The "Sandia Cooler" - fanless spinning heatsink

T

Tisca

Limp Gawd
Joined
Sep 4, 2010
Messages
244
http://www.extremetech.com/extreme/89710-the-fanless-spinning-heatsink-the-heatsink-is-the-fan
http://www.newscientist.com/blogs/onepercent/2011/07/new-heat-sink-could-slash-us-e.html

cooler3.jpg
 
"The Sandia Cooler may also be the technology that smashes down the “Thermal Brick Wall” that is preventing computer chips from moving beyond 3GHz." - The article described in the topic.

Ummmm do these people at ExtremeTech live under a rock or something?

Anyways its just a fan thats a heatsink. The term fan simply means:

any device for producing a current of air by the movement of a broad surface or a number of such surfaces. -Dictionary.com
 
it looks like it would need to be hooked directly to the power supply rather then a motherboard fan header to use it. also the thing looks huge and I worry that it will vibrate and damage the motherboard and or parts on it. When they release a computer cpu friendly version of this concept, I will really enjoy reading the reviews about it.

But it does seem like the concept is late to the party seeing as sandy bridge cpus do not give off as much heat like older cpus did. I am sure the new sandy bridge revisions or ivy bridge or newer, will lower the thermal heat and energy use even more and allow for higher clock speed and multitasking.
 
the problem is that a moving heatsink wont be able to have the solid contact with the CPU. maybe someone should invent a CPU socket that can rotate, lol.
it would definetly by quieter, than a typical setup, but my guess is that a big ol' good quality heatsink and a 500rpm fan will cool the same, sound the same, use less power and be less expensive.
i would think if it could be made in such a way so that there was a layer of thermally conductive viscous liquid that the base instead of a layer of air alot could be done with it, but sealing it would be an issue.

the advantage i see to this product is that it should blow away other low profile heatsinks out of the water...

problems is see are that this would be way expensivedue to the tight manufacturing tolerances it would need, it would have to be balanced perfectly, bending a fin could through that off. the air slice is so thin that you would need a really high quality motor, im not sure if FDB would work, it might make the case vibrate more, which can lead to a amplification factor of the noises that other components make...

i for one am impressed if the data is true...but i guess well have to wait and see. this is definetly an interesting concept.
 
Last edited:
tangoseal said:
Anyways its just a fan thats a heatsink. The term fan simply means:

any device for producing a current of air by the movement of a broad surface or a number of such surfaces. -Dictionary.com
Click to expand...

the difference is that this "fan" doesnt direct a current of air at another body...it induces a windchill factor on itself.

is it fan? yes. but stop nitpicking.
 
The Sandia Cooler may also be the technology that smashes down the “Thermal Brick Wall” that is preventing computer chips from moving beyond 3GHz.
Click to expand...
What?
 
I'm reading the research paper.. I like it.

For those of you blindly bashing it, go read the paper. For those of you not interested in reading the paper, grow up. This thing has it's uses, and it certainly is an interesting idea.

And let me remind you all, he started this project in 2005. Those of you who can't count in years, think Pentium 4... At that time, there certainly was a 3Ghz limitation.
 
The first step towards optimus prime, go go moving parts :)

I would love my vid card to change shape for each diff game... im sure when trying to run metro max triple screen my vid card's would form the shape "hell no".
 
The concept of how it moves air seems very plausible. The heat exchange between the CPU and the cooler is my concern. How do you maintain a thermal exchange when one piece is spinning and the other is fixed? It looks like a new type of thermal interface material would need to be used. This kind of technology looks promising, but it will need another new technology to make it work if I am understanding this correctly.
 
Direfox said:
The concept of how it moves air seems very plausible. The heat exchange between the CPU and the cooler is my concern. How do you maintain a thermal exchange then one piece is spinning and the other is fixed? It looks like a new type of thermal interface material would need to be used. This kind of technology looks promising, but it will need another new technology to make it work if I am understanding this correctly.
Click to expand...
According to the whitepaper, heat conductivity across the very thin air gap ("air bearing") that the system maintains between it's moving and static part is quite efficient.
 
