Sandia Shows Off Prototype of Its Impeller Cooler

Subject: Cases and Cooling | June 25, 2012 - 05:45 PM |
Tagged: sandia, impeller, heatsink, cooling, cooler, air bearing

A white paper by Sandia National Laboratories caught the attention of the media last year with big claims for high performance cooling. The researchers had claimed to invent a new type of heatsink based on a impeller design that was allegedly 30% more efficient at heat transfer while being smaller and quieter than traditional air coolers.

Dubbed the Sandia Cooler, the team has come up with an updated prototype that is nearly ready to come to market. Shown off in a recent video, the cooler is a small heatsink based on three relatively simple parts. A stationary disk acts as the base and area that comes into contact with the Integrated Heat Spreader (IHS) of a CPU. Then, a spinning array of curved fins resembling an impeller design is spun up by a small motor mounted in the center of the cooler.

View Full Size

During an industry day, they reportedly signed two license option agreements with two companies to bring the product to market in the areas of solid state lighting (LEDS, et al) and computer hardware cooling, implying that it is getting closer to a final product that it was last year.

Interestingly, the cooler uses an “hydrodynamic air bearing” such that the spinning part of the cooler is spun up to 2,000+ RPM such that the top part separates from the bottom stationary part and rides (they use the analogy of a car hydroplaning) on a very thin layer of air. (Update: as KngRider noted, there is still some friction from the motor spinning the upper part of the cooler, however.) That thin layer of air is what facilitates heat transference from the stationary part to the spinning fins. It does raise questions of efficiency, however. How a layer of air is more efficient than thermal interface material, for example. Reportedly, the air bearing is not an issue that will impact cooling performance but it is a difficult concept to grasp considering TIM and metal-to-metal contact has always been touted as the best cooling situation.

Sandia explains that cool air is drawn into the center of the impeller as heated air is forced outwards through the spinning fins, which reportedly enables efficient heat transfer. In the video, they demonstrate that it is capable of being extremely quiet (nearly silent) despite spinning at an extremely fast rate – the noise in the first part of the video is due to the prototype motor that is not covered. They claim that the final design will use a brush-less motor that will be much quieter.

It’s an intriguing design because of its simplicity and form factor. It is reportedly able to cool more efficiently than some of the best air coolers on the market, which use such techniques as heatpipes that come into direct contact with the CPU IHS, larger fin arrays, and multiple fans. Compared to those coolers, the Sandia prototype is much smaller and simpler in its construction.

The company has further released a white paper (PDF) and has an area of its website dedicated to more information on the Sandia cooler. While I cannot vet the fluid dynamics they detail, it certainly looks good on paper. I’m excited to see this come to market and whether or not it will live up to its promise of more efficient (and quiet!) cooling. It could be an important asset in cooling computer hardware in everything from desktops to server rooms. Also, it might just be the advancement that air coolers have been looking for as far as the next jump in performance – more than simply adding additional heatpipes or fins (and dealing with weight, size, and diminishing returns as a result) can do alone.

I’ll say that I’m skeptically optimistic on this one, but I do hope that it’s the real deal. What do you think of the impeller cooler? Does it appear promising?

Source: Sandia
June 26, 2012 | 08:29 AM - Posted by jcboy (not verified)

Very very interesting indeed. I had a similar idea to this but it involved a standard 120mm fan that conducted heat. If it can truly be quiet then this is a massive step in cooling.

June 26, 2012 | 09:04 AM - Posted by Billy Mathieu (not verified)

I wouldn't stick my finger in a computer with that as a heatsink. My water cooling does fine...

June 26, 2012 | 09:23 AM - Posted by Tim Verry

Yeah, ouch! You'd prolly need stitches if a finger came in contact with it!

June 26, 2012 | 10:07 AM - Posted by Chaitanya Shukla (not verified)

I remember Msi had shown off a chip-set cooler which was much more interesting few years back. I would start when the chip was hot enough to push hot gasses to propel the fan.

