Three Marvelous PC Cooling Technologies That You’ve Never Heard Of

Ads by Google

New PC cooling technologies use almost no energy and employ methods 50 times more efficient than conventional designs. Some come in a form factor as thin as cardboard. Others look like a hockey-puck. Unfortunately, of the many technologies promised by the Internet, few actually come to fruition. We call it “vaporware” when gadgets fail to materialize.

While hitting CPU speeds over 5 GHz on air may soon exist, right now these technologies don’t appear anywhere, outside a laboratory. For whatever reason, the following three technologies failed to make it to market. Although their manufacturers produced them in prototype form, they never entered mass production.

Sandia’s Supercooler

sandia cooler

The most interesting, and radical, out of the bladeless coolers hails from the US government-backed Sandia Labs. The Sandia fan made quite an impression on PC enthusiasts by cramming 50-times greater cooling ability into a hockey-puck sized heat sink. It achieves its power based on an extremely unusual phenomenon: Boundary layer effect.

The effect is difficult to explain. Rather than incorporating a copper or metal heat sink, it uses air. To be precise, Sandia’s cooler uses a vortex of air. If you examine the cooler’s central cooling mechanism, it appears to be a metal, serrated disk. This disk spins, creating a miniature vortex directly above the CPU’s integrated heat-spreader, where all the heat issues from. The vortex provides more efficient cooling than even the biggest of consumer heat sinks. This device ironically cools more efficiently than direct metal-to-metal contact by using air.

Sandia’s invention makes almost no noise, takes up hardly any space and cools more efficiently than the best heat sinks on today’s market. An anonymous company even purchased the licensing rights more than a year ago, but unfortunately, never produced any device. It’s possible that either the technology somehow proved unfeasible or it was shelved pending additional refinement.

Ads by Google

You can watch video of Sandia’s cooler in action below:

Ionic Wind Cooler

Ionic wind uses no moving parts to generate airflow for central processing units. It’s also completely silent and capable of cooling even large wattage CPUs. Anyone can build their own cooler, as well, given that it uses off-the-shelf parts. One blogger, Jared Bouck, even posted a guide on how to build your own iconic wind generator.

ionic wind cooled system

The ionic wind effect remains one of the most fascinating cooling technologies ever made. It employs two charged, metal plates. One plate receives a negative electrical charge and the other a positive. This causes charged air particles to move from one direction to the other, generating a gentle, but powerful, breeze.

You can watch video of the effect below:

This cooling method is useful in devices where not much space is available, like laptops, or where silence is required, such as HTPCs. It also provides the same effect as an air ionizer: the ionized particles will remove pollutants from the air.

There’s a huge number of patents focusing on the use of ionic wind as a cooling agent. Apple patented a large number of fanless technologies to cool their laptops — ionic wind generators remain among the most ideal solutions for silently cooling aesthetically clean designs. However, Apple failed to implement any such method in any product. It’s a shame, because some developers found that ionic jets can cool devices 30% better, with half as much energy consumption.

On the downside, there’s been a number of technical limitations facing ionic wind devices. Oxidation of the charged plate remains the biggest impediment to production. Essentially, when a charge is run through a metal plate, the plate tends to bond with oxygen particles more readily. Unfortunately, ionic wind coolers will lose efficacy over time, as its plates oxidize.

Piezo Bellows-Style Cooling

Piezo cooling methods utilize a bellows-type cooling device, in place of a fan. Instead of using axial rotation combined with an airfoil to generate airflow, a bellows blows air. The device can fit into extremely tiny places, oftentimes not that much thicker than a a few pieces of paper.


It also requires virtually no energy to run, while cooling near what a regular fan can cool. The technology can also scale dramatically, with larger devices cooling substantially greater volumes of heat.

One of the researchers for General Electric’s dual jet piezo cooler did an interview on Reddit. GE’s dual jet piezo unit can cool up to 18 watts processors with a relatively low sound generation at 20 dBA.


None of the three technologies presented within this article, at present, made it into markets. While you can make your own ionic wind cooler or eco-friendly PC, you can’t buy one with these technologies presented in this article. However, there are alternatives. For example, you can build a fanless PC of your own (although it may not work properly). Or, you can silence a traditionally-cooled laptop.

Anyone know of any other radical cooling technologies? Please share in the comments!

Join live MakeUseOf Groups on Grouvi App Join live Groups on Grouvi
Technology Explained
Technology Explained
152 Members
Comments (20)
  • Dave C

    This looks like it has potential for use where certain semiconductors cannot fit due to limited space for a heatsink, but many people seem to be thinking of it as a replacement for existing cooling for things like GPU or CPU.

    That thinking is in error because the CPU or GPU designer is designing their chips within the limits of what legacy cooling tech can accommodate, and in the case of retail product, even including the tested and proven heatsink solution.

