Overclock.net › Forums › Cooling › Cooling Experiments › Helium results!!! Window air con chill box build.
New Posts  All Forums:Forum Nav:

Helium results!!! Window air con chill box build. - Page 6

post #51 of 118
Thread Starter 
48 deg C delta is pretty much what I was getting before I put them in the box so I don't think its a box air flow issue.....but yeah the cabling is a mess. I already have the cards oced to the limit from 675 to 820 mhz and they are in sli which doesn't help temps.

Those fans are wolves in sheep's clothing mind, 120mm scythe gentle typhoons 5400rpm models each with 150cfm. So that's 450 cfm in total which as my unit is 12000btu seems to fit with the details of other units you've found.
I may reverse the flow on them and have them pushing through evap and blowing directly on the mobo, but that may conflict with the cpu fan flow....not sure have to check which way its blowing.

Tbh I'm not sure there is an air flow issue. If there was the evap would not hold the temp as low as it does under load. With the overclock on the gpu's and cpu the system is shoving out about 500watts, and yet the chamber temp is only going up about 4 deg C above the no load temp, if there was an air flow problem it would not deal with that heat load so well. But anything that will improve performance is a bonus of course.

Having said that, I love the idea of tacking a whopping great fan onto the evap.....may well do that.
Edited by technogiant - 7/1/12 at 2:00pm
post #52 of 118
Quote:
Originally Posted by technogiant View Post

48 deg C delta is pretty much what I was getting before I put them in the box so I don't think its a box air flow issue.....but yeah the cabling is a mess. I already have the cards oced to the limit from 675 to 820 mhz and they are in sli which doesn't help temps.

Those fans are wolves in sheep's clothing mind, 120mm scythe gentle typhoons 5400rpm models each with 150cfm. So that's 450 cfm in total which as my unit is 12000btu seems to fit with the details of other units you've found.
I may reverse the flow on them and have them pushing through evap and blowing directly on the mobo, but that may conflict with the cpu fan flow....not sure have to check which way its blowing.

Tbh I'm not sure there is an air flow issue. If there was the evap would not hold the temp as low as it does under load. With the overclock on the gpu's and cpu the system is shoving out about 500watts, and yet the chamber temp is only going up about 4 deg C above the no load temp, if there was an air flow problem it would not deal with that heat load so well. But anything that will improve performance is a bonus of course.

Having said that, I love the idea of tacking a whopping great fan onto the evap.....may well do that.

Yeah, perhaps your right, im just thinking that the cabling may restrict the air flow direction from cooling your components. Perhaps a great deal of the air is deflecting off of the cables and reentering the condenser. Lol, its hard to tell, and im just speculating at this point. Either way, even with the results you have, you have a great idea that seems to be working quite well. It certainly beats any ideas of using a refrigerator as a cooling option.
post #53 of 118
Thread Starter 
I appreciate your input Matt, I may well try and push those cables behind a curve of some material....something flexible.....would help airflow and can't be a bad thing.
post #54 of 118
Thread Starter 
Just did some more testing this morning, using a solid cpu heat sink I started the system at above zero and monitored the cpu temp in the bios hardware monitoring screen.

The temp came progressively down to 1 deg C and then jumped to 254, believe this is called wrapping, below zero the counter jumps to 255 and then counts down from there.

I turned off the ac unit and the temp reversed in the exact same way.

So this shows that the bios temp monitoring is functioning properly if if you allow for this wrapping feature.

Sadly it also shows that the issue I had with the heat pipe design heat sink was in fact a real issue and heat pipe coolers cannot be used with this set up.

I would imagine that vapor chamber coolers would also be affected in the same way as they use the same principle.

While that isn't too much of a problem for the cpu as there seems to be a fair choice of solid heatsinks available it is a problem with gpu choice as most now use vapor chamber or heat pipe designs.

So it looks like I will have to install a water loop in the chamber to cool the gpus at least........................unless anyone can think of an alternative?
post #55 of 118
Quote:
Originally Posted by technogiant View Post

Just did some more testing this morning, using a solid cpu heat sink I started the system at above zero and monitored the cpu temp in the bios hardware monitoring screen.

The temp came progressively down to 1 deg C and then jumped to 254, believe this is called wrapping, below zero the counter jumps to 255 and then counts down from there.

I turned off the ac unit and the temp reversed in the exact same way.

So this shows that the bios temp monitoring is functioning properly if if you allow for this wrapping feature.
Sadly it also shows that the issue I had with the heat pipe design heat sink was in fact a real issue and heat pipe coolers cannot be used with this set up.

I would imagine that vapor chamber coolers would also be affected in the same way as they use the same principle.

