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.
It's not often I agree with mel in an unqualified way, but this is one of those times.
I had this suggestion back in August, looks like someone is finally stealing my idea. It'll probably cost them millions just to patent this and thus the payoff will never be worth it.
I had this suggestion back in August, looks like someone is finally stealing my idea. It'll probably cost them millions just to patent this and thus the payoff will never be worth it.
I had this suggestion back in August, looks like someone is finally stealing my idea. It'll probably cost them millions just to patent this and thus the payoff will never be worth it.
I would imagine this being used as a self-adjusting cooling system. You connect one chip who takes heat and outputs electricity and connect it to an another chip who takes electricity and cools. If it gets hot, then it will be automatically adjusted! The warmer, the more electricity to cooling! No fans!
I would imagine this being used as a self-adjusting cooling system. You connect one chip who takes heat and outputs electricity and connect it to an another chip who takes electricity and cools. If it gets hot, then it will be automatically adjusted! The warmer, the more electricity to cooling! No fans!
What do you think?
You get a downward spiral. Remember, everything has inefficiencies. It isn't possible to avoid them.
What will happen, without outside input, is that the inefficiencies of both devices will lead to less output as time goes on, leading to a very low level of work. All the excess produced by the inefficiencies will exit as heat.
Total gain?zero.
Also, don't forget that you will need fans to exhaust that heat, requiring more power, and heat, being produced.
It's the fans that do the actual work of removing the heat from the unit.
As long as fans are required, this doesn't result in real benefit, except for what they are claiming, longer battery usage before recharges.
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.
Say what? How does one remove heat from a system (by converting into electricity) and NOT cool the environment it exists in? Wherever the optimum placement (before or after the heat sink, fan or both) you are making the job easier for cooling system as a whole.
For Migo's scenario, rather than hooking up two of these things (one in power conversion and the other in cooling mode) you hook the power conversion one to the fan (as suggested earlier). The hotter it is the faster the fan spins and 16% of the total heat inside the laptop is converted to mechanical work moving the fan blades around (assuming around 90% efficiency for the fan).
Edit: Removed reference to electrical potential energy when the electricity is stored in the battery...there is no battery if the thing works as claimed. Any waste heat converted gets used right away anyway. Heat reduction in the system is gained from burning less fuel in the primary power pack to power the device.
Say what? How does one remove heat from a system (by converting into electricity) and NOT cool the environment it exists in? Wherever the optimum placement (before or after the heat sink, fan or both) you are making the job easier for cooling system as a whole.
For Migo's scenario, rather than hooking up two of these things (one in power conversion and the other in cooling mode) you hook the power conversion one to the fan (as suggested earlier). The hotter it is the faster the fan spins and 16% of the total heat inside the laptop is converted to mechanical work moving the fan blades around (assuming around 90% efficiency for the fan).
Edit: Removed reference to electrical potential energy when the electricity is stored in the battery...there is no battery if the thing works as claimed. Any waste heat converted gets used right away anyway. Heat reduction in the system is gained from burning less fuel in the primary power pack to power the device.
Because you aren't "removing" the heat. You are converting it into power, which when used, produces heat. From the losses in the system, heat is always produced. You are removing it with fans, which use power, and produce heat.
So the operation of the fans do remove heat (From the cpu, as well as their own, but not that produced by the battery for the power used for the fans.), but the conversion process does not.
There's an assumption you are making that the device will power the fans directly. I'm not aware of any information that's been given out that suggests that can be done. Perhaps that's why they intend to funnel the power produced back into the battery.
Also, there is an approximate 20% power loss upon conversion. We don't know how theoretical that is. Will the final product be better or worse?
Migo's suggestion is less efficient because there would be two conversions going on, lowering the overall efficiency even further. 20% to convert to power, and 20% from the other to cool.
But, the cooling unit does not produce power, so it won't work anyway.
Because you aren't "removing" the heat. You are converting it into power, which when used, produces heat. From the losses in the system, heat is always produced. You are removing it with fans, which use power, and produce heat.
If you convert 20% of the heat into electricity and then use that energy to drive a motor that is 90% efficient then you've converted that 20% heat into 18% mechanical work and 2% heat. Heat IS produced but less than the input heat. And its driving the cooling system to boot. The energy remains constant, the heat does not.
Well sort of constant...you're pushing energy out in two forms: heat transfer to the air and then moving that air out of the system.
Quote:
So the operation of the fans do remove heat (From the cpu, as well as their own, but not that produced by the battery for the power used for the fans.), but the conversion process does not.
If you convert 20% of the heat into electricity and store that in a battery then you've converted 20% of the heat into electrical potential energy - inefficiencies of charging a battery lost as heat. How much waste heat is generated later depends on what I do with that stored energy. If expended to do mechanical work then some lesser amount of energy is released as heat.
Quote:
There's an assumption you are making that the device will power the fans directly. I'm not aware of any information that's been given out that suggests that can be done. Perhaps that's why they intend to funnel the power produced back into the battery.
So you have an electrical source that can't at least help drive an electric motor? And there likely wouldn't BE a battery if the technology works. The power density of this device is much higher than LiOn or LiPos. With a 20% efficient thermal diode they expect a run time of 12 hours @ 80W for a .6L volume pack in comparison to 3 hours with a battery. The primary issue there is the dispersal of waste heat from the power pack within a volume as small as a laptop.
Quote:
Also, there is an approximate 20% power loss upon conversion. We don't know how theoretical that is. Will the final product be better or worse?
