What should a good dry ice container design consist of?

[fareastgq]

I'm totally new to it guys, I just read through everything here, hehe. the more I read, the more interested I became.. soo... I especially like some of your guys' designs. I see slight variations in designs from the mouse pot to some holes, to cylindical bands on the inside. I would like to autocad one but have no idea what a good design would be other than put some holes in it and maybe a few bands, hahaha. The purpose of that is surface area, no? if so, why not make somthing that has more holes or more surface area? cnc limited?

[afireinside]

Make sure it's solid copper. Brazed can not compete.

[gclg2000]

Make sure it's solid copper. Brazed can not compete.

I'm willing to bet its not that much mo' better....(i know i'm going out on a limb) but if we machined a thick walled copper tube and a thick base and "brazed" (with brass) and NOT silver solder....the copper alloy would do pretty good i bet.

But what's the point of machining a chunk of copper into a tube to just braze it up right???

I have a chilly1 dry ice tube i made myself w/ one of his blocks...the alcohol was covering the block up...(a chunk of copper in itself) and i saw 1 C drop on the mobile chip M 735 i did it to....i ran it just over 3.0ghz....byt a dothan don't get near as hot as some chips.

[afireinside]

Dothan will hardly heat a tube up, especially at low voltages and clocks like that one. Anyone who has used a solid unit can tell you how much better it is. If you take a look around you'll notice that NONE of the top dice/ln2 benchmarks have been done on brazed tubes. They can NOT hold a load at all. I saw a video of the temps on a brazed DI unit with a pretty good base shoot up like a rocket during 1M PI while my solid unit here sits at -62 to -60c through an entire 32m run at much higher heatloads.

[felinusz]

I'm guessing that's because the full copper tubes conduct heat consistantly throughout, allowing the entire tube's surface area to act as a heatsink of sorts, while the brazing prevents efficient heat transfer from the basepiece to the vertical sides of your tube?

[harleybro]

I would also agree that the solid tube is definately the way to go. As far as designing one thing to take into consideration is the cpu you plan on cooling and the method of cooling (DI or LN2) I am sure one design can do both well but I think w/ LN2 the surface area plays a greater roll in cooling capability. I personally can't comment on holes or base design as to which is best but once again over the past year I have used a bunch of differant designs and IMO solid is the way to go. :)

[k|ngp|n]

I'm totally new to it guys, I just read through everything here, hehe. the more I read, the more interested I became.. soo... I especially like some of your guys' designs. I see slight variations in designs from the mouse pot to some holes, to cylindical bands on the inside. I would like to autocad one but have no idea what a good design would be other than put some holes in it and maybe a few bands, hahaha. The purpose of that is surface area, no? if so, why not make somthing that has more holes or more surface area? cnc limited?

On a turning machine you are somewhat limited as to what you can do in the base area....especially with a container over 7-8inches in length. All solid-non brazed containers with at least 3lbs of mass probably will do the job pretty well. IMO the best solid design has the most mass and working surface area closest to the core of the cpu.

[speed bump]

I don't think brazed tubes are that bad if you design them properly the big advantage of a solid container is surface area so if you had good way to to build and just used the tube as a holder which if you isolated it you could use just about any material as long as your contraction works okay and doesn't cuase the joint to break.

[Eldonko]

Dothan will hardly heat a tube up, especially at low voltages and clocks like that one. Anyone who has used a solid unit can tell you how much better it is. If you take a look around you'll notice that NONE of the top dice/ln2 benchmarks have been done on brazed tubes. They can NOT hold a load at all. I saw a video of the temps on a brazed DI unit with a pretty good base shoot up like a rocket during 1M PI while my solid unit here sits at -62 to -60c through an entire 32m run at much higher heatloads.AFI the brazed tube I got from you holds temps fine under load. I measure the temps with a Fluke 51 II at the base of the pot (where it contacts the CPU) and I haven’t saw load temps above -64C @ 1.7-1.8v. Idle is about 69-70C. 5-6C variation in temps is not too bad, however, a well made solid unit should give less temp movement (2-4C?). I guess that can be the difference between getting a WR and not getting a WR, but I don't think that is something I ever have to worry about haha. ;)

[ilkkahy]

Solid container is simple way to get superior DI container compared to almost any brazed one. As long as they are made as simply as they are (boring big hole to chunk of copper), i personally dont find them intresting.

However when Ln2 comes to picture i believe it is indeed possible to get better performing container by making it from two piece. Simply because therefore surface area can be added much closer to cpu. Perhaps even with something as simple as this:

http://koti.mbnet.fi/ilkkahy/ln2&co2%20kupit/ln2%20container%201.jpg

But then again i have a kinda narrow weird point of view to this subject because i dont even use those containers myself :confused2 (but some beatiful day when i have the time and energy i will test few models with reliable conditions!).