Liquid Hydrogen Cooling?

DruSi3r

Gawd
Joined
Nov 4, 2002
Messages
937
Has anyone ever tried liquid hydrogen cooling. 5 kelvin must be able to squeeze a few more MHz out of a chip possibly leading to a new world record. I havnt looked it up but i think liquid nitrogen is about 12 15K. I hear H(l) is less than $100.00 a liter. Not cheap but Id pay to see it.
Drew
 
I think 5 K would be far to dangerous for amateurs to handle properly. Not only that, when electrical devices get that cold I imagine the conductivty would change.

5K is ~ -270* Celsius
 
The major problem is not just cooling the chip to extreme levels, but not to damage any components while doing it. LH if not handled properly can shatter, crack and totally fubar a CPU. There's also diminishing returns. You could chill the CPU down to 0K or 1K, but at that point you'd be slowing down the electrons and that would be counterproductive. The goal is to remove resistence, not kill the CPU.

I've seen fluorinert used in conjunction with LH to get a whole computer chilled and cranking. Outside of setting records, just or just seeing if you can do it, there really isn't much of a point to using LH or other extremes to achive unheard of overclocking.
 
obivously not for the beginner, intermediate, or even expert. only crazy kids like kyle and brent
and i would imagine superconductivity would become an issue.
d
 
I think the biggest problem would be all that hydrogen after phase change.

One good spark and it's goodbye ocerclocking enthusiast. :)
 
I think the labs here at Truman use liquid nitrogen to cool their liquid hydrogen for the NMR so they don't lose it as quickly. There are other limiting factors in computers other than core processor temperature, so getting down below a certain level won't really result in much of a benefit.
 
lol thats a good point but once it evaporates it will diffuse with the air, not sure concentrations though. Maybe it could be piped over the cpu similar to water cooling? Unrelated but check out this hydrogen powered gas grill
D

jpmkm we just covered the lovely subject of nNMR today in Organic, hence the ideas....
 
u prob mean liquid nitrogen as hydrogen is extremely flammable and with one spark could blow your whole rig.
 
Originally posted by xbreaka
u prob mean liquid nitrogen as hydrogen is extremely flammable and with one spark could blow your whole rig.

No, obviously he doesn't. Read the post before replying.
 
Well, super computers that are kept at the super cool phase- so cool that they can do as much work with them as they want and there'd be no problem, they can do it, why wouldn't hydrogen be feasible?

Well, I'm pretty sure it would be except for the instability.

~Adam
 
I just got this reply back from my professor:
Liquid hydrogen is not used in NMR, only liquid nitrogen and liquid helium.
The latter two are non-flammable.

So maybe Liquid helium would solve the stability problems.
d
 
Originally posted by Torgo
You could chill the CPU down to 0K or 1K, but at that point you'd be slowing down the electrons and that would be counterproductive.
heh. No, electrons keep moving regardless of what you do to them, higher temperatures means a greater percentage of them are in excited states, *not* that they speed up.

As previously mentioned, you'd reach the limits of the chip easily with liquid nitrogen, there is nothing more liquid hydrogen can do for you besides blow up.
 
well 0Kelvin is still technically impossible. Further at true 0K all molecular motion would stop. I think scientist have come to like 1/10 of 1K using some crazy plasma setup. Im not familar with superconduction or the principles behind it. The only reason i mention this topic is because my prof talk about NMR today which supercools a 7 mile long strand of neodynium (spelling) metal allowing this machine (using radiowaves) to detect the difference in up and down spin of protons. Or something like that I wasnt really paying attention as I was reading the newe issue of Home Theater Magazine :(
D
 
Originally posted by Torgo
The major problem is not just cooling the chip to extreme levels, but not to damage any components while doing it. LH if not handled properly can shatter, crack and totally fubar a CPU. There's also diminishing returns. You could chill the CPU down to 0K or 1K, but at that point you'd be slowing down the electrons and that would be counterproductive. The goal is to remove resistence, not kill the CPU.

