LN2 Evaporator Design and Reference Links

[I.M.O.G.]

I like no uncertain terms. So we're basically talking about what sort of layout we want in the base, and what sort of depth/mass we should center there.

Really, I'd wager a guess that it doesn't matter too much. A big hunk of copper sitting on top of the CPU with LN2 being poured into it is going to work pretty well regardless. Would like to get a better feel for the ranges of thickness between the base and the bottom of the drilled out holes, as well as how long the holes are from top to bottom.

Looking around at the party, I'd guess that its maybe an inch or inch and a half from the base to the top of where the holes were drilled? And I'd guess it looked like the holes went nearly all the way down to the base, save a quarter of an inch maybe where the probe was drilled. I think there is more specific information out there regarding how to drill a hole in the base, and how much room there is to drill into there for the probe placement without risking putting hole right to the inside of the pot. I haven't looked around enough to find anything like that, but I'd guess it has probably been found.

 

[m0r7if3r]

just a thought, if you wanted some real measurments, you could a) post in the benching team lounge asking for some measurements or b) get the team pot and get some measurements for yourself.

 

[El<(')>Maxi]

Gautam pretty much summed it up, while a center pole works it's not what's being done today or ideal. You can see enough around to answer all your questions Matt, keep looking at the new stuff and you'll have a good idea. You can also start off with a slow base and drill more holes in it if you think it needs to be faster. In that sense one pot can do all the testing you need without several seperate designs/revisions.

 

[Gautam]

That's also exactly what I'd suggest, but I don't know if you have the means to do it or not. I'd start with a container (whether two-piece or one) with a ~2" thick base, and a few ~1" holes drilled into it, then gradually add more or deepen the holes until it's as responsive as you'd like it to be.

 

[I.M.O.G.]

Thanks for all the input.

I hadn't thought of reworking a completed design, thanks. You can always go deeper with the holes though, that'd be pretty easy to tweak down the line but it hadn't crossed my mind.

I'm shooting for a single piece design, as that does an end around on any issues with the matings and it just seems simpler. From what I'm seeing, the 2 piece designs out there are done for economical reasons (cheaper material costs) and machining reasons (harder to drill holes in a base thats 6 inches or more down inside a pot). This being a one-off, production costs or convenience don't matter much - I'm not making a dozen, maybe 2 or 3 depending on the metal stock used.

So chance, can you do me a container with the following characteristics?

- 3" height
- 2" thick base
- 4 1" deep centered holes, sloped or stepped to the walls funneling the LN2 towards the holes (later, holes could be deepened or multiplied out towards the walls)
- walls 1" high
- hold down bracket - designed to accommodate AMD/Intel
- backplate not needed, going to follow 3oh6's approach, with an insulated wood mounting that acts as the backplate

I'll send that to my brother and get feedback on the feasibility of milling it. Seems a lot shorter than most other designs, but then I don't see much motivation for making them any taller. With only 4 holes in the bottom, that keeps a lot of mass right in the base where it counts more, and maybe that can balance out the shorter walls. I'd also like to see the base with a lower centered region, or sloped up towards the walls to allow clearance around the socket area - I think most designs are like that.

 

[m0r7if3r]

you want a square base too...that makes it so much easier to mount

and 3" tall is probably not enough to keep ln2 from not bubbling over and you'd have to baby sit it as you pull it down.

 

[I.M.O.G.]

you want a square base too...that makes it so much easier to mount

and 3" tall is probably not enough to keep ln2 from not bubbling over and you'd have to baby sit it as you pull it down.

Cool, thanks.

I didn't see anyone pulling down really - was probably too busy babysitting the livestreamers.

But when running, pretty much everyone seemed to be running a nearly dry pot except when on the CPU tests, and even then it was usually a consistent slow pour it seemed like, even when the temps were being held really low... Dolk was the only person I saw with any volume in his pot, and even then it was problematic with the board getting too cold and having issues. I could see this being a bit more of a pain when pulling down, but then thats only at the very start of a benching session which typically is going to last a couple hours at least until done for the session.

I'm not saying you are wrong though, I was pretty distracted through the party, only poured a bit for a couple people while watching the probes and not really the contents of the pot, and thats my only first hand experience. Any other opinions on that?

 

[El<(')>Maxi]

I wouldn't go any shorter than 6". Machining a chunk of CU that tall get's to be a bit tricky so I'd go two piece. Square, Round, doesn't make a difference.

 

[m0r7if3r]

I wouldn't go any shorter than 6". Machining a chunk of CU that tall get's to be a bit tricky so I'd go two piece. Square, Round, doesn't make a difference.

square base is easier to mount imo...at the very least easier to insulate for I'll totally bow to someone with more experience on this though

 

[I.M.O.G.]

I wouldn't go any shorter than 6". Machining a chunk of CU that tall get's to be a bit tricky so I'd go two piece. Square, Round, doesn't make a difference.