    Unfortunately this design looks to be costly despite its simplicity. The reason is that it requires much more precise machining, of both sides of the base for CPU contact and the uniform air gap, as well as machining the fins so they are as perfectly balanced as possible. Otherwise the motor bearings take a beating and wear down much faster.

    All else being equal with regards to balancing and bearing tech, the motor bearing will still wear down faster due to the much greater mass of the fins relative to a traditional plastic bladed fan.

    A larger bearing and shaft could be used but this makes it an unusual assembly rather than a standard DC fan BOM which requires a separate much smaller production quantity thus raising cost even more, as well as the (currently) small target market.

    It “could” be used in many areas it would not be used in due to the higher cost and sheer # of competitive products already in the market.

    The irony is that we’re seeing this tech now. Soon transistors will be so small that a shift in materials is needed to combat leakage, meaning that prior to that tech we are nearing the limit in process size, and die size reduction. They could simply make a larger die with the thermal density decreased but this greatly reduces yield per wafer.

    The irony is that we have had enough prior art and competition the heatpipe cooler sector that it is now rather inexpensive to purchase a heatpipe based sink capable of cooling existing consumer oriented processors.

    The more efficient of these tend to be rather large but as with CPUs being designed per environmental requirements, so are heatsinks designed to fit within a typical case form factor and provide clearance for adjacent objects such as the motherboard chipset and memory module slots. On the other hand it does suggest the potential for further shrinking of the case form factor for CPUs above a certain wattage threshold, perhaps around 40W would be a good, current threshold.

    • Kannon Yamada

      Cooler Master is releasing the Kinetic line of HSF CPU coolers, in three different low-form factor sizes. They’ll probably cool somewhere around 100 watts and down, depending on the size of the unit. But you’re totally right. This is designed specifically for small form factor systems.

    • Dave C

      Cooler Master products are crap. They had terrible quality bearings and now they’re putting more mass on them to decrease lifespan. It’s the worst possible move for them.

      Their heatsink does not work on the Sandia/kinetic principle, rather it’s just a traditional heatsink with pipes and fins that merely replaced the plastic fan blades with a squirrel cage metal blade assembly for no good reason except to use marketing to trick people.

      Will it do the job? Sure, as will many other heatsinks that cost less and are more reliable, making it irrelevant.

    • Kannon Yamada

      I’ve only used their low end products, but they were good for the value. Someone mentioned in the comments of the Kinetic Cooling Engine video that CoolerMaster loves using sleeve bearings in these kinds of products. But they say it’s low RPMs. Perhaps it might not break down right away.

      What do you think about the initial price? My bet is $100 for the larger model.

    • Dave C

      $100 seems very expensive for what it is. There have been low profile heatsinks in rack systms and laptops for many years that can’t have added even $40 to the BOM.

      Calling it a $100 MSRP product then marking it down might work, but I suspect the best way to get market penetration would be to make them practically free after rebate during an introductory period so buzz about them grows on the internet.

      Cooler Master has used not just sleeve bearings but low quality ones. At least they didn’t pull the stunt that Thermaltake used to, putting one narrow sleeve and one ball bearing in a fan so they could call it “ball bearing” and save a few pennies, then you couldn’t even relubricate the sleeve bearing without risking getting some oil in the ball bearing and making it very noisy.

      Low RPM wiould help a lot, and yet the mass of metal is bound to be higher so it may not be enough to promote long life, especially if it is not *perfectly* balanced. Another concern is that if the bearing is not thermally isolated from the metal then running at higher temperatures that will be another source of early wear if the lubricant isn’t properly spec’d.

    • Kannon Yamada

      That’s an interesting theory regarding potential bearing failure if there’s waste heat gets distributed into the bearing. From the shots we’ve seen it’s impossible to tell how it’s rotating. Sandia’s spec sheet mentioned it uses a hydrodynamic air bearing (which I’ve never heard of), and they emphasize that there’s no physical bearings involved. Rather, it’s a spinning heat sink fan combo floating on a tiny cushion of air. The physics escape me entirely, but you can read about it here:

      Essentially, it’s using the boundary layer effect to transfer heat efficiently through an air gap. It sounds like science fiction to me, although you might understand it with greater fluidity than I.

      My hope is that it enables a small form factor, poor-man’s VR workstation. Although that seems like something for 2017 or 2018.

  • Ahmed K

    I wonder if the GPU alone can be Dunked in the Oil! ?

    • Guy M

      If you could build a containment device for it, you might be able to. The trick would be to have it not leak where it plugs into the motherboard and where the monitor cable plugs in.

    • Dave C

      Leaking to the extent of losing the fluid so it doesn’t do its job would be a problem, but most oils, certainly one picked for the purpose, are not conductive and wouldn’t pose a problem getting on the motherboard slot or monitor cable. There have been entire systems submerged in oil that ran fine, though it was more of a novelty to show off on the internet rather than a practical thing to do.

      However, oil does not conduct as well as metal. You’d still want a heatsink and where is all that oil going to be? The issue already present with GPU heatsinks is making them small enough to fit in the available space.

