Take the power supply out (just don't show it in the photos). Take everything else out too, and you have a proprietary system that is very small indeed, that you can patent. One wonders who did not think of it first.
- There is nothing new with taking out the power supply (Compulab has been doing it for many years, so does Intel with NUC and countless other). It actually gives the advantage of feeding the computer from a variety of DC sources. - Taking everything else out is a mistake that Apple did with the new Mac Pro. The result is a messy octopus on the desk. Compulab did the opposite with Airtop, it is packed full with internal devices - 4 HDDs, full-height GPU in 80mm wide chassis etc. - Thinking about miniaturization is one thing. Achieving it in a product is a little more difficult. Better take a closer look at how Airtop is organized internally. You may be pleasantly surprised. - The patents have nothing to do with minituarization. They are all about thermal design.
> Taking everything else out is a mistake that Apple did with the new Mac Pro.
And yet it still overheats like no other machine I've ever used.
It's really a brilliant thermal design, if you look at it from the right perspective: while simultaneously maxing out performance on all cores, memory, storage, and GPUs the outer shell stays cool to the touch and the fan is silent!
*Amazing* how they were able to effectively trap all that heat inside to keep it from bothering you while you work...
Well, the convection works well enough in their SFF systems. We use them at work. They're not perfect, but they get the job done as long as you give them some breathing space.
I can't vouch for their attempts at dissipating a larger amount of heat with this larger box, however. I think above a certain thermal footprint you should just use a silent industrial-grade fan and possibly a filter.
When someone claims to invent convection and the chimney it only makes me laugh.
Also, the tubes in that particular case are not only unnecessary, but actually impeding the ability to displace heat, as hot air expands inside the tube this will hamper the airflow.
I am willing to bet regular radiator with vertical fins would do a better job, but hey, then you don't get to make ridiculous claims of innovation.
Please see www.fit-pc.com, this is a Compulab where you will find many versions of finned computers from 5 to 50W. The rule of thumb in convection is stay under 0.1W per cm^2 or you'll burn your hand. Airtop has a wall with an area of approx 500 cm^2 that dissipates 100W and you can touch it.
Compulab invented neither convection nor chimneys, but we designed the chimneys in such a way that doubles its cooling capacity. 3 years ago when we started I personally thought its a waste of time, after running Airtop with FurMark in parallel to Prime95 overnight for several months at power consumption of over 150W I admitted it works. You are entitled to think this is not patentable, our patent attorney tends to disagree.
Orientation is key, convection will work either way, but how efficiently it will displace heat depends on the orientation of the radiator and the fins. When people hear finned radiator, they imagine the typical tower radiator used on CPUs or chipsets, whose design and orientation make up for extremely poor ability to displace heat without a fan. I am sure that is also he case of your previous products performing poorly.
I am willing to bet there is absolutely nothing special about your chimneys - because I have over a decade experience in design, engineering and simulations - there is nothing you can possibly do to improve that. You might have made it more "touchable" only by thermal insulation on the outside, this would drop the skin temperature, but would hamper the cooling capacity as do the enclosed tubes - there is a trade-off here - reduce cooling capacity for the sake of reducing the skin temperature. There is nothing new or even remotely innovative about this, if that took your engineers 3 years to figure out you should honestly consider getting someone better. I could have told you that in 2 minutes and rigged up a simulation in another 10. 3 years... just wow :)
As I said, I would not be surprise if you patent that, but that would not prove your "invention" - it will only prove what a complete joke the patent system is ;)
Furthermore, in this form factor an actually cleverly designed cooling solution could easily manage to displace 400++ watts. And even though it will require a little more complex manufacturing than what you are used to, it can still be done at a small fraction of the price you ask for your existing products.
I am not a native English speaker, so your argument is that of a straw man. The fact I reverse engineered that "3 year worth of genius invention" in my mind in an instant is testament to my experience. That being said, I am fairly certain I have encountered "skin" in engineering books as referring to the outer covering of a parts assembly, which is what dictated my choice, as that is more specific than "surface" which is ambiguous due to the large number of surfaces it may refer to in the particular case.
If you were into engineering, you'd know what I am talking about. There is a bell shaped curve for everything, and it has only one peak. The same way nobody in a several thousand years managed to find a better shape for the wheel than the circle - is it as good as it gets.
Their "revolutionary invention" is heat insulation to mask how hot the radiators really are and using a better orientation to move more air compared to their previous products which barely moved any. That's all there is to it.
