MSI ECOlution motherboard transforms chip heat into fan power
By Ryan Block 
posted February 29th 2008 3:31AM
posted February 29th 2008 3:31AM
Update: CG video demo posted after the break, via TweakTown. Thanks, Lin.
posted February 29th 2008 3:31AM
A look back on popular stories from today in a specific year.
Now that we've thrown 'em off the trail, use the form below to get in touch with the people at Engadget. Please fill in all of the required fields because they're required.
Did some of you read the article and take a detailed look at the picture?
The CPU socket is to the NW of the cooler, the ram is to the NE and the PCI/AGP slots are to the SE of it...making it a northbridge chipset cooler, NOT a CPU cooler.
It's not a whole bunch but it's better than powering a fan using the board. This tech applied elsewhere could lead to lower PC power requirements, which is always a good thing, so hopefully they'll continue and find more robust solutions that can be applied to CPUs, memory and graphics cards.
will it though?
If energy cannot be created or destroyed, only transferred from one form to another then the overall energy requirements will be the same for the system, because you still need the electricity to generate the heat, to generate the movement.
it comes down to whether this system can be more efficient than a standard heatsink and electric fan. If it can, then it'll save money, if not, it wont.
But I salute them for trying, there must be a merit to this system somewhere - otherwise they wouldnt have done it. So go MSI!!
G
G:
The energy is already being generated as heat. The difference is instead of being spent heating the air around it the heat powers a fan. The fan isn't going to draw extra power from the PSU just use what's already being "wasted".
OMG, HEAT-POWERED FANS! 1337 H4X!
Cool idea, but how long before the fan seizes up?
Erm...Generic:
Physics denotes that electrons move faster when things are cooler (heat = resistance) this is why when you use 30awg wire to wire your house, it all melts.
cooler is better in all chips that require fast refreshes of data. Thats why there are cryo-coolers for the extreme users...
*awaits flames*
G
In a sense it's true that keeping it constantly cool would increase the lifetime of the chip due to less 'wear', but with systems being outdated in 6 months currently that's not much of an issue, if it last 10 instead of 12 years doesn't matter.
As for the maximum attainable speed itself, not an issue with a northbridge surely, since it's fixed in functionality and OC'ing it is not really useful.
Since the thing has a proper heatsink attached I'm guessing it won't turn at all unless you are pushing the system like with raytracing or playing games and then it'll nicely come into action.
So nice and silent?
I don't see any practical use for this, you're still limited by the ambient temperature in the case. It also has to be fairly expensive, sterlings require some tight tolerances and even if you subtract the cost of a fan what's the point?
so you're saying there is no practical use for fans inside the case, and that we should only concern ourselves with dropping the ambient air temperature inside the case?
as far as expense:
there are alot of computers out there with chipsets that need cooling.
as a result, this type of cooler would be mass produced.
as a result, single unit cost is mitigated.
I didn't mean to imply we don't need fans only that this solution is a non starter. Depending on where the cost comes in you're going to be competing with some viable & proven technologies. If noise is the only concern then there are other ways to get there. Reliablity is a big concern, look at the fans in your case after a year or so, they are caked with lint, dust whatever, what would happen to this if that happened, is it powerful enough to overcome that added mass & drag? Look what Sony did with the PS3, they used a really big, low RPM fan, much more reliable than this & quiet. I think the goal of any design engineer is to use as little technology as possible to get the job done, this while cute, is over engineered for the job.
"Depending on where the cost comes in you're going to be competing with some viable & proven technologies."
"I think the goal of any design engineer is to use as little technology as possible to get the job done, this while cute, is over engineered for the job."
True enough, but there was a time when the automobile had to compete with the viable and proven horse drawn buggy technology. The automobile was overly complex but was upheld by the benefits it provided.
I wouldn't go so far as to say it is over engineered. Sterling engine technology was around before electricity was powering peoples homes. Indeed it is more mechanically complex just as the automobile was, but bearings in DC fans are also subject to wear.
"Is it powerful enough to overcome that added mass & drag? Look what Sony did with the PS3, they used a really big, low RPM fan, much more reliable than this & quiet."
Electric fans have to deal with the same problems with mass and drag. As far as what sony did .. there's nothing to stop you from putting a bigger fan on this.
A few other things to take into consideration with a sustained system like this is...!
Even when the system is shut down the fan would theoretically keep going no ?
And even better , when the system is started there won't be that much of a what i would call A..
" Cold boot "shock" "
...see.... Everytime a electronic ?"gadget"? .. ehem**... is started there is a brief "shock" to the system... which does 2 bad things...