They show 40% of ideal Carnot limit in their current prototypes. That's mighty impressive. And its 20% efficient at 400C/20C gradient for 80% losses. That's why I would be concerned for the heat dispersion in their power pack in small volumes.
The gradient requirement is also why you can't just layer these things as suggested in an earlier post.
They've been talking about this for 4-5 years now and if the science was hinkey someone would have called them on it. Folks still remember cold fusion.
Quote:
Migo's suggestion is less efficient because there would be two conversions going on, lowering the overall efficiency even further. 20% to convert to power, and 20% from the other to cool.
But, the cooling unit does not produce power, so it won't work anyway.
One reason to use the cooling option is to supercool the CPU for better performance. Connect the hot side to a heat pipe and you might as well scavenge that heat for (electrical) energy production. So if you use two units, you've increased the cost of the laptop but reduced power consumption of the cooling thermal diode by some 20% (well sorta but you get the idea). While the total energy gain is still negative the performance boost may be worth it. Plus if you really have 500W-h you have power to burn relative to today's notebooks.
If you convert 20% of the heat into electricity and then use that energy to drive a motor that is 90% efficient then you've converted that 20% heat into 18% mechanical work and 2% heat. Heat IS produced but less than the input heat. And its driving the cooling system to boot. The energy remains constant, the heat does not.
Well sort of constant...you're pushing energy out in two forms: heat transfer to the air and then moving that air out of the system.
I'd love to see one of these fan motors at 90% efficiency, but it's closer to 70%.
Quote:
If you convert 20% of the heat into electricity and store that in a battery then you've converted 20% of the heat into electrical potential energy - inefficiencies of charging a battery lost as heat. How much waste heat is generated later depends on what I do with that stored energy. If expended to do mechanical work then some lesser amount of energy is released as heat.
Sort of.
Quote:
So you have an electrical source that can't at least help drive an electric motor?
Motors require a constant source of current. The cpu won't be giving off heat that is convertible evenly during use. The voltage must also be correct. Motors generate a back EMF, which would require additional circuitry to protect the device, and regulate the voltage for the motor.. That will add to the inefficiency.
Quote:
And there likely wouldn't BE a battery if the technology works. The power density of this device is much higher than LiOn or LiPos. With a 20% efficient thermal diode they expect a run time of 12 hours @ 80W for a .6L volume pack in comparison to 3 hours with a battery. The primary issue there is the dispersal of waste heat from the power pack within a volume as small as a laptop.
That would be impossible. You must see that! Even if you were correct that a battery would be unnecessary during operation (I hate auto comparisons!), just like an auto, it would need a battery to start the machine.
What you are forgetting is that the efficiency losses result in less power as the cycle continues. Without using calculus, it's difficult to show how that works, but I'll try to use an example that does give the idea.
If one cycle is one second, that is, the cpu puts out heat which is then converted to power every second for the full turnaround, then, at best, 20% of the heat is converted.
That 20% is funneled back to the computer for running power.
But it must supply the entire machine! The cpu, plus avery other element. The RAM, the HD, the display, etc.
By the beginning of the second cycle, the machine is out of power. Poof!
They show 40% of ideal Carnot limit in their current prototypes. That's mighty impressive. And its 20% efficient at 400C/20C gradient for 80% losses. That's why I would be concerned for the heat dispersion in their power pack in small volumes. [/quote]
40% is good, but it's also a laboratory device. Mass production devices might not get to that level.
20% efficiency at 400c/20c is scary. I wouldn't want that thing NEAR my computer.
Quote:
The gradient requirement is also why you can't just layer these things as suggested in an earlier post.
They've been talking about this for 4-5 years now and if the science was hinkey someone would have called them on it. Folks still remember cold fusion.
There's nothing wrong with the science. It's just not a savior. That's why the are talking about as supplimenting the battery.
Quote:
One reason to use the cooling option is to supercool the CPU for better performance. Connect the hot side to a heat pipe and you might as well scavenge that heat for (electrical) energy production. So if you use two units, you've increased the cost of the laptop but reduced power consumption of the cooling thermal diode by some 20% (well sorta but you get the idea). While the total energy gain is still negative the performance boost may be worth it. Plus if you really have 500W-h you have power to burn relative to today's notebooks.
I have no problem with that. As long as they can get heatpipes in a thin laptop. They are somewhat bulky.
I'd love to see one of these fan motors at 90% efficiency, but it's closer to 70%.
Mkay...replace that with 70% if you like. Still you are removing heat from the system and converting it to work. There certainly are electric motors that run at 90%. No clue on FAN motors.
Quote:
Sort of.
Sort of what? Energy remains constant. You certainly don't generate the same amount of heat again as you had before or you end up with the same amount of heat + work for a net energy gain. The energy leaves the system from the air you've imparted velocity to.
Quote:
Motors require a constant source of current. The cpu won't be giving off heat that is convertible evenly during use. The voltage must also be correct. Motors generate a back EMF, which would require additional circuitry to protect the device, and regulate the voltage for the motor.. That will add to the inefficiency.
Are you saying you don't think you can add a little circuit to regulate this thing? Inefficiency simply reduces the cooling effect not eliminate it as you claim.
Quote:
That would be impossible. You must see that! Even if you were correct that a battery would be unnecessary during operation (I hate auto comparisons!), just like an auto, it would need a battery to start the machine.
Yes, this is because Solar Cells also need a battery to start generating electricity from sunlight. What you need is a heat source and they're talking about ethanol cartridges and catalytic micro burners.
