While we are being pompous asses, I not only have a degree in physics, I teach thermodynamics. The maximum thermal efficiency of a system is the Carnot efficiency:
eff = (Th-Tc)/Th
where Th is the Kelvin temperature of the heat source and Tc is the Kelvin temperature of the cold reservoir, which would be the fins of a heat sink or some such thing. The only way that this system can have any effect is to use a Seebeck Effect device to replace the computer's fans. You cannot refrigerate the heat sink because the same electricity is used for refrigeration is used to create the heat in the first place.
Yes. Refrigeration is a lossy process. All work is accompanied by inefficiency. The laws of thermodynamics ensure that.
If we really want to be pompus, we can also state that everything in the universe is information, which cannot be created or lost. Energy is simply one manifestation of that, that we can see, and use for work.
It doesn't sound like you're disagreeing with me. I basically said that you could prevent excess heat from escaping the system by having an infinite array of these things... meaning that they would act as a cooling device. No, not a cooling device, since they aren't really cooling anything. Just insulating against heat loss.
It's akin to having a pressurized balloon and poking a tiny pinhole in it as opposed to a pencil sized hole. The device just contains power better, not generate it out of nothing.
And if the processor was just turning electricity directly into heat, it would be a perpetual energy machine, given an infinite array.
-Clive
It isn't insulating either. It's merely translating. The device, as far as I can ascertain right now, doesn't do anything except to absorb heat and turn it into electricity. No heating or cooling.
Even an infinite array wouldn't be a perpetual energy machine, because with the required losses, each chip would have to incur, it would also have an infinite loss for the infinite array.
No, I am not disagreeing with you--at least. This device is supposed to convert heat that would have been vented to the air into electricity. This necessarily makes it a cooling device. It would maintain the computer's processor at a high constant temperature. If it were to lower the temperature of the processor, the device's conversion efficiency efficiency. Assuming that the processor is kept hot enough to boil water (Th=373 K/100 °C) and the cold temperature is about room temperature (Tc=300 K/27 °C), the maximum thermal efficiency of the device would be slightly less than 20%. Real conversion efficiency as high as 15% would be a major breakthrough.
Consider this: Intel has made remarkable progress is reducing the heat produced by its processors and thus their cooling requirements. This technique recovers lost heat by not losing it in the first place. 100% of Intel's recovered heat goes to a useful purpose. The best that Eneco can hope for is about 15%. You do the math.
I don't believe it would cool in a direct way. A certain percent would be emitted as lower frequency waves, which would still contribute to heating. In addition, any electricty that might be directed back to the battery would heat that up as well, from the losses in the recharging process. If it's use to power the machine, the same thing applies.
The main purpose would be to give the batteries greater service life, or help to power the fans for better cooling.
Personally, I think this is a wash as far as cooling goes, unless it's used as a thermionic cooler on the chip, as I mentioned earlier.
I don't believe it would cool in a direct way. A certain percent would be emitted as lower frequency waves, which would still contribute to heating. In addition, any electricty that might be directed back to the battery would heat that up as well, from the losses in the recharging process. If it's use to power the machine, the same thing applies.
The main purpose would be to give the batteries greater service life, or help to power the fans for better cooling.
Personally, I think this is a wash as far as cooling goes, unless it's used as a thermionic cooler on the chip, as I mentioned earlier.
Actually when they made the statement in the article that
Quote:
Alternatively the chip can refrigerate down to -200 degrees celsius when electricity is applied, Brown said.
my first thought is that it would be used for cooling CPUs and the ability to get CPUs to run faster than 4Ghz. After all IBM and other CPU manufacturers (as well as computer engineering nerds) have always bragged how they have gotten CPUs to run 10x or 40x faster in super cooled environments.
Actually when they made the statement in the article that my first thought is that it would be used for cooling CPUs and the ability to get CPUs to run faster than 4Ghz. After all IBM and other CPU manufacturers (as well as computer engineering nerds) have always bragged how they have gotten CPUs to run 10x or 40x faster in super cooled environments.
I said in two posts, that if it were to be used as a thermionic device to directly cool the chip, that could work.
But, the other function, that of translating heat energy into electricity won't lead to cooling by itself.
What has to be understood, in these conversations, is that eventually ALL of the energy expended is turned back into heat. The heat is merely moved about, and/or released at different times.