I've seen fluorinert used in conjunction with LH to get a whole computer chilled and cranking. Outside of setting records, just or just seeing if you can do it, there really isn't much of a point to using LH or other extremes to achive unheard of overclocking.

Not 0k.
 
Originally posted by DruSi3r
well 0Kelvin is still technically impossible. Further at true 0K all molecular motion would stop. I think scientist have come to like 1/10 of 1K using some crazy plasma setup. Im not familar with superconduction or the principles behind it. The only reason i mention this topic is because my prof talk about NMR today which supercools a 7 mile long strand of neodynium (spelling) metal allowing this machine (using radiowaves) to detect the difference in up and down spin of protons. Or something like that I wasnt really paying attention as I was reading the newe issue of Home Theater Magazine :(
D

They use a series of lasers to slow the molecular vibration. It's exceptionally hard to explain, but the results are obvious. (REALLY cold).
 
Originally posted by jagec
heh. No, electrons keep moving regardless of what you do to them, higher temperatures means a greater percentage of them are in excited states, *not* that they speed up.

As previously mentioned, you'd reach the limits of the chip easily with liquid nitrogen, there is nothing more liquid hydrogen can do for you besides blow up.

0k they wouldn't move :p

Then again, that's impossible. Just like travelling at the speed of light (note: I said NOTHING about going faster, just going AT).
 
I was wrong a billionth of a degree above absolute zero and no lasers just magnets...
d
 
Originally posted by DruSi3r
well 0Kelvin is still technically impossible. Further at true 0K all molecular motion would stop. I think scientist have come to like 1/10 of 1K using some crazy plasma setup.
D
"molecular"motion, but certainly not the electrons. 0K is defined as the point when all modes of quantum mechanical excitation are in their ground state (translation, rotation, vibration, and electronic). To get the uber-low temperatures (500 picokelvin) you use laser cooling, which "tricks" a bunch of atoms to jump into the ground state by a pretty nifty method where you excite the atoms a bit with the laser, then they emit photons with a bit more energy than you put into the system.
 
Originally posted by DruSi3r
I just got this reply back from my professor:
Liquid hydrogen is not used in NMR, only liquid nitrogen and liquid helium.
The latter two are non-flammable.

So maybe Liquid helium would solve the stability problems.
d
Yeah now that you mention it, I think I was thinking of liquid helium.
 
Originally posted by jagec
"molecular"motion, but certainly not the electrons. 0K is defined as the point when all modes of quantum mechanical excitation are in their ground state (translation, rotation, vibration, and electronic). To get the uber-low temperatures (500 picokelvin) you use laser cooling, which "tricks" a bunch of atoms to jump into the ground state by a pretty nifty method where you excite the atoms a bit with the laser, then they emit photons with a bit more energy than you put into the system.

Thank you for the correction and the clarification. I wrongly assumed that at 0 the eletrons would stop as well.

Then again, they still might. We will never know.
 
I'm no expert on this stuff, but wouldn't a super low temp cooling system in a room at a normal temp cause an assload of frost in your computer?
 
Thanks to everyone who corrected me. My brother is the physics and electronics genius. I just do games.

Goldbug: Frost or condensation is definitely an issue. Basically it boils down to risk management. Is it worth it to have a setup that could potentially fry your system? This could be applied to any cooling system: air, water or some other exotic method. At what point does the cost outweight the benefit?
 
Originally posted by Torgo
Thanks to everyone who corrected me. My brother is the physics and electronics genius. I just do games.

Goldbug: Frost or condensation is definitely an issue. Basically it boils down to risk management. Is it worth it to have a setup that could potentially fry your system? This could be applied to any cooling system: air, water or some other exotic method. At what point does the cost outweight the benefit?
well, if you were using liquid helium you would just be trying to get screenshots of CPUID, there's no WAY you'd use that for a day-to-day computer. You would either have to keep adding helium (expensive, hard to keep that much helium around) or have a massive refrigeration system to keep the helium you do have liquid.