I just hope no user ever decides to put that thing horizontally, because it will cook in an instant.
If that's indeed the case why is that you refer to such a trivial concept as a "real breakthrough"?
Here is the chain of reasoning that needs to be traversed in order to solve this engineering problem:
problem - we need to cool stuff down solution - we use a heat sink problem - heat sink can't displace the heat fast enough solution - put the heat sink radiator in a way to create more air flow problem - radiator is too hot to the touch solution - cover radiator with heat insulator leaving enough room for air to circulate
Just a few painfully obvious steps which should not be a problem even for a self taught enthusiast, much less for an experienced engineer.
Granted, it may well be the first time it is applied in a commercial PC, but that most certainly doesn't make it new.
The list of problem-solution you describe above is a healthy engineer process, not unsimilar to our development process. There are many correct observations, some are missing. There is one glaring (and common) misunderstanding about insulation.
We worked hard to conceptualize and develop Airtop. There are many companies in the industry that copy technology. We applied for patents to protect our investment.
You are entitled to believe you could develop it sooner. Maybe so. Many developments are obvious in hindsight. That doesn't make them stupid.
I never said your product is stupid, what is stupid are your claims on "inventing" stuff. You need to dial it down, because it is ridiculous. Also consider hiring a better designer and a little less absurd product pricing.
There's a difference between patenting the chimney and patenting specific (quite possibly application-specific) enhancements to the chimney. I have a feeling that at least some of the patents are the latter. I'm personally not super fond of the patent system, but there's a pretty big difference there.
Patented or not patented is not the key question. The question is whether that technology is really capable to dissipate so much heat for its specified dimensions. If it does, there is a real breakthrough here. For specified dimensions, 200 watts cannot be dissipated by cooling fins.
Exactly. Just an anecdote - heat exchange in a slow-moving air system is extremely difficult to calculate analytically. A combination of the boundary layer behavior, rate of airflow, varying deltaT along the path - all with mutual effects and high sensitivity to miniscule changes. Bottom line, we planned using GTX 750 Ti, but Airtop did not break a sweat with it so we upgraded Airtop to GTX 950.
Right. For cooling, it's just simple as airflow and surface area. The tubes are nothing but more surface area, though as already stated, dirt or debris might reduce its effectiveness faster than a fin design with similar surface area.
It is not correct. And it is not that simple as looks at the first glance. I will try to explain briefly.
Intuitive assumption is that in system with vertically arranged fins, the air flows up in parallel to the fins, and therefore efficiently extracts the heat from fin's surface. In reality it is not so. Due to expansion, the heated air obtains certain component of horizontal velocity and therefore moves not exactly up but rather diagonally. It is partially expelled from fins area and continues to raise in open space, contributing nothing to system cooling.
Confining the raising air into channel prevents its escape, so air continues to rise along the channel and to exchange the heat with channel’s walls.
Another important thing: during the raising the air heats, so the temperature difference between channel's walls and the air decreases. If channel is long, the temperature difference in the upper section wouldn't be sufficient for heat exchange at significant rate. To overcome this problem the air must flow fast. In confined channel air flows faster, which assures that its temperature will be lower and therefore heat exchange will continue along the entire channel's length.
There are several other important factors which should be taken into consideration, like the shape and dimensions of channel cross-section.
The simplistic attitude like “there is nothing new and nothing can be improved” is wrong all along.
While I usually find higher priced computers unappealing, there's a lot to like about this system that makes it's price point less of an obstacle. After I started playing with totally passive Bay Trail systems, I found that I'm willing to give up a lot in exchange for a completely silent PC. The fact that the Airtop can pull this off with more than enough processor and graphics power to make it capable of gaming without having to muddle along with a ~2 watt, ultra low-end chip with integrated graphics is an impressive feat of engineering. The one generation older CPU isn't much of a concern considering the small gains Intel made with their 6th gen chips and a 950 is more than enough for gaming at 1080. I think the only downside is that it's price and relative power puts it in competition with gaming notebooks that have the disadvantage of active cooling, but offset that by being easier to transport or use on the go.
I'd like to see Apple buy them out and make the next Power Mac this way. Imagine, an i7 6700, with roughly a GTX 970 or equivalent, SSD for drive, and no noise at all!
