Solid 4kg copper container

[ilkkahy]

Here is my latest creation. I had to change to plan-B since lack of proper lathe blades but im pretty confident this heavy thing will perform quite well. Container still havent got that finishing touch but here it is anyway:



Still that copper junk inside but im too tired to clean it now..


Holddown will be on top of the container.


Diameter is 80mm and height 200mm. I could lighten the container pretty much by latheing material from outer diameter (~8mm -> -1,5kg).

Ill propably finish it next weekend but im thinking about using it in the 160w heatload tester already before that.

edit. weight is now 2720g

[Dani]

WHOAH, giant monster

..but do outside of botton like this -> \_/
Because it will not fit on mobo (P5B/P5W etc.) if bottom is like this -> |_|

[Johnny Bravo]

ilkkahy your drilling the pot is something I've been thinking of for along time, quite interesed in how this works out. Should be good for maximising surface area.

[ilkkahy]

The Stilts container has 76mm diameter, otherwise similar bottom and it fit on P5B. Besides i dont make cone type bottoms because it reduces cooling efficiency (difference might be marginal though).

This desing doesnt really increase surface area (it may seem so but compared to one big hole it really doesnt) However it makes heat conduct better inside container so im pretty exited to test it.

edit. Here is 3d-model from the internals:

[TaPaKaH]

Nice tube

[ilkkahy]

I think i will test the container in bench with 80mm diameter and after that i reduce it to 72mm just for the poor cpu. I dont think reducing wall thickness from 8 to 4mm will have much effect on this particular desing. Also container shouldnt need 5 liters of Ln2 just to get it cold.

Anyway i simply want to make as efficient container as possible as long as it fits on most motherboards (when thinking p4 coolers this normal bottom model should go to most mobos) Visual aspects are secondary things

Perhaps i get to test it tomorrow. Today i could do some of those "water pre-tests" that i have done with previous containers.

edit. The very first water (13c to 70c @ 160w) test gave 6min 45s which is very promising indeed Result would get better with several attempts with better contact but ive had already enough information from that one test.

edit2. I did water test again with better contact and result was 9 min 45s which is simply awesome.

[Dani]

Ok, we tested an pot with flat bottom (outside) and it didn't fit mobo at all, because those caps around socket on P5W/P5B..
But you have a little step there so maybe it will fit then, should be atleast 3-5mm

[Johnny Bravo]

This desing doesnt really increase surface area (it may seem so but compared to one big hole it really doesnt) However it makes heat conduct better inside container so im pretty exited to test it.

why wouldn't the extra finning provide extra surface area compared to that of a plain bored hole? I guess it depends on how much LN2 you intend to put in it really

[The Stilt]

My container fits on every single motherboard I have tested.
Most recently I used it on P5B and it fitted perfectly.

An there even was 1mm spare space for the solid capacitors around the socket

[ilkkahy]

why wouldn't the extra finning provide extra surface area compared to that of a plain bored hole? I guess it depends on how much LN2 you intend to put in it really

I made little mistake due simplification how i calculated this in my head but the difference is still quite small as i thought:



However if you look at the hatched area there is much bigger percentual difference in that area which im obiously too lazy to start calculating it right now

This increases heat condcuting, not only due larger area inside copper but also because radius from cpu to "pipe section" is almost 10mm shorter. Disadvantage is that this makes container quite heavy.

[dinos22]

can you tell us what you plan to achieve with your pattern compared to say 'traditional' design >> hollow tube/even walls/flat or slightly stepped bottom

[ilkkahy]

Hehe.. Hopefully few miserable degrees colder temps than plain bottom

This is quite insane hobby if one thinks it with reason. I like to study how heat works.

I hope to get Danis solid central pole container (perhaps some other containers too?) for my test bench some day.

[Dani]

Yes you will, but first I most get it

[J0d3r]

Buf, 4kg container! thats sick!

