I think the BIGGEST stoppingstone in this "cooler" is something called the Second Law of Murphodynamics: if it looks too good to be true, its probably a scam. Maybe. If you ever paid attention in physics class, you know that air has one lousy heat capacity. The top part may make sense, but how should heat travel through the "dynamic air suspension"? From the baseplate to the radiator fins? PLUS, there is something called "deadloop" in thermodynamics. Where heated up or "spent" air is reused in the coolingcycle. When hot air is expelled from the sides, its also sucked back in through the top. Talking about 30x more efficency...well, the rotary radiator might, but the isolation from the baseplate and very high chance of deadloops, the efficency of the overall cooler is pointless.
***** Yeah, maybe the 1/10th of a mm with some "cold plasma" like energy transfer (compressed air) could do the trick. But still, its too sci-fi for me. Even ordinary coolers need thermal paste to fill in microscopic cracks... And I stand by my deadloop thing. I see a high potential of air getting re-used :/
***** You can easly test this on your own computer: take off you cooler and install it without any thermal paste. Run some temp program. Install the cooler again with thermal paste... compare temps with and without paste. There is also a VERY VERY thin layer of air there. Ok, to give some thought to it: the "cold plasma" thing i mentioned. There could be a channel for heat-energy through high pressure gas. Maybe even motion can contribute at "grabbing" heat and transfering it. But still. I think such transfer is very inefficent :/
I agree, when I was watching this I thought to myself, how the hell is the heat from the chip getting to that heatsink, since it's basically floating above a base plate... Or are they planning on floating the cpu around too lmao xD
I'm just making an educated guess, but it looks like the challenge they couldn't overcome is that the thing has to be oriented "up". They gloss over the "air bearing" -- which is identified as the key to transferring heat from the base to the fins -- but the spinning bit /hovers/ over the base when it gets up to speed. That's because gravity holds it "down" ... which it can't do if you tip the heat sink on its side. As a consumer product, that's a real limitation. *Edit:* the first commenter seems to have found a statement released by Sandia, explaining that a compression spring is used to hold the impeller and base together (not gravity, like it appears in the video).
It works independent of orientation The most common misconception seems to be that the Sandia heatsink only works in a horizontal orientation. In the RUclips video released by Sandia the cooler is demonstrated like this, so it seems reasonable to assume the air bearing heatsink only operates on a horizontally plane. Yet as we all know, most computers have the motherboard vertically oriented... According to the developers, the Sandia cooler is in fact orientation agnostic - meaning the air bearing is not affected by gravity and the spinning aluminum fins will operate vertically, horizontally, upside-down or at an angle. A compression spring is used to hold the impeller and base together, not gravity.
Hey, alysdexia? Hard drives work sideways because they have **a solid shaft running up the middle**. This thing is floating on a cushion of air. Trying being a nicer person next time. teh_r4serei says Sandia released a statement somewhere else saying the device uses a compression spring to hold the rotor in place, and it seems reasonable enough to me that I'm not going to double-check his source.
Oh yea, another useless thing they added I guess, the brushless motor's PWM duty cycle and mosfet frequencies are modulated with varying music amplitudes/frequencies. I.E. Fan is singing
I think I have fascinated by this exact video many years ago, and here I am now to the present, I got fascinated yet by it again, and then I look at the date of this video.
For those wondering, Thermaltake made Engine 27, which is based on this design. SKU CL-P032-CA06SL-A They have it on Amazon for 75USD. It runs on PWM and uses 8 pole motor, so RPM feedback is skewed.
Its amazing how something that looks so simple is of course so highly complex to design. Once it is designed of course, and it operates simply and efficiently, its hard to imagine why no one thought of it before. This is really and elegant design. Form and function mesh wonderfully here. Well done!
Yeah the actual product was shown off in the CoolerMaster suite. Both Linus Tech and JayzTwoCents did videos on the cooler from the CoolerMaster suite. the final commercial cooler will be sold under the CM brand.
Applebloom Nope. Not at all. Coolchip basically just rented Cooler Master's suite on CES, to introduce it's technology, while Cooler Master made a contract with Coolchip to release the very first commercial version of the product utilizing this technology. But Coolchip doesn't belong to Cooler Master. They just have contract to release first Coolchip-designed cooler under the Cooler Master's name/brand. David Dewis Yes, they did rent Cooler Master's suite's space to showcase their product at CES, but they don't belong to Cooler Master. They just have a contract with Cooler Master to release the very first commercial version of this cooler under the Cooler Master's name and brand.
According to Sandia's report on this device, it can be run in any orientation (including both vertical and even upside-down). This is a pretty amazing idea.
30 times more efficient and yet its 2014 and I have not seen this any where... so it has failed. Its all good seeing this kinda videos but they always talk about positives but never about the negatives.
do you realize that it usually takes around 10 years for new technologies to reach the market? That said, they compare it to regular heatsinks, and not stuff like cpu tower coolers with heatpipes, so I doubt they'd prove beneficial in regular computers for quite a while. Application in the industry is probably already happening at a small scale.
Dmitri Felbinger that must be why we had heatpipes in computers in 1960.. OR, we didn't, which shows that in some cases it takes a MUCH longer time. especially when it is things that were made for industrial application and not consumer products. You can expect that this will be in some industrial use now(which would then be considered a corporate secret, and you wouldn't know about it), OR within a few years, when efficiency has been further improved. And you can expect it to take a LONG while before we see these in consumer products.
Dmitri Felbinger also in most consumer products you will probably never see this, since piezoelectric fans will probably be available before this, at least for laptops. There are already being researched on by Fujikura etc. which make a large amount of the cooling systems for laptops now. watch?v=Hm5fXj-hUpk#t=191
I did my NDI testing there at Sandia National Labs and there are some really good people there. I think they could have emphasized a little better how this is a little more of concept than FINAL design. It does these men and women working there a great disservice.
LimitedWard He obviously didn't look to far in the "literature" because Nikola Tesla describes the "boundary effect", adhesion etc. in his Fluid Dynamics and Tesla Turbine Patents, that was in 1909.
+enzawashballs Tesla talked about diminishing the boandary effect by using a rotating object? Or did he just describe the boundary effect in general? Also, do you know why so many uneducated people think Tesla is hot-shit? No mention of Maxwell, Newton, Euler, Helmholtz...Maybe because they are also conspiracy nuts? I mean, Tesla is definitely a smart guy, but he is a dwarf of a genius compared to most other famous scientists. It's very strange
ab cd Maybe because all modern scientists have to refer to Tesla patents, not Einstein's, not Edison's, not Maxwell's, not Newton's, nor Euler or Helmholtz. Thats unless you include Edison's rock crusher and electric chair, using GE's A/C patents. (The electric chair used by Edison was not Tesla's A/C patents, GE's A/C is different patents) And GE's initial A/C patents didn't work so good electrocuting people.
ISFiYIywAFIBc6qAIIIIIIIIIIIIIIIIQrXTJiCtY3Asd4WF did you watch the video? The heatsink/fan spins at 2000rpm. cpu/case fans spin at upto 3000rpm. most commonly around 2000rpm for high static pressure fans. And so what if my fans are plastic? I can buy a 212 evo which has a smooth finish on the plastic and it'll still be covered in dust after a few weeks or so.
It's in the design. The typical fan blade cuts through air, each time pushing a slice of air forward, perpendicular to the blade, this creates lots of friction and static electrostatic, trapping dust on the blades themselves. With this design, air is centrifugally flung out of through the gaps between the blades with higher velocity so even at 200 rpm, the exit velocity is much higher. But I do see two problems with this though. First being that the viscosity of the air doesn't change and fluid dynamics would suggest that a layer of turbulence will still form along the surface of the blades, trapping some dust along with static created from the friction. But because the velocity is much higher, larger dust particle should be pushed through more easily. Second, dust have to land somewhere, if not on the blades, then on the heatsink down the line. The way the fan is implemented (ie, by Cooler Master) would mean that dust fling off the fan will just get caught by the heatsink. Another interesting thing, however, is the switch in how air is pulled into the fan, from a planar nature to a turbine-ish system, this would make it easier to filter the air beforehand if needed.
No dust collection because dust at 2,000rpm gets "flung off"? I see someone hasn't ever been in IT and had to clean a computer before. If you think your measly 2,000rpm is going to stop dust accumulation you are going to be in for quite a surprise.
im lucky to try out server grade stuff certificated for 24/7, and thats cooler is spins at 7000rpm, louder than a vacuum cleaner, but able to blow the dust off from itself and the heatsink fans (the normal fans i tried are dusting and needs regular cleaning ~1500-2500rpm)
Either that, or matte plastic surfaces designed to constrict airflow along the axis of spin aren't comparable to smooth aluminium surfaces designed to encourage said airflow.
The difference probably isn't the RPM, but the fact that those gaps are so narrow and there aren't diagonal slopes. The air is forced through at presumably higher speeds and more uniformly (less weird aerodynamic effects happen at the blade surfaces and allow dust to collect).
Though this will take years to make it into my PC I thank you people for your hard work! The best of luck to you all, I cannot wait to own a version of this in the future.
