Fanless, spinning heatsinks may be in our future!

[dealmaster]

Hey guys, some coworkers showed me this and I thought you guys would like to hear about it:

http://www.extremetech.com/extreme/89710-the-fanless-spinning-heatsink-the-heatsink-is-the-fan

 

[Tokae]

That is interesting! Like comments are saying, I don't see how that would reduce the thin layer of air.. But if it does indeed make cooling that much better I would be all over that!

 

[fall]

The Sandia Cooler may also be the technology that smashes down the “Thermal Brick Wall” that is preventing computer chips from moving beyond 3GHz.

I may have misread this but it sounds like these guys are a few years behind... Unless they meant 4ghz which there still kinda behind...

It does sound interesting... however i dont see this ecer being used as an enthusiest standpoint... maybe OEM machines (dell, HP) but other than that I dont see it

 

[Automata]

How is that spinning chuck of metal with fins not a fan?

 

[mbentley]

what i want to know is how exactly that conducts heat while it is spinning...

 

[jsharpley]

How is that spinning chuck of metal with fins not a fan?

Yea ... it connects with a baseplate and still leaves a thin layer of air ... sounds a lot like our current heatsink-fan combos. I wonder how efficient a conductive pipe setup + total vacuum would be as a cpu cooler ... wtb a PDF showing the details of that setup

 

[petteyg359]

How is that spinning chuck of metal with fins not a fan?

Yea ... it connects with a baseplate and still leaves a thin layer of air ... sounds a lot like our current heatsink-fan combos.

Per above, this is BS "innovation".

 

[TorqueRanger]

Lol This guy is completely ridiculous because the heat is being transferred to the spinning metal disc that looks like an impeller fan but is not a fan and heat is suppose to transfer through a 0.03-mm-thick circular disk of air.. Yeah ok ... Well then how is the base plate suppose to cool down with no air moving on it or around to help cool ??? I though that you want as little air as possible to help improve the heat exchange to the heat sink.. But if i read his finding correct then all I need to do for the ultimate cooler is to mount a thick block of copper to the cpu and then mount a spinning cooler on top with a 0.03-mm-thick circular disk of air between the to and we are set to go..

 

[Dooms101]

I think this is an interesting idea... however whats the point? An extruded alu heatsink and chuck of plastic fan costs an OEM like $1 to make and is effective enough to cool an average cpu (65w - 95w). Plus a PWM fan on a cpu sink barely uses any energy as the fan blades themselves are plastic and very lightweight. Trying to rotate a giant heavy chunk of metal is far less efficient and uses more energy. I don't see air as being a very good thermal interface between the base plate and spinning heatsink (aka fan), but maybe at a super thin layer it works okay. Although having a layer of air probably just creates more resistance and makes this design even less efficient...
As far as I am concerned OEM heatsinks and fans are far from amazing but they definetly get the job done cheaply and effectively.

 

[Archer0915]

Great idea. Too bad that .001 of air screws it up. I just did a search of the document and found nothing on overcoming the rotten thermo-conductivity of air.

Essentially it is working with radiant heat which would be fine for a lot of things but with the rapid changes in heat that happen inside a CPU I think the IHS to tim to lower plate transfer will be good but no better than current HSF setups. Moving the heat through an insulating layer of poor thermo conductive material (air) to another chunk of metal is where the true limitations of this will come into play.

Great for slow consistent heat, bad for rapid variations in heat output.

Who here would sit their HSF on .001 feeler gauges around the perimeter of the IHS with absolutely no clamping pressure?

 

[Bobnova]

I just read through it too.
There's bottled high pressure air going through a bazillion little passages in the base. That'd be where their cooling is coming from.
0.03mm is simply too large, plus their own graphs show lousy cooling at all, and only vaguely sortof decentish cooling with the thing running at >10,000rpm and eating >30w spinning the thing.
We all know what 30w high RPM fans sound like, not quiet!

It's one big fat disguised scam.

 

[MattNo5ss]

Yeah...don't think it'll work too well for CPUs or GPUs...

Who here would sit their HSF on .001 feeler gauges around the perimeter of the IHS with absolutely no clamping pressure?

I'd do it with my P4 631, it would be fun! Thermal shut down would happen almost instantly...or the P4 would explode...

 

[Archer0915]

Yeah...don't think it'll work too well for CPUs or GPUs...



I'd do it with my P4 631, it would be fun! Thermal shut down would happen almost instantly...or the P4 would explode...

Put it on video

 

[repilce]

That thing looks like a f'd up squirrel cage.

 

[RGE]

From reading their performance specs, they are aiming at mass market/vendors, ie replacing intel stock coolers, and not in any way competing with aftermarket coolers.

Looks like they are claiming at around 2000 rpms, it only uses about 2-3W of "fan" power and can be made cheaper, smaller, less dust, and yet beat intel stock cooler in performance, though that remains to be seen.

Maybe useful to intel or not, but clearly useless to anyone with an aftermarket air cooler or any enthusiast.

They admit they dont perform up to high end air, which for example with just single 2500 rpm fan using 4W, prolimatech mega has 0.1 C/W performance, and with faster fan the C/W would be even better. And their contraption uses up to 20-30W for less performance of 0.2 C/W and 10,000 rpms speed, ie probably more noise with much more energy for much worse performance. But they clearly state they are only aiming at mass market, small form factor, though remains to be seen whether they can even compete with that.

 

[Bobnova]

Intel stock 1366 fan uses 1.9w, so 2w isn't exactly an improvement. I just tested the Intel cooler last weekend.

 

[Psycogeec]

what is next? hard drives that the whole case turns instead of the platters, Spinning electronics in PSUs with needed electical tracts.
why not have the whole computer shoved in one of them turbines for the top of a roof, and try and spin the whole thing weee haaa
besides what has already been discussed the bearings themselves would be taking on a lot more work, in mobile situations you have a bit of a flywheel going, so gyroscopic things would come into play with any weight .
what would make it immune to dust? it flinging stuff off? if there is any air moving on the air cushion woulldnt dust be an even bigger problem?

.

 

[RGE]

yeah, hard to say....I lost all interest when i realized it was just aimed at mass market stock coolers, and not enthusiast cooling. And like Bobnova pointed out, I dont get the power savings claims either based on their own figures.... as to whether it can be made cheaper, cool better than typical stock coolers, or any better at preventing dust, no way to know unless it gets to manufacturing stage and independently tested.

 

[bing]

Say even it works, just curious if the current metal manufacturing is good enough that can make this chunk of rotating metal "cheaply" that can beat the stock heatsink in cost, while still has good precision on it's base for good heat transfer and the weight balance across the whole wheel, otherwise I'm afraid it will be just another vaporware.

 

[pwnmachine]

TBH I don't see this being used in our line of performance computing but I think this has huge implications for laptops and small form factor performance pcs.