Quote:
What you are forgetting is that the efficiency losses result in less power as the cycle continues. Without using calculus, it's difficult to show how that works, but I'll try to use an example that does give the idea.
If one cycle is one second, that is, the cpu puts out heat which is then converted to power every second for the full turnaround, then, at best, 20% of the heat is converted.
That 20% is funneled back to the computer for running power.
Thus far is correct.
Quote:
But it must supply the entire machine! The cpu, plus avery other element. The RAM, the HD, the display, etc.
This part is incorrect. The CPU is simply a source of waste heat that could be scavanged. A pretty decent one if you have a hot running CPU. But its not the primary heat source envisioned by them.
Quote:
By the beginning of the second cycle, the machine is out of power. Poof!
Yes, because perpetual motion machines don't work. However, since the laptop runs on a power pack that uses some ethanol heating element to heat the thermal diode we aren't trying to build a perpetual motion machine. Just one that lasts 12 hours.
Quote:
20% efficiency at 400c/20c is scary. I wouldn't want that thing NEAR my computer.
Yes, but they talk about using a 200c gradient as well. Oh...and I was wrong. Looking at the graph it appears they claim 28% efficiency at 400C for 30kW-80kW power generators (about 50% Carnot efficiency).
Presumably 200c would be much lower and I guess closer to that 40% range they claim elsewhere.
Quote:
There's nothing wrong with the science. It's just not a savior. That's why the are talking about as supplimenting the battery.
Not if you look at their website. As a portable power source they claim higher power densities than fuel cells and much higher than batteries. Sure, first phase is running the fan (which happens to be their scenario) to increase battery life. But if the technology pans out it really will be disruptive. Savior? Not so much but very very big.
But it's years away as near as I can tell. The talks with Apple and Dell are very preliminary.
Mkay...replace that with 70% if you like. Still you are removing heat from the system and converting it to work. There certainly are electric motors that run at 90%. No clue on FAN motors.
There are high efficiency motors for industrial use. They cost over $500, and weigh 40 pounds or more.
Quote:
Sort of what? Energy remains constant. You certainly don't generate the same amount of heat again as you had before or you end up with the same amount of heat + work for a net energy gain. The energy leaves the system from the air you've imparted velocity to.
I was agreeing with your basic premise, but the way you worded it was not exact.
Quote:
Are you saying you don't think you can add a little circuit to regulate this thing? Inefficiency simply reduces the cooling effect not eliminate it as you claim.
Regulation is a waster of a great deal of energy. They heatsink regulators. Regulators dump a great deal of power. For example; To get a regulated 5 volts, you must start with at least 7.5 volts. The difference is thrown away. For 12 volts, you need at least 15, though 18 is better.
Quote:
Yes, this is because Solar Cells also need a battery to start generating electricity from sunlight. What you need is a heat source and they're talking about ethanol cartridges and catalytic micro burners.
Yes, and that is energy as well. But, this still can't run the machine by itself.
Quote:
This part is incorrect. The CPU is simply a source of waste heat that could be scavanged. A pretty decent one if you have a hot running CPU. But its not the primary heat source envisioned by them.
This is the source they are talking about. These devices won't be cheap. Where are you planning on using them other than on the CPU?
You would need to place them arounf the HD. A large sheet would be needed behind the LCD light source. RAM, battery, chiopset, etc. would all have to be sinked with these. And that still won't recover more than 20% of the total heat, and thus only a small amount of power. Not enough to run the machine by itself.
Quote:
Yes, because perpetual motion machines don't work. However, since the laptop runs on a power pack that uses some ethanol heating element to heat the thermal diode we aren't trying to build a perpetual motion machine. Just one that lasts 12 hours.
It's not a perpetual machine at all. It uses more power than is released back by four times.
There is one kind of perpetual motion machine they have working.
Quote:
Yes, but they talk about using a 200c gradient as well. Oh...and I was wrong. Looking at the graph it appears they claim 28% efficiency at 400C for 30kW-80kW power generators (about 50% Carnot efficiency).
Efficiency always goes up as the generating device gets larger. It's the cube law.
Quote:
Presumably 200c would be much lower and I guess closer to that 40% range they claim elsewhere.
Perhaps.
Quote:
Not if you look at their website. As a portable power source they claim higher power densities than fuel cells and much higher than batteries. Sure, first phase is running the fan (which happens to be their scenario) to increase battery life. But if the technology pans out it really will be disruptive. Savior? Not so much but very very big.
Power density is not the same as total power. One atom of anti-matter, plus one atom of matter, has the highest known power density in the universe, but one atom of that, plus one atom of matter brought together, still makes for a very small amount of power.
Quote:
But it's years away as near as I can tell. The talks with Apple and Dell are very preliminary.
You're right, it's years away, if it works at all.
There are high efficiency motors for industrial use. They cost over $500, and weigh 40 pounds or more.
Well fine, I went and googled on fan motors. Brushless DC motors (like those used in computer fans) are on the order of 85-90% efficient. These certainly are not $500 and 40+ lbs.
Quote:
I was agreeing with your basic premise, but the way you worded it was not exact.
Ah, okay...so we are or aren't in agreement that when you convert heat into electricity that there is less heat in the system but the same amount of energy?
Quote:
Regulation is a waster of a great deal of energy. They heatsink regulators. Regulators dump a great deal of power. For example; To get a regulated 5 volts, you must start with at least 7.5 volts. The difference is thrown away. For 12 volts, you need at least 15, though 18 is better.
