Sandia Lab’s New “Cooler” Technology

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
Heatsink with a spinning fan? Screw that, Sandia Labs spins the heatsink! Thanks to J.C. Watson for the link!

The “Sandia Cooler,” also known as the “Air Bearing Heat Exchanger,” is a novel, proprietary air-cooling invention developed by Sandia researcher Jeff Koplow, who was recently selected by the National Academy of Engineering (NAE) to take part in the NAE’s 17th annual U.S. Frontiers of Engineering symposium.
 

TwistedAegis

[H]F Junkie
Joined
Oct 7, 2009
Messages
8,958
Well the idea is to get air moving across the surface...I guess it could make sense to cut out the middle man (fan) and just spin the whole damn thing. The only thing I'm wondering is how efficient the transfer of heat from the stationary base is to the spinning cooling unit.
 

Mungojerrie

[H]ard|Gawd
Joined
Jun 19, 2011
Messages
1,940
another thing is how much more energy is required to spin the heatsink than to spin a fan....
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
Sounds hot actually. Crossing a 0.001" air gap is like crossing a 1/4" thick slab of good thermal goo. Not to mention producing something like this is going to be quite expensive. Label me a skeptic until I see some hard numbers.
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
Ok, just read the white paper. It looks like it's going to be hard pressed to be better than a standard True 120. It's also going to require a external supply of pressurized air for the air bearing. I'm a good bit worried that they felt they had to make it out of 7075 (some of the strongest aluminum avaliable). It looks like it would be cost prohibitive for the PC market, but might have some usage in HVAC, though with something that large, a <0.001 flatness spec is going to murder you.
 

grinch

Weaksauce
Joined
Dec 6, 2004
Messages
81
With the Sandia Cooler, heat is efficiently transferred across a narrow air gap from a stationary base to a rotating structure.

This part puzzles me a bit. How narrow is the air gap? Is there any thermal substance between the two? Is there any lubrication between the two? What happens if dust starts collecting in this gab. What happens if the axle/bearing is bent a little? Seems like a lot of stuff could go wrong here, and it also seems like you would lose a lot of your heat transfer because of that gap. But maybe the effect of reducing the boundary layer gap helps enough to counteract that. Cool theory, but I guess we'll see.

Mungojerrie also brought up a good point about energy differences.
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
This part puzzles me a bit. How narrow is the air gap? Is there any thermal substance between the two? Is there any lubrication between the two? What happens if dust starts collecting in this gab. What happens if the axle/bearing is bent a little? Seems like a lot of stuff could go wrong here, and it also seems like you would lose a lot of your heat transfer because of that gap. But maybe the effect of reducing the boundary layer gap helps enough to counteract that. Cool theory, but I guess we'll see.

Mungojerrie also brought up a good point about energy differences.

It's all covered in the white paper that's linked from the article.
 

fairlane

Limp Gawd
Joined
Jun 18, 2004
Messages
297
It says the unit will use "significantly less" energy to cool the cpu, but makes no reference (unless its in the white paper Vengeance?) to actual performance compared to other HSF's. I'd also like to see how it fares
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
It claims 0.2 C/W for the prototype which is on the order of a true120/venomous X, etc. It's 4" in Dia and 0.6" high.

The white paper reads too much like an advertisement for my tastes.
 

cyclone3d

[H]F Junkie
Joined
Aug 16, 2004
Messages
14,728
Look at thos super thick fins... that sucker is going to weigh a decent amount and offer only mediocre cooling performance.

The "air gap" should be replaced by a "sealed" liquid chamber with heat transfer fins being on both the stationary and the spinning part. You would probably need a small reseviour with a float and pump to keep liquid in the "sealed" chamber all the time.

And them looking at HVAC applications... HAHAHAHAHAHAHA. All that really needs to be done to make current units a whole lot more efficient is replacing the sucky regular bladed fans with huge squirrel cage fans. That alone would cut power usage AND noise down a significant amount. The motors would also last longer due to their being a whole lot less force being put on the bearings by the way the fan pulls air.
 

velusip

[H]ard|Gawd
Joined
Jan 24, 2005
Messages
1,579
http://prod.sandia.gov/techlib/access-control.cgi/2010/100258.pdf

There is no consideration of failure points in the 48 page document released by Sandia. It seems like this could easily gum up over time. Disc drives heads run on air bearings as well, but in an enclosed space with an air filter to ensure the air is clean. It would be difficult to enclose this without greatly reducing the desired efficiency.

See page 17 for a diagram of the delicate internals. It seems liket he tiniest amount of dust would do a number to this.

Also page 19 shows the theory behind the mass flow rate of air (mostly nitrogen) through the thin air bearing.

It looks expensive! I like it, but I'm concerned with keeping it clean. I need to pull my ordinary HSF off and wash it every few months by simply taking off the fan and dunking the rest in hot water. Not so sure this would be easy to clean.
 

Necere

2[H]4U
Joined
Jan 3, 2003
Messages
2,767
It claims 0.2 C/W for the prototype which is on the order of a true120/venomous X, etc. It's 4" in Dia and 0.6" high.

Equivalent to a True 120 - but fits in a 1U rack, with low noise? Sounds pretty impressive to me.
 

gunderwood

Weaksauce
Joined
Sep 25, 2006
Messages
111
The "air gap" should be replaced by a "sealed" liquid chamber with heat transfer fins being on both the stationary and the spinning part. You would probably need a small reseviour with a float and pump to keep liquid in the "sealed" chamber all the time.

My thoughts too...except why not make the liquid side of the heat exchange the liquid pump as well? One motor, an air pump/exchanger and a liquid pump/exchanger.
 