When the fins start moving, they create a low pressure area on their 'back' side (the one pointing away from the direction of rotation) at the tips, which pulls air outward along the fins. That results in the pressure dropping in the center which pulls air down. Also, the fins are thicker near the center than they are at the edges which will help the air accelerate away from the middle too.
I'm not quite a "random youtube user". I've many years as an end user working with coolers and believe me, if I genuinely thought that this was going to make a difference I'd be cheerleading for it like the rest of you. the problem is that this video is extremely light on actual facts and manages to say a lot without actually saying anything. Again, I'd love for this to be real, but the whole thing triggers warning bells for me and it could just end up like the TMD fan.
With respect to orientation, they could easily add a spring to provide the counter force to the air bearing. The issue I'm having is that they totally gloss over the fact that they need to transfer heat across a nice insulative layer of air...
@krap101 The insulative air as you call it is reduced by the spinning effect of the central column aka a mini vortex or mini tornado. Cool air drawn in from the top hits the base expands and is expelled by the fins. Like stirring a tall glass of water. the faster you churn the water in one direction the deeper the vortex goes relative to the rate of initial turn.
I have the same question. The way Cooler Master uses this (in another video) is to have another heatink surrounds the fan.. same old design we see in stock gpu cooler.
I work as a aviation life support technician IRL, an aircraft I'm rated to maintain and repair has a duel RAM/engine-bleed air cooling turbine for its environmental systems. This turbine uses an air bearing instead of a classic oil cooled bearing. The air bearing turbine on this aircraft is mounted vertically, is rated to rapidly expand and cool air at temperatures up to 850F, and can can support itself above the max G rating/overstress condition of the aircraft (10-14G's). I think we're ok.
A simple heatpipe system will do fine for most uses. In my experience air bearings are great for stuff that stays put but if it is subject to shock... crash 😵
***** Well I hope you are wrong, because I wouldn't mind something like this. That being said a simple closed loop has so much performance per dollar, what really is the point right?
That's the reason they use an air bearing. It's the same concept your hard drive uses to float the heads over the platters- air is carried along the surface of the moving object and in doing so forms a thin film that provides both thermal conduction and mechanical cushioning. And I know air bearings work- a large machine that I work with uses a similar mechanism called air lift, enabling over 10 tons to be moved one-handed through a simple crank and screw arrangement to push it along the guide
@Hoshimi Draws because heat transfer coefficient of air is vastly less than that of metal, normal heat sinks have very large surface area and a thermal paste on the contact pad. even a thin layer of air at the contact pad significantly reduces the entire device's capacity to transfer heat, that's why traditional heat sinks practically stop working if you leave a gap in your thermal paste layer. in this proposal there's an air gap without thermal paste at all.
Graphics cards have what is called a squirrel cage blower. These are common in many things and is tech that was created well over 30 something years ago. This is something different, not a squirrel cage at all. A squirrel cage is just a blower that blows air through a conventional heatsink. This device in itself IS a heatsink, rotating itself to optimize cooling in a similar fashion to a centrifugal blower. It moves air like a centrifugal blower but displaces heat all by itself at the same time.
Did he ever say how the heat is transferred from the cpu to the cooler? If I am not mistaken, he said air bearing, and if memory serves me correct, air isn't the best conductor of heat.
Well, that's the great thing about it. It's so efficient that you can use a wide variety of metal materials. A mixture of Aluminum, Copper and (Silver, Albeit expensive, it has the best thermal conductivity at 418.0 W/mk, had to research that number :D )
I have one patent on a heat sink, however, I thought of this idea a while back and have a significant improvement on this. I combined a Tesla Turbine with an electric motor and wala. The world most efficient electric motor principle. It both cools and pumps air.
One of these made out of graphene, using a set of stanene-fluoride motor coils, would be really kickass. The fins would be extremely light, almost completely transparent, incredibly durable and would dissipate heat [almost] like nothing else, and the stanene-fluoride motor would be superconducting all of the way up to 100 Celsius, making it extremely powerful AND energy efficient. It sucks that graphene and stanene-fluoride have to be so expensive to build.
I could certainly hear it, even after the motor was turned off. It may be a good cooling solution for overclocking, but not so much for those who seek a really silent machine.
They pointed out in the video @3:18 that the noise was due to the motor running without a cover, and would not be present in the finished product. Also, this sound to me like it's being aimed more at industrial applications, meaning that the downtime of the motor would be minimal, which translates into little time being stationary and collecting dust.
GoogleMinus that, we do not know. The sound it does in the video seems typical to an early stage product without proper shielding. What I'm trying to say is that we dont know anything except from that they more or less promise it'll make less of a sound than regular fans. Modern quality regular fans hardly makes any sound a part from the "waterfall air noise" they speak about in the video so we can assume it has a pretty attractive performance to noise ratio. also, you have to take into account that due to it "hovering" it wont cause vibrations which causes A LOT more noise than most people think.
mtohrw The difference is with regular fans you can put rubber vibration isolators on the fan contact points because it doesn't need to conduct the heat...
It works just fine. At that thin of a gap the air acts as a great exchanger of heat. They said they're looking at .5 C/W difference between this and a normal air cooler.
3:10 the high pitched sound is just a prototype for the motor. They will be manufactured with better motors so it will only make the sound that it shows at 3:50
That's the main concern I've actually spotted so far. Modern top-of-the-line god tier air coolers like Noctua's NH-D14, Cryorig's R1 Ultimate, and other massive towers, have their fans oriented that way so they exhaust the hot air at the back of the case, where the actual exhaust is. This technology's design (and even Coolchip's improved model's), on the other hand, obviously just throws the hot air out everywhere, in all directions, which is an actually extremely bad thing if you have shitty air cooling chart in your PC case by default, and especially if you're putting something like this in a SSF factored systems like ITX or HTPCs. That's a massive problem. And how these guys (or hell - even Coolchip itself) are going to fix this problem, if they're able to anyway, which I doubt, is beyond me, until I'll see the goddamn thing working by my own eyes, in a consumer build.
Kawayolnyo Most non reference design GPUs also just eject heat everywhere, as long as you have good airflow through the case it shouldn't be a problem. I'm more concerned about dust getting stuck in the air bearing. (I know he talks about dust but he doesn't mention the probablility of dust getting stuck in the bearing).
Denmon >Most non-reference GPUs also just eject heat everywhere You're saying this in such of a tone like it's an actually alright and absolutely tolerable thing for hardware manufacturers to do. Which it isn't.
Kawayolnyo You're completely missing the point here though, as long as you have good airflow over the mobo area most heat should be exhausted anyway. It doesn't even take much, two fans in the front for intake and one in the back for echaust will move enough air that whatever heat gets blown around in there will be exhausted without much issue anyway.
Very Impressive Technology form SANDIA NATIONAL LABS. Congratulations! If I understand correctly, heat from the CPU or whatever, flows through the air bearing and heats the rotor. Then compression of the boundary layer against the cooling vanes happens by means of artificial gravity (i.e. rotating flow). Reduced thickness of the flow's boundary layer increases diffusive heat transport across the boundary layer thus cooling the spiral-shaped vanes. No wonder ANSYS-CFD is needed!
Dust gets stuck on both heatsink and fan, some on the underside of the blades and some on the cutting edge of the blade but also on and in the heatsink fins. It's why i usually keep a small toothbrush in my computer kit, so i can get between the blades and brush the dust from there as well and a can of air duster for the heatsink.
@QuantumBraced most modern fans are plastic. This is Metal. Less grip for dust particles and or static resistance may also inhibit the dust. If you run a metal desktop fan and a plastic equivalent and the same speed for the same length of time in the same environment you will find the plastic one gets clogged/coated easier / faster than the metal one.
Smoother surface. It's like a pearl forming in an oyster or a raindrop forming in a cloud. The more polished the metal is, the less chance particulates will have a chance to grip on to it.
Look at the workings of a firetruck pump. It works on the same principle. If you put a housing around it with a downward ventura in the center and make the discharge a little smaller it would draw in more air and exhaust out even more at a higher pressure. You could vent it to the out side of the case to disperse the heat. Build it like a squirrel cage fan design. The dynamics of a pump is to work with volumes of water or air and it builds it's on pressure pulling the heat away from the center.
Yep, it works similarly to a needle-vinyl player. The music is ineligible when it's on, due to high RPMs, everything gets really sped up. As it slows down, integrated needle slowly guides the top along vinyl-like engravings at the base, which can be customized to play any track you like.
It may look like an old intel pentium cooling fan but it has a lot more spinning inertia to it. It now has the same finger shreading potential of a wood router blade. A lexan shroud would be a good idea for this.
***** I'm planning to build a horizontal PC (the true desktop computer), either by modding an old case, or with something like a Node 202. Or maybe even build it in my broken Xbox 360's case. But I really need something like this, a low profile cooler with the highest TDP cooling.
I forgot to mention that I'll be building my next PC in 2017, when AMD launches Raven Ridge. So I'm hoping by then they would have launched their cooler and would have an AMD mounting solution. I don't intent to go that extreme with liquid cooling, I prefer a low profile heat sink and add small case fans for airflow. If it can run under 65°C, I'm happy (I don't live in a hot climate).
