Submersion : Thinktank Revision 2.0

Brainy Brainerson strikes again!
Next project should fold space time and fitted sheets. :)
Cheers R.
 
Thx HP ! :D

-----

I finished welding up my saddle unions , so I managed to get a 'tape up' done of my solvent recovery tank....the idea here , is the 4 " sheetmetal line gets connected to the suction side of a 300 cfm inline fan.....the three 2" lines are from the thinktank vents and the coolant res.

tt2_120.jpg


----

The basic idea is , the inner tank gets filled with dry ice.The fan sucks the co2+n2 outside , and any IPA is recovered in the recovery tank.
I'll have more tomorrow.

:D
 
[LYL]Homer;1033160850 said:
This looks like it could end badly.
Click to expand...

Hahaa , thanx for the vote of confidence. :p

------


Here's a couple of current shots , should help illustrate.....

Inside tank next to outside tank ;

tt2_130.jpg


One inside the other....

tt2_131.jpg


The 2" lines are inputs from the tank and the keg.

more soon.

:D
 
Thx everyone !!

I did a bunch of work on the n2/vent system.....it's 90% complete :

tt2_132.jpg


-----

:D
 
I'm halfway expecting a big boom.

Or... heck, what's beyond insane overclocking? "Oops we broke the time continuum"?
 
Jack Hammer said:
I'm halfway expecting a big boom.

Or... heck, what's beyond insane overclocking? "Oops we broke the time continuum"?
Click to expand...

It's the future of computing! Once they can't make the chips faster, they'll just need to make time slower. Just create a tidy little temporal displacement where time approaches infinity and your performance will be through the roof!
 
We have to be thinking of different IPAs. :p

All this talk of kegs and input lines and co2 and n2 doesn't make me think computers...

Does Wild Rose Brewery make an India Pale Ale? You might need to swap out your keg.
 
With all of that invested in an enclosure. I can't wait to see what kind of folding numbers that thing will put down.

And how sustainable is it?
 
I know CO2 is marginally less insulating, but why not use it instead of N2? You'll have a constant supply of CO2 gas and the leaks won't be as catastrophic since its not under higher pressures. Also, I'm not sure if you do, but you should allow an air gap to form in the internal chamber to allow disolved CO2 to evaporate out of the isopropanol. This will prevent a large build up gas that could seriously damage things if released, and to serve as an additional buffer in case of a catastrophic failure of the internal and external seals (a thin fog of CO2 on top of the alcohol). Maybe I don't understand your reasoning behind some of your design choices; it would be cool if you expounded them. This is honestly an insane mod and I do not understand completely your impetus, but I really want to see this in action.
 
-Yes , Wildrose brewery makes an India Pale Ale. :D

-There is no liquid nitrogen used....just nitrogen gas.

-It's not designed to be sustainable.It's for OC / benchmarking runs , each 'session' will probably cost about 200$.

-I'm using nitrogen because I just happened to have a nice expensive N2 regulator.It will allow me more control over pressures than trying to use the second hand co2 from the DI.
There is in fact an air gap for allowing co2 out of solution , and , the co2 is fully vented as well.

---------------

I had hoped to get a picture up of the lighting I will be using , but alas the bastards shipped me the wrong damn color luxeons....

------

I'll be out of town for a couple weeks working so there will be no updates for a bit.

I'll get some explanations up later in the day.

:D
 
:D

I'll explain my design a bit further.(some of this is a repeat)

The main 2 design considerations are ( for the 2nd version submersion rig) ;
A) "cold" (and not just a little)
B) As cost effective as possible.

So A + B totally rules out any of the nice safe synthetic fluids (fluorinert/galdol/novec etc...) Isopropanol will fit the bill quite nicely.It'll do dry ice.perfect !!

This then introduces the third design consideration : Safety.

