Because a superconductor conducts with literally zero resistance, you can create a ring of superconducting material, pump as much current into it as it will tolerate, and just let the current cycle forever. No degradation whatsoever. Then when you want power, you just tap into the ring and pull it out on demand.
The problem is that storage density is limited both by the maximum magnetic field your superconductors can tolerate, and by the tensile strength of your coil (interaction between the field and the current causes very large outward pressure on the solenoid).
I did the calculations assuming a material with the tensile strength of carbon nanotubes and no field limit at all, and still wound up with an energy density a couple of orders of magnitude below that of chemical fuels (by weight and by volume).
You might replace batteries with something like this, but fuel cells are already leaning on this market and have higher energy density than an inductor can achieve.
In summary, while it's a neat idea, it turns out to not be useful in most practical scenarios.
Yikes. If you try to "tap in" to an inductor, it will produce an enormous voltage and immediately arc to close the circuit. The only way to get energy out of a superconducting solenoid is through some magnetic interaction.
If you pick the number of windings carefully, tapping directly into the inductor works just fine.
The inductor wants to maintain the current flowing through the coil. If that is the amount of current you expect to draw for your load, both load and coil will be perfectly happy in the new configuration. If you wish to draw less current (or tolerate interruptions without arcing), drop a resistor in parallel with the load. This will limit voltage across the load to the amount needed to push the coil's current through the resistor.
When you aren't using the load, of course, you short across it so as to reduce resistive power loss. Typically this switching is actually performed by having a closed coil, and heating the part you want to cut out above the superconducting breakdown temperature, if I understand correctly.
The only design difficulty is that this requires a large number of windings (sheet current is typically millions of amps or more, which means you need millions of windings for a load that draws 1A).
exceptions are numerous, including sapphire which (in some range of temperatures) is actually a better heat conductor than copper, yet is an hard insulator. A slightly worse heat conductor, alumina, is also a hard insulator.
This parallel is not surprising, as sapphire is alumina doped with anything other than chromium or vanadium (those two count as "ruby").
At what point do we call something a "person" for purposes of rights?
Some time this century we'll likely be able to produce artificial intelligent creatures, be they machines or tailored organisms. Where do we draw the line between "person" and "non-person", and how do we assess this in practice?
What ethical/moral concerns, if any, are appropriate/mandatory to consider before creating an artificial person?
If the previous point is a concern, this one will be too.
In a society where information may be freely exchanged anonymously and without cost, what are appropriate and inappropriate models of ownership and rights of control over things that are now considered owned information?
E.g. works of art, algorithms/code, ideas/concepts, pictures of people, medical records. Justify from both a moral/ethical and a practical viewpoint.
How will or should the ability of anyone to undetectably conduct surveillance of anyone/any location affect privacy rights as they are currently known?
We arguably have this _now_.
All of these are going to have to be dealt with sooner rather than later, and none have cut-and-dried answers, no matter what position you take. Enjoy.
Also, he IS testing liqud fuel rockets. Maybe you should look at his homepage sometime.
No photos in the non-flash version of his site.
He fed the columnist in this article a line of vapour that his design team quickly downplayed. Until he actually demonstrates a liquid-fuel rocket, he doesn't have one.
Solid fuel won't get any reasonably sized manned capsule into orbit. Suborbital, maybe.
He's only aiming for sub-orbit (that's all the X-prize requires, and what he claimed in the article).
The problem is that he's trying to do it with model rocket engines that Were Not Specced For That. It would be like me trying to build a rocket using D engines that could hit Detroit from Chicago. It doesn't matter how many I'm going to tape together - it won't work. Saying how beautifully my latest duct-tape contraption lofted a brick to a thousand feet won't bring me closer to that goal, and that's about how Bennet's test flights compare to his stated plans.
He claims to have a more appropraite engine design in mid-stage development, but hasn't shown any evidence of such a thing (and his associates say things like "that? oh yes, we hope to build something like that some day").
He reminds me an awful lot of my boss from a small startup a few years back - operating in his own little reality, with everyone just looking at him funny when he makes his grand pronouncements.