Regulated DC-DC switching voltage regulators are said to be 85%+ efficient. Linear VRegs aren't the only option...
Quote:
Yes, and that is energy as well. But, this still can't run the machine by itself.
If the heat source is a catalytic microburner in the main power pack then the technology can run the machine.
Quote:
This is the source they are talking about. These devices won't be cheap. Where are you planning on using them other than on the CPU?
In their CPU example the CPU thermal diodes would only drive the fans. Not the entire machine. That will initially be batteries or fuel cells but the end goal is to use thermal diodes as the primary power supply driven by some kind of heat source. They suggest this might be a catalytic microburner using ethanol.
Quote:
You would need to place them arounf the HD. A large sheet would be needed behind the LCD light source. RAM, battery, chiopset, etc. would all have to be sinked with these. And that still won't recover more than 20% of the total heat, and thus only a small amount of power. Not enough to run the machine by itself.
I would imagine that those sources would be too diffuse to generate the heat gradient desired. I'd think that you'd do the CPU and GPU only.
Quote:
It's not a perpetual machine at all. It uses more power than is released back by four times.
Again, you miss the point. They postulate that the energy required to drive a laptop comes from some kind of microburner not from waste heat from a CPU. The only thing the CPU diodes would do is recover some small fraction of the power back.
Quote:
Perhaps.
They claim to have NIST reports that confirm this. And DARPA is involved. I certainly hope their claims are true because it would be a major step forward in energy production.
Quote:
Power density is not the same as total power. One atom of anti-matter, plus one atom of matter, has the highest known power density in the universe, but one atom of that, plus one atom of matter brought together, still makes for a very small amount of power.
Yes, but they aren't talking about 1 atom worth of volume and mass but that of a typical laptop battery. In their case they compare vs a 0.6L 0.9kg laptop battery. In which case power density does determine the total power available in that particular package. They say current batteries hold 88.8 W-h and their solution would hold 500 W-h.
Or are you trying to say there is no relationship between power density and the total power available in real world systems?
Quote:
You're right, it's years away, if it works at all.
As I said, I hope it pans out because it would be a great thing for everyone.
Well fine, I went and googled on fan motors. Brushless DC motors (like those used in computer fans) are on the order of 85-90% efficient. These certainly are not $500 and 40+ lbs.
Ok, that's pretty good.
Quote:
Ah, okay...so we are or aren't in agreement that when you convert heat into electricity that there is less heat in the system but the same amount of energy?
It's ticklish. The reason is that all energy ends up as heat, which then has to be removed, or radiated away. The total heat in the system is the same, but the localized heating MIGHT be less. What is happening is that the heat is being stored as energy in the battery, which will be later used, and converted to heat. It really is difficult to completely expain here.
Quote:
Regulated DC-DC switching voltage regulators are said to be 85%+ efficient. Linear VRegs aren't the only option...
That's only under ideal conditions. The reality is that they are more like 70%.
Quote:
If the heat source is a catalytic microburner in the main power pack then the technology can run the machine.
That still requires an external input of fuel.
Quote:
In their CPU example the CPU thermal diodes would only drive the fans. Not the entire machine. That will initially be batteries or fuel cells but the end goal is to use thermal diodes as the primary power supply driven by some kind of heat source. They suggest this might be a catalytic microburner using ethanol.
Right. I've no argument with that. But it's the ethanol that is providing the power. The heat, converted to energy is only supplying the running power for the fuel cell, which is a battery, of sorts. Battery, meaning a contained supply of energy, in this case, though, using ethanol, it's really a power plant.
Quote:
I would imagine that those sources would be too diffuse to generate the heat gradient desired. I'd think that you'd do the CPU and GPU only.
It could help, but it would be too complex, and expensive.
Quote:
Again, you miss the point. They postulate that the energy required to drive a laptop comes from some kind of microburner not from waste heat from a CPU. The only thing the CPU diodes would do is recover some small fraction of the power back.
I'm not missing the point. It's the point that I've been making. The recovered heat cannot run the machine, as has been said in this thread several times. The microburner is a fuel cell. That would be needed, or any other kind of battery that would be used.
Quote:
They claim to have NIST reports that confirm this. And DARPA is involved. I certainly hope their claims are true because it would be a major step forward in energy production.
Are you talking about the fuel cell, or the conversion technology they are working on?
Quote:
Yes, but they aren't talking about 1 atom worth of volume and mass but that of a typical laptop battery. In their case they compare vs a 0.6L 0.9kg laptop battery. In which case power density does determine the total power available in that particular package. They say current batteries hold 88.8 W-h and their solution would hold 500 W-h.
Or are you trying to say there is no relationship between power density and the total power available in real world systems?
I'm assuming here that you are talking about the fuel cell.
Fuel cells are already being proposed for laptops. One company will have one next year. I don't remember if it is Toshiba, or someone else.
But, power density, and total power does not always go hand in hand.
That doesn't mean that a high power density source can't supply large amounts of power, just that one doesn't equal the other.
One of the biggest peoblems with rechargable lithium cells is that they have a very high density/power ratio. That makes them dangerous. The slightest defect, and poof!
Fuel cells have had the same problem. They are very difficult to make in small, high power density versions. Large fuel cells have been operating for quite some time.
But, for a small fuel cell to be practical, they must have a much higher power density than larger models because of the cube law. Therefore the engineering requirements have made them notoriously difficult to design. Several breakthroughs in the past three years have ameliorated those problems, but the engineering is still difficult, much more so than in batteries.