Aluisious

2[H]4U
Joined
Mar 6, 2009
Messages
2,189
Considering how mercilessly difficult it would be to get that thing to actually work with those tolerances, pressurized air, etc, and how utterly pointless it is when it can't beat technology from a $40 cpu cooler, you may as well go whole hog and just have a line of unicorns ready to lick a diamond heatsink.
 

kzaske

Limp Gawd
Joined
May 12, 2010
Messages
211
I want to read the reviews of how well it works. For some reason, I don't expect to see much improvement.
 

murray13

n00b
Joined
Oct 7, 2009
Messages
21
Another engineering marvel from our national labs. They fix one problem, stagnant air around the fins, and introduce many more. They have produced a prototype that would be expensive to make, difficult to implement (servers don't have a compressed air supply), be prone to failure, and not improve the end result (except on paper).

When is the server world going to standardize liquid cooling? That's where we are headed, and need to go. Instead of chilling air with cooled liquid why not remove air all together. Air is listed in many places as a great insulator. Water on the other hand is one of the best at transporting heat. The biggest problem is getting all those involved to agree on a standard. <rant mode off>
 

AaronGant

Limp Gawd
Joined
Feb 21, 2010
Messages
391
I know that dust is nasty for hard drives, but hard drives work with much smaller gaps than this thing calls for (something like .0000002 inches). It's not like this thing has a head to crash or data to get scratched if dust did get in there, it's quite possible that it just pushes it out of the way and causes no problems.

I've got an opened up hard drive that I have lying around to show people the internals, it's covered in dust, but when I plug it in it'll keep on spinning and the heads just clean the dust off. Yeah, it wont keep data, but it still moves air.

You might need to clean the fins out, but what fan doesn't?

http://prod.sandia.gov/techlib/access-control.cgi/2010/100258.pdf

There is no consideration of failure points in the 48 page document released by Sandia. It seems like this could easily gum up over time. Disc drives heads run on air bearings as well, but in an enclosed space with an air filter to ensure the air is clean. It would be difficult to enclose this without greatly reducing the desired efficiency.

See page 17 for a diagram of the delicate internals. It seems liket he tiniest amount of dust would do a number to this.

Also page 19 shows the theory behind the mass flow rate of air (mostly nitrogen) through the thin air bearing.

It looks expensive! I like it, but I'm concerned with keeping it clean. I need to pull my ordinary HSF off and wash it every few months by simply taking off the fan and dunking the rest in hot water. Not so sure this would be easy to clean.
 

RadCliffeX

2[H]4U
Joined
Oct 11, 2005
Messages
3,375
/looks at giant metal heatsink in my computer...



/looks at big scar on finger from plastic fan on said heatsink



...Fuck that
 

Phoenix333

2[H]4U
Joined
Nov 27, 2009
Messages
3,510
I just want to know two things:

1) The results of Kyle putting it through a [H]ardOCP torture test, and
2) WILL IT BLEND?
 

cerwinvega

Limp Gawd
Joined
Jun 15, 2004
Messages
374
It's amazing that none of you will be attending NAE’s 17th annual U.S. Frontiers of Engineering symposium, since you all seem to know more than they, or anyone directly involved with this project, how well it will work.

Oh wait, no, this is the internet, and you likely just think you're far smarter than you actually are.

My bad. :eek:
 

vengence

Level capped
Joined
Nov 7, 2007
Messages
18,471
It's amazing that none of you will be attending NAE’s 17th annual U.S. Frontiers of Engineering symposium, since you all seem to know more than they, or anyone directly involved with this project, how well it will work.

Oh wait, no, this is the internet, and you likely just think you're far smarter than you actually are.

My bad. :eek:

That's a pretty overrated conference IMHO.
 

drakken

[H]ard|Gawd
Joined
Aug 19, 2004
Messages
1,196
if you would like to go boo this pointless device here are the requirements just to get a chance to go to the symposium.

Getting invited to FOE

Attendance at the meeting is by invitation-only following a competitive selection process.

A nominations mailing is sent each February to CTOs, professors, NAE members, heads of government labs, and others. We also get names of winners of "early career" engineer awards from professional engineering societies, NSF, ONR, and foundations. FOE alums can also nominate individuals to attend. Nominees should possess the following characteristics:

1. be between the ages of 30 and 40, or at most, 45;
2. have demonstrated accomplishment in engineering research and technical work with recognizable contributions to advancing the frontiers of engineering;
3. be interested in engineering developments in other fields and able to consider how advances, techniques, and approaches in those areas relate to the nominee's own field;
4. possess the potential to be a future leader in the U.S. engineering endeavor.


The nominated individuals are contacted and asked to send an application if they are interested in being considered for invitation. Once the applications are in, the organizing committee assists in the selection of the invitees, considering the research of applicants and ensuring an appropriate balance of engineering fields, institutions, geographic representation, etc.

If you are interested in applying, you must first be nominated by someone (a nominator), preferably a more senior individual who is familiar with your research or technical work.
 

Cbshahji

[H]ard|Gawd
Joined
Jul 3, 2010
Messages
1,970
now that's something to look in the future.

What is the Noise lvl on this??!!
 

Weenis

I said WEENIS, not...
Joined
Apr 10, 2006
Messages
4,807
At least they are trying something different.

This. We need things that push new ideas out for things rather than just accepting things as is.

The very first look equated to, wow that see.... wait no that's never going to work that well.

That's as far as I had to go..
 

Taldren

Gawd
Joined
Nov 28, 2006
Messages
521
no need to reinvent the wheel.

we already have good cooling solutions.

You know, people use this expression is it never makes any sense. The wheel has been reinvented so many times over and each time was a great improvement. Stone rollers to solid wooden wheels, to spoke wooden wheels, to spoke wooden wheels with steel bands, to steel solid wheels with rubber tires, to spoke wheels with rubber tires, ect.

Thank god people kept reinvented the wheel.
 
Top