Direfox said:
The concept of how it moves air seems very plausible. The heat exchange between the CPU and the cooler is my concern. How do you maintain a thermal exchange when one piece is spinning and the other is fixed? It looks like a new type of thermal interface material would need to be used. This kind of technology looks promising, but it will need another new technology to make it work if I am understanding this correctly.
Click to expand...

The measurements for the thermal resistnace of the air gap are right there in the whitepaper (they are quite low)

I still have my doubts though, a major basis for what's claimed as an advantage of this design is reduction of the boundary layer of airflow past the fins. This boundary layer is still there regardless if the HS is moving or a fan is moving across the HS. And reduction of this boundary layer (air flow velocity) is accociated with increased noise.
 
Hey its still a cool innovation non the less. But what we need is CPU's made from organic materials, i.e. brain matter. All we have to do is tube feed it water and food an it will get smarter and faster as it grows in your motherboard but it will not generate heat at all.

Hell the human brain is so multithreaded and complex that all the computers in the world added together do not have the processing power of a 3 year old child. This was according to some Discovery heath show I watched not too long ago.

I still think its amazing concept. Maybe we should put this spinning heat sink on our heads to keep cool in this southern Atlanta Georgia heat where I live at.
 
tangoseal said:
Hey its still a cool innovation non the less. But what we need is CPU's made from organic materials, i.e. brain matter. All we have to do is tube feed it water and food an it will get smarter and faster as it grows in your motherboard but it will not generate heat at all.

Hell the human brain is so multithreaded and complex that all the computers in the world added together do not have the processing power of a 3 year old child. This was according to some Discovery heath show I watched not too long ago.
Click to expand...

So what happens when it becomes self-aware?
 
okashira said:
The measurements for the thermal resistnace of the air gap are right there in the whitepaper (they are quite low)

I still have my doubts though, a major basis for what's claimed as an advantage of this design is reduction of the boundary layer of airflow past the fins. This boundary layer is still there regardless if the HS is moving or a fan is moving across the HS. And reduction of this boundary layer (air flow velocity) is accociated with increased noise.
Click to expand...

You seem to imply that there is no real difference, in terms of the boundary layer, between standard designs, and this design, and i would disagree.

The boundary layer exists because of general forces occuring between the metal and the air. I'm not going to pretend to know what those forces are, but they are there: i assume mostly friction.

The problem with moving air onto the heatsink is that what you are essentially trying to do is bombard the heatsink with enough air that you hope it dislodges some of the boundary layer. The upper parts of the boundary layer will get dislodged easier than the bottom areas, as they are better protected by the air above it. You end up with some sort of a force gradient; upper parts of the boundary layer are given a lot of force by air rushing past, but the lower parts get barely any forces applied to them at all. Furthermore, the boundary layer in this design depends completely on where that air being blown. Dead spots in the heatsink mean there is a constant thick boundary layer.

Wouldn't it be fantastic if instead of trying to push the boundary layer away with more air, there was some natural force that simply pulled and tugged on each air molecule equally away from the metal, regardless whether that air molecule was in the boundary layer or freely floating between the heatsink fins?

That's where this design comes in, instead of trying to push the boundary layer away with more air, something that, according to the designer, is ineifficient, he has found a way to tug away at all the molecules equally by using centrifugal force, something that does not occur on any conventional heatsink or radiator design.

So I would argue there is a major difference in the way the boundary layer is treated in this design, and also, the dimensions and shape of that layer are probably vastly different.

I'm still amazed though that that .001" air gap the heatsink floats on causes minimal heat resistance, when even a tiny air gap between the chip and a regular heatsink can cause terrible performance; perhaps another showcase of how bad boundary layers in static environments can get?
 
This might actually make a decent GPU cooler. The proof-of-concept design is just about the right size for that, and should be able to dump at least a couple hundred watts of heat.
 
serpretetsky said:
You seem to imply that there is no real difference, in terms of the boundary layer, between standard designs, and this design, and i would disagree.

The boundary layer exists because of general forces occuring between the metal and the air. I'm not going to pretend to know what those forces are, but they are there: i assume mostly friction.

