Mother of all Cascades, -100ºc Dual Evap that Cools the Motherboard also!!

[vlad]

Nice work and good luck with it.

[-(CH)-]

OMG , that is soo amazing...

Good luck with the rest of it

[IYP]

that is so awsome man....wish i had an enviroment to work in....i like ugly naked guys....err UnGs idea of replacing the caps with military grade ones...that would work well

BEST OF LUCK...BRING HOME SOM WRs

[extremecorvette]

1st Update 7-29-05

Today I worked on a way to mount the expansion tanks and the flat plate heat exchangers. I know some one will point this out, in the pictures the heat exchangers are not insulated yet, but they will be.

[extremecorvette]

I have yet to insulate the HX.

[Waus-mod]

goin great! kees us updated

[wdrzal]

be thoughtful when plumbing up, don't let any of the copper tube touch the aluminium.

[craig588]

I destroyed a chiller that way. I used a alu HX which got a hole in it and the compressor then sucked in water. This was around 5 years ago when I first started building refridgeration setups, I didn't think about the copper in the water blocks.

[extremecorvette]

Holy Sh*t,i got very impressed at all the buiding,the box,the project itself,the idea for cooling motherboard and all,will be refreshing this page every 5 minutes to see if something new appeared hehe,btw chilly doing a great job suplying parts for that kind of projects
---EDIT---
Will u use some kind of controller for this beast?If so,what will be?
Thank you

Builders of the Controller are Mariano Porta AKA Ssilencer & Hernan Martinez Reina AKA Golden Eagle Mail: msporta@2vias.com.ar

The controller it self is going to be a work of art. Just wait guys you haven't seen the best yet.

[extremecorvette]

When I started this project I didn’t know anything about phase. I just started reading about it on all the different forums. I have a lot of experience with air and water cooling. Last year I had a project were I had the top 10 water blocks on the market and I set out to build on better. All the blocks on the market are set up for low flow and high turbulence. 100-300 GPH, I wanted to make a block that could handle 500-900 GPH. It would have one huge radiator and a second one about half the size. I had two 172mm fans on the huge radiator and two 120mm on the smaller one. This was all made to fit in a large case. The block had a weight of 2lbs and cost just in machine time around $500, so it had no marketability. But it worked great, I was 6C lower than the silver cascade water block, with a 900 GPH pump. I also made a 9800Pro water block that cooled the memory also, and a NB water block. I tried a few set ups. The main gain came from using bigger tubing. I found out most pumps are rated for using 1 ½� tubing and not ½� so I used 1� and had made special flare fittings.

When I started this project I went out and found all the people that are helping me, I emailed, called, and went to them. I told them all what I was doing, and why. I’m receiving no funding from the Iowa State University. It has been great to find people that are willing to help and donate there time.

I want to thank all the people that are helping me.
Reggie Wing AKA Chilly1 at http://xtremesystems.org
Phil E. Tuma at 3M http://mmm.com/
Patti Mikula at ATI http://ati.com/
Jon Carvill at ATI http://ati.com/
Jennifer Barron at ATI http://ati.com/
David Taue at Antec http://www.antec.com/InternationalWelcome.php
Carol S. Archer at Western Digital http://wdc.com/en/index.asp?Language=en
Jake Ward at Centon Memory http://www.centon.com/
Randy Torio at Centon Memory http://www.centon.com/
Alex Mei at OCZTechnology http://ocz.com/
Darin Stoll at Mechanical Comfort Inc
Adam Lintott at Matrix Orbital http://matrixorbital.com/
Fred Namei at Danfoss http://danfoss.com/
Brian Flynn at Ritchie Engineering http://ritchieengineering.com/
Dave Berry at WTT America http://wttamerica.com/
James Nelson at Refrigeration Research http://refresearch.com/
Edward Kerins, Jr. at Cambridge-Lee Industries http://www.camlee.com/
Gene Ziegler at Sporlan/Parker http://sporlan.com/
Julian Parker at Henry Technology http://henrytech.com/
David Metcalf at Honeywell International http://honeywell.com/
Benjamin Benulis at AMD http://www.amd.com/us-en/
Jim Bogart at http://flatplate.com/
JoAnn Buonomo at http://www.monmouthrubber.com/
Danner Mfg. at http://dannermfg.com/
Builders of the Controller are Mariano Porta AKA Ssilencer & Hernan Martinez Reina AKA Golden Eagle Mail: msporta@2vias.com.ar

[GuGaCoSa]

With that guys building the controller im pretty sure it will be the best you can get,i recieved my g-froster last week and im impressed with the quality of the product and Ssilencer and Golden caring and helping me with set up and all

[Gryph]

Looks awesome, it is a work of art.

Gryph

[extremecorvette]

With that guys building the controller im pretty sure it will be the best you can get,i recieved my g-froster last week and im impressed with the quality of the product and Ssilencer and Golden caring and helping me with set up and all

I'm sure they want to shoot me by now. The contoller for this sytem has to do so many things. AIt will monitor 8-10 temperatures with probes, compressors, and the computer. When you want to start the system you push a button and the controller powers up starting the first stage compressor, and condenser fan. Once the two brazed plate heat exchangers reach a set temperature. The controller starts one of the second stage compressors, after 15 seconds it starts the other second stage compressor. Once the two-evaporator block reaches a set temperature, the controller starts two pumps in the dielectric circulating it over the motherboard. Once the two-evaporator block reaches a lower set temperature, the controller starts the computer. After 24 hours of run time the controller shuts off the computer and cascade system. It starts a delog cycle. In the delog cycle the controller starts up a heat probe that is mounted in the evaporator blocks. This raises the evaporator blocks temperature from -100ºc to 0ºc, once this is done the controller shuts off the heat probes in both blocks and starts up both compressors one the second stage to do a flush cycle for 30 seconds. This helps remove any oil that has become trapped in the evaporator blocks. That is the purpose of the delog cycle. Once this is done the controller starts the system back up in order. The controller also has fail-safes so that if temperatures get out of operation ranges it sets alarm and shuts it all down.