It isn't insulating either. It's merely translating. The device, as far as I can ascertain right now, doesn't do anything except to absorb heat and turn it into electricity. No heating or cooling.
....
This statement violates the Second Law of Thermodynamics. No process can convert heat into work--in this case electrical work--at a single temperature. To convert heat to work, you must lower its temperature.
This statement violates the Second Law of Thermodynamics. No process can convert heat into work--in this case electrical work--at a single temperature. To convert heat to work, you must lower its temperature.
I'm not violating the second law. That's what is happening. It just doesn't RESULT in cooling the unit, because it turns back into heat somewhere else inside.
First, the heat turns into power, which then turns back into heat. Nothing lost, nothing gained in the long run from those particular quanta.
Forgive me if this isn't precisely on topic, but would this technology be more useful in Hybrid Cars?
I mean, combustion engines create heat, couldn't some of that heat be recycled to batteries?
Well a few years ago Garrett had a prototype turbocharger that contained an electric motor/generator. A turbo's turbine sits in the exhaust stream and is spun primarily based on the heat flowing out from the engine (well, by the super-heated air expanding out from the engine anyhow). In place of the typical conventional wastegate (used to dump the excess pressure/heat and control turbine speed), it used essentially "regenerative braking" to keep the turbine speed down and recharge the car's battery at the same time. When first spooling the turbine the electric motor was used to accelerate it faster than just the exhaust gas could -- generating boost sooner and increasing the outflowing exhaust much more quickly, reducing turbo lag.
Hmmm... I've also read about on-chip micro-machines like turbines that have been built in the last few years. Perhaps these guys are just building an electrically assisted turbo-charger for your heat sink.
Hmmm... I've also read about on-chip micro-machines like turbines that have been built in the last few years. Perhaps these guys are just building an electrically assisted turbo-charger for your heat sink.
They've also built perpetual motion machines that work by harnessing Brownian Motion.
Apple can NOT be happy about this story. When startups show their technologies to companies like APPLE, they often sign NDA (Non Disclosure Agreements) to protect the privacy of these discussions.
This start up is simply using Apple and Dell's names to get a boost (most likely looking for 2nd round investment) I'm sure they have received a direct comment from Apple, as not to use their name anymore (or NEVER do business with Apple)
That said, the technology (if it works) sounds great. Would really change how people think about cooling. That said, why aren't they quoting Intel? If their chip has to be added a processor, they would have to work VEYR closely with Intel, AMD to integrate. perhaps, they're hoping Intel will just buy them out.
Hope they make a few million. (and they can spend it on a better website. http://www.eneco.com/)
Hmm, Apple doesn't own this technology and I doubt they signed a NDA with apple, why would they?. It's their technology and if apple decided not to use it, they could jump off a building for all this company is concerned. Ever heard of cell phones?. pda's, other computer companies?.. If this company can make their tech small, they have a potential market of millions regardless of whether apple is on board or not. I'm sure it would be nice to have apple as a customer (who doesn't want more money) but this company can hype their technology as they see fit, irrespective of apple feelings. This is not the ipod market guy, apple doesn't dominate the battery or power market (heck, this tech could be applied even to automobiles, etc.. forget potential market of millions, biliions possibly, and this is without apple participation).
[The only way that this system can have any effect is to use a Seebeck Effect device to replace the computer's fans. You cannot refrigerate the heat sink because the same electricity is used for refrigeration is used to create the heat in the first place.[/QUOTE]
It seems to me the most likely way to engineer this is to place "The Device" between the CPU and the heat sink (physical heat sink, not the physics virtual heat sink). It then has access to the highest temperature, and then it is up to the thermal engineers to figure out how deliver the lowest temperature in the ways they already do (fans, placement etc.) to maximize the current output of The Device. The Device itself becomes part of the cooling system, and the biproduct is current which most likely would go towards charging the batteries, and be distributed by normal means.
A previous post suggested putting these in series to improve overall efficency, which is a reasonable idea.
The big drawback to this, is that The Device will impede heat removal, which is a risk for the CPU. Unless...
