LN2 Evaporator Design and Reference Links

[I.M.O.G.]

Steps have more surface area than a slope for pull down, but either aim that once pulled down and managing temps/slow pouring the LN2, its working from the center of the mass directly above the die outwards.

kingpin's and sf3d's bases both look like they spread out their drillings across the base, to ensure there is a lot of mass in the base but pull it down as quickly as possible when desired. Having a flat base and simply drilling is surely more economical regarding production costs. Maybe its ideal design-wise also.

I haven't looked closely enough or seen a good enough picture to judge how thick the base is between the bottom of the drillings and the base in either, but it looks like sf3d's final revision is notably closer to the base than kingpins. sf3d's drillings also ensure any LN2 slow poured is going directly to the base and flowing across the internal floor, whereas kingpins could be spread across various holes but considerably less LN2 could be actually in direct contact with the internal floor since it can't flow from one hole to another.

All said and done, I don't expect this pot to be that great with the design as is. It should work more or less however, even if it does so poorly. From Mark's suggestion, I could get it reworked and have more holes drilled next to the ones already there, or deepen the existing holes. Once I see how it works initially, whats annoying about it, then I can see how changing it makes it better or worse. If its slow to pull down or too stable initially, its conservative and leaves room to expand without starting over.

 

[Bobnova]

It's worth noting that a flat base with holes drilled like the f1ee is much, much, much quicker and easier (read: cheaper) to do then a bunch of little steps. That's a large part of the commercial pots (and arguably the main drive behind the Gemini): cheapness.

 

[cold.nut]

i need a hookup with a cnc lathe.. what plans do you have to get some lathe time ?

 

[I.M.O.G.]

What ever happened with sno's lathe? Did he get any pots out of it or abandon that?

I sent a PM to dylskee about getting something worked up, he's still in the milling business and doing well currently so thats good - if I don't have better options, I'm hoping he may be able to do me a favor. Prefer not to lean on him if I don't have to though. My primary hope is going through my brother-in-law at his work, as that could possibly be the cheapest if its doable - left him a vm tonight so we could talk a bit more about specifics. I've also sent out some feelers to local cnc shops, and I'm waiting to hear back to see if that is economically feasible - I offered free beer and an exhibition of the final product to sweeten the deal.

It's worth noting that a flat base with holes drilled like the f1ee is much, much, much quicker and easier (read: cheaper) to do then a bunch of little steps. That's a large part of the commercial pots (and arguably the main drive behind the Gemini): cheapness.

It makes sense when trying to get reasonably priced pots into the hands of the masses as quickly as possible, and make a reasonable living off of it. He has a lot of experience and he's done a lot of testing however, so maybe thats also best functionally.

There doesn't seem to be much to go on as far as people posting their design ideas or motivations between revisions publicly however - mostly just presentations of final revisions from what I've found looking around. Doesn't help that most the people doing this are not native english speakers and I don't read any other languages. If I get anywhere, everyone will be able to read what I changed and what the effects were. Hopefully I know a little more about it when I'm done, it doesn't get ridiculously expensive, and it doesn't just totally blow.

 

[Bobnova]

If you want it square (which seems like the best option to me) you need a mill, not a lathe.
Unless the outside is square and the inside is round, at least.

 

[I.M.O.G.]

Yup, I'm looking at milling options. Square or round seems like a matter of taste mostly.

 

[Gomeler]

I'd design a pot that is most efficient at where your major expenses lie. This should be machine time and copper costs. We've had two piece pots forever for this reason. The less boring you have to do and the fewer intricate steps you create the faster the pot will be to create and therefore the cheaper it'll be. For simplicity Vince has it right with the Gemini. The only improvement I could see would be a mechanism for sealing the sleeve to the base but that's hard stuff when you can't simply thread the two like the Koolance CPU-LN2 Rev2.

 

[I.M.O.G.]

Thanks. I'm not real concerned with cost, but that makes sense to me. From a noob perspective... Starting out paying $250/300 for a proven performer doesn't seem too bad. It will probably be twice as expensive and perform half as well trying to do something on my own. It should be interesting seeing how it goes tho, especially if I'm able to get a couple revisions done and there's any considerable difference in the way the pot handles after making changes.

 

[hokiealumnus]

I know little to zero about this, but Giraffe Pot has quite a lot of holes in its base. Lots of surface area down there. No idea how it performs b/c I don't have a temp probe, but the seller thought it was good. It seemed to treat me well under DICE.

 

[m0r7if3r]

I know little to zero about this, but Giraffe Pot has quite a lot of holes in its base. Lots of surface area down there. No idea how it performs b/c I don't have a temp probe, but the seller thought it was good. It seemed to treat me well under DICE.

Thing is though, dice and ln2 are 2 completely different beasts...the Koolance pot is fantastic for dice but it's one of the worse ln2 pots out there. The thing about dice is, you don't have to worry about coldbugging at all, so you just want tons and tons and tons of surface area so your temps stay as low as possible. With ln2 you want a reasonable amount of surface area with a lot of mass to hold in the cold so that you can hold temps as close to the cold bug as possible.