Amateurs and small startups _can_ make useful contributions. Bennet just isn't one of them.
Brightness we can easily measure but distance for anything but the closest stars is very difficult to determine. They based their conclusion on a guess that this star is 20,000 light years away. While this is definitely an interesting event, if that guess is wrong it might not be nearly as dramatic as they say.
According to the article, they determined the distance by looking at the companion star in the pair, which is of a well-known type, with a well-known temperature/luminosity relation. That gave them a reasonable distance estimate for the companion, and so for the giant.
This event goes contrary to everything what is known about the star life cycle so far.
Not really. As far as I understand, it's actually pretty typical of the unstable time when a star either enters or leaves the red giant phase. We're seeing a "planetary nebula" being born.
It was meant as a joke. But maybe you can make a weapon with which you can expode a sun. Shouldn't be too hard considering the sun is one giant atomic bomb anyway...
If it's easy, it should happen all over the place already through natural processes. This does not seem to be the case (novae and supernovae are quite rare in the grand scheme of things).
Stars are very good at being self-balancing systems. As reaction rate increases, so does photon pressure, which makes the star less dense, which reduces reaction rate. This breaks down only in special cases.
Unstable giant stars, like this star appears to be, are one of those cases. Our sun may end up doing something not very different from this in a few billion years as its core runs out of fuel.
Violent explosions only occur when something overrides fusion-produced photon pressure and the star starts collapsing. This mainly happens when a star runs out of fuel, and stops again when either a new fusion stage starts, or when degeneracy pressure takes over.
The device he wrote the paper about works in the millimetre-wave regime, if I understand correctly (a bit above microwaves). It's relatively easy to build negative-index materials here, because you can do it by building oddly-shaped configurations of wires that interact in easily-controlled ways with the electric and magnetic components of the microwaves/mm-waves. To do this at optical wavelengths, you'd either need to use micromachining or find exotic compounds that have the properties you want. If I understand correctly both approaches are currently being followed.
The lower-frequency experiments are still interesting, though. The physics for the effect itself is the same, and it's easier to both build the devices and do measurements.
No electronic device, licensed or unlicensed my interfere with federally allocated spectrum. You don't need to yank a license to shut someone down any more than I need to build my own kW system (the orignial 802.11 alocation was supposed to be 100W!). If it were legal to use the spectrum, reputable builders would come to the rescue.
You miss my point. I'll state it more clearly:
If you have a reason to transmit signals long distances, you can do so cheaply by filling out the appropriate forms and following the rules.
Reputable builders already exist - check out the ham community and the suppliers they buy from. If you want to *use* a 1 kW transmitter, you'd better have a license for it.
But the most important thing you miss is that the spectrum WILL BE USED where TODAY IT CAN'T BE!
And this is a problem why?
If you're running short-range to wired hubs, there is no shortage of data capacity in the bands already alotted.
If you for some reason want to operate a higher-power transmitter and use it to flood large parts of the spectrum with your data... then I'm glad you don't have a transmitter. You and everyone else in your neighbourhood would be stepping on each others' toes. You have nicely proved my point.
That's about as silly a notion as they come. "Because things are bad they always will be and we should not change."
If you believe that there is a magical way to run a wired backbone at half the cost to the user, sell the business plan to the highest bidder. Last I checked the backbone providers were pretty deep in the red.
Limit wireless to short range, and put hubs everywhere. Problem solved (for urban areas; rural areas are an entirely different problem with different constraints).
So you would have no free long range high power spectrum at all?
There would be a few bands open for hobbyists, just like there are now. Want to build a 1 kW transmitter? Go ahead - just get your ham license first. Decide you're not going to play nicely in the community? Your license gets revoked.
Without management, anything longer than short-range will cause too many people to step on each others' toes.
The cost of all that badwith you want would be considerably less if more spectrum was given over to 802.11B type freedom. The equipment is cheap enough that people would build the infrastructure and run it as a free public service.
The free public service would then start charging a modest fee to support its overhead, and then the core of people running it would slowly drift to the dark side as bureaucracy started fossilizing, and you'd end up with something indistinguishable from the bandwidth providers we currently have.
