1366. :

 

I also quickly lapped the bottom, however I didn't lose my time to make it shiny :

 

Finally, I insulated it using pipe insulating and hot glue:

 

So I'll install it tonight. If everything goes well, hopefully, you'll get some results soon.

 

 

oct 27, 2008

 

Here is the temperature. It's important to note that for the first time, I preferred not to mess with acetone. I know the CPU can skyrocket if you put too much. So the temps were not as cold as they could have been.

 

 

Here is, after the experiment, ice on the socket. Should have I put more insulation? I touched with my fingers the motherboard on the exterior of the insulation and it only got barely colder...

 

 

Every comments and suggestions are welcome.

Edited October 28, 2008 by The Smith

[Crazy_Nate]

$14 is a deaaal for 7.2 lbs of Al! Good luck man!

 

6061 T6? Just a guess...

[The Smith]

6061 T6? Just a guess...

I think. I'm not sure though, because I bought this piece from a... scrap yard? I'm not sure if it's the right word lol.

 

edit: I know it's a quite hard aluminum alloy...

Edited October 5, 2008 by The Smith

[Crazy_Nate]

Yeah, scrap yard works. I've used a many in my undergrad days. I could find random bits of steel for $0.50 a lb. Nice find

 

I'll be watching this.

 

Edit: T6 is just a general aviation alloy...that I've seen just about everywhere. Slight variances in composition will mean it machines differently. With that said, Aluminum is very very nice to deal with

 

You can probably turn the outside and bore the inside with a really long boring bar...then hit the bottom on the mill to make it fit well with whatever mounting solution you engineer.

[gotdamojo06]

im interested to see how well this goes! good luck man

[PeterStoba]

This will be interesting, on a side note, Phil made one from plastic pipe didn't he?

[kitfit1]

Hope you have some good clocks out of it. Will be keeping a watch on this thread. What mobo are you going to be bolting it to ?

[The Smith]

Hope you have some good clocks out of it. Will be keeping a watch on this thread. What mobo are you going to be bolting it to ?

Well I had the choice between the X48-DQ6 I just got for the reviews and my P35-DS3R. Finally I will go for the P35-DS3R, because if something goes wrong and it gets damaged, the loss will be less important. And this one can still provide me with 500+ FSB so...

[suchuwato]

Looks like a straight roller gear from something that's been cut... aluminium should work fine... what's a few degrees difference when using dry ice?

[Andrewr05]

Pretty neat, the ridges on the outside (as suchuwato pointed out they are the teeth as this is a straight cut roller gear) look cool and could possibly aid in cooling...

[Crazy_Nate]

I could say something about the "gear," but it's all speculative... Aluminum would not be my material of choice for a gear (think fatigue...high cycle applications in a gear)

 

Since you want to transfer heat between the CPU, and the medium (Dry Ice...which sublimates...and heads into the atmosphere - although not very far, because it is denser), you probably want to reduce the amount of heat transfer along the sides. I'd insulate it with foam (on the outside) or some other appropriate material. It'd also help reduce the condensation along the outside.

 

PS. Keep 'em comin'! I'm here to help :thumbs-up:

[suchuwato]

I could say something about the "gear," but it's all speculative... Aluminum would not be my material of choice for a gear (think fatigue...high cycle applications in a gear)

 

Yeah, true. I've seen something similar used in aircraft though (although on a smaller scale). It depends on the type of alloy it is.