Quote:
As I said, I hope it pans out because it would be a great thing for everyone.
Comments
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.
It's not often I agree with mel in an unqualified way, but this is one of those times.
http://forums.appleinsider.com/showthread.php?t=65972
I had this suggestion back in August, looks like someone is finally stealing my idea. It'll probably cost them millions just to patent this and thus the payoff will never be worth it.
http://forums.appleinsider.com/showthread.php?t=65972
It can cost tens of thousands to get a patent, even 100 thousand or so, depends. But, most patents cost a few thousand, some even less.
I had this suggestion back in August, looks like someone is finally stealing my idea. It'll probably cost them millions just to patent this and thus the payoff will never be worth it.
http://forums.appleinsider.com/showthread.php?t=65972
Exactly what is a "thermocoupling"? A thermocouple is a sensor that detects temperature changes.
What do you think?
I would imagine this being used as a self-adjusting cooling system. You connect one chip who takes heat and outputs electricity and connect it to an another chip who takes electricity and cools. If it gets hot, then it will be automatically adjusted! The warmer, the more electricity to cooling! No fans!
What do you think?
You get a downward spiral. Remember, everything has inefficiencies. It isn't possible to avoid them.
What will happen, without outside input, is that the inefficiencies of both devices will lead to less output as time goes on, leading to a very low level of work. All the excess produced by the inefficiencies will exit as heat.
Total gain?zero.
Also, don't forget that you will need fans to exhaust that heat, requiring more power, and heat, being produced.
It's the fans that do the actual work of removing the heat from the unit.
As long as fans are required, this doesn't result in real benefit, except for what they are claiming, longer battery usage before recharges.
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.
Say what? How does one remove heat from a system (by converting into electricity) and NOT cool the environment it exists in? Wherever the optimum placement (before or after the heat sink, fan or both) you are making the job easier for cooling system as a whole.
For Migo's scenario, rather than hooking up two of these things (one in power conversion and the other in cooling mode) you hook the power conversion one to the fan (as suggested earlier). The hotter it is the faster the fan spins and 16% of the total heat inside the laptop is converted to mechanical work moving the fan blades around (assuming around 90% efficiency for the fan).
Edit: Removed reference to electrical potential energy when the electricity is stored in the battery...there is no battery if the thing works as claimed. Any waste heat converted gets used right away anyway. Heat reduction in the system is gained from burning less fuel in the primary power pack to power the device.
Vinea
PS Their latest paper: http://scitation.aip.org/journals/do...902_1.html#F12
Say what? How does one remove heat from a system (by converting into electricity) and NOT cool the environment it exists in? Wherever the optimum placement (before or after the heat sink, fan or both) you are making the job easier for cooling system as a whole.
For Migo's scenario, rather than hooking up two of these things (one in power conversion and the other in cooling mode) you hook the power conversion one to the fan (as suggested earlier). The hotter it is the faster the fan spins and 16% of the total heat inside the laptop is converted to mechanical work moving the fan blades around (assuming around 90% efficiency for the fan).
Edit: Removed reference to electrical potential energy when the electricity is stored in the battery...there is no battery if the thing works as claimed. Any waste heat converted gets used right away anyway. Heat reduction in the system is gained from burning less fuel in the primary power pack to power the device.
Vinea
PS Their latest paper: http://scitation.aip.org/journals/do...902_1.html#F12
Because you aren't "removing" the heat. You are converting it into power, which when used, produces heat. From the losses in the system, heat is always produced. You are removing it with fans, which use power, and produce heat.
So the operation of the fans do remove heat (From the cpu, as well as their own, but not that produced by the battery for the power used for the fans.), but the conversion process does not.
There's an assumption you are making that the device will power the fans directly. I'm not aware of any information that's been given out that suggests that can be done. Perhaps that's why they intend to funnel the power produced back into the battery.
Also, there is an approximate 20% power loss upon conversion. We don't know how theoretical that is. Will the final product be better or worse?
Migo's suggestion is less efficient because there would be two conversions going on, lowering the overall efficiency even further. 20% to convert to power, and 20% from the other to cool.
But, the cooling unit does not produce power, so it won't work anyway.
Because you aren't "removing" the heat. You are converting it into power, which when used, produces heat. From the losses in the system, heat is always produced. You are removing it with fans, which use power, and produce heat.
If you convert 20% of the heat into electricity and then use that energy to drive a motor that is 90% efficient then you've converted that 20% heat into 18% mechanical work and 2% heat. Heat IS produced but less than the input heat. And its driving the cooling system to boot. The energy remains constant, the heat does not.
Well sort of constant...you're pushing energy out in two forms: heat transfer to the air and then moving that air out of the system.
So the operation of the fans do remove heat (From the cpu, as well as their own, but not that produced by the battery for the power used for the fans.), but the conversion process does not.
If you convert 20% of the heat into electricity and store that in a battery then you've converted 20% of the heat into electrical potential energy - inefficiencies of charging a battery lost as heat. How much waste heat is generated later depends on what I do with that stored energy. If expended to do mechanical work then some lesser amount of energy is released as heat.
There's an assumption you are making that the device will power the fans directly. I'm not aware of any information that's been given out that suggests that can be done. Perhaps that's why they intend to funnel the power produced back into the battery.