Do you think that Cthulhu came to earth and decided to found UUNet to torment the mortals? Large-scale utility providers _naturally_ evolve to become this way!
If you want cheaper bandwidth, start a letter-writing campaign to get better government regulation of the industry. It's a utility, just like phone and power and water and so forth. Manage it like one.
think about what you are saying. If only a fraction of the currently restricted bandwith were so well utilized! As it is, you hear silence. Which is preferable? The possibility of a clog or enforced silence and frustration?
Clogging is a certainty without imposed limits; people are greedy that way. Removing band restrictions just guarantees that *all* parts of the spectrum are clogged.
Band restriction is a quality of service issue - if you want to be able to use your cell phone, or to put up an antenna and hear music from your favourite radio station, there must be a guarantee that the person next door isn't cluttering up that section of the spectrum for you. This is especially true for emergency services, and for bands that interfere with important equipment (radio beacons at airports come to mind).
What is it that you stand for?
Wired backbones. All the bandwidth you can eat, and much less contention for it. Limit wireless to short range, and put hubs everywhere. Problem solved (for urban areas; rural areas are an entirely different problem with different constraints).
This is assuming no triangulation of source. If I treat my input as a point, I get interfernce at any given frequency because I treat all transmission as coming from and arriving to this imaginary space. (this is the way traditional broadcasting is treated)
If on the other hand I treat my input as a three dimensional space all of a sudden I can have as many broadcast sources as my ability to process them can tolerate. I can distinguish signal based on directionality, though I have new concerns like multi-path signals and two sources coming one behind the other.
If you have a receiver that is smaller than the signal wavelength (i.e. is an omnidirectional point receiver, moving or non-moving), you will end up not having enough information to disambiguate sources when the system is beyond the capacity limit that I outlined previously. This is easy enough to demonstrate; the total amount of information available to you is the amount that one transmitter could produce, assuming that signal generation and reception fidelity are equivalent. Trying to distinguish between multiple sources dumping as much information as they can into the environment requires pulling extra information out of nowhere, which you aren't able to do.
Knowledge of the location of the sources, or of where signal-affecting surfaces are in your environment, or of the motion of your receiving unit (for synthetic aperture tricks) makes signal processing easier when you're below the saturation point, but doesn't help you when the total amount of information you're trying to extract is greater than that received.
If you feel I'm overlooking something, then let's consider an artificial case that's easy to analyze. Assume straight sampling of signal data from zero to a given frequency with a given number of detectable data levels (linear for easy analysis). Assume sources and receivers are point sources (which they are if they're smaller than a wavelength). Do whatever you like with the environment and with data processing, and show me how you'd get one receiver to reconstruct the continuously-streaming signals from two sources.
If there's a way to do this, great; I'll have learned something today. If you feel that it is only possible by changing one of the problem constraints, we'll negotiate (IIUC it's only possible by changing the problem to make the total data received at least equal to the total data transmitted, by any of a variety of means).
This comment follows a rant which ironically ignores most modern radio breaktrhoughs: packet routing and frequency hopping on low power devices to create a network with far greater bandwith than a single transmitter per frequency set up that's current.
Mesh routing schemes break down in highly populated areas - you end up with too many messages needing to be routed by any given node, and the fraction of node bandwidth used for that node's messages dropping like a rock.
Relation is a fun exercise in calculation that takes about 2 minutes.
The only way around this is to link to a backbone and strongly limit transmitter power, which sort of torpedoes the "let's stop regulating the spectrum" argument.
You can do point-to-point without a backbone, but only with a large dish or a large *wired* array of transceivers.
[TDMA] is an immensely powerful technique, and one that is infinitely scalable. It's only limitation is the speed of our electronics, which can and should maintain it's exponential speed curve. This is why the spectrum is underutilized.
You do realize that as sampling rate goes up, spectrum use (bandwidth) goes up, right?
Any given region of spectrum can only carry so much data, any way you slice it. Power, noise/clutter, and bandwidth combined determine (and limit) data rate.
