magnetic refridgeration

[the]

probably news for the custom phase guys, but pretty interesting none the less. a few guys seem to have gotten they're hands around a magnetocaloric effect with a new alloy. i don't know how much more efficient it'd be than iron, (4 degrees cooler per Tesla) but they're talking about making a refrigeration units up to 40 percent more efficient using this idea. I cant seem to find exact yeilds of the new material, but the article says its pretty promising. If anyone one could provide numbers, that'd be great to look at. I wouldn't mind a -40 unit that didnt run up my energy bill.

Then all we'd need is a few good objective minds to slap this on a heat sink. Instead of just powering a fan, you;d need to power an elctro magnet, too. could be enuff to tame the heat yeilds we'll see with quad cores next year. maybe even sub room temperature overclocks. that;d be pretty nice. We might even be able to replace this messy, hi maintence water wiht a cheaper, cooler alternative.

http://physorg.com/news64851465.html

[Thrilla]

OMG I've been waiting for this day so badly!!!

Place it in your WC loop and it'll be the 1337!!

[JamesAvery22]

OMG I've been waiting for this day so badly!!!

Place it in your WC loop and it'll be the 1337!!

So instead of a radiator with fans you have a radiator with magnets? No heatdump into the room was be a great plus

[WeStSiDePLaYa]

doesnt sound good for EMI.

[the]

im not to sure about either of those ideas. when i chilled my water with TEC's all those years ago it wasnt near as good as cooling the TEC with water. yoru asking an engery transfer to take place twice, in less than ideal conditions. however im no physics expert and my favorite rule still applies -->only one way to find out.

as far as the no heat dump, im positive that wont work. law of conversvation of energy says energy cant be created nor destroyed, jsut transferred. in this case the magnetci cooler would act as a heat pump. so it still be heating up, then dumping that heat.

[VulgarHandle]

what if the energy created by cooling the water can be used to pump the water....

[STEvil]

interesting applications could result from this.. maybe a layer on top of the CPU/GPU core that soaks up EMI to create the effect?

Assuming it could be made to work with EMI or that the EMI being generated could be altered to create the correct type of energy being emitted.

[Fairydust]

Since it basically "only" changes the thermal capacity of the material, you need a cooldown phase after the material has warmed to coolant temp, during which it's thermal capacity is reduced and thus releases additional heat, after heat has dissipated, material gets thermal capacity boost again and cools down. So you would need a warm (cooldown phase) and a cold circuit (cooling phase) with probably more than 1 Unit to continously cool something. Guess we have to wait until we know more about cost, energy consumption, field strenght, thermal range and performance before anyone can say it it's of any use for PC environment.

[Thrilla]

EMI is not a problem at all, the cooler doesn't have to be direct-on-die, but rather in a wc loop where chilled water can be carried to a waterblock.

To shield EMI, a simple faraday cage can be used, and based on its operating frequency, there can be holes in the cage to allow water to pass through.

This stuff is gonna be the cooling revolution (if it works and isn't very expensive)
http://img85.imageshack.us/img85/4635/electromag7it.png

[Fairydust]

That loop will not work for continious cooling.

[OmegaMerc]

Sounds great, but I don't see this being applied too soon. Maybe in a few years.:confused:

I don't think the EMI created will be too strong, nothing some more metal can't fix! :cool: (at least for CPU applications)

[MaxxxRacer]

So instead of a radiator with fans you have a radiator with magnets? No heatdump into the room was be a great plus

no heatdump into the room.. lol.. physics doesnt work that way.. if we could magically make heat disapear we would.

essentiaally its a heat pump.. the the enegry sucked up by the magnets gets put to use moving the heat around and also from creating the magnetic field itself. but IIRC eletromagnets heat up quite a bit.

[Fairydust]

...essentiaally its a heat pump.. the the enegry sucked up by the magnets gets put to use moving the heat around and also from creating the magnetic field itself. but IIRC eletromagnets heat up quite a bit.

What energy sucked up by magnets? The material changes it's thermal capacity depending on magnetic field in magnetocaloric materials or according to voltage in electrocaloric materials. More thermal capacity = material cools down as the stored thermal energy is spread among the higher capacity. Once the cooled material get's close to ambient again it has to have a cooldown phase with thermal capacity decrease where it dissipates the collected thermal energy to be able to cool again.

[the]

any one know how fast the phases would be? the heat up and cool down? if its in a matter of seconds or lower (ms mabye? that'd be ideal), it seems like it could be pheasable. a matter of minutes or higher seems a little out of reach. as far as emi? wouldnt you have to have a pretty stong field to start messing with the transitors in the chip? if this stuff is effeint enuff i doubt we'd need more than a tesla. i'll have to read up on what emi has been doing to processors. the computers at the maglab are within feet at a 42 telsa magnet and have never had a problem.

[MaxxxRacer]

What energy sucked up by magnets?

the electromagnets... the power they suck up from the wall outlet...

[nn_step]

I hate to tell you but Check Popular Mechanics about Two and a half years ago they covered this.. Clever technology but it really will only find its greatest advantages in space for the next couple years to come...