***** Yes, my current build is an unlocked and overclocked Sempron 145 (now dual core 3.2 GHz), 2 GB of RAM and an Asus 6670 GDDR5 edition. For my small gaming rig I'll use the best 65W APU with highest clock RAM and no dGPU (anyway, an APU's iGPU with HBM will be an improvement over my 6670). I like small and silent builds. I had a monster tower PC with FX 4100 (OC @4.2 GHz and turbo to 4.4 GHz), 8 GB of RAM and the same GPU, but sold it and bought a Windows tablet (Toshiba Encore WT8-A - which I'm typing from right now). It was pretty noisy because of case fans and it ran pretty cool (never exceeded 58°C). But I don't play hardcore games anymore, just light things like Asphalt 8. At first, I was pretty tempted to buy an Intel Skull Canyon NUC and use AMD X-Connect, but I like building PCs myself (I feel more accomplished). Anyway, thanks for your feedback. Much appreciation. :D
***** I thought I said I _had_ a FX (I sold it to buy my tablet and kept my old rig). I'm not really planning to upgrade the Sempron, as it has 2 GB of RAM and it's motherboard has a broken SATA port (for whatever reason, it won't recognize anything plugged in it). Also, being SATA II it won't help, since I want an SSD. Upgrading it might save me some money in the short run, but I plan to do it for the long run. Indeed, the Athlon II x2 have great unlock rates, but my motherboard was cheap to begin with, I'm not really trusting it anymore. Also, it's HDDs are dying as well. It has a 320 5400 RPM laptop HDD and it loads very hard, like 10 minutes to boot and use Windows 10 (it was the same with 8.1). When opening apps, it takes around 2 minutes to open things like Opera. After the apps have been opened, they run fairly well though. On the bright side, it's other 80 GB 5400 RPM laptop HDD runs Ubuntu and it's pretty fast except for boot times. After boot, it loads things almost instantly (and this one is at least 5 years older than the 320 GB one). On the bigger picture, I'm not living where my PC is anymore (I moved in another city), so I'll make sure that PC is running smoothly for my mom to use a small accounting software in Windows, while I build another rig (and I want to get a bigger resolution and dimension monitor. I used 17 inches 1280x1024 and now I want something around 22 inches or bigger 1080p or 1920x1200 (I like 16:10 aspect ratios). My plans are for the best 65W Raven Ridge APU (hopefully it will have HBM), 16 GB of DDR4 3200 MHz or higher (I'd love the juice of up to 4133 MHz RAM, but I'm not sure the motherboard will support it), a mini-ITX motherboard and a 256 / 512 GB SSD (if the motherboard will have M.2, I'll get something like a 64 GB for the OS and a 256 GB SATA SSD for storage), a low profile air cooler and 1 or 2 case coolers, depending on the space of the case (and some dust filters while I'm at it). No dGPU. Money aren't really a problem (of course I can't afford $3000 builds, but something around $800 is fair enough for me to spend).
***** Well, I kind of want a 40 in TV :D. I plan on 16 GB just for quad channel that I hope Zen will support (4 x 4 GB DDR4). And hopefully there will be mini-ITX motherboards with 4 DIMM slots (there are some on Intel's part). Also, quad channel is _almost_ double the bandwidth of dual channel DDR4, which will help an APU's iGPU a lot. I'm hoping it has HBM, but I too take it with grains of salt. HBM will drive the cost a lot on AMD's side, so I don't think they want to risk it. Quad channel DDR4 3200 is ~100 GB/s bandwidth, which is very close to what 1 stack of HBM gen 1 can achieve (128 GB/s). Paired with something like a 4 x86 cores and 12 gpu cores (for a total of 16 CU), without taking into consideration Zen architecture and Polaris improvements, I think 1080p high settings AAA gaming is achievable without anti-aliasing. Also, being sick of 2 GB of RAM only and seeing that games these days require 8 GB of RAM, I kind of want to future proof this build. I know it's overkill, but I might run some VMs on it (my main OS will be Linux), so 16 GB might be worth it (considering I won't be changing this build for ~7 years).
People have misinterpreted what this is all about. A later clarification by a Sandia spokesperson states "What we measured in lab was 30X as much heat transfer per unit fin area compared to a conventional heat sink ; this illustrates the boundary layer thinning effect of the sandia cooler. Somehow this result was misinterpreted as "30X more efficient" and went viral." Essentially, this means nothing of importance. You could take any conventional heatsink, which for comparison was probably a heatpipe sink since those have the most fin area to compare against, then get rid of most of the fins and run an fan at higher RPM (their fan runs at quite high RPM, easily over double the RPM of any CPU heatsink fan I use) to result in multiple times more heat transfer per unit fin area. It might not do AS well, but clearly our computers are running still. Essentially it's a shiny expensive toy. It DOES work, but the advantages in most situations are not all that great, and all else being equal, it will wear out the bearing faster so for all the expense you end up with a shorter lived product. Plus with a conventional heatsink if the fan eventually does wear out, you can usually replace it in half a minute with a standard $5 fan available at hundreds of merchants if you don't already have a spare fan in your parts bins. The last thing the PC industry and most industrial spaces need is a higher cost, shorter lived fan-sink substitute.
There is a Q&A they put out there that discusses it. It works better than thermal paste because of the laws of thermodynamics. "quantitatively, the numbers (gap distance, gap area, thermal conductivity of air, enhancement of conductivity by convection) work out quite well. For a 10 cm diameter device, an air gap resistance of 0.02 C/W is certainly feasible....the violent shearing of the air residing in the gap region between the stationary and rotating surfaces generates substantial convection"
Air is an insulator, sure, but its thermal resistance is directly proportional to how much air you have. The air bearing is only one thousandth of an inch and it's being stirred by the movement of the upper sink. Their prototype airgap had a measured thermal resistance of 0.02C/W; a tiny fraction of the thermal resistance of heatsinks on the market now.
Richard Smith don't really have ac in the uk but i get what you mean, they have these type of fans in many things like hoovers and water pumps just with different blade designs.
Patrik Manni I dont understand how exactly that works with a pocket of air between the processor and blower tho... that would make an excellent insulator, the opposite of a heat transference medium.
On the contrary, the speed at which it's spinning brings the size of the gap down tremendously, making for excellent heat interchange. This is a really cool solution to cooling, actually. But there are some big engineering hurdles to get past. Like how to make it work when it isn't standing flat on a table. And how to best use the blades to bring air further down the impeller without creating lift, like most blade solutions would.
@maddogfarg0, you are forgetting the fact that you can transfer heat between the gap between the heat sink impeller and the base plate. Remember there is air between that gap that serves as a thermal barrier between the two objects. As long as there is a thermal gradient between the air and the base plate and the air and the impeller, there will be heat transfer and thus cooling. Note that the thermal barrier in the Sandia design is smaller than conventional heat exchangers and more efficient.
cheif10thumbs I did. I'm not saying that the connecting part will be making the noise, I'm saying that the blades will cut the air and make a familiar airy sound.
jusk i think you are right because the video at a certain point was claiming it could potentially be 'silent' in it's final itereation, though overall in the video I think they were claiming it is significiantly quieter than traditional fans, while also being more efficient at cooling - not that it would be completely silent.The CoolChip cooler prototype that was shown at CES this year in this video: ruclips.net/video/ahgnVrULqUs/видео.html, which liscences the Sandia technology claims to be 20db quieter than a commercial fan at the same RPM while removing more heat. There's various reports online about this prototype which confirm that it does seem significantly quieter.Not really fair to nail the design by interpreting the claim of being quiter as claiming it will be silent, and that accumulating less dust as accumulating no dust. I'm excited about a consumer cooler that is 20db quieter, while being say 50% more efficient.
The idea is that first spin up is loud, then it coasts with moderate boosts from the motor. Moderate meaning enough to keep it spinning but not enough to be loud. So, when you boot a computer, for example, you'll hear a loud whine for a few seconds akin to normal fans, then it will be silent the rest of the time.
I still want more detail on how they couple the heat transfer from the stationary base to the spinning blade section. Is all of the heat transferred via the air bearing? If so, I wonder what other fluids with higher heat carrying capacity than air could be used?
No, you guys were not paying attention. They said that this was an unfinished cooler, with the motor uncovered. When he turns it off, that is what it would sound like in a finished, and insulated product.
I just don't understand how heat is transfered from the heat source to the spinning fins. It looks very inefficient to me as they're not touching each other.
considering most cases have a fan positioned to dissipate cpu heat I imagine a blower like design could easily direct the air out the back of the case the same way gpu blower fans work?
***** Correction: the air gap acts like a very, very bad thermal conductor. That's why we need thermal paste in conventional CPU coolers, where the gaps are in micrometers.
They explain it at 1:45 - they make that layer extremely thin (the same method as getting that tiny distance from a hard drive's disk to its read/write head).
Hλwk978 Yes.. actually. Been building custom PCs for 15 years.. been through a lot of cooling solutions from back in the day when pond pumps were used for water cooling. I am sensitive to high pitched noises and you can tell how loud it is in the video......
In WW-1 they need to cool an engine without using fluids and be very light weight. They designed a "Rotary" engine to cool it so they rotated the whole engine.
