• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

LN2 CPU Chiller

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
My sister-in-law lives out in SF, I loved visiting there a couple years ago, but I can't imagine it being possible for me to do so without a ton of planning/scheduling vacation (and bringing the wife and 1yr old along).

I will make a post in our team section to see if anyone capable lives in the area and would be able to give a report. I received your PM, will send more information to you in the morning/when I hear more.
 
Awesome!!! I'll talk with my partner and make sure a ship-out is legit and get back with you. All I'd need to ship is an OCC and a purge case. I'll send you a PM. I'm in Cali. Maybe you want to come see the Golden Gate? Do you know any reputable folks in the Bay Area? I'd be happy and ready now to hook up nearby. I am totally more interested though in seeing you set a world record with a finicky chip like the 5960 with a cold bug. Do you realize the OCC can be set at whatever temp down to -190-196C and the OCC will maintain it? So if the cold bug on the 5960 is -130C then we just set the OCC at -129C and forget it!!!!!! And there's NO ICE!!!!!! NO INSULATION!!!!! On top of all that, everything in the box is super-cooled with (gaseous nitrogen) N2!!!! In fact, I had it down to -190 for like 5 hours once and the only place there was any ice was the outside of the box!!! Everything inside was freeze dried!!!!! Damn, I should have put a banana in there and I could have sent it with my kids as space food the next day. . . JK

This looks like some next level cooling. I would for sure like to see it put through the paces with heavily overclocked 5960x.

What were the temps inside the box and how much ln2 was consumed after 5 hours of run time? Also is it possible to run one chuck for the CPU and one for the GPU and regulate the temps independently?
 
That is one coo looking contraption you have there Mr.Loomis. Great concept.
 
Question:
Would the distance from the supply to the OCC have any effect?
Meaning, could I keep a dewar in a remote location say 40 feet from where I'd actually use the OCC?
 
My sister-in-law lives out in SF, I loved visiting there a couple years ago, but I can't imagine it being possible for me to do so without a ton of planning/scheduling vacation (and bringing the wife and 1yr old along).

I will make a post in our team section to see if anyone capable lives in the area and would be able to give a report. I received your PM, will send more information to you in the morning/when I hear more.

Awesome! I live nearer to Yosemite - in Sonora. But I go to the Bay Area quite a bit. I found a post on another site from a guy who lives in modesto... Only about 45 minutes away...
 
This looks like some next level cooling. I would for sure like to see it put through the paces with heavily overclocked 5960x.

What were the temps inside the box and how much ln2 was consumed after 5 hours of run time? Also is it possible to run one chuck for the CPU and one for the GPU and regulate the temps independently?

Not sure of the ambient temps, but the next time (this coming week) I run it on the 8350 I will be using a temp sensor. I didn't realize the Formula Z has three jumpers specifically for temps. So its just a matter of using another sensor. This board makes things crazy easy for LN2. Also unknown to me before the last session, it also has a switch for "LN2 Mode" and a select jumper for LN2 mode.

I can't say exactly how much LN2 it used during the 5 hour session, but the sight glass on the Dewar only moved very little.

Yes you could run as many OCC's as needed as long as there are ports through the purge case. Right now that would be two with the current chuck design. The GPU OCC enters through the side of the case. But running both off of one Dewar is only a matter of using a splitter adapter before the OCC's.
 
Question:
Would the distance from the supply to the OCC have any effect?
Meaning, could I keep a dewar in a remote location say 40 feet from where I'd actually use the OCC?

That is certainly possibly, but you would need to pipe and insulate from your dewar to the OCC. You must be thinking about sneaking some of your neighbor's LN2? ;) Many labs/hospitals do this with the giant Dewars outside piped in for equipment like MRI's and aerospace equipment. They got deep pockets tho!

- - - Updated - - -

That is one coo looking contraption you have there Mr.Loomis. Great concept.

Thanks a bunch, here's a better look at it: 3.jpg NoIce.jpg

- - - Updated - - -

That is one coo looking contraption you have there Mr.Loomis. Great concept.

And a couple more: IMG_20150922_234844.jpg IMG_20151021_210235.jpg
 
That is certainly possibly, but you would need to pipe and insulate from your dewar to the OCC. You must be thinking about sneaking some of your neighbor's LN2? ;) Many labs/hospitals do this with the giant Dewars outside piped in for equipment like MRI's and aerospace equipment. They got deep pockets tho!
I figured about the plumbing and insulation.
LOL, no, not "borrowing" from somebody else.
My issue is space and location related. I cannot keep a dewar in the house.
 
Wow! That is indeed the most impressive LN2 cooling apparatus I've ever seen! Makes me want to come back to extreme cooling again! :rofl:

Sadly that would be impossible as I now live in an apartment as opposed to a house. I'd sure like to see the benching team leaders put this badboy through it's paces with a more demanding overclocked system.

Mad props! :salute:
 
Wow! That is indeed the most impressive LN2 cooling apparatus I've ever seen! Makes me want to come back to extreme cooling again! :rofl:

Sadly that would be impossible as I now live in an apartment as opposed to a house. I'd sure like to see the benching team leaders put this badboy through it's paces with a more demanding overclocked system.