I see this type of design as the future of home computing. After a few more litography changes, and architecture efficiencies are further refined, I can see this being completely viable. The hurdle remains the price. At $1200 barebones this would likely be the most expensive barebones on the planet, excluding the barebones that have auto-quality paint jobs. I understand that a lot of engineering went into this, and you want to recoup that investment, but I just don't see the market for this being large enough to have any impact on the PC market at all. Maybe at 7 nm things will be entirely different.
I'm looking into putting together a fanless gaming PC. Does anyone know of any nice looking passively cooled horizontal cases? I'd like to put it in shelf under my TV. I'll probably use a 150 W GPU.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
35 Comments
Back to Article
versesuvius - Thursday, January 14, 2016 - link
Take the power supply out (just don't show it in the photos). Take everything else out too, and you have a proprietary system that is very small indeed, that you can patent. One wonders who did not think of it first.Compulab - Thursday, January 14, 2016 - link
- There is nothing new with taking out the power supply (Compulab has been doing it for many years, so does Intel with NUC and countless other). It actually gives the advantage of feeding the computer from a variety of DC sources.- Taking everything else out is a mistake that Apple did with the new Mac Pro. The result is a messy octopus on the desk. Compulab did the opposite with Airtop, it is packed full with internal devices - 4 HDDs, full-height GPU in 80mm wide chassis etc.
- Thinking about miniaturization is one thing. Achieving it in a product is a little more difficult. Better take a closer look at how Airtop is organized internally. You may be pleasantly surprised.
- The patents have nothing to do with minituarization. They are all about thermal design.
ninjaburger - Thursday, January 14, 2016 - link
> Taking everything else out is a mistake that Apple did with the new Mac Pro.And yet it still overheats like no other machine I've ever used.
It's really a brilliant thermal design, if you look at it from the right perspective: while simultaneously maxing out performance on all cores, memory, storage, and GPUs the outer shell stays cool to the touch and the fan is silent!
*Amazing* how they were able to effectively trap all that heat inside to keep it from bothering you while you work...
Alexvrb - Sunday, January 17, 2016 - link
Well, the convection works well enough in their SFF systems. We use them at work. They're not perfect, but they get the job done as long as you give them some breathing space.I can't vouch for their attempts at dissipating a larger amount of heat with this larger box, however. I think above a certain thermal footprint you should just use a silent industrial-grade fan and possibly a filter.
Alexvrb - Sunday, January 17, 2016 - link
Whoops sorry I didn't mean to reply to your comment on the Mac. This was strictly in reference to Compulab convection cooling setups.toronado455 - Friday, January 15, 2016 - link
Speaking of the power supply, can anyone tell me where power goes into this? From the photos I can't figure out where the power supply plugs in.BlueTortoise - Friday, January 15, 2016 - link
It uses the Mini-DIN connectors (that look like old Keyboard/Mouse PS2 ports) on the back for power.I'm not 100% sure why two are required though.
ddriver - Thursday, January 14, 2016 - link
When someone claims to invent convection and the chimney it only makes me laugh.Also, the tubes in that particular case are not only unnecessary, but actually impeding the ability to displace heat, as hot air expands inside the tube this will hamper the airflow.
I am willing to bet regular radiator with vertical fins would do a better job, but hey, then you don't get to make ridiculous claims of innovation.
ddriver - Thursday, January 14, 2016 - link
But given what a pathetic mockery the patent system is, they could totally patent it LOL it will not surprise in the slightest.Compulab - Thursday, January 14, 2016 - link
Please see www.fit-pc.com, this is a Compulab where you will find many versions of finned computers from 5 to 50W. The rule of thumb in convection is stay under 0.1W per cm^2 or you'll burn your hand. Airtop has a wall with an area of approx 500 cm^2 that dissipates 100W and you can touch it.Compulab invented neither convection nor chimneys, but we designed the chimneys in such a way that doubles its cooling capacity. 3 years ago when we started I personally thought its a waste of time, after running Airtop with FurMark in parallel to Prime95 overnight for several months at power consumption of over 150W I admitted it works.