For AM transmissions, theoretically a single, exact frequency can suffice. Assuming the transmitter is truly on the expected frequency, all you need is a very narrow bandpass filter.
If you try to send an AM signal across a 1 Hz band, you will get a 1 Hz bandwidth signal out at the other end. Not very useful if you were trying to play music. Definitely not useful if you were trying to transmit data.
The number people are interest in is data rate. Data rate is bandwidth times the log to base 2 of the number of levels you can distinguish. Different encoding schemes (FM, wide-spectrum coding) express the relation differently, but the same limit applies.
You can narrow the bandwidth, but as soon as you hit noise limits, your data rate starts going down too. *That's* the problem. Low-noise electronics doesn't help if the noise is from other users.
The only way to avoid user clutter is to switch to something other than a broadcast system, which involves either large dishes or short-range transceivers and hubs connected to a _wired_ backbone.
Re:Even predicting the recent past can be tricky.
on
5595 Days and Counting
·
· Score: 1
But at present making longer fibers is literaly "only a matter of time;" with a continious deposition process, the longer you let it run, the longer the fibers are.
My understanding is that this is not correct. The nanotube fibers cap themselves after a certain average length (consistent with this being a random process with low but significant probability). Leave it on, and you get more fibers, not longer. Last I heard they were experimenting with different additives, as various metals inhibit capping for poorly understood reasons (which is how they produced long tubes in the first place, as opposed to spheres and short tubes).
If you have a citation that demonstrates otherwise, though, I'd be glad to look it over.
if they are too well aligned they either act as fracture features or "pull out"
My understanding was that this was the failure mode of concern. For a space elevator, you want aligned fibers for maximum tensile strength (you don't care as much about other stresses).
Re:Even predicting the recent past can be tricky.
on
5595 Days and Counting
·
· Score: 1
The present problem is the old "how do they stick the teflon to the pan" problem; getting them to play nicely in composites, and finding ways to manufacture the composites.
When I asked a polymer chemist friend about this, he said that longer fibers would solve the problem. This suggests that we cannot yet manufacture fibers long enough to be useful in composites.
Re:I'm willing to bet $$$ it will never work
on
5595 Days and Counting
·
· Score: 2, Insightful
You're missing something here: the difference between science and engineering. Space elevator advocates often point out that most of the remaining problems are not in the realm of science but instead tech and financing. So progress is not dependent on some long haired genius in a basement lab having a brainwave. You can make confident predictions that technology will improve and that the material with the required tensile strength will be constructed soon in the future.
I agree except for the "soon" part.
You can make exactly the same arguments about whisker fibers, which have been around for quite some time and would be a wonderful material if we could solve the degradation and production problems.
You can also make the same argument about controlled fusion. Few doubt that it's possible; there's just been a history of very optimistic estimates for when we'll finally have all of the engineering problems solved.
If we have nanotubes in quantity tomorrow, I'll be the first to cheer, because you can do many interesting things with them on a smaller scale than building space elevators. However, I'm not holding my breath.
Conduction from the much larger surface area of the oil resovoir to the surrounding environment should work. Think of this a better way of coupling the cpu to the case heat-wise (use the whole case/resovoir basin as a heat sink to dump heat into the environment).
I've thought about doing something like this off and on for years.
The reason: I hate CPU fans.
They're loud. They die with distressing regularity. They're louder *as* they die - the death rattle can last for a year or more.
Put the motherboard in a bin of vegetable oil, keep the drives and power supply out of it (or even put the power supply into it), and you get convection cooling with heat sinks and no fans at all.
The only catch is that you're going to have to either filter the oil or change it regularly even if it _is_ in a sealed container, and have a working procedure for draining the box and cleaning the components if you ever want to swap out a card or perform other reconfigurations.
Still awfully tempting. My second fan is on its last legs at the moment. And don't get me started about chipset fans.
It's also a gun easily disabled by an electro-magnetic pulse, which is especially relevant since the military now has EMP bombs.
The military will always have their own weapons; they won't be stuck using something so failure-prone. This is great for civilian use. It doesn't hamper legitimate uses while making illegitimate ones more difficult to execute and get away with.