So you have an electrical source that can't at least help drive an electric motor? And there likely wouldn't BE a battery if the technology works. The power density of this device is much higher than LiOn or LiPos. With a 20% efficient thermal diode they expect a run time of 12 hours @ 80W for a .6L volume pack in comparison to 3 hours with a battery. The primary issue there is the dispersal of waste heat from the power pack within a volume as small as a laptop.
Also, there is an approximate 20% power loss upon conversion. We don't know how theoretical that is. Will the final product be better or worse?
They show 40% of ideal Carnot limit in their current prototypes. That's mighty impressive. And its 20% efficient at 400C/20C gradient for 80% losses. That's why I would be concerned for the heat dispersion in their power pack in small volumes.
The gradient requirement is also why you can't just layer these things as suggested in an earlier post.
They've been talking about this for 4-5 years now and if the science was hinkey someone would have called them on it. Folks still remember cold fusion.
Migo's suggestion is less efficient because there would be two conversions going on, lowering the overall efficiency even further. 20% to convert to power, and 20% from the other to cool.
But, the cooling unit does not produce power, so it won't work anyway.
One reason to use the cooling option is to supercool the CPU for better performance. Connect the hot side to a heat pipe and you might as well scavenge that heat for (electrical) energy production. So if you use two units, you've increased the cost of the laptop but reduced power consumption of the cooling thermal diode by some 20% (well sorta but you get the idea). While the total energy gain is still negative the performance boost may be worth it. Plus if you really have 500W-h you have power to burn relative to today's notebooks.
Vinea
If you convert 20% of the heat into electricity and then use that energy to drive a motor that is 90% efficient then you've converted that 20% heat into 18% mechanical work and 2% heat. Heat IS produced but less than the input heat. And its driving the cooling system to boot. The energy remains constant, the heat does not.
Well sort of constant...you're pushing energy out in two forms: heat transfer to the air and then moving that air out of the system.
I'd love to see one of these fan motors at 90% efficiency, but it's closer to 70%.
If you convert 20% of the heat into electricity and store that in a battery then you've converted 20% of the heat into electrical potential energy - inefficiencies of charging a battery lost as heat. How much waste heat is generated later depends on what I do with that stored energy. If expended to do mechanical work then some lesser amount of energy is released as heat.
Sort of.
So you have an electrical source that can't at least help drive an electric motor?
Motors require a constant source of current. The cpu won't be giving off heat that is convertible evenly during use. The voltage must also be correct. Motors generate a back EMF, which would require additional circuitry to protect the device, and regulate the voltage for the motor.. That will add to the inefficiency.
And there likely wouldn't BE a battery if the technology works. The power density of this device is much higher than LiOn or LiPos. With a 20% efficient thermal diode they expect a run time of 12 hours @ 80W for a .6L volume pack in comparison to 3 hours with a battery. The primary issue there is the dispersal of waste heat from the power pack within a volume as small as a laptop.
That would be impossible. You must see that! Even if you were correct that a battery would be unnecessary during operation (I hate auto comparisons!), just like an auto, it would need a battery to start the machine.
What you are forgetting is that the efficiency losses result in less power as the cycle continues. Without using calculus, it's difficult to show how that works, but I'll try to use an example that does give the idea.
If one cycle is one second, that is, the cpu puts out heat which is then converted to power every second for the full turnaround, then, at best, 20% of the heat is converted.
That 20% is funneled back to the computer for running power.
But it must supply the entire machine! The cpu, plus avery other element. The RAM, the HD, the display, etc.
By the beginning of the second cycle, the machine is out of power. Poof!
They show 40% of ideal Carnot limit in their current prototypes. That's mighty impressive. And its 20% efficient at 400C/20C gradient for 80% losses. That's why I would be concerned for the heat dispersion in their power pack in small volumes. [/quote]
40% is good, but it's also a laboratory device. Mass production devices might not get to that level.
20% efficiency at 400c/20c is scary. I wouldn't want that thing NEAR my computer.
The gradient requirement is also why you can't just layer these things as suggested in an earlier post.
They've been talking about this for 4-5 years now and if the science was hinkey someone would have called them on it. Folks still remember cold fusion.
There's nothing wrong with the science. It's just not a savior. That's why the are talking about as supplimenting the battery.
One reason to use the cooling option is to supercool the CPU for better performance. Connect the hot side to a heat pipe and you might as well scavenge that heat for (electrical) energy production. So if you use two units, you've increased the cost of the laptop but reduced power consumption of the cooling thermal diode by some 20% (well sorta but you get the idea). While the total energy gain is still negative the performance boost may be worth it. Plus if you really have 500W-h you have power to burn relative to today's notebooks.
I have no problem with that. As long as they can get heatpipes in a thin laptop. They are somewhat bulky.
I'd love to see one of these fan motors at 90% efficiency, but it's closer to 70%.
Mkay...replace that with 70% if you like. Still you are removing heat from the system and converting it to work. There certainly are electric motors that run at 90%. No clue on FAN motors.
Sort of.
Sort of what? Energy remains constant. You certainly don't generate the same amount of heat again as you had before or you end up with the same amount of heat + work for a net energy gain. The energy leaves the system from the air you've imparted velocity to.
Motors require a constant source of current. The cpu won't be giving off heat that is convertible evenly during use. The voltage must also be correct. Motors generate a back EMF, which would require additional circuitry to protect the device, and regulate the voltage for the motor.. That will add to the inefficiency.
Are you saying you don't think you can add a little circuit to regulate this thing? Inefficiency simply reduces the cooling effect not eliminate it as you claim.