A.) Dramatically decreases the life of the part.
B.) Uses up allot of electricity on its start. (much like when you start a non fuel injected vehicle, you would be using up allot of fuel on startup)
And a few other things too i just thought of... , I don't think this was made to be an "efficient" design by msi more of a "green" design.
So did they hit there mark or not ?
I'm gonna bet on yes.
Whatya guys think about my cold start and shutdown observations ?
this is very cool, but unfortunately I think it falls under the catorgory of "gimmick"
V__I disagree see below.__V
Now that all the narrow minded people have spoken, i would like to add think about this on a Large scale say at a power plant you could use the excess heat generated to spin the fan that is hooked to a different generator to make more power off the unused heat, or even use the same idea to make a nuclear reactor cooler by hooking it to the top of the rod banks. and even if no one ever uses it i may cause and innovation some where else in a different field of science.
First of all, it really should be said that there are no thermodynamic laws being broken here. All of the energy to power both the chip set and the fan come from your local power company. I know that the thermo comment in the article was likely a joke, but this is very different than the "Steom" project or whatever it was where he was making MORE heat. This is just using the excess heat from the chip.
That's all, I'll stop pointing out the obvious now.
A microgenerator could be attached to the fan, thereby powering the CPU. And any surplus generated energy could power the memory, screen and hard disk.
Wow...this kind of like a suicide machine. It's constantly trying to kill itself, no?
I'm dumbfounded by all the people who seem to be strongly against ANY attempt to actually improve efficiency.
Is this some kind of misplaced backlash against global warming hype? The points that this doesn't save MUCH energy--and that the device takes more energy to build than what it will save--are really beside the point. From an engineering standpoint, the idea of using heat energy to help dissipate itself is actually pretty cool (pardon the pun).
While this particular development may not be perfect, the engineers will doubtless find a way to improve on it.
Its humans 'what we dont know about, be scared of' nature.
Global warming is affecting us, but because we dont know which way it will swing (Or when and how it would) people fear it and attack it (and people that talk about it) with brute force insted of thinking.
This has been going on for centurys, many a great mind got killed for 'blasphamy' and 'crimes against nature' or 'the state' for saying something like 'the earth revolves around the sun'.
But the idea is starting to stick in peoples minds, people ARE changing - we now have recycle pickups in the UK, Energy saving lightbulbs are being installed everywhere and people are thinking about how they live affects other people.
The Global Warming bashing will subside once they accept the idea, of course there will always be a energy saving version of the KKK or people telling gils to 'get back to the kitchen' but as a whole teh world has rose above that.
I'd like to thank the Greenpeace and other Org's memebrs that are in prison, been stalked, sued and attacked by big copmpanys for trying to get our attention, the inventer who got trodden on and told to drop his idea because it was energy efficient yet not 'cool enough for modern day markets' and that one guy somewhere that just turned around one day and just said 'we're killing ourself, i must tell someone'.
This seems like a great gimmick to displace consumers of their money in yet another green effort!
To drive the piston you are relying on expansion and contraction of the cooling gas/fluid which equates to a temperature fluxuation (excludes state changes and enthalpy of formation). This fluxuation would invariably change the efficiency of the chip likely in the +/- efficiency range. So the chip would likely draw .5 to 1 more watt if it used 20 watts.
Current fans running at full efficiency use .1-.2 watts. So this technology would likely use more energy than just adding a fan. At best the technology would be a energy wash with significant more cost in building and energy use to create the item!
Temperature fluxuations are dampened by the heat sink. You cannot blindly assume the electric fan maintains a lower average temperature. In a system where both implementations provide equal cooling, the heat powered fan will certainly be more efficient.
The savings are obvious .. a fan powered by your PSU vs a fan powered by wasted heat. While it might personally save you a negligible amount of electricity, there are alot of computers in this world besides yours.
To ameridan:
Temperature fluxuations are not dampened by the heat sink. Go back to you thermo book and look up sterling cycles. The liquid coolant cools absorbs the heat which is is not transfered until isothermal expansion of the cooling liquid. The liquid is then cooled at which time the heating blocks latent heat is building awaiting for the cooled liquid to transfer back. So there very much will be a build up of energy on the chip followed by cooling. Varibilty in temperature equates to variability in energy consumption.
