LN2 Evaporator Design and Reference Links

I know a lot of design stuff and home made pots/evaporators are being done across the hardware community, but I haven't ever gone looking for it. If you have, or if you want to help me look, I could use your help.

I'm interested in having my own pot/evaporator milled. I want to take any information I can find on the net, and anything we can infer or learn from existing commercial designs, and then I'll figure out exactly what I want to have milled. Once we compile some reference material, I'm going to put together some dimensions, state some guidelines, list desirable features, then recruit some people from the forums who can do autocad/solidworks stuff as prototypes, which we can tweak and I'll then hand it off to the people who will create it.

Anything contributed in the replies below, I'll compile into this first post, and possibly use as an article on the frontpage where I'll solicit further input from our readers. I'll be doing my own looking around also and I'll reply below with my own findings also.

Ultimately, my goal is to avoid the drawbacks of commercial designs (they are designed to be good, and be economical to produce/sell), and have an "open" community designed pot/evaporator that anyone can take to the mill and create for themselves.

Fundamental values (suggest additions to this list):
1. Higher mass stabilizes temp swings
2. Reasonable size/dimensions around socket eases insulation and reduces likelihood of interfering with socket area caps on some boards
3. Solid copper eliminates concerns of interfacing between copper base/aluminum upper
4. Increased internal base surface area helps pull down temps, uses LN2 more efficiently
5. Mounting hardware for present/future AMD/Intel sockets

Links/References:

1. ?
2. ?
3. ?

Useful Specifications:

1. Dimensions for hold down brackets
775 is 72mm square (center to center)
1366 is 80mm square

Fundamental values (suggest additions to this list):
1. Higher mass stabilizes temp swings
2. Reasonable size/dimensions around socket eases insulation and reduces likelihood of interfering with socket area caps on some boards
3. Solid copper eliminates concerns of interfacing between copper base/aluminum upper
4. Increased internal base surface area helps pull down temps, uses LN2 more efficiently
5. Mounting hardware for present/future AMD/Intel sockets

Click to expand...

1. This is very true, but it is also very important where the mass is. The dragon evo 2.4 isn't a "huge" pot, but it can hold its own still.
2. This is a big factor in my pot choice as well. I don't want to get a Venom-sized pot like Dolk has because truth is that it does interfere with the board. A tapered base helps avoid this.
3. This is true with bases and tops that are removable. The Dragon Evo 2.4 has an Alu. top, but it is press fitted to avoid any possible leaks. This may be a good way to go because you get much more height, without the raw costs of copper.
4. Steps (see Dragon Evo 2.4 and Koolance V2), holes (see Venom, F1EE, Gemini), and posts (see Dragon Evo 2.4) are very effective and efficient ways of maximizing the small amount of space and thus increasing the surface area.
5.A square backplate that could be drilled and tapped later fixes the mounting issue, as well as a delrin or metal top bracket that could be drilled for later mounting configurations.


Ps. I'm working on something in 3ds Max 2010.

Thanks for the input.

Does 3ds max 2010 output files in a format compatible with autocad or other industrial milling stuff? If not, it will be good for prototyping, but ideally the final we'll have done in something the guy operating the mill can use directly.

There might be such a thing as too much surface area. The koolance pots certainly had a mass problem but I also think that the amount of surface area they had added to the temp swing problem.

That could be true, in that the temps would drop too fast because the Dry Ice of Ln2 was being sublimated too efficiently.

As for the output formats, I don't use Cad, but here is a screenshot of all the files I can output as:

Awesome, that should work great I think then.

What sort of copper should I be looking at purchasing? What size prior to milling it? I need to figure out where I want to buy the actual metal from. This could be one option: http://www.onlinemetals.com/

if there is too much mass it will take a TON of ln2 to pull down.

you may want to go to a local welding shop or sheet metal shop. they can usually get discounts on orders of metal. i looked into building my own pot and my friends dad told me i could get better "buys" through this local sheet metal company. so you may want to look into this

I.M.O.G. said:

Awesome, that should work great I think then.

What sort of copper should I be looking at purchasing? What size prior to milling it? I need to figure out where I want to buy the actual metal from. This could be one option: http://www.onlinemetals.com/

Click to expand...

That place works great, but I would wait until we work out the design. Which brings me to my next point, what kind of combination of holes, steps and poles do we want to use?

cold.nut said:

you may want to go to a local welding shop or sheet metal shop. they can usually get discounts on orders of metal. i looked into building my own pot and my friends dad told me i could get better "buys" through this local sheet metal company. so you may want to look into this

Click to expand...

Great point.

Not sure, I was hoping to come up with links to build/design logs for other pots prior to talking much about the design. I did a bit of looking earlier on XS to see what they have, but I've only come up with this LN2 container gallery so far:
http://www.xtremesystems.org/forums/showthread.php?t=102575

We really need to look at some build logs or discussions to get an idea about optimum base thickness, wall thickness, mass, etc... I'm sure other people have shared more details about their design, though many don't share much because it takes a bit of milling and testing to figure out optimal designs on your own.

Anyone else find useful references yet?

EDIT: Here's a decent one showing some of his design ideas:
http://www.xtremesystems.org/forums/showthread.php?t=251508

I read through just about all of those. Lol

Here's my take:

1. A rod is essential.
2. 8-10mm base thickness
3. 5-7.5mm wall thickness.
4. Steps from the rod, to the walls, and NOT from the walls to the center (rod).
5. Mass in the center of the design, so that the heat from the processor is almost absorbed into the mass without a large spike.

Now I'm off to 3ds Max.

What do you mean by a rod - protrusion coming up from the center? Can you link me to further details on that? Makes sense in my head if thats what you mean, but I didn't see much about that. I also noticed many popular pots go with a thick base with holes drilled into it - I haven't seen anything with the rod design.

I think the base thickness you are working with may be half as thin as it should be - this is an LN2 pot primarily, and thinner bases were suggested for DICE from what I've read, while a bit thicker for LN2. Your wall estimate seems about right from what I've seen, but I'm not nearly well read enough yet.

See here for the rod idea. Also, check out the Kingpin Dragon Evo 2.4 that I mentioned. It's a lighter pot, but it has mass in the correct areas (directly over the cpu) to stablize temp swings.

As for base thickness, we can always double it.

Edit: I totally forgot to link this:
http://www.xtremesystems.org/forums/showpost.php?p=1076941&postcount=13

Linky?

Kingpin doesn't really give out pictures of his internals. The best I could do is a picture of my pot, but that might not even jive. Check out the link I just edited in.

http://www.xtremesystems.org/forums/showpost.php?p=1076941&postcount=13

Edit: I just rendered this so you could visualize my idea:



Double edit: Just some late night renderings!

Have you seen the f1ee chance?

Not the internals. Kingpin doesn't do internal shots, so unless I saw one in person (which I haven't) I wouldn't be able to. Did I copy it exactly, but in a round tube?

there are a million internal shots of Vince's pots. let me dig one up for you guys..

people will always spoil the internals; no matter how hard someone tries to keep them secret!

this is one of the new Gemini bases.

Does he actually use this internal design?

That's the original F1. The F1EE's base is a thicker, (much) less fancy version of it.