Anyone Ever Attempted Liquid Helium?

[kouch]

Hi guys,

I am no expert by any means but at our school physics department, they use some liquid helium for some superconductor experiments. I was just wondering if anyone has tried using this stuff (around 10 deg K) or at those temps do the physical properties of silicon change too much to work anymore?

[Jort]

isn't it very expensive?

Liquid helium boils at -268.93 Centigrade

who doesn't want that.
maybe it is to cold like your saying.
you need very thick insulation too i guess :p

[Stewie007]

Not satisfied with liquid nitrogen? hehe

[SoddemFX]

With liquid nitrogen, you spill it on yourself and get a nasty burn at worst. Liquid helium would penetrate your skin, expand inside the body -and you'd die

Ignoring the cost, its just too dangerous.

Tom

[Muffe]

Maybe someone should borrow a laboratory with the correct suits and see if there is any borders for a WR that can be moved on some CPU/GPU's

Would probably be VERY expensive

[dippyskoodlez]

Maybe someone should borrow a laboratory with the correct suits and see if there is any borders for a WR that can be moved on some CPU/GPU's

Would probably be VERY expensive

Very cool nonetheless

[Dissolved]

i don't think processors would run that cold.

[blinky]

theyd definitely not boot up

[9mmCensor]

Once long ago, Fugger concidered (and bought a helium Compressor), but it seems the project never got off the ground.

[D_o_S]

Yeah, my dad said he could get me a whole container of liquid helium (well, how much I needed anyway, he knows a guy who works with it every day). Should I give it a shot? Where can I get that big "chimney" from, to pour it into?

[dippyskoodlez]

Yeah, my dad said he could get me a whole container of liquid helium (well, how much I needed anyway, he knows a guy who works with it every day). Should I give it a shot? Where can I get that big "chimney" from, to pour it into?

I'd talk to chilly1 about his ln2 containers.... and use tons of insulation... tons and tons of insulation....

[D_o_S]

OK, I'll send a PM to chilly1 then.

[Disposibleteen]

i don't think processors would run that cold.

i thought electricity worked better the closer it got to absolute 0. Isnt that what all of the experiments with superconducters are doing?

[Aphex_Tom_9]

yes, but there comes a point where the semiconductors in a silicon chip will simply stop conducting altogether, and just become insulators. LHe2 might be too cold...
only one way to know for sure

[Disposibleteen]

yes, but there comes a point where the semiconductors in a silicon chip will simply stop conducting altogether, and just become insulators. LHe2 might be too cold...
only one way to know for sure

ah, gotcha, didnt know that, thanks for the headsup

[Muffe]

Highly recommended to use some protective Gloves, and maybe clothing, whe don't want anyone to lose a finger, or get som flesh severely burned!

IF you really try this out, D_o_S then remember to take, more than enough pictures! No.. Better yet, get a mate to come over and take alot of pictures

[Charles Wirth]

My Helium rig is complete but the capacity is so low that is not not a good option. We are working on a practicle use for it. One problem that exists is the freezing of capacitors near the CPU socket. I found once fired up they stay running but beyond -50c they are deaded when I reboot.

[Muffe]

I know this might sound like the ravings of a madman, but why not heat the capacitors with some electrical heating, something that would keep the capacitors above -50? - Something that would not affect the CPU cooling or any other things!

[dippyskoodlez]

I know this might sound like the ravings of a madman, but why not heat the capacitors with some electrical heating, something that would keep the capacitors above -50? - Something that would not affect the CPU cooling or any other things!

How about start the system with maybe a dry ice mix in the container, or ln2, and then slowly step up to ln2 (if starting with DI) and then finally for the grand finalle, the helium? Or possibly insulate them with something heated, as suggested above? like heating blanket, etc...

[Magnj]

lol so cold that you need to heat it up, this some how sounds counter productive :P good luck

[Kazoo]

Once long ago, Fugger concidered (and bought a helium Compressor), but it seems the project never got off the ground.

Well the helium compressor and the Cryocooler and that is sitting in my workshop. The real issue with it is capacity @ 70 K it is about 40 watts
Maybe a little dothan action? There is also the vibration issue and the fact that the head weighs 15 lbs. And then there is the fact that the head needs near perfect vacume for insulation, the heat gain at 70k from air only on the head is more than 40 watts.

[oublie]

I read somewhere that below -120c electrons stop flowing properly, in other words the processor would stop working properly, that doesnt mean that a super cold system couldn't be used, just that the processor would need to be kicking out enough heat to stay above 120 ish.

[Aphex_Tom_9]

I read somewhere that below -120c electrons stop flowing properly, in other words the processor would stop working properly, that doesnt mean that a super cold system couldn't be used, just that the processor would need to be kicking out enough heat to stay above 120 ish.

LN2 gets down to -196c though...of course that doesnt mean a die temp of -196, but it might be below -120

[oublie]

Not sure on this tom, i read it on a uk over clock forum, this came from their expert, he had a problem with the idea of using cascades. But if someone can confirm this as wrong or right id appreciate it.

[SoddemFX]

Hi,

I posted this before, but what the hey, its relevant:

Consider the forward drain current:

Id = Z/L (electron mobility)x(Oxide capacitance)x((Gate voltage-Threshold voltage)^2)/2

Decreasing the temperature will decrease the electron mobility as the electrons move from the conduction to valence bands (decrease in thermal excitation energy). This term will be a major factor as the mobility dependency is to temp^-3/2. This will be compensated to a small degree as the threshold voltage reducing and the increase in oxide capacitance (as reducing temp will increase the relative permativity of the dielectric).

So all in all I can only guess that reducing temperatures will decrease the performance of the actual silicon part of the device. I assume that most increases in switching speed with reduced temperatures come from reduced propagation delays on the metallic interconnects, which should respond very well to lower temperatures.

So i think we have a gain vs. loss situation. Gains in speed from reduced temps for the interconnects vs. losses in speed from the silicon under reduced temperatures.

This also explains why increases in voltage are necessary when a CPU is running at reduced temperatures, as increasing voltage will increase forward drain current -keeping the silicon "happy". The gains in speed will probably still come from the reduced metallic interconnect propagation, the extra voltage will probably just keep the semiconductor working.

All of this is just speculation from the basic microelectronics/fabrication stuff ive done. We obviously dont deal with things this complex and everything is moddelled operating at room temperature

I'd say that liquid helium is too cold...

Tom