 

[hokiealumnus]

Oh, there is mass. It weighs at least 6 lbs (shipped weight) even with all those holes. Took some quick measurements - Copper part is 13.5cm from base to tip (not including the CPU contact point). From the copper top (heh) to the holes is 8cm and the holes themselves are 5cm, leaving 1/2cm of copper at the bottom....and it still weighs 6lbs.

It's a custom Duniek pot and AFAIK is designed for both LN2 and DICE, but with all that weight, I hope it controls temps decently. Only way to find out is to get out of the LN2-gouging place. Guess I have to convince my wife to move so we can find cheap LN2.

 

[ChanceCoats123]

One thing to remember here (that means you, nonbelievers) is that this pot isn't a commercial pot. Milling time (while making it cheaper to have a shorter time) is much less of a factor because IMOG isn't going to be selling these by the day.

With that said, I'm still wondering. How did Hokie get a shot of the bottom of his pot? That's like a 7 foot deep pot!

 

[m0r7if3r]

One thing to remember here (that means you, nonbelievers) is that this pot isn't a commercial pot. Milling time (while making it cheaper to have a shorter time) is much less of a factor because IMOG isn't going to be selling these by the day.

With that said, I'm still wondering. How did Hokie get a shot of the bottom of his pot? That's like a 7 foot deep pot!

flashlight

 

[xoke]

One thing to remember here (that means you, nonbelievers) is that this pot isn't a commercial pot. Milling time (while making it cheaper to have a shorter time) is much less of a factor because IMOG isn't going to be selling these by the day.

With that said, I'm still wondering. How did Hokie get a shot of the bottom of his pot? That's like a 7 foot deep pot!

iTelescope....

 

[I.M.O.G.]

Good news, bro-in-law believes time on the mill will not be a problem... Quite likely this will only cost materials.

 

[nzaneb]

it will be near impossible to get the dimensions behind the Gemini. unless you sweet talk Vince over at kingpincooling forums, you will only be able to guess...

Or I could measure mine?

For a direct transfer to CAD/CAM software you will not be using 3dsMax. You'll need Inventor, Solidworks, etc...
If I get some free time, I might be able to model something tomorrow at work. You'll want the software to be compatible, that way the CNC operator doesn't have to code it by hand. Unless, of course, you plan on milling it by hand... which is quite possible.

 

[I.M.O.G.]

So far no one I've talked to has asked about a direct transfer... The place I called for a quote requested a PDF with dimensions, and my brother-in-law was looking for the same. I mentioned I can provide preferred formats... Not sure if thats what they are used to starting from.

3dsmax wouldn't work? Looks like it can save in some autocad filetypes, thought that may do the trick.

Once he looks at the design exactly, I'll see if there's a specific filetype that may make this easier or if its easy enough to do it themselves. Thanks.

 

[nzaneb]

Yeah, sorry, I guess I had a typo and wasn't too clear. For a one-off part, they can work from anything (dimensioned PDF, napkin sketch, etc...) 3dsmax is fine for many things, it could even provide you with a dimensioned isometric drawing, which is very helpful for the mill operator. CAD/CAM software is a sure fire way to get EXACTLY what you submit to them, they drop the file in and do nothing. The software optimizes the code for passes, speed, and tooling to produce the exact part that was submitted, in the fastest possible time. I guess I was simply trying to say that there is less operator interaction, less labor, less room for error if provided with the correct model file format. However, this all depends on the whether or not the company even uses CAD/CAM and the program that they use. I guess it could be possible, that 3dsMax transferable into some software. Just not any that I know of. It would be a good question for the shop you send it to.
Ask if they utilize a CAD/CAM software, would it help to provide a model in that format, and if so, what file format. Most companies don't require it, because they actually want work, and don't want to exclude people that do not have access to $5000+ modeling software. If you were going to actually hire someone to make the part for you (not your brother-in-law ) I would suspect significant cost savings by providing a plug-n-play model, that requires nothing on their part. Other than tooling changes.

I can measure the Heavy and Speed bases from the Gemini (when it gets back from the party), if you want to compare to it.

 

[I.M.O.G.]

That would be great if you wouldn't mind. I also wanted to do the math on the weight, and compare that to see if I'm in the ballpark I want to be in. I'll ask about the cad/cam thing once I have a copy of the design I can send over.

 

[dejo]

one thing to note on design. the larger diameter pots such as the koolance v2, when you get condensation, if it does run down the side of the pot it is harder to keep off the board. the smaller diameter ones are easier to keep in on the foam underneath.

I would also try to keep the direct copying of someone elses stuff down. I would say its fine to look for the best ideas of what is available and make use of those ideas. But dont think you want the people thinking that you just copied Vinces stuff directly. There seems to be alot of people that are sensitive to that, and I understand that. But if you come up with a design that has the best of what is currently known, I could be an easy sell, rather than everyone just thinking that you stole your design from the "godfather" of cooling (Vince).