[ilkkahy]

Okay i did quick test (only one mount to bench) with dryice today and result was -36c which is 7c better than any previous container



I would propably get temps go even more down a bit by lapping the bottom properly and reducing base thickness a bit. I will propably do more testing later after little changes.

Now i really want to get other solid containers for that bench too

[Revage]

-36 . How big was the heatload ?

[dinos22]

Hehe.. Hopefully few miserable degrees colder temps than plain bottom

This is quite insane hobby if one thinks it with reason. I like to study how heat works.

I hope to get Danis solid central pole container (perhaps some other containers too?) for my test bench some day.

hehehe reminds me of SuperPI benchmarks

[prosk8]

wow its big man

[Johnny Bravo]

I made little mistake due simplification how i calculated this in my head but the difference is still quite small as i thought:



However if you look at the hatched area there is much bigger percentual difference in that area which im obiously too lazy to start calculating it right now

This increases heat condcuting, not only due larger area inside copper but also because radius from cpu to "pipe section" is almost 10mm shorter. Disadvantage is that this makes container quite heavy.

Very true ilkkahy, I stand corrected

My design was more to do with increasing surface area by drilling multiple holes, something along the lines of this:



This way the copper area has maximum exposure to the LN2 when immersed. The method of construction is either stitch drilling or gun drilling, due to the length of the required drill bit. I shall be watching your results with great interest to this end.

[ilkkahy]

-36 . How big was the heatload ?

Heatload is 160w. Another key element is contact surface area where heat comes from and this bench doesnt have much of it (about 3,14cm^2). Temps can be compared only to ones that are run with this same bench. Here is picture from earlier testings:

I will make some day summary thread about dryice tests with this bench where are some results with different containers. Before that there are still some models and things i want to try out (most importantly forced convection dryice cooling which i was supposed to do half year ago).

My design was more to do with increasing surface area by drilling multiple holes, something along the lines of this:

This way the copper area has maximum exposure to the LN2 when immersed. The method of construction is either stitch drilling or gun drilling, due to the length of the required drill bit. I shall be watching your results with great interest to this end.

I would propably continue that cross-thingie all the way to pipe section so there would be extra concucting from cpu to pipe. I dont know does this make temp controlling harder with ln2 since there would be many little chambers instead of one big chamber.

I have to say drilling those 16,5mm diameter 18cm deep holes to copper with lathe wasnt all the time like eating pudding. Its easy if the hole is in the centre of cylinder but when its on the outer-radius of cylinder things start to get slow.

[ilkkahy]

Okay i lightened the container a bit. Diameter is now 72/76mm and mounting can be done from bit lower part of the container too. There is also gpu container in the pictures but i still need to make mounting system for that one (its going to be bit different but should work fine).

[Cpt Twitchy]

I was hoping you would have gone with this design

I know this italian guy who is planning to start mass produce solid containers. So i hacked to his hard drive and found his drawing which is exposed here:



Might take a while before you can purchase one but i think that container will rock as they arriwe. It has the greatest surface-area/mass ratio ive seen so far.

[Stapler]

The pot looks good, I've actually been working on a semi solid container with similiar internal design. Do you think we could get some higher res pics of the GPU container?

[gloatlizard]

nice containers ilkka

[ilkkahy]

The pot looks good, I've actually been working on a semi solid container with similiar internal design. Do you think we could get some higher res pics of the GPU container?

Here:
http://koti.mbnet.fi/ilkkahy/sekalaiset/gpu%201.JPG
http://koti.mbnet.fi/ilkkahy/sekalaiset/gpu%202.JPG
http://koti.mbnet.fi/ilkkahy/sekalaiset/gpu%203.JPG

I was hoping you would have gone with this design

Yeah that would have been sweet. Unfortunately i dont have the ability to follow that path.. that superior desinging requires superior machinist.

edit. Here is couple of pics from the mounting system. Its pretty easy to replace those sideplates to different ones when hole measures change. I think small 4mm threads and soft metal is very bad combination so i did bit different approach and im pleased how stif it turned out to be.