LN2 Evaporator Design and Reference Links

[ChanceCoats123]

Gosh darnit! Lol fail on googling skillz.

 

[Gautam]

A design like the ones you sketched up might possibly have the capacity to handle a modern chip (like a Gulftown), but having a thick rod like that effectively puts your pouring into slow motion. For some that might be desirable, but you won't get an instant response from every pour with a design like that. You get a delay.

 

[m0r7if3r]

no...no that's not it...it's just a solid block of copper with holes drilled in it, and it's probably the single best pot ever made...the whole post thing with holes would be a HUGE pain to make. A base with just holes in it would be more than adequate...the evo also doesn't keep up with modern heatloads too well (oc'd gt, etc)...not enough mass

 

[Gautam]

Yep, the key is just getting the base thickness and the surface area/amount of holes right. And to a point, "right" is subjective. Some people like a sluggish pot that doesn't need to be poured too often, others want one that's fast-acting and responsive, but also needs to be baby-sat.

 

[Bobnova]

On copper type: The purer the metal the better the heat transfer, but the softer and easier to scratch it is.

 

[m0r7if3r]

heat transfer doesn't matter too much with ln2...it's more about speed of transfer

EDIT: on second thought, would nickel plating some really good copper be feasable, it works for waterblcoks.

 

[ChanceCoats123]

I think after reading G's post, we need IMOG in here. I mean, it's his pot, we should probably see what he wants out of it.

 

[I.M.O.G.]

First pot, I'd take a sluggish design over one that needs babysat. I need a bit of stability while I learn the ropes, and get a feel for pouring and the benches. Down the read, when I know things more instinctively, a faster pot may be more appropriate.

 

[m0r7if3r]

you're gonna want a lot of mass and not a ton of surface area then...did you lift either f1ee that was at the party?

 

[ChanceCoats123]

So possibly implementing a rod in the center then. Like G said, a rod will put your pour into slow motion because of the large mass directly over the cpu. That also means that you get a slow increase in temperatures when a heat load is applied.

 

[I.M.O.G.]

you're gonna want a lot of mass and not a ton of surface area then...did you lift either f1ee that was at the party?

Yes, when I was straightening up that center table hah... Went to pick it up at arms length and was a lot heavier than I expected it to be looking at it. I'm guessing it was an f1ee honestly, but from the weight compared to the other pots I handled there, I'm pretty sure thats what it was.

 

[Bobnova]

Heat transfer is still important, the part of the pot that touches the cpu gets warmer then the rest of the pot, the better the heat transfer the less that base portion can change temp wise without taking the rest of the pot with it.

 

[ChanceCoats123]

Heat transfer is still important, the part of the pot that touches the cpu gets warmer then the rest of the pot, the better the heat transfer the less that base portion can change temp wise without taking the rest of the pot with it.

That's why I'm thinking we could use the center post design, with steps on the outside, and inside, but minus all the holes. The holes would decrease the mass by more than people realize, and they might make the pot too responsive.

 

[cold.nut]

look at a design like SF3D's Inflection pot. http://www.xtremesystems.org/forums/showthread.php?t=257300

this pot can retain a very stable temperature under load with the most heat intensive processors. it has a fairly decent amount of surface area, but its mass far outweighs the surface area.

here is a video of it performing...

i considered getting this pot a while back because of its ability to hold solid temperatures under load!

 

[I.M.O.G.]

In the new gemini base, how deep are the holes drilled? How thick is the distance from the base to the bottom of the those holes?

How many people are currently milling pots with center post designs? I'm not real eager to do the timewarp. I'd prefer avoiding basing the design off a single example from 5 years ago. In the same way, I'm not real eager to copy Vince's exactly either.

Basically, if we could get an idea of how thick the base is, maybe a post in the center to get mass where we need it for stability, then steps from the center post/walls and holes drilled in a circle between the outer and inner steps? From what I was watching at the party, the pots spend a lot of time dry, and the idea is to hold the pot temperature controllably, and bring the temps down a bit when you want to.

Overall, the walls of the pots were nearly useless from what I saw, as you never really were holding LN2 except for at the very base - if anything, the walls were just adding to the total mass.

 

[m0r7if3r]

in the pots i used, the ln2 hardly ever came above the level of the holes that were drilled in the base (cpu pots at least) except when pulling down the pot

 

[cold.nut]

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...

 

[Bobnova]

That's the real genius behind the F1ee.
It has pretty decent surface area if you fill it, but if you do anything less you're likely only pouring ln2 into one or two of the drilled holes. Small surface area means a slow cold swing, which means it's easy to avoid the CB.

That's why the koolance is stepped, or part of it. A small pour only fills the outer ring, far less surface area then a larger pour covers.

The result, in both cases, is that the more LN2 you pour in the more surface area you get and the faster it pulls down.

 

[Gautam]

I'll put this in no uncertain terms; do not use a center post. The walls are mostly useless. SF3D's design is in essence very similar to the slow base of the Gemini and the F1EE.

 

[ChanceCoats123]

Lol