DescentBB

[youtube]JWQZNXEKkaU[/youtube]

This new cooler design is 30X better than an equivalent heatsink and fan combo. That's right, not 30% better. 30X better!
Ok, so you spin the heatsink - seems like a great idea initially right? But then you realise the problem of howto thermally connect a SPINNING heatsink to a STATIONARY heat source.

Their solution? the heatsink hovers on a tiny thickness "air-bearing", much like the read-write heads of HDDs hover above the platters.
Apparently the superfast air of the air-bearing gives great thermal conductivity, coz it acts "more like a fluid" than slower air. *shrug*

It confuses me greatly. Because one wonders what the point of the METAL heatsink is at all - the whole apparatus is essentially making the heat travel from the heat-source, THROUGH AIR, to the metal heatsink, AND THEN INTO AIR AGAIN. They are going through air TWICE. That seems so daft!!!

Okay, now this is cool stuff.

Roid, from what I can tell, the fluid effect of the air-bearing does the job of drawing the heat directly from the chip...but that heat still needs to move away from the chip. The reason we use heatsinks over things like CPUs, as opposed to just blowing a fan straight over the chip, is because of the increased surface area involved: the more space the heat has to "spread out," the more air can flow over it, and the more effectively the structure can be cooled. In this case, the rotation is pulling double-duty as both heatsink and fan, giving the heat the large surface area over all those fins to spread out, and simultaneously drawing air over them. It's a remarkably elegant design.

If the heat first has to go into the air of the air-bearing, then why don't they just right-then-and-there exhaust that air. I mean, the heat has now moved from the heat-source into the air - the goal is reached, problem over, no?
I guess i'm confused as to why the fins are on the top of the heatsink, instead of being on the bottom, positioned to renew-&-exhaust the air-bearing's air at max efficiency.
Do you get what i mean? If an air-bearing can move heat from a source into air (ANY air) at such a staggering speed, then surely that's where the design should concentrate for all of it's gains. To me - this whole thing is an air-bearing, everything else is superfluous and is probably REDUCING it's efficiency, I don't get why they even need those fins on the top other than to supply air to the bearing.

it's so obvious... that i'm sure they must have already ruled that path out for some reason, Sandia boffins must know what they are doing, i just wish i knew why too, i'm frustrated in my ignorance :-/

I am just so blown away at the thermal efficiency of that air-bearing. *gush*


edit: when i think about it, this air-bearing seems a lot like a single-platter Tesla turbine. hmmmmm, this is part of why i wonder why they bother with the fins.
(ps: no i am not a Tesla fanboy/conspiracy theorist, i find the whole quasi-religious devotion that ppl have to Tesla to be very frustrating)

Apparently the superfast air of the air-bearing gives great thermal conductivity

I guess the velocity of the air changes it thermal conductivity characteristics.

EDIT: The fins must create enough top turbulence to keep the heat from being trapped and then retained by the air bearing. I bet if the fins were on the bottom you could do the opposite and create higher temperatures.

roid wrote:If the heat first has to go into the air of the air-bearing, then why don't they just right-then-and-there exhaust that air. I mean, the heat has now moved from the heat-source into the air - the goal is reached, problem over, no?
I guess i'm confused as to why the fins are on the top of the heatsink, instead of being on the bottom, positioned to renew-&-exhaust the air-bearing's air at max efficiency.

Well yeah, the heat has moved into that air-bearing layer, but it's still right next to the heat source: I'd wager that the air layer is probably less than a millimeter thick. You're going to need to get it much further away in order to properly cool the source. As for why the fins are on top, for one, I'd be willing to bet that you need a flat surface on the bottom in order to generate that air-bearing effect in the first place...if you flipped it upside-down, you'd generate all kinds of crazy turbulence in that boundary, and it might not even be able to hover at all. For another, the fins on top allow for the cooler external air to be drawn into the fin enclosure, which wouldn't work if the opening was facing down right against the heat source.

To put it more simply, if you inverted the fin assembly, you wouldn't get the proper effect that makes this whole thing possible in the first place.

If they are talking about an air-bearing like what a hard drive uses, the height is even less than you think: The fly height of the heads on a modern hard drive is roughly 3 nanometers (0.000003 millimeters). Even a fingerprint is much thicker than that.

If we were to scale this up so a modern 3.5" / 7200 RPM disk was instead the size of the earth (a scale of 143,489,314 to 1): The read heads would fly at a height of just under 17 inches off the ground at a velocity of 10.7 billion miles per hour (or just over 16 times the speed of light).

Top Gun wrote:Well yeah, the heat has moved into that air-bearing layer, but it's still right next to the heat source: I'd wager that the air layer is probably less than a millimeter thick. You're going to need to get it much further away in order to properly cool the source.

ah, ok. So the boundary layer is likely both key to the bearing's thermal efficiency AND key to why it would be difficult to exhaust the air.
Kinda cool that both the top and bottom of the heat-sink utilise innovative manipulations of boundary layer physics.

...argh but it's still so tantalising, i'm gonna be pondering ways to get around this bottleneck forever.
And it's so similar to a Tesla turbine design. oh god oh god i'm going to turn into one of those people i just know it.