Look at the technology before you make nonsense comments. You should not waste words and energy of your speech by not understanding first the technology. My family has a background of technologists in machine development and construction machinery. We have replaced common bearing designs with custom made air-bearings for our machines because they are that much better and maintenance free. We didn't go through PhDs to understand this. We just saw something that works and implemented it. This technology works. The fact that its not gotten out yet means that the company has either received a lot of money to not produce the technology; if it does ever get into production, it'll be very inefficient and so people will not buy it. Hey if I was given $50+ million to shut up, I would. This would probably retail for $50 considering the manufacturing cost of less than $10. 1 million units could easily be sold, profiting of $40 million. When you take something from prototype to production, you actually make the product better and not the other way around. This is a simple design and the company claims to have had started production but with only 7x performance compared to the 30x stated, which is non-sense. There are so many companies selling CPU coolers. This would make them all obsolete. Understand the technology and then make your claims about what the company has gone through. The company doesn't matter. It's the technology. Understand it and use it. My elder brother is a mechanical engineer specialized in the production of automotive parts. The company he was hired by through a HR team required him and the team to make parts for vehicles. They came up with a design to last 15-20 years on commonly used motor parts with an increased cost, but not that high. However, the plans were refused and were told to use materials and designs that would last for 2 years as the company couldn't sustain the increased cost. Which is total bullshit. We live in a very different world now. Go figure.
They are taking advantage of the fact that the boundary layer gets thinner if the object is rotating, thus making for a better heat transfer. Think of it as an conventional heatsink that's split into two separate parts. The tricky thing is how to accomplish the heat flow from the stationary part of the heatsink to the rotary one. Since a metal on metal bearing would mean to much friction, their approach is to use an air bearing.
It about time for a new fan tech to come. The idea is fascinating by itself, however general public is already moving their computation loads towards portable devices where rotating fans are not an option. So in the end, there is a need for more efficient chips. Again, the combination of levitating fan an a heat exchanger is darn cool !!
I don't know anything about your other points, but I do know that it works in any orientation, due to magnetic attraction between the plate and the floating part.
i'll take two of these coolers for my graphics cards please. two 570's fans running at full speed cooling with virtually no noise? OH HELL YES. GIMME ONE. NOW.
For those bitching about the noise already, He clearly stated that is because the motor has no covers on it, with the covers the noise will be dulled. Even WITH the noise is is quieter than the crap I already have IN my case! 580 leaf blower anybody?
It's the prototype. Watch the rest of the video when they turn off the motor. The running noise is almost nothing, of course due to the dynamic air bearing.
I see the grand amount heat displacement in the design. But I bet your still in the prototype stages because your not pulling enough of the loss from the base. I would not worry about what is drawn from the base that fancy fan is going to keep things cool. :)
I can't see this working very well in a standard computer tower where the cooler is going to be sitting vertically. Now perhaps a GPU might benefit from it, since those are typically horizontal.
Heat IS MEANT to be transferred to the fins! The crucial difference compared to conventional coolers is that the fins are SPINNING. There's nothing "blown" on the heatsink but it's rotating itself to create air movement. It's about boundary layer. That's a very thin layer of fluid/air that sticks to the surface of an object, in this case the metal of the heat sink. It prevents the heat from being effectively transferred to the surrounding air.
it cant work sideways with the air bearing which is where its currently at. Which is why I heard they only have these advertised for servers that sit horizontal, Which normal PC's dont do so dont expect this anytime soon
+Kevin Moondust Stole it? Use your brain... There's this little thing call LICENCING. What makes more business sense: investing millions to setup a production line, QC, marketing, customer service, maintenance/retooling, etc. from scratch all for a single product -OR- simply licencing the technology to an already established manufacturer and saving all that time, money and risk?
Zero necker I stand corrected guys; after digging a bit deeper I found out Thermaltake DID license the tech from CoolChip, whereas CoolerMaster did NOT, so the culprit is another company. Whatever the case, the "final" product is the Thermaltake Engine 27 1U cooler, of which I expect a review ASAP because I'm deeply interested.
If it was fully detached from the base it would also stop spinning from being disconnected from the motor. Imagine the air that forms from the separation as more of a solid bearing and it will make more sense. If was disconnected it would also probably spin wildly out of control unless it was perfectly balanced, thus it can be assumed that the spinning motion keeps it elevated off the surface but its held down and in place by motor connection and centripetal force respectively.
They REALLY need to reduce the air bearing gap as the current air bearing gap or tolerance it way too high, as you can see in the video the cooler is unstable in its rotation due to the air gap, if this is reduced then heat exchange efficiency will also be increased and also reliability may be increased under certain circumstances
what I meant was that if you are pissed because of your loud GPU cooling, you can easily get rid of your stock cooling and find something better for less than £15 >> like accelero for example, it generates no more than 23 dB of noise. I want a gtx680 too, a water cooled and hopefully stuttering-proof for at least 3 years :)
I'm guessing it would work fine regardless of how the device is oriented, since all it is doing is spinning really fast to suck in air. You don't need to be in a special position to move air around--fans work no matter what direction they're facing. Since its designed to spin at such a high rate, i doubt there would be any problems from operating it vertically, or even upside down.
They say at 3:20 that it's because they don't have anything covering up the sound from the motor. I think they should be able to get the sound down to something like a hard drive motor's noise + the noise you hear at around 3:45 (which is almost nothing).
I think the BIGGEST stoppingstone in this "cooler" is something called the Second Law of Murphodynamics: if it looks too good to be true, its probably a scam. Maybe. If you ever paid attention in physics class, you know that air has one lousy heat capacity. The top part may make sense, but how should heat travel through the "dynamic air suspension"? From the baseplate to the radiator fins? PLUS, there is something called "deadloop" in thermodynamics. Where heated up or "spent" air is reused in the coolingcycle. When hot air is expelled from the sides, its also sucked back in through the top. Talking about 30x more efficency...well, the rotary radiator might, but the isolation from the baseplate and very high chance of deadloops, the efficency of the overall cooler is pointless.
*****
Yeah, maybe the 1/10th of a mm with some "cold plasma" like energy transfer (compressed air) could do the trick. But still, its too sci-fi for me. Even ordinary coolers need thermal paste to fill in microscopic cracks... And I stand by my deadloop thing. I see a high potential of air getting re-used :/
*****
You can easly test this on your own computer: take off you cooler and install it without any thermal paste. Run some temp program. Install the cooler again with thermal paste... compare temps with and without paste. There is also a VERY VERY thin layer of air there.
Ok, to give some thought to it: the "cold plasma" thing i mentioned. There could be a channel for heat-energy through high pressure gas. Maybe even motion can contribute at "grabbing" heat and transfering it. But still. I think such transfer is very inefficent :/
I agree, when I was watching this I thought to myself, how the hell is the heat from the chip getting to that heatsink, since it's basically floating above a base plate... Or are they planning on floating the cpu around too lmao xD
I'm just making an educated guess, but it looks like the challenge they couldn't overcome is that the thing has to be oriented "up". They gloss over the "air bearing" -- which is identified as the key to transferring heat from the base to the fins -- but the spinning bit /hovers/ over the base when it gets up to speed. That's because gravity holds it "down" ... which it can't do if you tip the heat sink on its side. As a consumer product, that's a real limitation.
*Edit:* the first commenter seems to have found a statement released by Sandia, explaining that a compression spring is used to hold the impeller and base together (not gravity, like it appears in the video).
It works independent of orientation
The most common misconception seems to be that the Sandia heatsink only works in a horizontal orientation. In the RUclips video released by Sandia the cooler is demonstrated like this, so it seems reasonable to assume the air bearing heatsink only operates on a horizontally plane. Yet as we all know, most computers have the motherboard vertically oriented... According to the developers, the Sandia cooler is in fact orientation agnostic - meaning the air bearing is not affected by gravity and the spinning aluminum fins will operate vertically, horizontally, upside-down or at an angle. A compression spring is used to hold the impeller and base together, not gravity.
even if orientation were the problem, it wouldn't be too difficult to setup an appropriate mounting sollution
magnets
those might actually create to much distance...
Hey, alysdexia? Hard drives work sideways because they have **a solid shaft running up the middle**. This thing is floating on a cushion of air. Trying being a nicer person next time.
teh_r4serei says Sandia released a statement somewhere else saying the device uses a compression spring to hold the rotor in place, and it seems reasonable enough to me that I'm not going to double-check his source.
3:50 "Listen to how quiet it is..." Why would music be playing during a "quiet" test?
its actually a speaker
Hhahahaahhaah
Oh yea, another useless thing they added I guess, the brushless motor's PWM duty cycle and mosfet frequencies are modulated with varying music amplitudes/frequencies. I.E. Fan is singing
I think I have fascinated by this exact video many years ago, and here I am now to the present, I got fascinated yet by it again, and then I look at the date of this video.
So what happened after 4 years? It was so good so the fan flew away to outer space on its own?
That's exactly what happened, then it grew sentient and decided to go where boldly no fan has gone before.
i....is that an Amazo reference?! o.o
Coller Master use that. I forget the name of it but you can find it.