That would be impossible. You must see that! Even if you were correct that a battery would be unnecessary during operation (I hate auto comparisons!), just like an auto, it would need a battery to start the machine.
Yes, this is because Solar Cells also need a battery to start generating electricity from sunlight. What you need is a heat source and they're talking about ethanol cartridges and catalytic micro burners.
What you are forgetting is that the efficiency losses result in less power as the cycle continues. Without using calculus, it's difficult to show how that works, but I'll try to use an example that does give the idea.
If one cycle is one second, that is, the cpu puts out heat which is then converted to power every second for the full turnaround, then, at best, 20% of the heat is converted.
That 20% is funneled back to the computer for running power.
Thus far is correct.
But it must supply the entire machine! The cpu, plus avery other element. The RAM, the HD, the display, etc.
This part is incorrect. The CPU is simply a source of waste heat that could be scavanged. A pretty decent one if you have a hot running CPU. But its not the primary heat source envisioned by them.
By the beginning of the second cycle, the machine is out of power. Poof!
Yes, because perpetual motion machines don't work. However, since the laptop runs on a power pack that uses some ethanol heating element to heat the thermal diode we aren't trying to build a perpetual motion machine. Just one that lasts 12 hours.
20% efficiency at 400c/20c is scary. I wouldn't want that thing NEAR my computer.
Yes, but they talk about using a 200c gradient as well. Oh...and I was wrong. Looking at the graph it appears they claim 28% efficiency at 400C for 30kW-80kW power generators (about 50% Carnot efficiency).
Presumably 200c would be much lower and I guess closer to that 40% range they claim elsewhere.
There's nothing wrong with the science. It's just not a savior. That's why the are talking about as supplimenting the battery.
Not if you look at their website. As a portable power source they claim higher power densities than fuel cells and much higher than batteries. Sure, first phase is running the fan (which happens to be their scenario) to increase battery life. But if the technology pans out it really will be disruptive. Savior? Not so much but very very big.
But it's years away as near as I can tell. The talks with Apple and Dell are very preliminary.
Vinea
Mkay...replace that with 70% if you like. Still you are removing heat from the system and converting it to work. There certainly are electric motors that run at 90%. No clue on FAN motors.
There are high efficiency motors for industrial use. They cost over $500, and weigh 40 pounds or more.
Sort of what? Energy remains constant. You certainly don't generate the same amount of heat again as you had before or you end up with the same amount of heat + work for a net energy gain. The energy leaves the system from the air you've imparted velocity to.
I was agreeing with your basic premise, but the way you worded it was not exact.
Are you saying you don't think you can add a little circuit to regulate this thing? Inefficiency simply reduces the cooling effect not eliminate it as you claim.
Regulation is a waster of a great deal of energy. They heatsink regulators. Regulators dump a great deal of power. For example; To get a regulated 5 volts, you must start with at least 7.5 volts. The difference is thrown away. For 12 volts, you need at least 15, though 18 is better.
Yes, this is because Solar Cells also need a battery to start generating electricity from sunlight. What you need is a heat source and they're talking about ethanol cartridges and catalytic micro burners.
Yes, and that is energy as well. But, this still can't run the machine by itself.
This part is incorrect. The CPU is simply a source of waste heat that could be scavanged. A pretty decent one if you have a hot running CPU. But its not the primary heat source envisioned by them.
This is the source they are talking about. These devices won't be cheap. Where are you planning on using them other than on the CPU?
You would need to place them arounf the HD. A large sheet would be needed behind the LCD light source. RAM, battery, chiopset, etc. would all have to be sinked with these. And that still won't recover more than 20% of the total heat, and thus only a small amount of power. Not enough to run the machine by itself.
Yes, because perpetual motion machines don't work. However, since the laptop runs on a power pack that uses some ethanol heating element to heat the thermal diode we aren't trying to build a perpetual motion machine. Just one that lasts 12 hours.
It's not a perpetual machine at all. It uses more power than is released back by four times.
There is one kind of perpetual motion machine they have working.
Yes, but they talk about using a 200c gradient as well. Oh...and I was wrong. Looking at the graph it appears they claim 28% efficiency at 400C for 30kW-80kW power generators (about 50% Carnot efficiency).
Efficiency always goes up as the generating device gets larger. It's the cube law.
Presumably 200c would be much lower and I guess closer to that 40% range they claim elsewhere.
Perhaps.
Not if you look at their website. As a portable power source they claim higher power densities than fuel cells and much higher than batteries. Sure, first phase is running the fan (which happens to be their scenario) to increase battery life. But if the technology pans out it really will be disruptive. Savior? Not so much but very very big.
Power density is not the same as total power. One atom of anti-matter, plus one atom of matter, has the highest known power density in the universe, but one atom of that, plus one atom of matter brought together, still makes for a very small amount of power.
But it's years away as near as I can tell. The talks with Apple and Dell are very preliminary.
You're right, it's years away, if it works at all.
There are high efficiency motors for industrial use. They cost over $500, and weigh 40 pounds or more.
Well fine, I went and googled on fan motors. Brushless DC motors (like those used in computer fans) are on the order of 85-90% efficient. These certainly are not $500 and 40+ lbs.
I was agreeing with your basic premise, but the way you worded it was not exact.
Ah, okay...so we are or aren't in agreement that when you convert heat into electricity that there is less heat in the system but the same amount of energy?
Regulation is a waster of a great deal of energy. They heatsink regulators. Regulators dump a great deal of power. For example; To get a regulated 5 volts, you must start with at least 7.5 volts. The difference is thrown away. For 12 volts, you need at least 15, though 18 is better.