You are entitled to think this is not patentable, our patent attorney tends to disagree.
ddriver - Thursday, January 14, 2016 - link
Orientation is key, convection will work either way, but how efficiently it will displace heat depends on the orientation of the radiator and the fins. When people hear finned radiator, they imagine the typical tower radiator used on CPUs or chipsets, whose design and orientation make up for extremely poor ability to displace heat without a fan. I am sure that is also he case of your previous products performing poorly.I am willing to bet there is absolutely nothing special about your chimneys - because I have over a decade experience in design, engineering and simulations - there is nothing you can possibly do to improve that. You might have made it more "touchable" only by thermal insulation on the outside, this would drop the skin temperature, but would hamper the cooling capacity as do the enclosed tubes - there is a trade-off here - reduce cooling capacity for the sake of reducing the skin temperature. There is nothing new or even remotely innovative about this, if that took your engineers 3 years to figure out you should honestly consider getting someone better. I could have told you that in 2 minutes and rigged up a simulation in another 10. 3 years... just wow :)
As I said, I would not be surprise if you patent that, but that would not prove your "invention" - it will only prove what a complete joke the patent system is ;)
ddriver - Thursday, January 14, 2016 - link
Furthermore, in this form factor an actually cleverly designed cooling solution could easily manage to displace 400++ watts. And even though it will require a little more complex manufacturing than what you are used to, it can still be done at a small fraction of the price you ask for your existing products.royalcrown - Thursday, January 14, 2016 - link
If you have a decade of design experience then you should refer to it as surface temperature, which is affected by delta T.ddriver - Thursday, January 14, 2016 - link
I am not a native English speaker, so your argument is that of a straw man. The fact I reverse engineered that "3 year worth of genius invention" in my mind in an instant is testament to my experience. That being said, I am fairly certain I have encountered "skin" in engineering books as referring to the outer covering of a parts assembly, which is what dictated my choice, as that is more specific than "surface" which is ambiguous due to the large number of surfaces it may refer to in the particular case.GY - Thursday, January 14, 2016 - link
"I have over a decade experience in design, engineering and simulations - there is nothing you can possibly do to improve that"So, you spent 10 years and could not find improvement, and it means that none else can.
Healthy reasoning.
ddriver - Thursday, January 14, 2016 - link
If you were into engineering, you'd know what I am talking about. There is a bell shaped curve for everything, and it has only one peak. The same way nobody in a several thousand years managed to find a better shape for the wheel than the circle - is it as good as it gets.Their "revolutionary invention" is heat insulation to mask how hot the radiators really are and using a better orientation to move more air compared to their previous products which barely moved any. That's all there is to it.
I just hope no user ever decides to put that thing horizontally, because it will cook in an instant.
GY - Thursday, January 14, 2016 - link
Well, I'm deeply into engineering and physics of thermal design ...ddriver - Thursday, January 14, 2016 - link
If that's indeed the case why is that you refer to such a trivial concept as a "real breakthrough"?Here is the chain of reasoning that needs to be traversed in order to solve this engineering problem:
problem - we need to cool stuff down
solution - we use a heat sink
problem - heat sink can't displace the heat fast enough
solution - put the heat sink radiator in a way to create more air flow
problem - radiator is too hot to the touch
solution - cover radiator with heat insulator leaving enough room for air to circulate
Just a few painfully obvious steps which should not be a problem even for a self taught enthusiast, much less for an experienced engineer.
Granted, it may well be the first time it is applied in a commercial PC, but that most certainly doesn't make it new.
Compulab - Thursday, January 14, 2016 - link
The list of problem-solution you describe above is a healthy engineer process, not unsimilar to our development process.There are many correct observations, some are missing. There is one glaring (and common) misunderstanding about insulation.
We worked hard to conceptualize and develop Airtop. There are many companies in the industry that copy technology. We applied for patents to protect our investment.
You are entitled to believe you could develop it sooner. Maybe so. Many developments are obvious in hindsight. That doesn't make them stupid.
Sincerely,
Irad Stavi
Compulab
ddriver - Sunday, January 17, 2016 - link
I never said your product is stupid, what is stupid are your claims on "inventing" stuff. You need to dial it down, because it is ridiculous. Also consider hiring a better designer and a little less absurd product pricing.xthetenth - Thursday, January 14, 2016 - link
There's a difference between patenting the chimney and patenting specific (quite possibly application-specific) enhancements to the chimney. I have a feeling that at least some of the patents are the latter. I'm personally not super fond of the patent system, but there's a pretty big difference there.GY - Thursday, January 14, 2016 - link
Patented or not patented is not the key question. The question is whether that technology is really capable to dissipate so much heat for its specified dimensions. If it does, there is a real breakthrough here. For specified dimensions, 200 watts cannot be dissipated by cooling fins.Compulab - Thursday, January 14, 2016 - link
Exactly.Just an anecdote - heat exchange in a slow-moving air system is extremely difficult to calculate analytically. A combination of the boundary layer behavior, rate of airflow, varying deltaT along the path - all with mutual effects and high sensitivity to miniscule changes.