Sorry - I don't buy that. You're talking about an industry that has prospered over the past 20 years largely by creating the perception that there is a continual need to upgrade. They'd probably be quite cheerful as they sold 0.001% of their market a special, expensive, backwards-compatible chip for legacy systems.
Remember the Cyrix "486". A chip that doesn't support _everything_, is broken as far as the customer is concerned. Selling a "mostly" compatible chip would be a great way to torpedo their own sales.
Slashdot Mirror
Because a superconductor conducts with literally zero resistance, you can create a ring of superconducting material, pump as much current into it as it will tolerate, and just let the current cycle forever. No degradation whatsoever. Then when you want power, you just tap into the ring and pull it out on demand.
The problem is that storage density is limited both by the maximum magnetic field your superconductors can tolerate, and by the tensile strength of your coil (interaction between the field and the current causes very large outward pressure on the solenoid).
I did the calculations assuming a material with the tensile strength of carbon nanotubes and no field limit at all, and still wound up with an energy density a couple of orders of magnitude below that of chemical fuels (by weight and by volume).
You might replace batteries with something like this, but fuel cells are already leaning on this market and have higher energy density than an inductor can achieve.
In summary, while it's a neat idea, it turns out to not be useful in most practical scenarios.
Yikes. If you try to "tap in" to an inductor, it will produce an enormous voltage and immediately arc to close the circuit. The only way to get energy out of a superconducting solenoid is through some magnetic interaction.
If you pick the number of windings carefully, tapping directly into the inductor works just fine.
The inductor wants to maintain the current flowing through the coil. If that is the amount of current you expect to draw for your load, both load and coil will be perfectly happy in the new configuration. If you wish to draw less current (or tolerate interruptions without arcing), drop a resistor in parallel with the load. This will limit voltage across the load to the amount needed to push the coil's current through the resistor.
When you aren't using the load, of course, you short across it so as to reduce resistive power loss. Typically this switching is actually performed by having a closed coil, and heating the part you want to cut out above the superconducting breakdown temperature, if I understand correctly.
The only design difficulty is that this requires a large number of windings (sheet current is typically millions of amps or more, which means you need millions of windings for a load that draws 1A).
exceptions are numerous, including sapphire which (in some range of temperatures) is actually a better heat conductor than copper, yet is an hard insulator. A slightly worse heat conductor, alumina, is also a hard insulator.
This parallel is not surprising, as sapphire is alumina doped with anything other than chromium or vanadium (those two count as "ruby").
Some time this century we'll likely be able to produce artificial intelligent creatures, be they machines or tailored organisms. Where do we draw the line between "person" and "non-person", and how do we assess this in practice?
If the previous point is a concern, this one will be too.
E.g. works of art, algorithms/code, ideas/concepts, pictures of people, medical records. Justify from both a moral/ethical and a practical viewpoint.
We arguably have this _now_.
All of these are going to have to be dealt with sooner rather than later, and none have cut-and-dried answers, no matter what position you take. Enjoy.
Also, he IS testing liqud fuel rockets. Maybe you should look at his homepage sometime.
No photos in the non-flash version of his site.
He fed the columnist in this article a line of vapour that his design team quickly downplayed. Until he actually demonstrates a liquid-fuel rocket, he doesn't have one.
Solid fuel won't get any reasonably sized manned capsule into orbit. Suborbital, maybe.
He's only aiming for sub-orbit (that's all the X-prize requires, and what he claimed in the article).
The problem is that he's trying to do it with model rocket engines that Were Not Specced For That. It would be like me trying to build a rocket using D engines that could hit Detroit from Chicago. It doesn't matter how many I'm going to tape together - it won't work. Saying how beautifully my latest duct-tape contraption lofted a brick to a thousand feet won't bring me closer to that goal, and that's about how Bennet's test flights compare to his stated plans.
He claims to have a more appropraite engine design in mid-stage development, but hasn't shown any evidence of such a thing (and his associates say things like "that? oh yes, we hope to build something like that some day").
He reminds me an awful lot of my boss from a small startup a few years back - operating in his own little reality, with everyone just looking at him funny when he makes his grand pronouncements.