The problem with moving air onto the heatsink is that what you are essentially trying to do is bombard the heatsink with enough air that you hope it dislodges some of the boundary layer. The upper parts of the boundary layer will get dislodged easier than the bottom areas, as they are better protected by the air above it. You end up with some sort of a force gradient; upper parts of the boundary layer are given a lot of force by air rushing past, but the lower parts get barely any forces applied to them at all. Furthermore, the boundary layer in this design depends completely on where that air being blown. Dead spots in the heatsink mean there is a constant thick boundary layer.

Wouldn't it be fantastic if instead of trying to push the boundary layer away with more air, there was some natural force that simply pulled and tugged on each air molecule equally away from the metal, regardless whether that air molecule was in the boundary layer or freely floating between the heatsink fins?

That's where this design comes in, instead of trying to push the boundary layer away with more air, something that, according to the designer, is ineifficient, he has found a way to tug away at all the molecules equally by using centrifugal force, something that does not occur on any conventional heatsink or radiator design.

So I would argue there is a major difference in the way the boundary layer is treated in this design, and also, the dimensions and shape of that layer are probably vastly different.

I'm still amazed though that that .001" air gap the heatsink floats on causes minimal heat resistance, when even a tiny air gap between the chip and a regular heatsink can cause terrible performance; perhaps another showcase of how bad boundary layers in static environments can get?
Click to expand...

That's pretty long winded for someone who doesn't know what forces cause fluid flow. :p

You may not know, but I do. This is one of the cornerstone topics taught in a mechanical engineering program - boundary layers and heat transfer through boundary layers.

There is no difference between the HSF moving and air moving past the HSF in terms of heat transfer when talking about a particulir fine.

Boundary layers result from the difference in velocity between a solid surface and a free fluid. Fluid infinitesimally close to the surface does not move. If it did, it would be shearing at an infinitly fast rate (not possible). The boundary layer is simply the velocity profile of the fluid charted from the surface (zero relative to the surface) all the way to infinity (whatever the free velocity of the fluid is) Yes, the boundary layer is infinitely thick. For engineering purposes we define the actual boundary to be something like 90% of the free fluid velocity which gives us something finite.

That said, for our purposes, there are two things that infuence boundary layer thickness:

1. distance from the leading edge of the fin (assuming the fluid is flowing paralell to the fin, approaching the edge of the fin)

2. free fluid velocity.

the boundary layer thickness is zero (undefined) at the leading edge and grows to a constant value some distance down the fin.

The thickness of the boundary layer, and weather or not the flow is turbulent are the key (only) factors in determining the thermal resistance. (there are some additional effects that arise from different delta T's and actual average air temperature, but we can consider those constant.)


Oh, and the heat transfer across the air gap is so good because of turbulence. The shear rate of the fluid is very high (remember the fluid is stationary relative to each plate, so the velocity profile has to make a sharp change from one plate to the other over a short distance) This induces a lot of turbulence, which is a big driver of heat transfer.
 
okashira said:
That's pretty long winded for someone who doesn't know what forces cause fluid flow. :p

You may not know, but I do. This is one of the cornerstone topics taught in a mechanical engineering program - boundary layers and heat transfer through boundary layers.
Click to expand...
So the centifugal force applied by the fins onto the air has no bearing on it whatsoever?
 
Any "centrifugal force" is confined to the heatsink itself, otherwise I am not sure what you are referring to. There is certaintly (rotational) airflow relative to the fins on the OD of the heatsink. However, that's not the intent of the design. **note the figure in the OP shows the heatsink rotating in the wrong direction** It rotates and the fins are angled such that the fluid is guided out of the heatsink (the fluid will try to resist the motion of the heatsink; imagine the fluid flowing rotatonally relative to the heatsink within it. The blade angle guides the fluid out of the heatsink) so the rotational flow I think you are referring to is actually wasted in this circumstance (in a real centrifugal pump, the energy in this rotational motion of he fluid leaving the impeller can be recovered by the design of the housing which is obviously non existant here)

I think any advantage of this design will come from the fact that airflow across all of the fins is more uniform in velocity compared to a "blown" heatsink which may have more areas of stagnant flow (poor heat transfer) and areas of higher velocity (noise). There is a certain "sweet spot" velocity where you balance noise and heat transfer.



edit: I knew something was weird. They call it an impeller but they have the dang thing rotating the the wrong direction in the figure in the OP.
 