Cooler Master has a similar technology that it's working on nowadays.
Thermaltake Engine 27 1U
For those wondering, Thermaltake made Engine 27, which is based on this design. SKU CL-P032-CA06SL-A They have it on Amazon for 75USD. It runs on PWM and uses 8 pole motor, so RPM feedback is skewed.
Good reviews and very small
Its amazing how something that looks so simple is of course so highly complex to design. Once it is designed of course, and it operates simply and efficiently, its hard to imagine why no one thought of it before. This is really and elegant design. Form and function mesh wonderfully here. Well done!
Complexity itself is complex. Not being math able I cannot have a sufficient grasp on it. Neither can the teleprompter and cue cards team makers.
Coolermaster have just shown off a cooler based on this at CES 2015
Cooler Master didn't show jack so far. The first commercial attempt will be in mid/late 2015. It was Coolchip themselves.
Kawayolnyo
Theryre apparently a division of CM.
Yeah the actual product was shown off in the CoolerMaster suite. Both Linus Tech and JayzTwoCents did videos on the cooler from the CoolerMaster suite. the final commercial cooler will be sold under the CM brand.
Applebloom
Nope. Not at all. Coolchip basically just rented Cooler Master's suite on CES, to introduce it's technology, while Cooler Master made a contract with Coolchip to release the very first commercial version of the product utilizing this technology. But Coolchip doesn't belong to Cooler Master. They just have contract to release first Coolchip-designed cooler under the Cooler Master's name/brand.
David Dewis
Yes, they did rent Cooler Master's suite's space to showcase their product at CES, but they don't belong to Cooler Master. They just have a contract with Cooler Master to release the very first commercial version of this cooler under the Cooler Master's name and brand.
Kvarnholmen. The fan blades may look the same, but the underlying technology is completely different I believe.
I came here to find a way to cool my watermelons
I recommend water cooling a watermelon
Zajebex i came here to find something to cut my watermelons. and found it.
According to Sandia's report on this device, it can be run in any orientation (including both vertical and even upside-down). This is a pretty amazing idea.
30 times more efficient and yet its 2014 and I have not seen this any where... so it has failed. Its all good seeing this kinda videos but they always talk about positives but never about the negatives.
I know 1one negative, if you toch the coooler with your finger you will suffer a lot of pain
do you realize that it usually takes around 10 years for new technologies to reach the market? That said, they compare it to regular heatsinks, and not stuff like cpu tower coolers with heatpipes, so I doubt they'd prove beneficial in regular computers for quite a while. Application in the industry is probably already happening at a small scale.
10 years? were have you been living? As far as I know most great inventions blew up on to the market 2-3 years after first prototypes.
Dmitri Felbinger that must be why we had heatpipes in computers in 1960.. OR, we didn't, which shows that in some cases it takes a MUCH longer time. especially when it is things that were made for industrial application and not consumer products. You can expect that this will be in some industrial use now(which would then be considered a corporate secret, and you wouldn't know about it), OR within a few years, when efficiency has been further improved. And you can expect it to take a LONG while before we see these in consumer products.
Dmitri Felbinger also in most consumer products you will probably never see this, since piezoelectric fans will probably be available before this, at least for laptops. There are already being researched on by Fujikura etc. which make a large amount of the cooling systems for laptops now. watch?v=Hm5fXj-hUpk#t=191
Guys, that's an early prototype. I suspect newer versions are much quieter.
I did my NDI testing there at Sandia National Labs and there are some really good people there. I think they could have emphasized a little better how this is a little more of concept than FINAL design. It does these men and women working there a great disservice.
LimitedWard He obviously didn't look to far in the "literature" because Nikola Tesla describes the "boundary effect", adhesion etc. in his Fluid Dynamics and Tesla Turbine Patents, that was in 1909.
LimitedWard 2015 they already did one for inside a computer... it work pretty well and quiet
+enzawashballs Tesla talked about diminishing the boandary effect by using a rotating object? Or did he just describe the boundary effect in general? Also, do you know why so many uneducated people think Tesla is hot-shit? No mention of Maxwell, Newton, Euler, Helmholtz...Maybe because they are also conspiracy nuts? I mean, Tesla is definitely a smart guy, but he is a dwarf of a genius compared to most other famous scientists. It's very strange
ab cd
Maybe because all modern scientists have to refer to Tesla patents, not Einstein's, not Edison's, not Maxwell's, not Newton's, nor Euler or Helmholtz. Thats unless you include Edison's rock crusher and electric chair, using GE's A/C patents. (The electric chair used by Edison was not Tesla's A/C patents, GE's A/C is different patents) And GE's initial A/C patents didn't work so good electrocuting people.
"dust will fling off" yeah the same way my fans get buried in dust...
exactly what I thought
Your fans are plastic and spin much slower.
ISFiYIywAFIBc6qAIIIIIIIIIIIIIIIIQrXTJiCtY3Asd4WF did you watch the video? The heatsink/fan spins at 2000rpm. cpu/case fans spin at upto 3000rpm. most commonly around 2000rpm for high static pressure fans. And so what if my fans are plastic? I can buy a 212 evo which has a smooth finish on the plastic and it'll still be covered in dust after a few weeks or so.
It's in the design.
The typical fan blade cuts through air, each time pushing a slice of air forward, perpendicular to the blade, this creates lots of friction and static electrostatic, trapping dust on the blades themselves.
With this design, air is centrifugally flung out of through the gaps between the blades with higher velocity so even at 200 rpm, the exit velocity is much higher.
But I do see two problems with this though. First being that the viscosity of the air doesn't change and fluid dynamics would suggest that a layer of turbulence will still form along the surface of the blades, trapping some dust along with static created from the friction. But because the velocity is much higher, larger dust particle should be pushed through more easily.
Second, dust have to land somewhere, if not on the blades, then on the heatsink down the line. The way the fan is implemented (ie, by Cooler Master) would mean that dust fling off the fan will just get caught by the heatsink.
Another interesting thing, however, is the switch in how air is pulled into the fan, from a planar nature to a turbine-ish system, this would make it easier to filter the air beforehand if needed.
Get a proper case with good dust filters. I've got no problems with dust in my Fractal Design Define R3.
No dust collection because dust at 2,000rpm gets "flung off"? I see someone hasn't ever been in IT and had to clean a computer before. If you think your measly 2,000rpm is going to stop dust accumulation you are going to be in for quite a surprise.
I have tons of dust on my fans. Just cleaned some off today.
im lucky to try out server grade stuff certificated for 24/7, and thats cooler is spins at 7000rpm, louder than a vacuum cleaner, but able to blow the dust off from itself and the heatsink fans (the normal fans i tried are dusting and needs regular cleaning ~1500-2500rpm)
Either that, or matte plastic surfaces designed to constrict airflow along the axis of spin aren't comparable to smooth aluminium surfaces designed to encourage said airflow.
The difference probably isn't the RPM, but the fact that those gaps are so narrow and there aren't diagonal slopes. The air is forced through at presumably higher speeds and more uniformly (less weird aerodynamic effects happen at the blade surfaces and allow dust to collect).
I we seen 8000 rpm fans with dust on their blades
2000 rpm stands no chance
so he shuts off the motor so we can hear how it sounds yet you keep the music playing
Though this will take years to make it into my PC I thank you people for your hard work! The best of luck to you all, I cannot wait to own a version of this in the future.
if that ever came loose while on, imagine all the damage inside your pc lol
Can't wait to see benchmark results in the CPU cooling area
When the fins start moving, they create a low pressure area on their 'back' side (the one pointing away from the direction of rotation) at the tips, which pulls air outward along the fins. That results in the pressure dropping in the center which pulls air down. Also, the fins are thicker near the center than they are at the edges which will help the air accelerate away from the middle too.
I'm not quite a "random youtube user". I've many years as an end user working with coolers and believe me, if I genuinely thought that this was going to make a difference I'd be cheerleading for it like the rest of you. the problem is that this video is extremely light on actual facts and manages to say a lot without actually saying anything.
Again, I'd love for this to be real, but the whole thing triggers warning bells for me and it could just end up like the TMD fan.
With respect to orientation, they could easily add a spring to provide the counter force to the air bearing. The issue I'm having is that they totally gloss over the fact that they need to transfer heat across a nice insulative layer of air...
@krap101 The insulative air as you call it is reduced by the spinning effect of the central column aka a mini vortex or mini tornado.
Cool air drawn in from the top hits the base expands and is expelled by the fins.
Like stirring a tall glass of water. the faster you churn the water in one direction the deeper the vortex goes relative to the rate of initial turn.
There's no hole at the bottom. Also look up what a boundary layer is.
krap101 is talking about from the stationary eg. CPU mounted part to the fin.
I have the same question. The way Cooler Master uses this (in another video) is to have another heatink surrounds the fan.. same old design we see in stock gpu cooler.