Any personal motivations for 6"? Shorter was the idea to get around machining issues with it being that tall.

Not sure how easy or difficult it is to mate a 2 piece design as close as necessary, and then 6" seems like a lot of material above the base typically doing nothing other than giving more area for the cold of the LN2 to migrate upwards. Then when you put a cup on top to help pour you are looking at a 9" protrusion off the motherboard.

I mean your suggestion is pretty much what everyone is mostly doing so it makes sense from that perspective... I'm just looking at it all wondering why. Maybe thats because I'm a clueless noob, but it seems like once things are pulled down all that vertical height isn't doing much of anything.

@m0r7: I could see the square base being easier to insulate for, without the trouble of cutting a nice round hole around the socket.

 

[El<(')>Maxi]

A 3" container will will require more pouring during pull down and you're gonna have nitro spilling if it's that short. I've got a R3 Mouspot (6") and it's already too short for my tastes, something that can hold a decent volume of LN2 just works better. You could maybe get away with 5" it it were a flat base drilled design. You're right once it's @ temp it doesn't need a wall really, it's what heppens before that or if you're fresh off a torch job and need to get back to benching temps...you need some internal volume for that.

 

[Gautam]

Yes, technically you don't need high walls at all. It's just that pulling down is annoying without them. The shorter they are, the more frequently you have to pour to pull it down. If you make it only 3" tall, you'll likely end up putting something on top in order to funnel the LN2 in.

But all they're needed for is pull-down, so technically the "walls" can be any material within reason, can be removable etc. I've never seen this done before, but a container that's nothing but a copper base with a long piece of foam pipe insulation attached to it would theoretically do just as well as the same base design with huge copper walls. I guess that's essentially what the Gemini is, except it's a piece of aluminum.

Edit, damn, I hadn't refreshed the browser window in over 30 minutes...oops.

RE round vs square, go with whatever is cheaper/easier to obtain. From what I hear, suppliers only tend to have stock of one or the other, and it really doesn't matter either way.

 

[I.M.O.G.]

Thanks guys, appreciate the input

 

[Gautam]

Also, I probably wasn't that clear, but I definitely second making it a two-piece. I'd say just work on making a base first, 2" tall/deep, deal with the walls later. You can end up making the top in virtually any material of your choice as I said...even a rolled up newspaper would do fine on something like the Gemini.

P.S. there are some crazy pics of some German guys having literally 3-foot-long tubes on top of their containers. Size matters, and I guess 6-9" isn't enough for everyone.

 

[ChanceCoats123]

Thanks for all the input.

I hadn't thought of reworking a completed design, thanks. You can always go deeper with the holes though, that'd be pretty easy to tweak down the line but it hadn't crossed my mind.

I'm shooting for a single piece design, as that does an end around on any issues with the matings and it just seems simpler. From what I'm seeing, the 2 piece designs out there are done for economical reasons (cheaper material costs) and machining reasons (harder to drill holes in a base thats 6 inches or more down inside a pot). This being a one-off, production costs or convenience don't matter much - I'm not making a dozen, maybe 2 or 3 depending on the metal stock used.

So chance, can you do me a container with the following characteristics?

- 3" height
- 2" thick base
- 4 1" deep centered holes, sloped or stepped to the walls funneling the LN2 towards the holes (later, holes could be deepened or multiplied out towards the walls)
- walls 1" high
- hold down bracket - designed to accommodate AMD/Intel
- backplate not needed, going to follow 3oh6's approach, with an insulated wood mounting that acts as the backplate

I'll send that to my brother and get feedback on the feasibility of milling it. Seems a lot shorter than most other designs, but then I don't see much motivation for making them any taller. With only 4 holes in the bottom, that keeps a lot of mass right in the base where it counts more, and maybe that can balance out the shorter walls. I'd also like to see the base with a lower centered region, or sloped up towards the walls to allow clearance around the socket area - I think most designs are like that.

I'll do this, but 6'' tall. And then possibly a 3-4 inch extension.

 

[I.M.O.G.]

Cool, thanks dude.

Go browns!

 

[ChanceCoats123]

It might be later tonight, or tomorrow afternoon. I DIced last night and my computer is at my Mom's house. I've got a bit of homework, but we'll see how it goes.

 

[I.M.O.G.]

Any luck Chance? Looking to get an idea on the feasibility of getting this done through my preferred channel, or possible alternatives.

 

[ChanceCoats123]

Well, homework has really caught up to me, but I'm also in need of more dimensions.

IE:

1. Pot diameter.
2. Hole diameter.
3. Wall thickness.
4. Updated heights for a taller pot (>3 inches. )

 

[I.M.O.G.]

2-3" square?
1/4" walls?
1/4" holes?
6" height

Height doesn't really matter, thats easy to tweak or say add/remove an inch or whatever. Homework comes first though, no pressure, just following up.