...you feed current into the device. If designed properly, it should refrigerate. So, like regenerative brakes on a electric car, you could maintain the CPU at a constant temperature. Under low use, the heat would supplement battery power. Under high use, battery power could be used for cooling. Perhaps this would allow for elimination of fans entirely, but only if you could still provide an adequate low temp to the heat sink.
No, I am not disagreeing with you--at least. This device is supposed to convert heat that would have been vented to the air into electricity. This necessarily makes it a cooling device. It would maintain the computer's processor at a high constant temperature. If it were to lower the temperature of the processor, the device's conversion efficiency efficiency. Assuming that the processor is kept hot enough to boil water (Th=373 K/100 °C) and the cold temperature is about room temperature (Tc=300 K/27 °C), the maximum thermal efficiency of the device would be slightly less than 20%. Real conversion efficiency as high as 15% would be a major breakthrough.
Consider this: Intel has made remarkable progress is reducing the heat produced by its processors and thus their cooling requirements. This technique recovers lost heat by not losing it in the first place. 100% of Intel's recovered heat goes to a useful purpose. The best that Eneco can hope for is about 15%. You do the math.
No, there is no desire to maintain the CPU temp or insulate it this way in any case. The whole point is to extract heat from the CPU at the maximum rate possible. All other goals are secondary to that unless you wish to melt a silicon layer and short the chip out.
Since the amount of reclaimed electricity is secondary to the cooling, we explicitly are accepting a lower heat-to-electricity efficiency as we cool the CPU.
No matter what, we need to maintain a heat gradient since the converter chip is not 100% efficient, so keeping the heat-sink and placing the converter between the CPU and heat sink guarantees a flow of heat out of the CPU. Whatever is recaptured to electricity is metaphorical gravy. We use the recaptured electricity to power the fans since fans will be needed most when the CPU is producing the most heat and electricity. Those fans help us cool the heat sink to maintain the heat gradient in the direction we want. The shed heat will still make the laptop warm, but a just little less warm than they are now. We don't create or eliminate heat, just store it back into the battery as charge, or minimally reduce the battery output, which either way causes slightly less new heat to be created to power the fan motors. The amount battery output per unit time is reduced will roughly equate to the lower heat output per unit time of the entire laptop system. No changes in the total amount of heat the battery would generate while discharging to zero, it just gets spread out over time. All laws of conservation are adhered to.
I'm going to speculate these folks are proposing their chip between the CPU and heatsink. If it was going to be part of the CPU package they would work with Intel directly and Apple/Dell would just get it from Intel.
[The only way that this system can have any effect is to use a Seebeck Effect device to replace the computer's fans. You cannot refrigerate the heat sink because the same electricity is used for refrigeration is used to create the heat in the first place.
It seems to me the most likely way to engineer this is to place "The Device" between the CPU and the heat sink (physical heat sink, not the physics virtual heat sink). It then has access to the highest temperature, and then it is up to the thermal engineers to figure out how deliver the lowest temperature in the ways they already do (fans, placement etc.) to maximize the current output of The Device. The Device itself becomes part of the cooling system, and the biproduct is current which most likely would go towards charging the batteries, and be distributed by normal means.
A previous post suggested putting these in series to improve overall efficency, which is a reasonable idea.
The big drawback to this, is that The Device will impede heat removal, which is a risk for the CPU. Unless...
...you feed current into the device. If designed properly, it should refrigerate. So, like regenerative brakes on a electric car, you could maintain the CPU at a constant temperature. Under low use, the heat would supplement battery power. Under high use, battery power could be used for cooling. Perhaps this would allow for elimination of fans entirely, but only if you could still provide an adequate low temp to the heat sink.
Phred
Be careful of loosing sight of the primary goal of a CPU cooling system, removing heat.
Period.
Putting these things in series will almost assuredly cause a heat backlog killing the CPU. But it makes a lot of sense to use these as part of the cooling system and just getting out of it whatever you can without jeopardizing CPU lifetimes.
Also remember that there already are piezoelectric coolers available, and they eat plenty of current. Hardly what you want for a laptop and exactly the opposite energy draw of what these new chips are supposed to accomplish. Keep it simple and don't try to build an Edsel here.
No, there is no desire to maintain the CPU temp or insulate it this way in any case. The whole point is to extract heat from the CPU at the maximum rate possible. All other goals are secondary to that unless you wish to melt a silicon layer and short the chip out.