Obviously , using a flammable solvent as a coolant , is not a task undertaken lightly.Here's the thing though;The 1500$ I've spent on safety systems etc for using IPA is still thousands of dollars cheaper than any 'safe' fluids.

-------

Ok so it's a submersion rig using IPA.

The cooling system is quite similar to a watercooling loop.There is a pump and resevoir , cooling lines/fittings/etc.Instead of a 'waterblock' tho , there's a 'tank' the whole rig goes into.The other physical difference is that there is no radiator.Cooling is provided by the DI in the resevoir.

Everything after that , is a system for insulation , or fire safety , or "air safety".

--------

The "A" consideration was the entertaining one , and this really only beame possible after the introduction of Polymer Capacitors to mainstream component manufacture.....
After deciding on Dry Ice + IPA , the research began.Spec'ing out stuff to run at -80 C is no joke .Aerospace grade all the way.. I have learned all about glasses , and silicones and insulation.

So some of this is me trying stuff out.

I just happen to have a Vacuum pump , so I wanna try out vacuum jacketing , hence the dewar resevoir (the big keg) and the vacuum jacket coolant lines.

The nitrogen system provides insulation where I don't wanna do vacuum.

The vent system sucks away all the nasty co2 and n2 so I don't suffocate.

:D
 
Bump.

I'm back from a month outta town werking.Cheer's to Fernie B.C. ....Been a nice visit.

My backplate should be ready , my luxeons are in , I have 40 l of Iso comin.....

Pics tomorrow.

:D
 
Wewt !

My backplate is finished.

3/4" NPT threads , 7/8 flare unions in ;

tt2_136.jpg


----

I'm starting plumbing for it tomorrow.

:D
 
Hawhaw.... :D

Nah , no booms involved.Long way from testing yet......

I've been bending the drain lines ;

tt2_157.jpg


Lotta werk still........rest assured I'm on it tho.

:D
 
Definitely the most industrial-looking mod so far :cool:
 
Thx !

While I do try for the industrial look , it's almost a by-product....It's basically a result of trying to spec stuff to the best grade I can afford , and the limits of my fabribation ability.

Good fun so far.

:D
 
:D

Here's a pic to stimulate good Christmas dreams , current planning setup ;

tt2_161.jpg


-The motherboard location is approximate.
-I ended up getting Luxeon III emitters....the tank has ~200 lumens of lighting....

I'll be doing some more work over the holidays.

:D
 
You're building the ultimate geek dream. Seriously :)
 
I'd be a liar if I said I wasn't impatient to see how this turns out!

Any plans for real-time updating while you're benchmarking? (twitter, or whatever)
 
Thx Everyone !

:D

-----

I've finished 95% of the coolant loop.I just have to finalize some valves...the bending is over.

tt2_170.jpg


-------

Thankfully.Bending those 2 drain lines as close as I got em probably took 40+ hours....soft copper my ass....I've got blisters and bruises from doing the bending.It's done now tho.....so it's all good.Next for these pipes : Insulation.

----

At this point It's time to take it all apart again , and start finishing the tank seals.This could take 30-40 days to complete....21 days at the very least.

I will be starting the 'actual computer work' at the same time though , so things like the motherboard/card mounts and the I/O wiring should be happening right away too.

:D
 
wee update :

Motherboard mounts and test fit mobo....(not the real one...)

tt2_323.jpg


--------

Pump enclosure welded up ;

tt2_340.jpg


--------------

Backplate sealed on : 'glued 'n screwed' with RTV 157........

tt2_347.jpg


----

More soon........final testfits/final assembly in the next 4 weeks.....

---

:D
 
It's not often welding is a part of a case 'mod' :D

Nice AT mainboard you got in there, BTW. What Slot CPU type is it? :)
 
Yeah, understandable :D Haven't seen my Slot 1 boards, to be honest. Try not to damage this one ;)
 
This is looking great, mate, can;t wait to hear about the final temps either...
 
You must log in or register to reply here.
Back
Top