Amateurs and small startups _can_ make useful contributions. Bennet just isn't one of them.
Brightness we can easily measure but distance for anything but the closest stars is very difficult to determine. They based their conclusion on a guess that this star is 20,000 light years away. While this is definitely an interesting event, if that guess is wrong it might not be nearly as dramatic as they say.
According to the article, they determined the distance by looking at the companion star in the pair, which is of a well-known type, with a well-known temperature/luminosity relation. That gave them a reasonable distance estimate for the companion, and so for the giant.
This event goes contrary to everything what is known about the star life cycle so far.
Not really. As far as I understand, it's actually pretty typical of the unstable time when a star either enters or leaves the red giant phase. We're seeing a "planetary nebula" being born.
It was meant as a joke. But maybe you can make a weapon with which you can expode a sun. Shouldn't be too hard considering the sun is one giant atomic bomb anyway...
If it's easy, it should happen all over the place already through natural processes. This does not seem to be the case (novae and supernovae are quite rare in the grand scheme of things).
Stars are very good at being self-balancing systems. As reaction rate increases, so does photon pressure, which makes the star less dense, which reduces reaction rate. This breaks down only in special cases.
Unstable giant stars, like this star appears to be, are one of those cases. Our sun may end up doing something not very different from this in a few billion years as its core runs out of fuel.
Violent explosions only occur when something overrides fusion-produced photon pressure and the star starts collapsing. This mainly happens when a star runs out of fuel, and stops again when either a new fusion stage starts, or when degeneracy pressure takes over.
Here's the prof's page:
m l
d ex.html
http://www.waves.utoronto.ca/prof/gelefth/main.ht
Here's the prof's publications list; the paper that these press articles are about is right at the top.
http://www.waves.utoronto.ca/prof/gelefth/jpub/in
The device he wrote the paper about works in the millimetre-wave regime, if I understand correctly (a bit above microwaves). It's relatively easy to build negative-index materials here, because you can do it by building oddly-shaped configurations of wires that interact in easily-controlled ways with the electric and magnetic components of the microwaves/mm-waves. To do this at optical wavelengths, you'd either need to use micromachining or find exotic compounds that have the properties you want. If I understand correctly both approaches are currently being followed.
The lower-frequency experiments are still interesting, though. The physics for the effect itself is the same, and it's easier to both build the devices and do measurements.
No electronic device, licensed or unlicensed my interfere with federally allocated spectrum. You don't need to yank a license to shut someone down any more than I need to build my own kW system (the orignial 802.11 alocation was supposed to be 100W!). If it were legal to use the spectrum, reputable builders would come to the rescue.
You miss my point. I'll state it more clearly:
If you have a reason to transmit signals long distances, you can do so cheaply by filling out the appropriate forms and following the rules.
Reputable builders already exist - check out the ham community and the suppliers they buy from. If you want to *use* a 1 kW transmitter, you'd better have a license for it.
But the most important thing you miss is that the spectrum WILL BE USED where TODAY IT CAN'T BE!
And this is a problem why?
If you're running short-range to wired hubs, there is no shortage of data capacity in the bands already alotted.
If you for some reason want to operate a higher-power transmitter and use it to flood large parts of the spectrum with your data... then I'm glad you don't have a transmitter. You and everyone else in your neighbourhood would be stepping on each others' toes. You have nicely proved my point.
That's about as silly a notion as they come. "Because things are bad they always will be and we should not change."
If you believe that there is a magical way to run a wired backbone at half the cost to the user, sell the business plan to the highest bidder. Last I checked the backbone providers were pretty deep in the red.
Limit wireless to short range, and put hubs everywhere. Problem solved (for urban areas; rural areas are an entirely different problem with different constraints).
So you would have no free long range high power spectrum at all?
There would be a few bands open for hobbyists, just like there are now. Want to build a 1 kW transmitter? Go ahead - just get your ham license first. Decide you're not going to play nicely in the community? Your license gets revoked.
Without management, anything longer than short-range will cause too many people to step on each others' toes.