Since the amount of reclaimed electricity is secondary to the cooling, we explicitly are accepting a lower heat-to-electricity efficiency as we cool the CPU.
....
You have not shed any new light on this device. The device is supposed to capture heat from the hot processor and convert it to electricity as an aid in cooling the processor. I don't believe that this device can help cool the processor any more efficiently than the passive and forced convection cooling mechanisms currently used. I believe that the opposite is true. Nothing you said so far makes me want to reconsider my position.
Be careful of loosing sight of the primary goal of a CPU cooling system, removing heat.
Period.
Putting these things in series will almost assuredly cause a heat backlog killing the CPU. But it makes a lot of sense to use these as part of the cooling system and just getting out of it whatever you can without jeopardizing CPU lifetimes.
Also remember that there already are piezoelectric coolers available, and they eat plenty of current. Hardly what you want for a laptop and exactly the opposite energy draw of what these new chips are supposed to accomplish. Keep it simple and don't try to build an Edsel here.
Hey Hiro:
Obviously, we need to keep the CPU below the max operating temp. Obviously, the primary goal of a "cooling system" is cooling. However, cooling is a secondary and contingent goal of the system known as "laptop".
Your entire premise is wrong:
If...
Quote:
Originally Posted by Hiro
The whole point is to extract heat from the CPU at the maximum rate possible. All other goals are secondary to that
as you state in your first post then we would add a trailer and a water tank to the laptop.
The primary goal of "laptop" is to provide maximum functionality which I take to mean maximum computing power allowable for a reasonable period of time, with acceptable weight and cost. If we have to keep the CPU just this side of Tmax to achieve that, then that is fine.
Contrary to your premise, the trend is to increase CPU temperature, not minimize it. That is the purpose of variable speed fans, on-chip temp circuitry etc. -- to unshackle us from worst-case cooling needs. We allow the temp of the CPU to be HOTTER during lower CPU demand, because in that situation hotter is still below Tmax, and we reduce the power consumption needed for cooling. All of that added technology and engineering is very expensive, but worth it once you are done with it because it improves the performance of Laptop against its PRIMARY goal.
There is nothing in my post that is not in direct accordance with the laws of physics. All I've done is propose one way that such a system might be engineered. I suggested that "The Device" go between chip and heat sink because I had read only other suggestions, which seem inferior to me. I see that we concur on that.
I suggested that if The Device supports cooling, which isn't a stretch, then we could use it to even out the power/cooling/temperature usage. This is the industry approach -- minimize unnecessary cooling to improve maximum cooling -- with a twist.
This has nothing to do with status quo ante piezo coolers. The premise of this thread is that there is new, much more efficient technology, so I fail to see the relevance.
As for building an Edsel... Dude! You are letting your inner Luddite show. Yes my '68 Firebird was easier to design and easier to work on than my WRX, but it gave less performance using more resources. Yes, that old school mechanics may scoff at all the fancy crap under the hood isn't important. Old school mechanics are only good to add texture to lame movies. Cool stuff is often complicated.
You have not shed any new light on this device. The device is supposed to capture heat from the hot processor and convert it to electricity as an aid in cooling the processor. I don't believe that this device can help cool the processor any more efficiently than the passive and forced convection cooling mechanisms currently used. I believe that the opposite is true. Nothing you said so far makes me want to reconsider my position.
This of course is the nut -- even if the technology works, does it provide an advantage? It might, simply because it provides a way to recover otherwise lost energy, improving the overall efficiency of the system, but possibly with more precision than fan speed. Same net effect.
The main purpose of this device is to provide a dump for the heat, some of which is then turned into electricity, the rest is radiated away.
By itself, the way they intend to use it, it will not remove any more heat than a heatsink and fans. Anyone who thinks that it will, is simply wrong.
However, there is a mode that will do that, it just doesn't create electricity in that mode. It's an electronic manifestation of the metallic thermionic devices around today that cool on one side, and in turn, heat the other. Apply power to these devices, and one side gets cold, and the other hot. This could do that as well, but then won't generate power.
That sentence describes the post it emanates from. You just don't get it or read very well, or is is some innate defensiveness that prevents you from thinking straight? Your entire post couldn't be any more successful at missing the points it was responding too. So much so there isn't anything factual I can even respond with.