June 26, 2012 | 10:48 AM - Posted by Tim Verry

hmm, so it didn't need to run on electricity? That sounds neat, it would sorta be self-cooling :).

June 26, 2012 | 10:20 PM - Posted by Chaitanya Shukla (not verified)

yes, it was a very interesting concept which never reached the market. it was displayed by MSI at CeBIT of year 2008. and I found a link to the coverage page: http://www.tomshardware.com/news/CEBIT-MSI-PASSIVE-FAN,4985.html

June 26, 2012 | 11:08 AM - Posted by Anonymous Coward (not verified)

You guys remember this old penny arcade comic featuring Romero?
http://www.penny-arcade.com/comic/1998/11/25/

I'd like to say the same thing to these guys. Stop blowing hot air up my asshole and start blowing it out of my computer. No device, no weiner, as it were. In stock at newegg or STFU.

June 26, 2012 | 02:43 PM - Posted by JMays (not verified)

Nice...Its not like they are pulling this out of no where. They obviously have a functioning prototype and people interested in the technology. If they want to show off an innovative, unreleased product or concept, then that is great! It may encourage or inspire others to design new cooling methods. Not to mention it gives me something to read/watch between classes. By the way, thanks for putting this up, Tim!

June 26, 2012 | 09:14 PM - Posted by Anonymous Coward (not verified)

I know, but last year they showed off the same damn thing. I got all excited, and this year it's more talk. It's becoming a thing in the tech industry and I'm sick of it. OLEDs have been a year away for 8 years now, SSD were supposed to replace hard drives how long ago? How about non-volatile RAM? Quantum computers? 10gbit consumer Ethernet? New wifi standard? Native Usb3 without third party hacks? External pic-e? Memristors?

I could go on forever. My point stands. These people ALL need to STFU until they're ready to deliver. Shit or get off the pot.

June 28, 2012 | 07:15 PM - Posted by JMays (not verified)

New tech doesn't fall from the sky. I get excited from seeing what people are coming up with, whether it comes out in a year or five or never. If it doesn't come to fruition in a year you would just rather not hear about it? If this turns out to be a viable cooler then that's awesome, but I wouldn't be surprised if it took at least another 2 years before it can be purchased from Newegg. Watching new tech come to life is half the fun...but maybe that's just me.

June 26, 2012 | 03:18 PM - Posted by Arb1 (not verified)

Hrm issue that i see with being used for cpu cooling is they only have it mounted 1 way where as it would have to mounted vertically to work with a majority of cpu's the concept if its as good as they say could lead to much more efficient coolers.

June 26, 2012 | 08:29 PM - Posted by pdjblum

So I guess it would have to lie flat so to be able to hydroplane without separating from the base. I prefer test-benches and cases where the mobo lies flat, so this would not be an issue for me; most people, though, have traditional cases with vertically oriented mobos. I hope my understanding is wrong and that it can be used in any orientation.

June 26, 2012 | 10:29 PM - Posted by Tim Verry

I'm not entirely sure how, but iirc the creator has stated that it will work vertically, but at angles greater than 90-degrees it might lose efficiency.

June 26, 2012 | 11:49 PM - Posted by pdjblum

Thanks Tim for the clarification. Can't imagine how that is possible, but I will take the creator's word for it.

June 27, 2012 | 12:05 AM - Posted by Tim Verry

heh, yeah I'm not sure either but I imagine that it's something they've accounted for :)

June 27, 2012 | 07:37 PM - Posted by Kris (not verified)

If I were designing this for vertical usage I'd put the rotational assembly on a bearing and have a spring pushing it against the base. The spring tension would have to be adjusted to give about the same force as if it were laying flat (essentially just the force would equal its weight).

The only problem with this is that you'd have to adjust the tension depending on the angle it was used at, with upside down requiring twice the force as using it vertically.

June 27, 2012 | 09:09 AM - Posted by cyow

never work in a system.

it would drop off as soon as you put your case up the right

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.