You must look at the thermal efficiencies of the chip vs. the thermal efficiency of the process to decide which approach is more efficient. This sterling engine is only efficient when large temperature gradients are present similiar to heat pump cycles. So this believed savings is often wasted and thats the point. You believe you are recovering energy but at what cost? Chips become less efficient with increased temperature and variability. So you trade chip efficiency for that of an electric fan. Once this calcution is complete you likely will find the electric fan a more efficient use of resources since you'll exchange for load on the chip.
http://video.google.com/videoplay?docid=4654319024851860269&q=stirling+engine+hand&total=28&start=0&num=10&so=0&type=search&plindex=3
Where is this large temperature gradient you speak of? That guys hand must be cycling from uncomfortably hot to icy cold according to you.
The area between the Sterling engine and the chipset is a heat reservoir and provides constant, consistent heat flux. The temperature of the material will cycle but certainly not to the point of affecting chipset efficiency. The cycles are close together and not defined by extreme differences in temperature. The material subjected to these cycles is not an ideal conductor of heat and thus will dampen any cyclic temperatures before they reach the surface of the chip. The heatsink which the fan is blowing on will also absorb / dampen the temperature fluctuations of the material between the chip and the Sterling engine.
Thisis saving on a lot of fronts.
>XDS's 'Cold Boot' idea (see above)
>XDS's 'Cooling after shutdown' (inturn extending CPU life)
>Saving energy from less fans in system
>Saving energy by incresing the efficiency of the CPU, thus getting more work per clock done.
Replying to other comments...
>>"Wow nice. The hotter the CPU the faster it spins, right?
Wrong. The hotter the CPU the sooner the BSOD comes"
nope, your wrong.
The hotter it becomes, the faster the fan spins, the more heat withdrawn, the faster heat is drawn AWAY from the CPU - stopping that BSOD - with the added benifit of not having to wait for a temperature guage to read that its getting hot or using any more energy.
>>"0.1 watt. Will burn more power to produce it than it ever saves..."
Currently fans cost money to make, and to run, this would cost money to make, but not to run. A energy saving lightbulb may be $5 more but saves $10 a year and lasts longer.
>>"If energy cannot be created or destroyed, only transferred from one form to another then the overall energy requirements will be the same for the system, because you still need the electricity to generate the heat, to generate the movement."
This is energy that is currently wasted and not used anyhow, the system wont need to produce more energy to run this, its just recycling the energy wasted of running the chips.
>>"In a sense it's true that keeping it constantly cool would increase the lifetime of the chip due to less 'wear', but with systems being outdated in 6 months currently that's not much of an issue, if it last 10 instead of 12 years doesn't matter."
Go look in a bank, a office or any large scale corp, there still running Pentium 3's and old hack computers cause they do the job they are for, and dont need more electric to do it. What about them peeps that run them for 24/7 yet only upgrade every 5-10 years? even then you can send your chip to a recycling centre to be redistributed at schools, librarys or even another country.
>>"I don't see any practical use for this, you're still limited by the ambient temperature in the case."
Erm.. so why is there fans on CPU's/Northbridge chips then? the chips need to be cooled, the heat from them need to be transfered to the local air first - then the PSU fan (or extra case fan some peeps have) would take it out the box. This design increases the efficiency of the chip to local air exchange.
This is a bad idea. Look at modern high-efficiency furnaces and water heaters for an example:
Heating air to make it move around is a poor use of energy. To get the same amount of air movement, you could spend a lot less energy to just spin a fan. A modern furnace extracts so much heat from the combustion gases that they won't reliably rise up a chimney. So, they use a fan to force the issue. It takes less energy to spin the fan than the additional heat extracted from the gases, so it's a net win.
In this case, you're letting the chip get extra hot in order to produce enough thermal differential to start the engine. If the ambient air happens to be very cold then the chip is still in a comfortable range when this happens, but in most cases, the chip's lifespan will be shortened by the extra heat, causing early failure and replacement, which costs energy well in excess of what you would've just spent to spin a fan in the first place.
When your goal is to move heat, you should do work to move heat. Using the heat to do work generally involves bad trade-offs. That's the case here.
The chip already gets extra hot. This is just converting the heat, which is already generated, to spin a fan.
i would like the heat of my cpu and gpu use converted steam power just like they are trying to do in cars...could have used something like this on by amd desktop processor powered HP laptop that would reach boiling point and shut down...instead of using notebook cooling pad, maybe the heat could be harvested by steam and put to good use
What if you tied all of the heat generating components together with a single heat sync matrix and cooled the whole mess with a larger one of these babies. It would cool the entire system. Good for proprietary desktops of specific design. Perfect for laptops.
Less electrons pumped into the system to move heat out = more efficiency.
The speed of the fan would instantaneously increase/decrease with amount of heat applied to it just as in a thermostatically controlled electric fan. The fan naturally spins at different speeds given the amount of total heat inside (and outside) the system because sterling engines work on temperature differentials.