Comments
While we are being pompous asses, I not only have a degree in physics, I teach thermodynamics. The maximum thermal efficiency of a system is the Carnot efficiency:
eff = (Th-Tc)/Th
where Th is the Kelvin temperature of the heat source and Tc is the Kelvin temperature of the cold reservoir, which would be the fins of a heat sink or some such thing. The only way that this system can have any effect is to use a Seebeck Effect device to replace the computer's fans. You cannot refrigerate the heat sink because the same electricity is used for refrigeration is used to create the heat in the first place.
Yes. Refrigeration is a lossy process. All work is accompanied by inefficiency. The laws of thermodynamics ensure that.
If we really want to be pompus, we can also state that everything in the universe is information, which cannot be created or lost. Energy is simply one manifestation of that, that we can see, and use for work.
It doesn't sound like you're disagreeing with me. I basically said that you could prevent excess heat from escaping the system by having an infinite array of these things... meaning that they would act as a cooling device. No, not a cooling device, since they aren't really cooling anything. Just insulating against heat loss.
It's akin to having a pressurized balloon and poking a tiny pinhole in it as opposed to a pencil sized hole. The device just contains power better, not generate it out of nothing.
And if the processor was just turning electricity directly into heat, it would be a perpetual energy machine, given an infinite array.
-Clive
It isn't insulating either. It's merely translating. The device, as far as I can ascertain right now, doesn't do anything except to absorb heat and turn it into electricity. No heating or cooling.
Even an infinite array wouldn't be a perpetual energy machine, because with the required losses, each chip would have to incur, it would also have an infinite loss for the infinite array.
No, I am not disagreeing with you--at least. This device is supposed to convert heat that would have been vented to the air into electricity. This necessarily makes it a cooling device. It would maintain the computer's processor at a high constant temperature. If it were to lower the temperature of the processor, the device's conversion efficiency efficiency. Assuming that the processor is kept hot enough to boil water (Th=373 K/100 °C) and the cold temperature is about room temperature (Tc=300 K/27 °C), the maximum thermal efficiency of the device would be slightly less than 20%. Real conversion efficiency as high as 15% would be a major breakthrough.
Consider this: Intel has made remarkable progress is reducing the heat produced by its processors and thus their cooling requirements. This technique recovers lost heat by not losing it in the first place. 100% of Intel's recovered heat goes to a useful purpose. The best that Eneco can hope for is about 15%. You do the math.
I don't believe it would cool in a direct way. A certain percent would be emitted as lower frequency waves, which would still contribute to heating. In addition, any electricty that might be directed back to the battery would heat that up as well, from the losses in the recharging process. If it's use to power the machine, the same thing applies.
The main purpose would be to give the batteries greater service life, or help to power the fans for better cooling.
Personally, I think this is a wash as far as cooling goes, unless it's used as a thermionic cooler on the chip, as I mentioned earlier.
I don't believe it would cool in a direct way. A certain percent would be emitted as lower frequency waves, which would still contribute to heating. In addition, any electricty that might be directed back to the battery would heat that up as well, from the losses in the recharging process. If it's use to power the machine, the same thing applies.
The main purpose would be to give the batteries greater service life, or help to power the fans for better cooling.
Personally, I think this is a wash as far as cooling goes, unless it's used as a thermionic cooler on the chip, as I mentioned earlier.
Actually when they made the statement in the article that
Alternatively the chip can refrigerate down to -200 degrees celsius when electricity is applied, Brown said.
my first thought is that it would be used for cooling CPUs and the ability to get CPUs to run faster than 4Ghz. After all IBM and other CPU manufacturers (as well as computer engineering nerds) have always bragged how they have gotten CPUs to run 10x or 40x faster in super cooled environments.
Actually when they made the statement in the article that my first thought is that it would be used for cooling CPUs and the ability to get CPUs to run faster than 4Ghz. After all IBM and other CPU manufacturers (as well as computer engineering nerds) have always bragged how they have gotten CPUs to run 10x or 40x faster in super cooled environments.
I said in two posts, that if it were to be used as a thermionic device to directly cool the chip, that could work.
But, the other function, that of translating heat energy into electricity won't lead to cooling by itself.