Last edited:
edit: I knew something was weird. They call it an impeller but they have the dang thing rotating the the wrong direction in the figure in the OP.
Click to expand...

Really? The similarly shaped impellers in my DDC pumps spin the same way.
 
Are you sure? It should be spinning counterclockwise when looking at the figure.
 
xDezor said:
I'm reading the research paper.. I like it.

For those of you blindly bashing it, go read the paper. For those of you not interested in reading the paper, grow up. This thing has it's uses, and it certainly is an interesting idea.

And let me remind you all, he started this project in 2005. Those of you who can't count in years, think Pentium 4... At that time, there certainly was a 3Ghz limitation.
Click to expand...

I had a socket 478 P4 HT that was stock at 3.6 Ghz in 2004....
 
okashira said:
Are you sure? It should be spinning counterclockwise when looking at the figure.
Click to expand...
that is the correct direction.
here is a video of a impeller spinning opposite direction with opposite angled fins:
[ame="http://www.youtube.com/watch?v=UrChdDwHybY"]Centrifugal Pump impeller (3d design )[/ame]

edit: uggh... i didn't want to imbed that video. Oh well, whatever.
 
okashira said:
Any "centrifugal force" is confined to the heatsink itself, otherwise I am not sure what you are referring to.
Click to expand...
I am referring to the fact that you can design an impeller that has straight, non angled, non curved vains radiating straight from the center, and you can spin this impeller in either direction, and it will stick suck air through the center of the impeller. The vains are constantly "escaping" from the linear travel from the air, because of the circle they must follow. From the point of the view of the air molecules, they are being pulled constantly outward relative to the impeller. A straight vained impeller wouldn't be as eifficient, but it would still work.

It's the exact opposite of us humans standing on earth, instead of us being pulled inward towards the center of the earth, imagine we're all being pulled out equally, and desperately trying to hold on to whatever we can find. The other option to remove humans from earth is to bombard the earth with other humans and hope the standing humans get dislodged by this bombardment.
 
serpretetsky said:
I am referring to the fact that you can design an impeller that has straight, non angled, non curved vains radiating straight from the center, and you can spin this impeller in either direction, and it will stick suck air through the center of the impeller. The vains are constantly "escaping" from the linear travel from the air, because of the circle they must follow. From the point of the view of the air molecules, they are being pulled constantly outward relative to the impeller. A straight vained impeller wouldn't be as eifficient, but it would still work.

It's the exact opposite of us humans standing on earth, instead of us being pulled inward towards the center of the earth, imagine we're all being pulled out equally, and desperately trying to hold on to whatever we can find. The other option to remove humans from earth is to bombard the earth with other humans and hope the standing humans get dislodged by this bombardment.
Click to expand...

I got a visual on the bombing of earth with humans and i Lol'd
 
I read the research paper and it sounds quite promising. Hopefully it'll make it to market soon so we can see if it'll work well in practice.
 
Its been a few months and I want one of these in my Xmas stocking. Anyone know the latest
 
xDezor said:
I'm reading the research paper.. I like it.

For those of you blindly bashing it, go read the paper. For those of you not interested in reading the paper, grow up. This thing has it's uses, and it certainly is an interesting idea.

And let me remind you all, he started this project in 2005. Those of you who can't count in years, think Pentium 4... At that time, there certainly was a 3Ghz limitation.
Click to expand...

In the early 2000s intel had northwood c, p4s at 3.4ghz
 
why would a motor spinning a metal sink be more reliable or quieter than a motor spinning a fan?
 
I find the idea fascinating. Since I do not know enough about fluid dynamics I will stay away from that discussion. I could ask my engineer brother, but he gets way too technical sometimes. (he was a nuclear technician in the Navy-he would know something about heat transfer)

But the idea of a quiet and efficient CPU cooler. I am for it. I just hope it will work on its side as a lot of case have vertical motherboards.
 
You must log in or register to reply here.
Back
Top