Finally some more info!! :D
Great achievements all involved!
I really would like to see how mineral oil + LHe plays together. Maybe next time Simon?
Next session I guess will be DX11 cards + 975BE ... Can't wait :)
Printable View
Finally some more info!! :D
Great achievements all involved!
I really would like to see how mineral oil + LHe plays together. Maybe next time Simon?
Next session I guess will be DX11 cards + 975BE ... Can't wait :)
fantastic work congrats all round.
while your benching do the temps of the mineral oil rise much would you see any benefit chilling it in some way.
have u ever tested the gd70 a mineral oil bath (since it is hard to insulate)
and not to be vain but did this post inspire the ln2 lhe or was it planned like that
http://www.xtremesystems.org/forums/...5&postcount=18
and how much are those clear pots are they cheaper to make than the copper
wow great show. Good work :p: 7088mhz!!
http://solotko.wordpress.com/files/2009/08/img_4505.jpg
http://solotko.wordpress.com/files/2009/08/img_4508.jpg
http://solotko.wordpress.com/files/2009/08/img_4515.jpg
http://solotko.wordpress.com/files/2009/08/img_4534.jpg
http://solotko.wordpress.com/files/2009/08/img_4543.jpg
http://solotko.wordpress.com/files/2009/08/img_4553.jpg
Chemists ,-) ?:D ...
What cooling this? Strange pot...
Sorry it took so long to post, but I've been trying to get my feet back underneath me since QuakeCon!
A little more about the AMD Abomination
The Abomination was a project system that is best labeled as "Revision 1" There are so many things to do differently/try! Though, anyone could be successful in replicating the simplicity of "Revision 1."
The "Revision 1" was well thought out, and it served its purpose perfectly! But everything can always be improved and more specialized. That being said, you always have to start somewhere and give the ideas a try -- regardless of level of refinement.
So, the primary purpose of longer-term overclocking proof of concept was a success! I think that the secondary purpose of the Abomination is to charge the system with simplistic innovation and give others a jump start to some interesting approaches.
Clear is SEXY
The pot used on Abomination 1.0 was an LN2Cooling Phantom evaporator (revision 1). Although, any evaporator/pot could be used! We used the the clear dual wall option that Nomis and I collaborated on, as it "matched" the transparency theme!
Forewarning: The transparency feature is a bit complex to use, a tad costly due to all the threading, materials, and such -- but it is VERY neat to watch! It's great for shows similar to QuakeCon or, overclocking parties. I would only recommend it for such. For hardcore benching of LN2, there is no real benefit. Also, to answer a fore-posted question, the evaporator is STILL copper at the base! It's not ALL clear.... :) The alternative clear materials to use as the base are WAY to expensive. (more to come on that later...)
Why a dual wall evaporator?
The original idea of the dual wall design came as a byproduct from an overclocking conference call with AMD last year, where it was mentioned that a dual wall evaporator could allow for the outer wall to be "buffered/insulated" with Ln2 (liquid nitrogen), and the main chamber could be charged with LHe4 (Liquid Helium). When Nomis presented the need for a clear pot -- we knew what we had to do! We had to use the dual wall concept with some sort of insulation material -- which just happened to be Everclear alcohol (due to it's extremely low freezing point).
If there had been more time and cooling fluids -- we could have tried the LHe4 on the system by using the same setup. The standard setup of the LN2 pot has the necessary features to support efficient use of the LHe4. There is a conversion kit to increase that efficiency.
Using the Abomination on LHe4?!?!?
So, in the case of converting the LN2 Abomination configuration, to the LHe4 Abomination configuration, we would have ALL (Nomis, Sami, Kingpin, Nemesis) had to take turns sipping the freezing cold Everclear from the outer wall of the evaporator. (that is a joke, the temperatures would have destroyed your teeth/mouth/esophagus). After vacating the outer chamber of alcohol, we would have then filled that outer wall with Liquid nitrogen and then gone full bore with the liquid helium in the center.... Hoping that we would be able to cool the processor even further!!! Or see frozen LN2? :rofl:
Not sure what would have happened...
Maybe a world record? Dunno!
Maybe a small recreation of the Exxon Valdez? if the container cracked... Dunno!
Maybe another fire? Dunno?
Maybe another COMPLETE power outage in the BYOC area of QuakeCon? Most definitely!
Who knows what would have happened?
But, if we had done it at QuakeCon -- you can only bet it would have been crazy, fun, and "Sick!" (all useage of "Sick" needs to accredited back to the owner of the phrase -- Kingpin).
Special People, Special Thanks
Special thanks goes out to AMD for being willing to try the crazy, abstract, brilliant, and sometimes stupid ideas that come up. They seemingly like to encourage innovation in the community -- regardless of whether it benefits them directly.
Special thanks also goes out to Nomis and his suspicious ability to adequately pace revelation of innovation. And all of us are in debt an enormous amount to ALL of those at QuakeCon who volunteered to make the event possible!!!
More crazy and cool computer stuff is coming!!! But, all good things in due time...