What has to be understood, in these conversations, is that eventually ALL of the energy expended is turned back into heat. The heat is merely moved about, and/or released at different times.
It isn't insulating either. It's merely translating. The device, as far as I can ascertain right now, doesn't do anything except to absorb heat and turn it into electricity. No heating or cooling.
....
This statement violates the Second Law of Thermodynamics. No process can convert heat into work--in this case electrical work--at a single temperature. To convert heat to work, you must lower its temperature.
This statement violates the Second Law of Thermodynamics. No process can convert heat into work--in this case electrical work--at a single temperature. To convert heat to work, you must lower its temperature.
I'm not violating the second law. That's what is happening. It just doesn't RESULT in cooling the unit, because it turns back into heat somewhere else inside.
First, the heat turns into power, which then turns back into heat. Nothing lost, nothing gained in the long run from those particular quanta.
Forgive me if this isn't precisely on topic, but would this technology be more useful in Hybrid Cars?
I mean, combustion engines create heat, couldn't some of that heat be recycled to batteries?
Well a few years ago Garrett had a prototype turbocharger that contained an electric motor/generator. A turbo's turbine sits in the exhaust stream and is spun primarily based on the heat flowing out from the engine (well, by the super-heated air expanding out from the engine anyhow). In place of the typical conventional wastegate (used to dump the excess pressure/heat and control turbine speed), it used essentially "regenerative braking" to keep the turbine speed down and recharge the car's battery at the same time. When first spooling the turbine the electric motor was used to accelerate it faster than just the exhaust gas could -- generating boost sooner and increasing the outflowing exhaust much more quickly, reducing turbo lag.
Hmmm... I've also read about on-chip micro-machines like turbines that have been built in the last few years. Perhaps these guys are just building an electrically assisted turbo-charger for your heat sink.
Hmmm... I've also read about on-chip micro-machines like turbines that have been built in the last few years. Perhaps these guys are just building an electrically assisted turbo-charger for your heat sink.
They've also built perpetual motion machines that work by harnessing Brownian Motion.
Apple can NOT be happy about this story. When startups show their technologies to companies like APPLE, they often sign NDA (Non Disclosure Agreements) to protect the privacy of these discussions.
This start up is simply using Apple and Dell's names to get a boost (most likely looking for 2nd round investment) I'm sure they have received a direct comment from Apple, as not to use their name anymore (or NEVER do business with Apple)
That said, the technology (if it works) sounds great. Would really change how people think about cooling. That said, why aren't they quoting Intel? If their chip has to be added a processor, they would have to work VEYR closely with Intel, AMD to integrate. perhaps, they're hoping Intel will just buy them out.
Hope they make a few million. (and they can spend it on a better website. http://www.eneco.com/)
Hmm, Apple doesn't own this technology and I doubt they signed a NDA with apple, why would they?. It's their technology and if apple decided not to use it, they could jump off a building for all this company is concerned. Ever heard of cell phones?. pda's, other computer companies?.. If this company can make their tech small, they have a potential market of millions regardless of whether apple is on board or not. I'm sure it would be nice to have apple as a customer (who doesn't want more money) but this company can hype their technology as they see fit, irrespective of apple feelings. This is not the ipod market guy, apple doesn't dominate the battery or power market (heck, this tech could be applied even to automobiles, etc.. forget potential market of millions, biliions possibly, and this is without apple participation).
It seems to me the most likely way to engineer this is to place "The Device" between the CPU and the heat sink (physical heat sink, not the physics virtual heat sink). It then has access to the highest temperature, and then it is up to the thermal engineers to figure out how deliver the lowest temperature in the ways they already do (fans, placement etc.) to maximize the current output of The Device. The Device itself becomes part of the cooling system, and the biproduct is current which most likely would go towards charging the batteries, and be distributed by normal means.
A previous post suggested putting these in series to improve overall efficency, which is a reasonable idea.
The big drawback to this, is that The Device will impede heat removal, which is a risk for the CPU. Unless...
...you feed current into the device. If designed properly, it should refrigerate. So, like regenerative brakes on a electric car, you could maintain the CPU at a constant temperature. Under low use, the heat would supplement battery power. Under high use, battery power could be used for cooling. Perhaps this would allow for elimination of fans entirely, but only if you could still provide an adequate low temp to the heat sink.