This doesn't appear to be a CPU cooling solution, but rather a Northbridge one (as implemented by MSI). Given that many NB chips are only cooled by passive heatsinks or pipes, this is a welcome change as long as the size is kept under reasonable limits.
Wow, this is a 'duh'' idea that I am pleasantly surprised to see someone work out. It'd speed up as the processor drew more heat and lose speed as it cooled down, meaning it'd help attain maximum efficiency. Bring this to the market and I will buy it.
In fact, i can see numerous apps other than processors: graphics cards would benefit from reduce power consumption, chipset cooling could work like this....
The possibilities to reduce power consumption here are endless. Good show MSI.
looks like we knocked out the fareastgizmos server... I wanted to read more. Sterling engines are pretty interesting:
http://en.wikipedia.org/wiki/Stirling_engine#Action_of_a_beta_type_Stirling_engine
You can make one out of a couple soda cans and some tubes:
http://www.rotarystirlingengines.com/rotacola.htm
I'm surprised no one thought of this before! Come on now, It's a Sterling Engine. It just moves thermal energy from one place to another. If nothing else, I imagine the steampunks will be crawling all over it!
This is a briliiant idea, and I always wondered why the hell anybody on this Earth could not design anything like this. (I dont mean a motherboard fan that uses chips energy particularly, but anything that uses the energy which is considered to be lost to create something useful efficiently)
Well I have thought this before, and this is just perfect. The world is approaching to a state that energy will be the most valuable gift to humanity. People are trying to use natural sources much less and recycled sources more to get the maximum efficiency out of anything, energy or product or simply waste!
Transferring heat energy which is something unwanted to mechanical energy would be a good place to start. If only someone could come with a efficient liquid cooling system like this...
People may find this product useless or inefficient, but anyone who calls the idea behind it stupid IS stupid.
"In this case, you're letting the chip get extra hot in order to produce enough thermal differential to start the engine."
there will always be a temperature differential, this system takes advantage of it while an electric fan does not. Stirling engines operate at a variety of temperature differentials and do not necessitate that your chip get hotter than it normally does with an electric fan.
I like the idea overall but I see 3 major problems with it.
1) Sterling engines (as this is obviously one) use some kind of a flexible membrane in their design if I recall right. This membrane represents the weakpoint in their design and will likely wear out faster than any of the other parts.
2) If you watch the video, the fan is pulling air in through the fins attached to the heatpipe and then across the top of the unit. Sterling engines operate by having a hot side and a cold side (even if this is only a few degrees difference). Putting the heatsink fins in frotn of the airflow means you will be pulling HOT air across the top of the sterling engine where we see more fins sticking straight up. This will heat up the top of the sterling engine and reduce the efficiency of the engine overall. They simply need to turn the unit around and reverse the direction of the fan (or its blades) which will draw cool(er) air across the top of the sterling engine, producing a colder temp on the top of the engine, and then push that air through the heatsink fins to cool the chipset itself.
3) Sterling engines usually need some kind of a little push to get them started spinning. MSI would need to figure out how to kick start the fan blades for this to work. And the kickstart may have to be delayed 30 seconds after boot since right at boot time the chipset may not be that hot and the fan could stall out if the kick start was sent too early.
For those who said earlier that heat is not being removed from the chipset only used to spin the fan, look at the heatpipes that go up into the fin assembly. That whole peice of this design IS cooling the chipset.
For those who said that the temperature fluctuations will kill the chip faster than a normal fan, I would ask where you think these fluctuations will come from? The chipset will reach a certain temp and the fan will spin at a certain speed and there may be a slight spike in the temp of the chipset while the sterling engine adjusts, but they will both very quickly reach equilibrium and will have a constant temp that I believe is overall lower than the temp of the chipset with a passive cooler attached. And this is all accomplished with NO additional electricity to power the FAN.
And for the Skeptics that say this will never work to cool a CPU, I think your wrong. It may not cool a Q6600 or QX9770, but think of the more energy efficient chips and the fact that you can upscale this design a bit and it should work just fine. Keep in mind that people used to passively cool low end P4s and those things were heaters in themselves.
combine this design with a bigger heatsink fin assembly and bigger fan, turn it around like I said earlier, and it would easily cool a AMD BE-2300 or Intel Mobile Penryn chip all day long.
-Casper42
While I admire the geekiness of the Sterling Engine heatsink, I can see one fatal flaw...
Sterling Engines are notoriously noisy. That defeats most people's desire for a quiet computer, but even more seriously - the mechanical vibration will probably destroy the chip it's attached to.