I work as a aviation life support technician IRL, an aircraft I'm rated to maintain and repair has a duel RAM/engine-bleed air cooling turbine for its environmental systems. This turbine uses an air bearing instead of a classic oil cooled bearing. The air bearing turbine on this aircraft is mounted vertically, is rated to rapidly expand and cool air at temperatures up to 850F, and can can support itself above the max G rating/overstress condition of the aircraft (10-14G's). I think we're ok.
might it be dual too? "duel RAM/engine-bleed air cooling turbine " ..... chose your weapon !! LOL
Actually, there are copper alloys that outperform silver significantly, and are just a bit cheaper. In fact, quite a few materials outperform silver.
4 years already ? this should be in store everywhere now !
A simple heatpipe system will do fine for most uses. In my experience air bearings are great for stuff that stays put but if it is subject to shock... crash 😵
This was 9 years ago dude
You are seeing next level cpu gpu cooling technology, in next years we will all going to use that kind of coolers.. thanks for the brilliant idea
who is watching in 2016 wondering where the hell this ever went?
+No Lucks Given , cooler master redoing it and probably making mad bucks on it
denisgsv well hopefully normal people like me can buy them wothout spending too much
+No Lucks Given its all about conspiracy, some kind like Gas Fueled engine VS Electrical Engine
***** Well I hope you are wrong, because I wouldn't mind something like this. That being said a simple closed loop has so much performance per dollar, what really is the point right?
www.google.com/#q=cooler+master+sandia+cooler&newwindow=1&safe=strict&tbm=isch
That's the reason they use an air bearing. It's the same concept your hard drive uses to float the heads over the platters- air is carried along the surface of the moving object and in doing so forms a thin film that provides both thermal conduction and mechanical cushioning.
And I know air bearings work- a large machine that I work with uses a similar mechanism called air lift, enabling over 10 tons to be moved one-handed through a simple crank and screw arrangement to push it along the guide
How does the air bearing achieve heat transfer?
exactly, this is the biggest question. it's like putting a heatsink without thermal paste.
@Hoshimi Draws because heat transfer coefficient of air is vastly less than that of metal, normal heat sinks have very large surface area and a thermal paste on the contact pad. even a thin layer of air at the contact pad significantly reduces the entire device's capacity to transfer heat, that's why traditional heat sinks practically stop working if you leave a gap in your thermal paste layer. in this proposal there's an air gap without thermal paste at all.
Graphics cards have what is called a squirrel cage blower. These are common in many things and is tech that was created well over 30 something years ago. This is something different, not a squirrel cage at all. A squirrel cage is just a blower that blows air through a conventional heatsink. This device in itself IS a heatsink, rotating itself to optimize cooling in a similar fashion to a centrifugal blower. It moves air like a centrifugal blower but displaces heat all by itself at the same time.
Did he ever say how the heat is transferred from the cpu to the cooler? If I am not mistaken, he said air bearing, and if memory serves me correct, air isn't the best conductor of heat.
that puts a whole new meaning on a 'cable management' from a PC perspective.
Could just imagine this thing jumping out my pc and rolling down the stairs, out of the front door onto the road and never coming back......
Well, that's the great thing about it. It's so efficient that you can use a wide variety of metal materials. A mixture of Aluminum, Copper and (Silver, Albeit expensive, it has the best thermal conductivity at 418.0 W/mk, had to research that number :D )
So where does the thermal paste go cuz I have no idea how that Spinning Disk is supposed to pull heat from the CPU
I have one patent on a heat sink, however, I thought of this idea a while back and have a significant improvement on this. I combined a Tesla Turbine with an electric motor and wala. The world most efficient electric motor principle. It both cools and pumps air.
What about two working in opposite ways to maybe cancel vibration and some noise
One of these made out of graphene, using a set of stanene-fluoride motor coils, would be really kickass. The fins would be extremely light, almost completely transparent, incredibly durable and would dissipate heat [almost] like nothing else, and the stanene-fluoride motor would be superconducting all of the way up to 100 Celsius, making it extremely powerful AND energy efficient.
It sucks that graphene and stanene-fluoride have to be so expensive to build.
Amaroqdricaldari We have the technology.
Would be insane cooling for small form factor. Maybe around $150 for like 200w tdp.
I could certainly hear it, even after the motor was turned off. It may be a good cooling solution for overclocking, but not so much for those who seek a really silent machine.
What about a vertical orientation? Is is possible or feasible?
shhh
They pointed out in the video @3:18 that the noise was due to the motor running without a cover, and would not be present in the finished product.
Also, this sound to me like it's being aimed more at industrial applications, meaning that the downtime of the motor would be minimal, which translates into little time being stationary and collecting dust.
This may be quieter than a fan, but that sound is awful.
+GoogleMinus listen to what they say in the video; "nearly noiseless with a ducting cover", its a prototype!
mtohrw Nearly noiseless, so it will be very quiet but still the same sound.
GoogleMinus that, we do not know. The sound it does in the video seems typical to an early stage product without proper shielding. What I'm trying to say is that we dont know anything except from that they more or less promise it'll make less of a sound than regular fans.
Modern quality regular fans hardly makes any sound a part from the "waterfall air noise" they speak about in the video so we can assume it has a pretty attractive performance to noise ratio.
also, you have to take into account that due to it "hovering" it wont cause vibrations which causes A LOT more noise than most people think.
mtohrw The difference is with regular fans you can put rubber vibration isolators on the fan contact points because it doesn't need to conduct the heat...
It works just fine. At that thin of a gap the air acts as a great exchanger of heat. They said they're looking at .5 C/W difference between this and a normal air cooler.
It's 2015 and I just heard about this thing. If there is an airgap, how does it transfer the heat to the fins? It doesn't. Air is a good insulator.
It can be mounted in any orientation and will be enclosed in a housing so it cant go flying off inside your case.
Air bearing in a dusty space to transfer heat?
3:10 the high pitched sound is just a prototype for the motor. They will be manufactured with better motors so it will only make the sound that it shows at 3:50
does it work on it's side? like it should in a pc?
That's the main concern I've actually spotted so far. Modern top-of-the-line god tier air coolers like Noctua's NH-D14, Cryorig's R1 Ultimate, and other massive towers, have their fans oriented that way so they exhaust the hot air at the back of the case, where the actual exhaust is. This technology's design (and even Coolchip's improved model's), on the other hand, obviously just throws the hot air out everywhere, in all directions, which is an actually extremely bad thing if you have shitty air cooling chart in your PC case by default, and especially if you're putting something like this in a SSF factored systems like ITX or HTPCs. That's a massive problem. And how these guys (or hell - even Coolchip itself) are going to fix this problem, if they're able to anyway, which I doubt, is beyond me, until I'll see the goddamn thing working by my own eyes, in a consumer build.
Good question, since it's powererd by heat convection i think a diferent design based on the same principal would be needed to work on it's side.
Kawayolnyo
Most non reference design GPUs also just eject heat everywhere, as long as you have good airflow through the case it shouldn't be a problem. I'm more concerned about dust getting stuck in the air bearing. (I know he talks about dust but he doesn't mention the probablility of dust getting stuck in the bearing).
Denmon
>Most non-reference GPUs also just eject heat everywhere
You're saying this in such of a tone like it's an actually alright and absolutely tolerable thing for hardware manufacturers to do. Which it isn't.
Kawayolnyo
You're completely missing the point here though, as long as you have good airflow over the mobo area most heat should be exhausted anyway. It doesn't even take much, two fans in the front for intake and one in the back for echaust will move enough air that whatever heat gets blown around in there will be exhausted without much issue anyway.
Very Impressive Technology form SANDIA NATIONAL LABS. Congratulations! If I understand correctly, heat from the CPU or whatever, flows through the air bearing and heats the rotor. Then compression of the boundary layer against the cooling vanes happens by means of artificial gravity (i.e. rotating flow). Reduced thickness of the flow's boundary layer increases diffusive heat transport across the boundary layer thus cooling the spiral-shaped vanes. No wonder ANSYS-CFD is needed!
Fans spin at 2000 RPM too and collect dust just fine...
Dust gets stuck on both heatsink and fan, some on the underside of the blades and some on the cutting edge of the blade but also on and in the heatsink fins.
It's why i usually keep a small toothbrush in my computer kit, so i can get between the blades and brush the dust from there as well and a can of air duster for the heatsink.
@QuantumBraced most modern fans are plastic. This is Metal.
Less grip for dust particles and or static resistance may also inhibit the dust.
If you run a metal desktop fan and a plastic equivalent and the same speed for the same length of time in the same environment you will find the plastic one gets clogged/coated easier / faster than the metal one.
Smoother surface. It's like a pearl forming in an oyster or a raindrop forming in a cloud. The more polished the metal is, the less chance particulates will have a chance to grip on to it.
Yes, that's why they call it the dust devil.
Always. Just ask the wifey
Look at the workings of a firetruck pump. It works on the same principle. If you put a housing around it with a downward ventura in the center and make the discharge a little smaller it would draw in more air and exhaust out even more at a higher pressure. You could vent it to the out side of the case to disperse the heat. Build it like a squirrel cage fan design. The dynamics of a pump is to work with volumes of water or air and it builds it's on pressure pulling the heat away from the center.
Yep, it works similarly to a needle-vinyl player. The music is ineligible when it's on, due to high RPMs, everything gets really sped up. As it slows down, integrated needle slowly guides the top along vinyl-like engravings at the base, which can be customized to play any track you like.