Condensation is the least of your worries, really...by that point you're doing liquid immersion or similar insanity.

lopoetve: I used to think the electrons stopped too, but my pchem prof assured us otherwise.
 
actually we dont know what will happen at 0Kelvin so its mute. And using liquid helium of course would only be for show and not everyday lol. Just wondering if any1 had used anything other than L2N.
d
 
Originally posted by DruSi3r
actually we dont know what will happen at 0Kelvin so its mute. And using liquid helium of course would only be for show and not everyday lol. Just wondering if any1 had used anything other than L2N.
d
we have good data points leading up to it, and a rock solid model that explains pretty much all the behavior the system exhibits. Taking the limit as T->0 works just as well as measuring what happens when T=0
 
i believe you to be incorrect, are you aware of what happens to an ideal gas as it undergoes extreme pressure or very low temperature? It does not follow PV=nRT. We really dont know what will happen only guesses, bt most assume all motion to stop. This would include electrons as movement would imply kenetic energy and entropy...I point you to this link rules two and three. I maybe wrong
 
Originally posted by jagec
we have good data points leading up to it, and a rock solid model that explains pretty much all the behavior the system exhibits. Taking the limit as T->0 works just as well as measuring what happens when T=0

Theoretically, but since when has ANYTHING in physics worked perfectly to theory?

Stupid physicists... all their assumptions... :mad:
 
Actually Quantum mechanics dictates that the lowest energy level possible for an electron, or anythign for that matter, is not zero energy-- this is called the ground state energy level.
 
Originally posted by TheArchitect
Actually Quantum mechanics dictates that the lowest energy level possible for an electron, or anythign for that matter, is not zero energy-- this is called the ground state energy level.

Yeah, and since when have we been able to prove quantum theory, eh?

It's all supposition at the extremes, welcome good old Heisenberg and teh glory of models.

Nothing is certain. Nothing.
 
Originally posted by Spewn
So far, death has been pretty certain :p

Maybe, maybe not. Maybe we enter a different quantum state? Maybe I'm just imagining you (gotta sell whatever I'm on, if THAT'S true).
 
Well I just talked to two of my teachers. First my organic lab instructor (no phd or doctorate) he said at 0K molecular motion stops like molecules and atoms but that electrons do not. If they did he said they would collapse into the nucleus. Later I asked my Chem 1 professor (smarted guy I know came from MIT) and he said 0K is an impossible state to reach but electrons would stop. He said the system would contain no KE but would have positional entropy. Such as no motion in a crystal :
ABABABABAB
ABABABABAB
AABABABABA
(the last row being transcribed shows positional entropy even though no movement)
Just try to keep our conversation going, I guess no one really knows...
D
 
We have no way of knowing. Both are theorizing out of their asses, since there are models that show both as possibilities. The electrons would most likely collapse inwards, and we don't know what would happen then.

We will never know.
 
I believe that that is true. At 0K there is no kenitik energy, but at slightly more than 0k I think that you would reach a superconductive state in which electron would flow with no electronegative attraction or repulion between other atoms. And isn't that exactly what you want to make your computer go fast.
Can you say 40V Vcore?
 
Below about 150K, silicon goes into a regime known as freeze-out.

The dopants in the lattice sites are no longer electrically active because there is almost no possibility of the electrons reaching the conduction band from the valence band. This means silicon starts acting like an insulator, and no longer a semiconductor. Devices no worky when so coldy.
 
Originally posted by jagec
we have good data points leading up to it, and a rock solid model that explains pretty much all the behavior the system exhibits. Taking the limit as T->0 works just as well as measuring what happens when T=0


not in this instance. were talking about the absense of any molecular motion. we dont even think its possible to hit by any methods. for all we know the particles could stop and never start again.
 
Back
Top