Phred
No, I am not disagreeing with you--at least. This device is supposed to convert heat that would have been vented to the air into electricity. This necessarily makes it a cooling device. It would maintain the computer's processor at a high constant temperature. If it were to lower the temperature of the processor, the device's conversion efficiency efficiency. Assuming that the processor is kept hot enough to boil water (Th=373 K/100 °C) and the cold temperature is about room temperature (Tc=300 K/27 °C), the maximum thermal efficiency of the device would be slightly less than 20%. Real conversion efficiency as high as 15% would be a major breakthrough.
Consider this: Intel has made remarkable progress is reducing the heat produced by its processors and thus their cooling requirements. This technique recovers lost heat by not losing it in the first place. 100% of Intel's recovered heat goes to a useful purpose. The best that Eneco can hope for is about 15%. You do the math.
No, there is no desire to maintain the CPU temp or insulate it this way in any case. The whole point is to extract heat from the CPU at the maximum rate possible. All other goals are secondary to that unless you wish to melt a silicon layer and short the chip out.
Since the amount of reclaimed electricity is secondary to the cooling, we explicitly are accepting a lower heat-to-electricity efficiency as we cool the CPU.
No matter what, we need to maintain a heat gradient since the converter chip is not 100% efficient, so keeping the heat-sink and placing the converter between the CPU and heat sink guarantees a flow of heat out of the CPU. Whatever is recaptured to electricity is metaphorical gravy. We use the recaptured electricity to power the fans since fans will be needed most when the CPU is producing the most heat and electricity. Those fans help us cool the heat sink to maintain the heat gradient in the direction we want. The shed heat will still make the laptop warm, but a just little less warm than they are now. We don't create or eliminate heat, just store it back into the battery as charge, or minimally reduce the battery output, which either way causes slightly less new heat to be created to power the fan motors. The amount battery output per unit time is reduced will roughly equate to the lower heat output per unit time of the entire laptop system. No changes in the total amount of heat the battery would generate while discharging to zero, it just gets spread out over time. All laws of conservation are adhered to.
I'm going to speculate these folks are proposing their chip between the CPU and heatsink. If it was going to be part of the CPU package they would work with Intel directly and Apple/Dell would just get it from Intel.
[The only way that this system can have any effect is to use a Seebeck Effect device to replace the computer's fans. You cannot refrigerate the heat sink because the same electricity is used for refrigeration is used to create the heat in the first place.
It seems to me the most likely way to engineer this is to place "The Device" between the CPU and the heat sink (physical heat sink, not the physics virtual heat sink). It then has access to the highest temperature, and then it is up to the thermal engineers to figure out how deliver the lowest temperature in the ways they already do (fans, placement etc.) to maximize the current output of The Device. The Device itself becomes part of the cooling system, and the biproduct is current which most likely would go towards charging the batteries, and be distributed by normal means.
A previous post suggested putting these in series to improve overall efficency, which is a reasonable idea.
The big drawback to this, is that The Device will impede heat removal, which is a risk for the CPU. Unless...
...you feed current into the device. If designed properly, it should refrigerate. So, like regenerative brakes on a electric car, you could maintain the CPU at a constant temperature. Under low use, the heat would supplement battery power. Under high use, battery power could be used for cooling. Perhaps this would allow for elimination of fans entirely, but only if you could still provide an adequate low temp to the heat sink.
Phred
Be careful of loosing sight of the primary goal of a CPU cooling system, removing heat.
Period.
Putting these things in series will almost assuredly cause a heat backlog killing the CPU. But it makes a lot of sense to use these as part of the cooling system and just getting out of it whatever you can without jeopardizing CPU lifetimes.
Also remember that there already are piezoelectric coolers available, and they eat plenty of current. Hardly what you want for a laptop and exactly the opposite energy draw of what these new chips are supposed to accomplish. Keep it simple and don't try to build an Edsel here.
No, there is no desire to maintain the CPU temp or insulate it this way in any case. The whole point is to extract heat from the CPU at the maximum rate possible. All other goals are secondary to that unless you wish to melt a silicon layer and short the chip out.
Since the amount of reclaimed electricity is secondary to the cooling, we explicitly are accepting a lower heat-to-electricity efficiency as we cool the CPU.