While that isn't too much of a problem for the cpu as there seems to be a fair choice of solid heatsinks available it is a problem with gpu choice as most now use vapor chamber or heat pipe designs.
So it looks like I will have to install a water loop in the chamber to cool the gpus at least........................unless anyone can think of an alternative?


When I was reading on heat-pipe designs earlier, I had read that manufacturers use one of several designs, the two that were mentioned were distilled water, and the other used alcohol and argon. I would imagine that if you had a heat pipe cooler that utilized alcohol/argon this would be a non issue, as the freezing point of alcohol is around -100c depending on what type of alcohol it is. Im not a chemist, and im not very strong in science so I don't know what the mixture of argon and alcohol does to the freezing/boiling temperature, however I would believe that a design using those two compounds would yield better results at sub 0 temps. I also don't know how the vacuum present within the tube will change the freezing/boiling point of whatever compounds are present. Where is Bill Nye when you need him rolleyes.gif

With that said, assuming my presumption is correct in that an alcohol filled heat-pipe cooler would work better at sub 0 temps, the hard part is to find out which coolers use water and which ones do not.

In terms of temp monitoring, you can get temp probes for multimeters (if you have one), or most fan controllers with temp monitoring capabilities have temp probes as well. You would just need one that can register sub 0 temps, and slap in on the CPU cooler. If your more brave, you could attach it to the corner of the CPU IHS.
Edited by Mattb2e - 7/2/12 at 6:59am
post #56 of 118
Thread Starter 
That's interesting Matt, I haven't done a great deal of background on heat pipes just read a wiki on them and didn't realize some used other evaporants.....alcohol, if they mean ethanol by that, then it has a freezing point of -114 deg C so the freezing issue would not be a problem, I guess the argon is just there due to its inert properties.

The other issue with heat pipes apart from freezing is that they are tuned by adjusting the pressure in the tube to get max evaporation at the desired temp range, the wiki however said that because as you lower the surrounding temp lower than the "optimal" range the pressure in the tubes drops, so further reducing the boiling point of the evaporant so although tuned for a specific range they remain effective down to a few degrees above the freezing point of the evaporant and below that they drop to about 1/80th of their performance.

So this really does bode well......I can see me emailing lots of companies asking for detailed specs.

I might even try and adapt my old heat pipe heat sink and replace the evaporant with methanol just to test this out. Just cut off the heat pipe tip, flush first with methanol to remove the water then put a small quantity in there, heat the pipe to drive out the air as the methanol evaporates and re solder the tip while still hot. As it cools the pressure in the tube drops....of course great care must be taken heating methanol and soldering with methanol vapor around.
Edited by technogiant - 7/2/12 at 9:34am
post #57 of 118
you cannot go below -3 or - 4 celcius past that everything solidify!and stop working!unless you find a medium to prevent solidification below freezing point!i would suggest antifreeze but it kind of defeat your purpose!you ll probably need something to create a vacuum in your cooling chamber.no air = no humidity = a lot lot lower possible temperature!but then you are at the leading edge here so you ll probably have to borrow your better half bag sealer or find a thing to create a vacuum.doesnt need to be a big negative just enough so there is no air!
post #58 of 118
Thread Starter 
Sorry Dr...not sure I understand your comment, are you referring to the chill box chamber or the heat pipes?
post #59 of 118
your balloon idea wont work!why?because external air is lets say 40 celcius and internal air is - 24 celcius this will create condensation and corrupt your system i suspect .but then i could be wrong!

basicly the heat is coming in by the baloon and by the cpu!me i would live with the vacuum you mention and try it like that !use plastic to plug the hose if you do.

as for the start and stop ,a/c are meant to start and stop they arent meant to run all the time!but you can probably bypass the termistor or whatever they call it in english
Edited by drbaltazar - 7/2/12 at 10:02am
post #60 of 118
Quote:
Originally Posted by drbaltazar View Post

your balloon idea wont work!why?because external air is lets say 40 celcius and internal air is - 24 celcius this will create condensation and corrupt your system i suspect .but then i could be wrong!
basicly the heat is coming in by the baloon and by the cpu!me i would live with the vacuum you mention and try it like that !use plastic to plug the hose if you do.

I do not think you understand the premise of the bag. Correct me if I am wrong OP, the bag is being used for pressure equalization and leak detection.

Additionally the OP has been monitoring for condensation from the beginning. After he had noticed that he had some, and it was suggested to get some desiccant, minimal condensation has been present since he introduced his desiccant into the system.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Cooling Experiments
Overclock.net › Forums › Cooling › Cooling Experiments › Helium results!!! Window air con chill box build.