      Adding a less effective thermal transfer substance seems like a step in the wrong direction instead of using water cooling with a remotely mounted radiator, although personally I think they’re acceptable already with heatpipe sinks providing the fan is thick enough for a good dual ball bearing system inside, which essentially means a double height card to provide enough space to do that.

      UGH! WHY does M.U.O corrupt my post structure and add characters that weren’t there when I wrote it? They need to fix that.

  • Guy M

    I keep thinking there must be a place for a Peltier Junction somewhere in electronics cooling. They’re not terribly energy efficient though.

    • Kannon Y

      That’s a good question. Thanks Guy!

      Someone showed me a Peltier cooled CPU once at a LAN party and explained the underlying tech. And I was like – that’s an ionic cooler, bro.

      Of course, I was wrong, and he very politely explained what Peltier cooling is. Yet I persisted in my original remark, claiming that using ionically charged plates formed the basis of ionic cooling. He eventually gave up trying to explain and very politely excused himself from the conversation.

      Heady with my perceived intellectual victory over someone with an engineering background, I then smugly bragged to my friends how I was smarter than an engineer.

    • Guy M

      I kind of like the idea of the ionic wind generator as well, but it seems like a lot of effort for what you get.

      Any Peltier junctions that I’ve found recently are fairly high wattage (45w or greater) Then you still need a significant heat sink to pull the heat off the hot side, possibly even a fan. The cool side will provide better cooling than most other methods – you can go into the negatives.

      The cooler you can get the warm side, the colder the cold side gets. Let’s say the junction is rated for a 50 degree delta (delta meaning difference) If you are putting out +40 degrees on the hot side, the cool side is at -10 degrees.

      I would go this route over water cooling though. Just the thought of water that close to my CPU makes me cringe.

    • Dave C

      Peltiers are only rated at their max wattage using a tested heatsink standard, or at a particular voltage. You can run them at less than that voltage for lower wattage, but practically speaking in an app where you feel a peltier is worth bothering to use, you’ll probably want to run it at full power.

      You definitely need a fan on a peltier cooling something that would otherwise need a fan if the peltier wasn’t there, or else a heatsink so massive that you still didn’t need the peltier plus a fair amount of passive airflow.

      Your numbers are fairly unrealistic for the thermal loads present with modern CPU or GPU. It would tend to be more like 90C on the hot side and 40C on the cold, though 50′ is a lot and that’s also a lot more heat to get out of a system case so you might just be moving your cooling problem and still need another fix.

  • Jon S

    The Sandia fan is really cool, pun intended.
    I could tell that it looks like a heat sink,
    but I don’t remember ever hearing about it.
    How old is this design anyway?

    • Kannon Y

      I believe this design originates back to ye olden days of 2012. What’s a shame is that the tech got licensed out – which means a company intended to produce these things, but then the technology never materialized. It makes me wonder if the design just got shelved.

  • bardwso

    Man, I really like the design of the Sandia Supercooler. A lot of air but I feel I could lose a finger or two if I wasn’t paying attention.

    That said, I am willing to risk it.

    • Guy M

      Because of the air cushion, there isn’t much torque to one of these fans. Should be as easy to stop as a regular CPU fan with a finger.

      Sandia says that dust isn’t an issue as it tends to spin too fast for dust to adhere to it. I’m not sure I entirely buy that though.

    • Dave C

      This is incorrect. First, they contrast this against CPU fans with high RPM and such fans can cut your finger if stick into one which is the opposite of easy. Second, with either a CPU fan or this, it is not the motor torque you are overcoming as much as the rotating mass, the inertia.

      Because this has a substantially higher rotating mass it will be substantially more friction required to stop it. Further, the air cushion does not bear on there being an especially low torque fan. The air cushion is the equivalent of there being no contact at all with a traditional fan blade, except to the motor shaft with either.

      Further because there is a greater rotating mass, the motor would be higher torque unless it is within design parameters for it to rise in RPM at a much more leisurely pace than a typical (computer) DC brushless fan does.

      With all this written, it would be hard to do more than shave off a thin layer of skin if touching the top of the fins which look to be more 90′ edged as a typical aluminum extrusion cutoff area is, because the fins are close together. You would need to hold the finger against the spinning fins with just the right pressure to continue shaving off skin layers but not enough to decelerate rotation too much, and more likely it would cause minor bruising.

Load 10 more
Affiliate Disclamer

This review may contain affiliate links, which pays us a small compensation if you do decide to make a purchase based on our recommendation. Our judgement is in no way biased, and our recommendations are always based on the merits of the items.

For more details, please read our disclosure.
Affiliate Disclamer

This review may contain affiliate links, which pays us a small compensation if you do decide to make a purchase based on our recommendation. Our judgement is in no way biased, and our recommendations are always based on the merits of the items.

For more details, please read our disclosure.
New comment

Please login to avoid entering captcha

Log In