Mad props! :salute:

Thanks. We've put a lot of time and belief that there is a better way into this. Actually, there likely is a way if you live in an apartment. I'm talking with a company now that hopefully be able to build us 15l and 25l dewars that put out regulated 50 psi liquid. That will solve everyone's no-Dewar-space problems!!! There is a solution!!
 
Actually, there is a way if you live in an apartment. I'm talking with a company now that hopefully be able to build us 15l and 25l dewars that put out 50 psi liquid. That will solve everyone's no-Dewar problems!!! There is a solution!!

Hmmm...interesting! ;)
 
Got it. I'll put it in my thinker. There is a solution!

I have found a company that will likely be able to provide 15 and 25l dewars at 50 psi liquid, which the OCC needs for optimal cooling. Their 15 and 25l dewars will regulate 25 psi, but that is just enough to chill (50 psi is optimal not too little - not too much). You're going to be able to do this!
 
Still had throttling issue, but temp swing and cooling is what we are shooting for. So here is a vid from our session last night with a Novabench test. New: Mainboard temp is shown, so everything was chilled around -30C while the chip was cooled to -180C to -190C. Temp swing during benchmark was about 4 degrees.
 
Thanks. We've put a lot of time and belief that there is a better way into this. Actually, there likely is a way if you live in an apartment. I'm talking with a company now that hopefully be able to build us 15l and 25l dewars that put out regulated 50 psi liquid. That will solve everyone's no-Dewar-space problems!!! There is a solution!!

That's what I thought, my 35L wouldn't work with this set-up. Janus and I were talking about "review" but too many strikes against me for that. I also live in Canada so shipping would likely be prohibitive.
I think you would notice more of a temperature swing if you were to go a bit faster. I can bench the FX at nearly those speeds on a cold water loop. I assume that's why you're looking for an extreme bencher to take it for a spin
 
Last edited:
Crazy stuff this OCC device you've come up with. I'd love to see some test runs of the device with more "taxing" benchmarks.

The tests you've shown are good to get an overall idea of what a computer is generally capable of (sort of like testing out to see if your every day drive car's mpg is really what the manufacturer advertises), but what we are into is closer to the funny car drag racing side.... Basically; run full bore till you either reach the finish line and get a sub 8s time or go up in flames half way down the track. ;)

Maybe do a couple runs of WPrime1024, PiFast and SuperPi 32M. I (we) would love to see that 8350 get a real workout and see how the OCC reacts/keeps up with a few minutes of 100% CPU load.
 
Crazy stuff this OCC device you've come up with. I'd love to see some test runs of the device with more "taxing" benchmarks.

The tests you've shown are good to get an overall idea of what a computer is generally capable of (sort of like testing out to see if your every day drive car's mpg is really what the manufacturer advertises), but what we are into is closer to the funny car drag racing side.... Basically; run full bore till you either reach the finish line and get a sub 8s time or go up in flames half way down the track. ;)

Maybe do a couple runs of WPrime1024, PiFast and SuperPi 32M. I (we) would love to see that 8350 get a real workout and see how the OCC reacts/keeps up with a few minutes of 100% CPU load.
Sure. We can do that. I ran Wprime 1024 for about five minutes and the temp didn't fluctuate at all, so I figured that novabench was the more strenuous. I will run W pifast and superPi 32 m this afternoon. The OCC is a freezebeast though, I'm not sure there will be much different in temp swings, and that's more the point of the vids than the processor. BTW The processor was bought used and could likely have lots of hours on it so its not a cherry picker or even new for that matter, so pushing 8 may not be an option.
 
If that thing was able to hold that CPU in stable temperatures while running WPrime 1024... You can color me even more impressed than I already was. Are you sure you had it running in all cores though? It runs on 1 thread by default.

wprime.jpg
 
If that thing was able to hold that CPU in stable temperatures while running WPrime 1024... You can color me even more impressed than I already was. Are you sure you had it running in all cores though? It runs on 1 thread by default.

View attachment 170960

I am using wPrime v 2.10. When you start it, it uses CPUz to access the CPU info and automatically sets to the number of threads to 1 per core. I OC'd the chip to 6.3 on four cores as it seemed that was as high as I could go. So wPrime the thread count was four cores. You gotta understand with this stuff, I am not familiar with benchmarking so, I could be setting records and not know the difference ... or I could be wasting time and likewise think I am setting records. What I know for sure is that the temp did not fluctuate under wPrime and only fluctuated 4-5C under novabench. I thought wprime was not putting enough stress on the CPU as the temp didnt move, after what I saw as no results, I quit wPrime. Now I can, this afternoon either run 8 cores at around 5.7 w wPrime and superPi32 or 4 cores at 6.3. I believe I have reached these limits because of either my lack of knowledge w voltages and Mhz's or I have pushed the limit of this used CPU. Would you like to set up a skype so and you can walk me through settings on the CPU? If so PM me. I'm gonna run the tests in about an hour. 330 pst
 
Back