....
You have not shed any new light on this device. The device is supposed to capture heat from the hot processor and convert it to electricity as an aid in cooling the processor. I don't believe that this device can help cool the processor any more efficiently than the passive and forced convection cooling mechanisms currently used. I believe that the opposite is true. Nothing you said so far makes me want to reconsider my position.
Be careful of loosing sight of the primary goal of a CPU cooling system, removing heat.
Period.
Putting these things in series will almost assuredly cause a heat backlog killing the CPU. But it makes a lot of sense to use these as part of the cooling system and just getting out of it whatever you can without jeopardizing CPU lifetimes.
Also remember that there already are piezoelectric coolers available, and they eat plenty of current. Hardly what you want for a laptop and exactly the opposite energy draw of what these new chips are supposed to accomplish. Keep it simple and don't try to build an Edsel here.
Hey Hiro:
Obviously, we need to keep the CPU below the max operating temp. Obviously, the primary goal of a "cooling system" is cooling. However, cooling is a secondary and contingent goal of the system known as "laptop".
Your entire premise is wrong:
If...
The whole point is to extract heat from the CPU at the maximum rate possible. All other goals are secondary to that
as you state in your first post then we would add a trailer and a water tank to the laptop.
The primary goal of "laptop" is to provide maximum functionality which I take to mean maximum computing power allowable for a reasonable period of time, with acceptable weight and cost. If we have to keep the CPU just this side of Tmax to achieve that, then that is fine.
Contrary to your premise, the trend is to increase CPU temperature, not minimize it. That is the purpose of variable speed fans, on-chip temp circuitry etc. -- to unshackle us from worst-case cooling needs. We allow the temp of the CPU to be HOTTER during lower CPU demand, because in that situation hotter is still below Tmax, and we reduce the power consumption needed for cooling. All of that added technology and engineering is very expensive, but worth it once you are done with it because it improves the performance of Laptop against its PRIMARY goal.
There is nothing in my post that is not in direct accordance with the laws of physics. All I've done is propose one way that such a system might be engineered. I suggested that "The Device" go between chip and heat sink because I had read only other suggestions, which seem inferior to me. I see that we concur on that.
I suggested that if The Device supports cooling, which isn't a stretch, then we could use it to even out the power/cooling/temperature usage. This is the industry approach -- minimize unnecessary cooling to improve maximum cooling -- with a twist.
This has nothing to do with status quo ante piezo coolers. The premise of this thread is that there is new, much more efficient technology, so I fail to see the relevance.
As for building an Edsel... Dude! You are letting your inner Luddite show. Yes my '68 Firebird was easier to design and easier to work on than my WRX, but it gave less performance using more resources. Yes, that old school mechanics may scoff at all the fancy crap under the hood isn't important. Old school mechanics are only good to add texture to lame movies. Cool stuff is often complicated.
Phred
You have not shed any new light on this device. The device is supposed to capture heat from the hot processor and convert it to electricity as an aid in cooling the processor. I don't believe that this device can help cool the processor any more efficiently than the passive and forced convection cooling mechanisms currently used. I believe that the opposite is true. Nothing you said so far makes me want to reconsider my position.
This of course is the nut -- even if the technology works, does it provide an advantage? It might, simply because it provides a way to recover otherwise lost energy, improving the overall efficiency of the system, but possibly with more precision than fan speed. Same net effect.
Phred
By itself, the way they intend to use it, it will not remove any more heat than a heatsink and fans. Anyone who thinks that it will, is simply wrong.
However, there is a mode that will do that, it just doesn't create electricity in that mode. It's an electronic manifestation of the metallic thermionic devices around today that cool on one side, and in turn, heat the other. Apply power to these devices, and one side gets cold, and the other hot. This could do that as well, but then won't generate power.
Love it. Any chance of a new smiley that looks like a donkey we could use?
BTW, this is a very informative discussion (not that I have anything useful to add here...
Now THAT's a great idea! !@!
Hey Hiro:
<snip>
Your entire premise is wrong:
<snip>
Phred
That sentence describes the post it emanates from. You just don't get it or read very well, or is is some innate defensiveness that prevents you from thinking straight? Your entire post couldn't be any more successful at missing the points it was responding too. So much so there isn't anything factual I can even respond with.