[Frambosie]

that is out of this world, i cant wait to see more progress. Its amazing how many people came out to help on this wounderful cause!

[pythagoras]

Nice idea with the delog cycle Of course it would be better if you could laser etch a non stick surface on the evap so that the oil cant stick no matter how dense it gets due to the cold.

Regards

John.

[Activate: AMD]

man, thats a nice list of "sponsors" you've got there



im not sure i want to ask what the price tag is thus far.... not to mention the final completed cost

[Ssilencer]

I'm sure they want to shoot me by now.

After seeing all your work, we forgot why we wanted to shoot you

[gokickrocks]

looks great




except for the tig welding

[bobbobson]

Dear mother of god!

[ShawnMcGrail]

That is amazing. When is it going into mass-production?

Anways, , lol.

Since your so good at sanding, Will the final outside of the case be perfectly sanded to a mirror finish?

[gclg2000]

Hey bud, awesome work!!

But i have a couples questions(concerns) about your design.

I dunno about doing the dual expansion valve thing on the 1st stage. I think one would be better for load balancing. I'm not the expert, but what happens when you have a large load variance on the twn HX's. Let's say you shut one of the stages off and now the heat from the 2nd HX is no longer around. The expansion valve that has no load on it now, should squeeze down as the temperatue drops and it's sensing bulb dropping the temp and causing a variance in pressure difference on the high side and could affect the other metering valve's function.

I think it might be easier and more effecient to just use one expansion valve for the 1st stage...let the 1st stage evap coil over both HX's, and then one sense bulb can adjust the capacity of the system without interferance from the other stage shutting off, or having a smaller video card load.

I just think that it may be more effecient on the 1st stage to have one expansion valve. I think the 2 valves in parallel would mess up each other's capacity functions.

If you were to do the dual expansion valve idea, i think it would prob be better to use a Constant Pressue expansion vavle instead. That way once you get the temp's were you want them with everything going, at the set pressure you assign it, the stage would be able to handle at least some ideal load @ a certain PSI or anyload less than that (even if there is no real load).

I've used a Parker CPEV (AA-A7 i think it was) on the 2nd stage of my 2nd cascade and it boom, all done for.

These are just a couple things that i've though off. And if your gonna run part #23 (the suction line HX) it would prob be best to have the sense(ing) bulbs after that HX so they can account for the load that add's too.

But by far, this is the coolest case/frame for any cascade i've ever seen.

I bow down to you for that!!

Awesome stuff

[Golden_Eagle]

This is what I call "A Big Proyect"
I´m glad to be part of it.

[Colin]

That is simply beautiful!

[placebo]

Darn, what a heck of a project... Even I am impressed...

Well, you seem to be working with the right type of people (Tell Randy I said hi...) and I am sure that the support you are getting from them is top notch.

Now, I have personally been messing with dielectric fluids, about 2 years ago. I attached a few pictures for you to look at, etc... Let me share a couple of experiences with you...

First, please be aware that I was using Dynalene MV as a fluid and no Hydrofluoroether 3M fluids. However, they are both kinda a pita to work with and here is why...

1.) Dynalene smells highly toxic (no concern with 3M HFE )
2.) THe pump is an absolutely crucial part. I believe that 3M most definately suggest an all stainless steel pump. This is very important as most other wont last long... A decent SS pump will run about $2.000 (or you find a sponsor....)
3.) Sealing gaskets should be of upmost concern. I was working with a modified GRI (Gorman Rupp Industries) 18651 pump, using special "Viton (TM)" gaskets and they were still leaking with the Dynalene.

Definately take the requirements, which 3M lists for their HFE fluid seriously. I would very much suggest an all SS pump with SS lines and SS heat exchangers for the HFE fluid.

Lastly, I am not quiet sure what you are going to be using the heat plate exchnagers for. However, they have pretty much proven to be unreliable over the long term, when dealing with very large TDs (e.g. cascade stuff). The large differences will cause cracks and the things will start leaking like a SOB... lol Point going for them is their much smaller size and that they should work fine for a couple years...

Anyways m8, great project, great pictures and thanks for sharing this with everyone! I will most certainly keep my fingers crossed and follow the progress on this one!

I attached a few pictures of my older Dynalene project. If you have any questions, do not hesitate to ask...
(Dynalene was chilled in the big bucket on top. It had a copper spiral in it, circulating the dynalene, and lots of dryice around it...)

[placebo]

Btw, funny thing was with my setup.... when I turned the pump off for a few seconds and wated to turn it back on, it wouldnt move... At first I was wondering what was going on but than I realized that the condensate on the magnets of the pump had frozen and as such they couldnt turn anymore. It took 10 minutes of "hair drying" to get them unfrozen again....