-Aaron
So, it was brought to my attention that I had no signature... I have attempted to fix this issue.
Nice looking setup there :clap:
I love my Phantom Pot as I can run it in either config, with center tube for shows and events, or without the inner tube for benching. It works perfect either way.
Plus I have the monster tall version HAHA :rofl:
Genghis, since the temperature differential is much lower on the outer wall in the dual wall configuration with Alcohol in the outer chamber, I would argue that we have significantly less condensation and no need to insulate the pot. We have not done the side by side test. Indeed, that is one of the factors that made it ideal for the mineral oil experiment - less icing in the vicinity of the lower pot diminishing the risk of voltage regulator & PWM failure.
Or stated differently
The Alcohol is the thing in which we catch the cold and diminish the icing...
We are all lucky to have Genghis working on CPU coolers - he has advanced the state of the art and created new directions in cooling.
:clap::clap::clap::clap:
Diamond used to be the girls' best friend. I think that, in the future, it will be the computer/electronic's best friend. How near future? I have NO clue. I'll just say that I've already began design and testing...
However, I see diamond/carbon only being useful at ambient temperatures!
: DISCLAIMER: This is a playful example
So, that being said -- if I lived in a far away space station, on the "cold" side -- I would be using an AMD, non-cold bugging processor, something like a Phenom2 X4, with a Copper cooler, with a circulating LN2 loop, at 6.0Ghz, 24-7. If my room in the space station got moved over to the "warm" side, i'd swap my cooler out from the copper one to a "diamond" one... Besides, I would be "baller" enough to afford a diamond grill for my processor if I had a apartment on the space station anyway.
: END DISCLAIMER:
The bottom line, copper likes to be COLD. And the more pure the copper is, the more it likes to pass things around... like heat, electricity, etc. The downside is that special copper is pricey and horrible to machine. But, back to conductivity -- At room temperatures copper is great, but there are better things. It's just unfortunate that they are out of the realm of economical at this point in time.
Now to Nomis' comment.
Insulation is not perfect. It is merely a resistor/buffer. In the case of the alcohol buffer for the Abomination, we are merely slowing down the inevitable. If you recall, I was pouring TON's of watts of energy into the alcohol to keep it clear during the event. There will always be a gradient. The only difference will be the distance that gradient has to spread across.
In the event of a conventional pot, it's just one wall and some insulation. In the case of the Phantom pot, it's one clear wall, followed by another clear liquid wall, followed by another clear wall. It doesn't eliminate the challenges of insulation. But, it does make for an excellent visual, longer-term insulation than "normal", and some incredible pyrotechnics.
If you want to trade the visual for more performance, replace the walls with the same overall thickness of styrofoam... Or, ideally, we'd seal the outer chamber, pull a vacuum, and insulate without the alcohol... But, that's not really economical either... Or is it economical -- just not practical?
Computer cooling (using ambient gasses and fluids) and Overcooling (using subzero C fluids/gasses) is very advanced in technology. However, at the same time, it's still very remedial.
There reason I say that is because it's like everything else. Manufacturing approaches and technologies become more cost effective and ideas that once were unobtainable become more mainstream ever day.
I'm excited to see where computer cooling and overcooling goes! I think i have some cool ideas that might put me in the front seat of that proverbial rollercoaster. If not, I'm still content to ride along with everyone else and see where it leads.
Since it's not my "day job", I try these ideas out when people give me the chance. As I get the ideas done and tested, i'll post them at Ln2Cooling.com. With any luck, I'll get a chance to do the following here soon:
- economical dice evaporator
- test chassis
- chipset coolers
- continued development of the existing Phantom pot assy and base
- cold-bug compatible LN2 evaporator
- revision 2 of the LHe4 evaporator?
- concept design cooler for "next years..."
- release of my long awaited water block (i'm the only one that has been waiting for it...)
I already have the designs done for most of them... I just need to cut them. Most of the designs use a CNC lathe to keep things cheap. A few of them obviously use the CNC mill due to the fact they are sorta orthogonal in nature. Some of them are already up at LN2Cooling.
Those are just the things I feel comfortable disclosing. But, as I get more time away from the job that "pays the bills", I'll spend more time on this more fun endeavor of computer cooling and design.
Or, maybe someone like Koolance, Kingpin Cooling, DangerDen, or someone of the like can sponser/partner me to do some of this work/research.
Either way, i'll get it done. :)
-Aaron
thx for answer, im now a little bit smarter :)
btw, its under NDA get some info about C3 revision Phenoms?:)
is there a practical use for machines like this? i mean could you put it to work folding or something to help cure cancer or something like that?
With a little bit more work, I think it could be very useful for gaming and more worthy causes, as you mentioned.
There have been submerged systems in the server market, in the past. If I recall correctly -- it allowed more density in the racks due to the increased efficiency over air cooling.
Abomination 2.0 | May 2010 | Exodous
whau, nice. Nomis, and with LHe? And with Thubans?:)
yesssss!
Thuban abomination? Shaweet!
wow, that's weird