If you want to build a fun version using some of the same principles, this little paper version takes 5 minutes to make and will garner lots of co-worker comments:
From MAKE: http://blog.makezine.com/archive/2008/02/build_a_heatdriven_monito.html?CMP=OTC-0D6B48984890
@ Generic
No offense, but it sounds like you're taking this the wrong way.
(1) This is not a CPU cooler. The picture shows a fan with a footprint too small for current processors. It'll probably cool the Northbridge/Southbridge, which usually don't need a lot of cooling.
"The CPU needs to be at a certain temperature range to function efficiently. This temperature is not the same for humans." —Generic
(2i) What?! That's definitely the other way around. I don't know about you, but we humans function best at 98.6°F.
"Chilling the CPU is bad and will require more energy to move the electrons (or holes for the physics guys) around the CPU." —Generic
(2ii) CPUs follow the typical physics of metals. Conductivity is inversely proportional to temperature, no exceptions. As a metal (or any substance) approaches absolute zero, the electric conductivity rises. Have you heard of superconductivity? Frozen, compressed helium would conduct better than room-temperature copper. The optimal temperature for energy efficiency and operating conditions is as low as possible (but above 0 K!).
Also, why are you talking about holes? Only when physicists write text books to explain why the current is visualized as running opposite the electron path are holes talked about. K.I.S.S!
And just a note to everyone: The northbridge normally doesn't need active cooling. Even in high-end boards, it'll use the ambient breeze from the CPU fan. Some high-performance-for-price (GA-P35-DS3L *AHEM*) sacrifice quality in that regard and the northbridge doesn't effectively wick away heat, but whatever.
http://www.newegg.com/Product/Product.aspx?Item=N82E16813128059
Looks like a toilet. :D
A couple of observations...
I have a basic understanding of how the sterling theory works.
Looking at what i see , i don't think this is based on that theory whatsoever.
So. I don't know who brought that up in the first place but they could have been a little more explicit on what they where referin to!!!!!
And damnit thunder.... Who ever said they where against the idea ?
Some people are just skeptical is all.
I chastise you for saying such a thing.
Just to comment about the person that mentioned this would need a "kick start" in order to help it work , I could see how this could be a user initiated "knetic" start feature.
EG. Where the power button could pull double duty. , By some form of spring loading action.
And to anyone making rude comments about it looking like a toilet and what not . , I motion that the editor ban your mac address forever. :)
Bye Bye.
Regards,
XDS
A simple low-tech Stirling engine that you can buy and build from a kit can easily spin at several hundred RPM with a big enough heat differential - like from a hot cup of coffee to ambient air temp. The overal efficiency of converting heat to mechanical energy of these machines is very low - a few percent. However a high tech manufactured one could be much higher, and really you're just looking for a few watts to spin a fan to send air over cooling fins which requires low torque. Suitably designed it could easily provide a the kind of RPM your average CPU cooler fan provides. The main problem is it would be way more expensive than a regular electric fan. There may also be problems with getting the fan to self-start - but it depends on the engine design.
Fan's spinning backwards.
The XBOX 360 could use one of these.
isnt that fan spinning backwards?
I'd like to see this on a lower power appliance, like my wireless gateway/router. It only runs on 12V 500ma, but the regulator and wireless chip set get hot. Since adding a powered fan would mean spending more for a bigger power supply, this would fit a really nice niche application. My router lives longer since it is now cooled, I don't spend more money on power supplies or electricity to run the fan, and all I'm out is the engine cost and the effort of taking the top of the case off.
Would make a lot of sense for something like this, where the device runs constantly 24/7/365, maybe for 2 years...
Lets see... small 50ma fan
100ma*12VDC=1.2W/1000=.0012KW/hr. My electricity costs .145 per KW/hr, so for 2 years use that's a savings of .145*.0012*24*365*2=3.05
If the router PS can handle the extra 100ma, I'm not out $20 for a power supply. So that's a negative... If the router lasts an extra 2 years because it's now cooled more effectively, I am saving $6+$70 for the doubled router life (either way).
It might be worth it. I guess to really figure this out, you would look at alternatives, like simply adding aluminum chip set heat sinks and pulling the top off the case.. those run $20 with adhesive, but I don't know how much heat they remove versus the fan, nor how long the life of the product is extended. Nor do I know the cost of the fan or if I can move it to another product when the router finally dies. Some Stirling engines last for years..
Tough call on the numbers alone. But having a Stirling engine fan, well the neato factor makes it an obvious choice!
Oops, I meant 100ma fan. the dinky 50's work but are noisy unless you spend a lot... 100 ma's are generally bigger and cheaper. I got the calculation right, just didn't update the text.
-=RB