Regulated DC-DC switching voltage regulators are said to be 85%+ efficient. Linear VRegs aren't the only option...
Yes, and that is energy as well. But, this still can't run the machine by itself.
If the heat source is a catalytic microburner in the main power pack then the technology can run the machine.
This is the source they are talking about. These devices won't be cheap. Where are you planning on using them other than on the CPU?
In their CPU example the CPU thermal diodes would only drive the fans. Not the entire machine. That will initially be batteries or fuel cells but the end goal is to use thermal diodes as the primary power supply driven by some kind of heat source. They suggest this might be a catalytic microburner using ethanol.
You would need to place them arounf the HD. A large sheet would be needed behind the LCD light source. RAM, battery, chiopset, etc. would all have to be sinked with these. And that still won't recover more than 20% of the total heat, and thus only a small amount of power. Not enough to run the machine by itself.
I would imagine that those sources would be too diffuse to generate the heat gradient desired. I'd think that you'd do the CPU and GPU only.
It's not a perpetual machine at all. It uses more power than is released back by four times.
Again, you miss the point. They postulate that the energy required to drive a laptop comes from some kind of microburner not from waste heat from a CPU. The only thing the CPU diodes would do is recover some small fraction of the power back.
Perhaps.
They claim to have NIST reports that confirm this. And DARPA is involved. I certainly hope their claims are true because it would be a major step forward in energy production.
Power density is not the same as total power. One atom of anti-matter, plus one atom of matter, has the highest known power density in the universe, but one atom of that, plus one atom of matter brought together, still makes for a very small amount of power.
Yes, but they aren't talking about 1 atom worth of volume and mass but that of a typical laptop battery. In their case they compare vs a 0.6L 0.9kg laptop battery. In which case power density does determine the total power available in that particular package. They say current batteries hold 88.8 W-h and their solution would hold 500 W-h.
Or are you trying to say there is no relationship between power density and the total power available in real world systems?
You're right, it's years away, if it works at all.
As I said, I hope it pans out because it would be a great thing for everyone.
Vinea
Well fine, I went and googled on fan motors. Brushless DC motors (like those used in computer fans) are on the order of 85-90% efficient. These certainly are not $500 and 40+ lbs.
Ok, that's pretty good.
Ah, okay...so we are or aren't in agreement that when you convert heat into electricity that there is less heat in the system but the same amount of energy?
It's ticklish. The reason is that all energy ends up as heat, which then has to be removed, or radiated away. The total heat in the system is the same, but the localized heating MIGHT be less. What is happening is that the heat is being stored as energy in the battery, which will be later used, and converted to heat. It really is difficult to completely expain here.
Regulated DC-DC switching voltage regulators are said to be 85%+ efficient. Linear VRegs aren't the only option...
That's only under ideal conditions. The reality is that they are more like 70%.
If the heat source is a catalytic microburner in the main power pack then the technology can run the machine.
That still requires an external input of fuel.
In their CPU example the CPU thermal diodes would only drive the fans. Not the entire machine. That will initially be batteries or fuel cells but the end goal is to use thermal diodes as the primary power supply driven by some kind of heat source. They suggest this might be a catalytic microburner using ethanol.
Right. I've no argument with that. But it's the ethanol that is providing the power. The heat, converted to energy is only supplying the running power for the fuel cell, which is a battery, of sorts. Battery, meaning a contained supply of energy, in this case, though, using ethanol, it's really a power plant.
I would imagine that those sources would be too diffuse to generate the heat gradient desired. I'd think that you'd do the CPU and GPU only.
It could help, but it would be too complex, and expensive.
Again, you miss the point. They postulate that the energy required to drive a laptop comes from some kind of microburner not from waste heat from a CPU. The only thing the CPU diodes would do is recover some small fraction of the power back.
I'm not missing the point. It's the point that I've been making. The recovered heat cannot run the machine, as has been said in this thread several times. The microburner is a fuel cell. That would be needed, or any other kind of battery that would be used.
They claim to have NIST reports that confirm this. And DARPA is involved. I certainly hope their claims are true because it would be a major step forward in energy production.
Are you talking about the fuel cell, or the conversion technology they are working on?
Yes, but they aren't talking about 1 atom worth of volume and mass but that of a typical laptop battery. In their case they compare vs a 0.6L 0.9kg laptop battery. In which case power density does determine the total power available in that particular package. They say current batteries hold 88.8 W-h and their solution would hold 500 W-h.
Or are you trying to say there is no relationship between power density and the total power available in real world systems?
I'm assuming here that you are talking about the fuel cell.
Fuel cells are already being proposed for laptops. One company will have one next year. I don't remember if it is Toshiba, or someone else.
But, power density, and total power does not always go hand in hand.
That doesn't mean that a high power density source can't supply large amounts of power, just that one doesn't equal the other.
One of the biggest peoblems with rechargable lithium cells is that they have a very high density/power ratio. That makes them dangerous. The slightest defect, and poof!
Fuel cells have had the same problem. They are very difficult to make in small, high power density versions. Large fuel cells have been operating for quite some time.
But, for a small fuel cell to be practical, they must have a much higher power density than larger models because of the cube law. Therefore the engineering requirements have made them notoriously difficult to design. Several breakthroughs in the past three years have ameliorated those problems, but the engineering is still difficult, much more so than in batteries.
As I said, I hope it pans out because it would be a great thing for everyone.
Vinea
I hope so too.