Road Trip to -300 °F anyone?
 

I thought this would warrant a separate topic. To those who are unaware, I've been chatting with some folks at http://www.300below.com/ about the possibility of using their Liquid Nitrogen Cryogenic freezer as a potential housing for overclocking CPUs.

The cliff notes version: They have a specialized process whereby the Liquid Nitrogen is somehow phased back into a "dry vapor", and they can cool the contents of the freezer to -300 °F (about -184 °C) and hold it there, for as long as the Liquid Nitrogen holds out.

Good news item #1: There is virtually NO CONDENSATION at all inside the cryogenic chamber.

Good news item #2: The Liquid Nitrogen consumption rate is mostly a function of the mass of the object(s) being cooled. With a typical motherboard being so light, the consumption rate is very low, and once in thermal equilibrium at the desired temp, it levels off even further.

Good news item #3: The cost of their Liquid Nitrogen tanks equates to only 10 cents per pound (and if I did my math right, about 51 cents per gallon).

Good news item #4: Their system is designed so that you can change out Nitrogen tanks while the cryogenic process is active. No need to shut down and switch tanks.

Good news item #5: There is a 1 inch diameter venting port through which we could probably snake the power supply cabling. No need to deep freeze the power supply.

So now, I guess, the question becomes, with the Power Supply external to the chamber, could we "get away with" freezing the crap out of everything else? RAM, Video Cards, etc? Would such low temps really mess us up when trying to tweak the voltage, as well as the other tests/optimizations that would need to be done to get the thing totally cranking?

These questions aside, the folks at 300 Below are willing to host an event, allowing us to "test" our motherboards in their cryo chamber.

For those that are interested, please reply here. It is all open-ended, and I let them know I'd have to synch any meeting date/time with at least 6 people that I know of that are interested in going with me. I say we make it one big "road trip" if anyone else would like to come along.

Add a caveat about location/venue of said experimental overclock.

Sounds interesting to say the least. Do you have photos of these ice boxes? You'll want to run a motherboard/cpu combo that doesn't have a cold bug and cold boot bug problem. The memory on the GPU tends to hate extreme cold. But, I'm not sure just how well the nitrogen/air heatsink transfer rate will be when running the system. Are these systems pressurized or just mostly contained with a vapor escape?

An experiment would be to start the process at ambient and log at which cold point the system bugs out at.

Sounds like an interesting experiment

This is simply a cold box, how ever a very cold one. Many things to consider before a test can even be started because I think several have done simular things on XS.

This box sounds like it can cool objects to very low temps, but what about objects that emit a large amount of heat. A CPU at high over clocks can output a great deal of heat, something like 700+watts if using AMD cpu's, and that does not take into effect anything else such as GPU's, NB & SB chips plus anything else.

Other things like componets on the MB like Caps, can they stand extreme cold, some cannot. Other pieces of hardware like HDD's cannot stand the cold, SSD's might be able to tho.

So the question then becomes if this box uses LN2 like what you say for a object that does not emit any heat, how much LN2 will it use keeping these hot objects cool. I would suspect a lot more and that brings up is it more effecient to just use a CPU/NB/GPU pot to run the test.

In this beast of a machine he still uses a chiller coolent system for the CPU and perhaps GPU's while everything is in a sealed chilled chamber.

http://www.xtremesystems.org/forums/...d.php?t=160872

Good point there Kal about CB's

Some motherboards have a CB effect, the AMD boards I have used do not and can stand max cold, I do not believe the Classy can do that. A lot of Intel CPUs CB at warmer temps than max LN2 temps, while AMD do not.

There are pictures on their website. Not too interesting. Unless staring at white rectangles is your thing :)

I believe the chamber is at standard 1 atmosphere (but what an atmosphere!) and the tanks are pressurized.

Valid point. It becomes a battle of thermodynamics. The mass of the vapor times it's specific heat capacity and temperature vs. the mass of the elements giving off heat times their specific heat capacity and temperature are dueling. The Liquid Nitrogen Vapor will win out, but that will increase the consumption rate, which I had almost completely overlooked! Good catch!


A complication. Migrating components outside the case sounds like it might be able to defeat this entirely. Unless an interface can be made from the outside to the inside, this ship is sunk.

Can anyone resurrect this with a solution?

You can use extenders for the psu

For the HDD just pick up a few long sata cables, i think SSD can still work even if cooled.

if you want the gpus out too you can get riser cards but i don't know if they will work in this scenario

Another "hypothetical" question. What would be the expected % of speed gained from, say, -110°C to -184°C? I know this is "wide open" but is there some graph/performance curve somewhere? I know the record-setting 8.2 GHz guys did some extreme cooling, wonder where they were on the temperature scale.