[largon]

What energy sucked up by magnets? The material changes it's thermal capacity depending on magnetic field in magnetocaloric materials or according to voltage in electrocaloric materials. More thermal capacity = material cools down as the stored thermal energy is spread among the higher capacity. Once the cooled material get's close to ambient again it has to have a cooldown phase with thermal capacity decrease where it dissipates the collected thermal energy to be able to cool again.Not quite.

thermal energy = atomic/molecular resonance

Magnetic cooling means that the atoms affected by the magnetic field are forced to align with the em flux, this means the resonance of the affected particle is prevented/inhibited. Less resonance is lower temperature. For every "immobilized" atom a certain amount of electrons need to pass the coil that creates the em field, and more heat production in the affected group of atoms (e.g a CPU) means that the coil needs more electrical current to maintaing the temperature. More current passing through the coil creates heat - unless you have a superconducting coil, of course.

Magnetic cooling doesn't really transfer heat, it enables it to appear in another location.
;)

[PL4YD34D]

so what your saying largon is like cooling via magnetic compression, but that isnt what the design of the material is going to allow is it? otherwise wouldnt the field have to be too high to maintain sufficient cooling in metal alloy

[largon]

:doh:
My mistake. Forget what I said. After actually reading the article I see this has nothing to do with magnetic compression cooling.

But, it will not be possible to use this method to cool computer parts. The device requires physical movement of the magnetocaloric material (alloy particles on a rotating tray) and atleast one fluid loop. Plus, the amount of the needed alloy would surely be unpractical to be used in a tight pc enclosure.

The system would easily be more complex compared to a traditional watercooling loop.

[MaxxxRacer]

:doh:
My mistake. Forget what I said. After actually reading the article I see this has nothing to do with magnetic compression cooling.

But, it will not be possible to use this method to cool computer parts. The device requires physical movement of the magnetocaloric material (alloy particles on a rotating tray) and atleast one fluid loop. Plus, the amount of the needed alloy would surely be unpractical to be used in a tight pc enclosure.

The system would easily be more complex compared to a traditional watercooling loop.

PLUS it uses MAGNETS!!! lol... one small problem.. HDD + magnet of this strength = reformat.

[the]

im pretty confident that this woulnd;t be a problem. magnetic storage media and magnetic feilds can co-exsist. a .5 telsa magnet wont touch a hard drives data even it was sitting rihgt on top of it. so even if we did need more than that, i dont think sheilding would be much of a problem.

on a personal note, we needed to destroy all the hdd;s of the old computers at the bank, and got an elcetor magnet designed for hte job. if you gotta vhs wihtin 6 feet of this it would erase it completely. this magnet on erased about 20 percent of the hard drives :( the rest still booted up and we accessed all the data just fine. we ended up have a huge bon fire and melting the rest. so i still consider it a vicotry for magnets :D

[JamesAvery22]

A much better description


The wheel is arranged to pass through a gap in the magnet where the magnetic field is concentrated. As it passes through this field, the gadolinium in the wheel exhibits a large magnetocaloric effect — it heats up. After entering the field, water is circulated to draw the heat out of the metal. As the gadolinium leaves the magnetic field, the material cools further as a result of the magnetocaloric effect. A second stream of water is cooled by the gadolinium. This water is then circulated through the refrigerator's cooling coils. The overall result is a compact unit that runs virtually silent and nearly vibration free, without the use of ozone-depleting gases, a dramatic change from the vapor-compression-style refrigeration technology in use today.


http://www.eurekalert.org/features/doe/2001-11/dl-mrs062802.php



The 'magnetic fridge' envisaged by the Cambridge team would use a phenomenon called the magnetocaloric effect (MCE), whereby a magnetic field causes certain materials to get warmer (a positive MCE) or cooler (a negative MCE).


That sure made it sound like they had a material that got cold when placed in a magnetic field...

[Fairydust]

Wohoo thanks for the link, was beginning to think I got confused.

[the]

enter the man wiht the asnwer. awesome, so its more maintence unfortunaly than i had hoped, but does seem to be a decent idea for computing. seems we could get sub water temps pretty easily, and the compact design idea leads me to bevile it;d be less area required than say a vapochill. we might even see units similar to the vapo/prometia idea only with magnetocaloric compnents. maybe save on energy costs too? what do you guys think?

[PL4YD34D]

sounds like you could have the same problems as you would get tryin to use a regular fridge to cool a water loop rad. if you force the temps up by constantly reheating the cool side of the cooling loop ie the 'magnetic fridge' in this case or the evap loop in a common fridge, then you negate the cooling effect which only works to ambient temps. dont forget that these processors can reach 70c easy if left uncooled for any length of time. so a constant heat draw that removes heat upto even 50c would over power a system designed to remove just ambient temps steadily.

[iddqd]

what if the energy created by cooling the water can be used to pump the water....
That's quite simple, really. It's called thermal convection - a phenomenon where heating a liquid unevenly causes it to circulate within its vessel. I'm not sure what you were going for here...