Bottom line, we planned using GTX 750 Ti, but Airtop did not break a sweat with it so we upgraded Airtop to GTX 950.
zodiacfml - Thursday, January 14, 2016 - link
Right. For cooling, it's just simple as airflow and surface area. The tubes are nothing but more surface area, though as already stated, dirt or debris might reduce its effectiveness faster than a fin design with similar surface area.GY - Thursday, January 14, 2016 - link
It is not correct. And it is not that simple as looks at the first glance. I will try to explain briefly.Intuitive assumption is that in system with vertically arranged fins, the air flows up in parallel to the fins, and therefore efficiently extracts the heat from fin's surface. In reality it is not so. Due to expansion, the heated air obtains certain component of horizontal velocity and therefore moves not exactly up but rather diagonally. It is partially expelled from fins area and continues to raise in open space, contributing nothing to system cooling.
Confining the raising air into channel prevents its escape, so air continues to rise along the channel and to exchange the heat with channel’s walls.
Another important thing: during the raising the air heats, so the temperature difference between channel's walls and the air decreases. If channel is long, the temperature difference in the upper section wouldn't be sufficient for heat exchange at significant rate. To overcome this problem the air must flow fast. In confined channel air flows faster, which assures that its temperature will be lower and therefore heat exchange will continue along the entire channel's length.
There are several other important factors which should be taken into consideration, like the shape and dimensions of channel cross-section.
The simplistic attitude like “there is nothing new and nothing can be improved” is wrong all along.
khon - Thursday, January 14, 2016 - link
It looks nice, but who's going to pay $2k for a "gaming" PC with outdated CPU and underpowered GPU ?wurizen - Thursday, January 14, 2016 - link
But, it looks like a Netgear router from a 2005.ddriver - Thursday, January 14, 2016 - link
The case design is as bad as the logo design. Seems like the company is short not only in the engineering department. And at those prices...BrokenCrayons - Thursday, January 14, 2016 - link
While I usually find higher priced computers unappealing, there's a lot to like about this system that makes it's price point less of an obstacle. After I started playing with totally passive Bay Trail systems, I found that I'm willing to give up a lot in exchange for a completely silent PC. The fact that the Airtop can pull this off with more than enough processor and graphics power to make it capable of gaming without having to muddle along with a ~2 watt, ultra low-end chip with integrated graphics is an impressive feat of engineering. The one generation older CPU isn't much of a concern considering the small gains Intel made with their 6th gen chips and a 950 is more than enough for gaming at 1080. I think the only downside is that it's price and relative power puts it in competition with gaming notebooks that have the disadvantage of active cooling, but offset that by being easier to transport or use on the go.Ken_g6 - Thursday, January 14, 2016 - link
I'd like to see Apple buy them out and make the next Power Mac this way. Imagine, an i7 6700, with roughly a GTX 970 or equivalent, SSD for drive, and no noise at all!fanofanand - Thursday, January 14, 2016 - link
I see this type of design as the future of home computing. After a few more litography changes, and architecture efficiencies are further refined, I can see this being completely viable. The hurdle remains the price. At $1200 barebones this would likely be the most expensive barebones on the planet, excluding the barebones that have auto-quality paint jobs. I understand that a lot of engineering went into this, and you want to recoup that investment, but I just don't see the market for this being large enough to have any impact on the PC market at all. Maybe at 7 nm things will be entirely different.Ktracho - Thursday, January 14, 2016 - link
Disregarding price for a moment, something like this could be used as a (silent) home theater PC that can also be used for casual gaming.modpr - Thursday, January 14, 2016 - link
Can this support the requirements for the Oculus Rift?modpr - Friday, January 15, 2016 - link
Looks like it doesn't or at least I'd rather not install the CPU and GPU myself (mostly the CPU).arjen32 - Friday, January 15, 2016 - link
I'm looking into putting together a fanless gaming PC. Does anyone know of any nice looking passively cooled horizontal cases? I'd like to put it in shelf under my TV. I'll probably use a 150 W GPU.I've seen some but they look weird.