It may look like an old intel pentium cooling fan but it has a lot more spinning inertia to it. It now has the same finger shreading potential of a wood router blade. A lexan shroud would be a good idea for this.
Where is CoolChip? What happened with the cooler that should have been launched already?
***** I'm planning to build a horizontal PC (the true desktop computer), either by modding an old case, or with something like a Node 202. Or maybe even build it in my broken Xbox 360's case. But I really need something like this, a low profile cooler with the highest TDP cooling.
I forgot to mention that I'll be building my next PC in 2017, when AMD launches Raven Ridge. So I'm hoping by then they would have launched their cooler and would have an AMD mounting solution.
I don't intent to go that extreme with liquid cooling, I prefer a low profile heat sink and add small case fans for airflow. If it can run under 65°C, I'm happy (I don't live in a hot climate).
***** Yes, my current build is an unlocked and overclocked Sempron 145 (now dual core 3.2 GHz), 2 GB of RAM and an Asus 6670 GDDR5 edition.
For my small gaming rig I'll use the best 65W APU with highest clock RAM and no dGPU (anyway, an APU's iGPU with HBM will be an improvement over my 6670).
I like small and silent builds. I had a monster tower PC with FX 4100 (OC @4.2 GHz and turbo to 4.4 GHz), 8 GB of RAM and the same GPU, but sold it and bought a Windows tablet (Toshiba Encore WT8-A - which I'm typing from right now). It was pretty noisy because of case fans and it ran pretty cool (never exceeded 58°C). But I don't play hardcore games anymore, just light things like Asphalt 8.
At first, I was pretty tempted to buy an Intel Skull Canyon NUC and use AMD X-Connect, but I like building PCs myself (I feel more accomplished).
Anyway, thanks for your feedback. Much appreciation. :D
***** I thought I said I _had_ a FX (I sold it to buy my tablet and kept my old rig). I'm not really planning to upgrade the Sempron, as it has 2 GB of RAM and it's motherboard has a broken SATA port (for whatever reason, it won't recognize anything plugged in it). Also, being SATA II it won't help, since I want an SSD. Upgrading it might save me some money in the short run, but I plan to do it for the long run. Indeed, the Athlon II x2 have great unlock rates, but my motherboard was cheap to begin with, I'm not really trusting it anymore.
Also, it's HDDs are dying as well. It has a 320 5400 RPM laptop HDD and it loads very hard, like 10 minutes to boot and use Windows 10 (it was the same with 8.1). When opening apps, it takes around 2 minutes to open things like Opera. After the apps have been opened, they run fairly well though. On the bright side, it's other 80 GB 5400 RPM laptop HDD runs Ubuntu and it's pretty fast except for boot times. After boot, it loads things almost instantly (and this one is at least 5 years older than the 320 GB one).
On the bigger picture, I'm not living where my PC is anymore (I moved in another city), so I'll make sure that PC is running smoothly for my mom to use a small accounting software in Windows, while I build another rig (and I want to get a bigger resolution and dimension monitor. I used 17 inches 1280x1024 and now I want something around 22 inches or bigger 1080p or 1920x1200 (I like 16:10 aspect ratios).
My plans are for the best 65W Raven Ridge APU (hopefully it will have HBM), 16 GB of DDR4 3200 MHz or higher (I'd love the juice of up to 4133 MHz RAM, but I'm not sure the motherboard will support it), a mini-ITX motherboard and a 256 / 512 GB SSD (if the motherboard will have M.2, I'll get something like a 64 GB for the OS and a 256 GB SATA SSD for storage), a low profile air cooler and 1 or 2 case coolers, depending on the space of the case (and some dust filters while I'm at it). No dGPU.
Money aren't really a problem (of course I can't afford $3000 builds, but something around $800 is fair enough for me to spend).
***** Well, I kind of want a 40 in TV :D. I plan on 16 GB just for quad channel that I hope Zen will support (4 x 4 GB DDR4). And hopefully there will be mini-ITX motherboards with 4 DIMM slots (there are some on Intel's part). Also, quad channel is _almost_ double the bandwidth of dual channel DDR4, which will help an APU's iGPU a lot. I'm hoping it has HBM, but I too take it with grains of salt. HBM will drive the cost a lot on AMD's side, so I don't think they want to risk it.
Quad channel DDR4 3200 is ~100 GB/s bandwidth, which is very close to what 1 stack of HBM gen 1 can achieve (128 GB/s). Paired with something like a 4 x86 cores and 12 gpu cores (for a total of 16 CU), without taking into consideration Zen architecture and Polaris improvements, I think 1080p high settings AAA gaming is achievable without anti-aliasing.
Also, being sick of 2 GB of RAM only and seeing that games these days require 8 GB of RAM, I kind of want to future proof this build. I know it's overkill, but I might run some VMs on it (my main OS will be Linux), so 16 GB might be worth it (considering I won't be changing this build for ~7 years).
People have misinterpreted what this is all about. A later clarification by a Sandia spokesperson states "What we measured in lab was 30X as much heat transfer per unit fin area compared to a conventional heat sink ; this illustrates the
boundary layer thinning effect of the sandia cooler. Somehow this result was misinterpreted as "30X more efficient" and went viral."
Essentially, this means nothing of importance. You could take any conventional heatsink, which for comparison was probably a heatpipe sink since those have the most fin area to compare against, then get rid of most of the fins and run an fan at higher RPM (their fan runs at quite high RPM, easily over double the RPM of any CPU heatsink fan I use) to result in multiple times more heat transfer per unit fin area. It might not do AS well, but clearly our computers are running still.
Essentially it's a shiny expensive toy. It DOES work, but the advantages in most situations are not all that great, and all else being equal, it will wear out the bearing faster so for all the expense you end up with a shorter lived product. Plus with a conventional heatsink if the fan eventually does wear out, you can usually replace it in half a minute with a standard $5 fan available at hundreds of merchants if you don't already have a spare fan in your parts bins.
The last thing the PC industry and most industrial spaces need is a higher cost, shorter lived fan-sink substitute.
They probably got confused because the video description literally says "30x more efficient"
How does the heat transfer from the top of a hot object like CPU? Which part comes with contact
the stationary base is mounted to the CPU, and the thin layer of air transfers the heat to the spinning heatsink.
The air bearing is just as efficient a thermal insulator as the boundary layer on a conventional heat sink.
That I do not understand. How can air conduct the heat efficiently to the fins yet we need massive amounts of fins to remove that heat from them.
Thank you, I was looking for this answer, will look it up some more.
There is a Q&A they put out there that discusses it. It works better than thermal paste because of the laws of thermodynamics. "quantitatively, the numbers (gap distance, gap area, thermal conductivity of air, enhancement of conductivity by convection) work out quite well. For a 10 cm diameter device, an air gap resistance of 0.02 C/W is certainly feasible....the violent shearing of the air residing in the gap region between the stationary and rotating surfaces generates substantial convection"
WHEN WILL THEY BE RELEASED!?
Any day now lol
Air is an insulator, sure, but its thermal resistance is directly proportional to how much air you have. The air bearing is only one thousandth of an inch and it's being stirred by the movement of the upper sink. Their prototype airgap had a measured thermal resistance of 0.02C/W; a tiny fraction of the thermal resistance of heatsinks on the market now.
Just looks like a centrifugal fan with many blades to me
Its a blower with extra surface area xD You have one in your ac units. There is no innovation here xD
Richard Smith don't really have ac in the uk but i get what you mean, they have these type of fans in many things like hoovers and water pumps just with different blade designs.
Richard Smith The innovation is using the surface area of the blades to dissipate heat. Instead of using a blower and a heatsink.
Patrik Manni I dont understand how exactly that works with a pocket of air between the processor and blower tho... that would make an excellent insulator, the opposite of a heat transference medium.
On the contrary, the speed at which it's spinning brings the size of the gap down tremendously, making for excellent heat interchange. This is a really cool solution to cooling, actually. But there are some big engineering hurdles to get past. Like how to make it work when it isn't standing flat on a table. And how to best use the blades to bring air further down the impeller without creating lift, like most blade solutions would.
so, 10 years later... where can i see this in a commercial product? inquiring minds want to know...
So still nothing after 8 years?
@maddogfarg0, you are forgetting the fact that you can transfer heat between the gap between the heat sink impeller and the base plate. Remember there is air between that gap that serves as a thermal barrier between the two objects. As long as there is a thermal gradient between the air and the base plate and the air and the impeller, there will be heat transfer and thus cooling. Note that the thermal barrier in the Sandia design is smaller than conventional heat exchangers and more efficient.
It's not silent and it will collect dust just like pc fans do.
it's silent as my ass.
***** And you may pretend that you don't fart, but you do.
Apparently you didn't catch the part about boundary layer effects.
Maybe you didn't listen at all?
cheif10thumbs I did. I'm not saying that the connecting part will be making the noise, I'm saying that the blades will cut the air and make a familiar airy sound.
jusk i think you are right because the video at a certain point was claiming it could potentially be 'silent' in it's final itereation, though overall in the video I think they were claiming it is significiantly quieter than traditional fans, while also being more efficient at cooling - not that it would be completely silent.The CoolChip cooler prototype that was shown at CES this year in this video: ruclips.net/video/ahgnVrULqUs/видео.html, which liscences the Sandia technology claims to be 20db quieter than a commercial fan at the same RPM while removing more heat. There's various reports online about this prototype which confirm that it does seem significantly quieter.Not really fair to nail the design by interpreting the claim of being quiter as claiming it will be silent, and that accumulating less dust as accumulating no dust. I'm excited about a consumer cooler that is 20db quieter, while being say 50% more efficient.
The idea is that first spin up is loud, then it coasts with moderate boosts from the motor. Moderate meaning enough to keep it spinning but not enough to be loud.
So, when you boot a computer, for example, you'll hear a loud whine for a few seconds akin to normal fans, then it will be silent the rest of the time.
at 2:49 - 2:51 any one elss seeing video crouption
kyle krone lol i thought it was me
+BlackWing Cyper I thought these were ads xd
I still want more detail on how they couple the heat transfer from the stationary base to the spinning blade section. Is all of the heat transferred via the air bearing? If so, I wonder what other fluids with higher heat carrying capacity than air could be used?
as neat as it is liquid cooling is more deficient
You mean "efficient"
haha yea that
feel like a dumb-ass for not proofreading that
taitaylor86 :D
No, you guys were not paying attention. They said that this was an unfinished cooler, with the motor uncovered.
When he turns it off, that is what it would sound like in a finished, and insulated product.
I just don't understand how heat is transfered from the heat source to the spinning fins. It looks very inefficient to me as they're not touching each other.
The air gets heated and moves outwards through natural convection. It then blows air through a heat sink that surrounds it.
considering most cases have a fan positioned to dissipate cpu heat I imagine a blower like design could easily direct the air out the back of the case the same way gpu blower fans work?
***** Correction: the air gap acts like a very, very bad thermal conductor. That's why we need thermal paste in conventional CPU coolers, where the gaps are in micrometers.
They explain it at 1:45 - they make that layer extremely thin (the same method as getting that tiny distance from a hard drive's disk to its read/write head).
YEY i want a meatgrinder in my computer :D
the solution to the dusts in the heat sinks thank you
Air cooling is loud? I don't think so buddy. Also that high pitch noise that thing is putting out sounds annoying as hell. Good luck.
The sound is explained in the video, that's the noise any motor makes when you take the shrouding off.
It's loud... anything turning that RPM is loud period.. good luck.
ISEE YOU
Not necessarily.
Hλwk978 Yes.. actually. Been building custom PCs for 15 years.. been through a lot of cooling solutions from back in the day when pond pumps were used for water cooling. I am sensitive to high pitched noises and you can tell how loud it is in the video......
He even says it's the motor making the noise! When he let's it coast it is no longer being driven.. well DUH its going to be quiet!
In WW-1 they need to cool an engine without using fluids and be very light weight. They designed a "Rotary" engine to cool it so they rotated the whole engine.
Look at the technology before you make nonsense comments.
You should not waste words and energy of your speech by not understanding first the technology.
My family has a background of technologists in machine development and construction machinery. We have replaced common bearing designs with custom made air-bearings for our machines because they are that much better and maintenance free. We didn't go through PhDs to understand this. We just saw something that works and implemented it.
This technology works. The fact that its not gotten out yet means that the company has either received a lot of money to not produce the technology; if it does ever get into production, it'll be very inefficient and so people will not buy it.
Hey if I was given $50+ million to shut up, I would. This would probably retail for $50 considering the manufacturing cost of less than $10. 1 million units could easily be sold, profiting of $40 million. When you take something from prototype to production, you actually make the product better and not the other way around. This is a simple design and the company claims to have had started production but with only 7x performance compared to the 30x stated, which is non-sense. There are so many companies selling CPU coolers. This would make them all obsolete.
Understand the technology and then make your claims about what the company has gone through. The company doesn't matter. It's the technology. Understand it and use it.
My elder brother is a mechanical engineer specialized in the production of automotive parts. The company he was hired by through a HR team required him and the team to make parts for vehicles. They came up with a design to last 15-20 years on commonly used motor parts with an increased cost, but not that high. However, the plans were refused and were told to use materials and designs that would last for 2 years as the company couldn't sustain the increased cost. Which is total bullshit.
We live in a very different world now. Go figure.
Not very humble are you.
They are taking advantage of the fact that the boundary layer gets thinner if the object is rotating, thus making for a better heat transfer.
Think of it as an conventional heatsink that's split into two separate parts. The tricky thing is how to accomplish the heat flow from the stationary part of the heatsink to the rotary one. Since a metal on metal bearing would mean to much friction, their approach is to use an air bearing.
this hurts my ears
My guess is that the fan they're using is a PoC (Proof of Concept)
would perfectly fit in my build. would buy it as soon as it hits the market.
L
It about time for a new fan tech to come. The idea is fascinating by itself, however general public is already moving their computation loads towards portable devices where rotating fans are not an option. So in the end, there is a need for more efficient chips.
Again, the combination of levitating fan an a heat exchanger is darn cool !!
I don't know anything about your other points, but I do know that it works in any orientation, due to magnetic attraction between the plate and the floating part.
What is the efficiency increase over conventional coolers?
Still not clear about the heat transfer process from a stationary hot surface to a moving-flying-rotating surface.
i'll take two of these coolers for my graphics cards please. two 570's fans running at full speed cooling with virtually no noise? OH HELL YES. GIMME ONE. NOW.
For those bitching about the noise already, He clearly stated that is because the motor has no covers on it, with the covers the noise will be dulled. Even WITH the noise is is quieter than the crap I already have IN my case! 580 leaf blower anybody?
Sounds the same as hard drives did circa 1995. Very simple concept, good implementation.
It's the prototype. Watch the rest of the video when they turn off the motor. The running noise is almost nothing, of course due to the dynamic air bearing.
I see the grand amount heat displacement in the design. But I bet your still in the prototype stages because your not pulling enough of the loss from the base. I would not worry about what is drawn from the base that fancy fan is going to keep things cool. :)
this reminded me of a high school science project. I bet this was impressive in the 1950s.
I can't see this working very well in a standard computer tower where the cooler is going to be sitting vertically. Now perhaps a GPU might benefit from it, since those are typically horizontal.
America's best and brightest "we put a big fan on it"
Heat IS MEANT to be transferred to the fins! The crucial difference compared to conventional coolers is that the fins are SPINNING. There's nothing "blown" on the heatsink but it's rotating itself to create air movement.
It's about boundary layer. That's a very thin layer of fluid/air that sticks to the surface of an object, in this case the metal of the heat sink. It prevents the heat from being effectively transferred to the surrounding air.
Right right, and where is it now in 2016?
it cant work sideways with the air bearing which is where its currently at. Which is why I heard they only have these advertised for servers that sit horizontal, Which normal PC's dont do so dont expect this anytime soon
Thermaltake stole it: Search it under Thermaltake Engine 27 1U.
+Kevin Moondust Stole it? Use your brain... There's this little thing call LICENCING. What makes more business sense: investing millions to setup a production line, QC, marketing, customer service, maintenance/retooling, etc. from scratch all for a single product -OR- simply licencing the technology to an already established manufacturer and saving all that time, money and risk?
+Kevin Moondust No, its CoolerMaster and some other company stole it first, they didn't say anything about it.
Zero necker I stand corrected guys; after digging a bit deeper I found out Thermaltake DID license the tech from CoolChip, whereas CoolerMaster did NOT, so the culprit is another company.
Whatever the case, the "final" product is the Thermaltake Engine 27 1U cooler, of which I expect a review ASAP because I'm deeply interested.
There's probably some sort of low friction rail with a stopper on the end in the center it.
If it was fully detached from the base it would also stop spinning from being disconnected from the motor. Imagine the air that forms from the separation as more of a solid bearing and it will make more sense. If was disconnected it would also probably spin wildly out of control unless it was perfectly balanced, thus it can be assumed that the spinning motion keeps it elevated off the surface but its held down and in place by motor connection and centripetal force respectively.
3:45 I hear music from a coasting impeller? That's a lot of noise...
It's already been discussed, it can run in any orientation.
They REALLY need to reduce the air bearing gap as the current air bearing gap or tolerance it way too high, as you can see in the video the cooler is unstable in its rotation due to the air gap, if this is reduced then heat exchange efficiency will also be increased and also reliability may be increased under certain circumstances
what I meant was that if you are pissed because of your loud GPU cooling, you can easily get rid of your stock cooling and find something better for less than £15 >> like accelero for example, it generates no more than 23 dB of noise. I want a gtx680 too, a water cooled and hopefully stuttering-proof for at least 3 years :)
I'm guessing it would work fine regardless of how the device is oriented, since all it is doing is spinning really fast to suck in air. You don't need to be in a special position to move air around--fans work no matter what direction they're facing. Since its designed to spin at such a high rate, i doubt there would be any problems from operating it vertically, or even upside down.
They say at 3:20 that it's because they don't have anything covering up the sound from the motor. I think they should be able to get the sound down to something like a hard drive motor's noise + the noise you hear at around 3:45 (which is almost nothing).