The cost of all that badwith you want would be considerably less if more spectrum was given over to 802.11B type freedom. The equipment is cheap enough that people would build the infrastructure and run it as a free public service.
The free public service would then start charging a modest fee to support its overhead, and then the core of people running it would slowly drift to the dark side as bureaucracy started fossilizing, and you'd end up with something indistinguishable from the bandwidth providers we currently have.
Do you think that Cthulhu came to earth and decided to found UUNet to torment the mortals? Large-scale utility providers _naturally_ evolve to become this way!
If you want cheaper bandwidth, start a letter-writing campaign to get better government regulation of the industry. It's a utility, just like phone and power and water and so forth. Manage it like one.
think about what you are saying. If only a fraction of the currently restricted bandwith were so well utilized! As it is, you hear silence. Which is preferable? The possibility of a clog or enforced silence and frustration?
Clogging is a certainty without imposed limits; people are greedy that way. Removing band restrictions just guarantees that *all* parts of the spectrum are clogged.
Band restriction is a quality of service issue - if you want to be able to use your cell phone, or to put up an antenna and hear music from your favourite radio station, there must be a guarantee that the person next door isn't cluttering up that section of the spectrum for you. This is especially true for emergency services, and for bands that interfere with important equipment (radio beacons at airports come to mind).
What is it that you stand for?
Wired backbones. All the bandwidth you can eat, and much less contention for it. Limit wireless to short range, and put hubs everywhere. Problem solved (for urban areas; rural areas are an entirely different problem with different constraints).
This is assuming no triangulation of source. If I treat my input as a point, I get interfernce at any given frequency because I treat all transmission as coming from and arriving to this imaginary space. (this is the way traditional broadcasting is treated)
If on the other hand I treat my input as a three dimensional space all of a sudden I can have as many broadcast sources as my ability to process them can tolerate. I can distinguish signal based on directionality, though I have new concerns like multi-path signals and two sources coming one behind the other.
If you have a receiver that is smaller than the signal wavelength (i.e. is an omnidirectional point receiver, moving or non-moving), you will end up not having enough information to disambiguate sources when the system is beyond the capacity limit that I outlined previously. This is easy enough to demonstrate; the total amount of information available to you is the amount that one transmitter could produce, assuming that signal generation and reception fidelity are equivalent. Trying to distinguish between multiple sources dumping as much information as they can into the environment requires pulling extra information out of nowhere, which you aren't able to do.
Knowledge of the location of the sources, or of where signal-affecting surfaces are in your environment, or of the motion of your receiving unit (for synthetic aperture tricks) makes signal processing easier when you're below the saturation point, but doesn't help you when the total amount of information you're trying to extract is greater than that received.
If you feel I'm overlooking something, then let's consider an artificial case that's easy to analyze. Assume straight sampling of signal data from zero to a given frequency with a given number of detectable data levels (linear for easy analysis). Assume sources and receivers are point sources (which they are if they're smaller than a wavelength). Do whatever you like with the environment and with data processing, and show me how you'd get one receiver to reconstruct the continuously-streaming signals from two sources.
If there's a way to do this, great; I'll have learned something today. If you feel that it is only possible by changing one of the problem constraints, we'll negotiate (IIUC it's only possible by changing the problem to make the total data received at least equal to the total data transmitted, by any of a variety of means).
This comment follows a rant which ironically ignores most modern radio breaktrhoughs: packet routing and frequency hopping on low power devices to create a network with far greater bandwith than a single transmitter per frequency set up that's current.
Mesh routing schemes break down in highly populated areas - you end up with too many messages needing to be routed by any given node, and the fraction of node bandwidth used for that node's messages dropping like a rock.
Relation is a fun exercise in calculation that takes about 2 minutes.
The only way around this is to link to a backbone and strongly limit transmitter power, which sort of torpedoes the "let's stop regulating the spectrum" argument.
You can do point-to-point without a backbone, but only with a large dish or a large *wired* array of transceivers.
[TDMA] is an immensely powerful technique, and one that is infinitely scalable. It's only limitation is the speed of our electronics, which can and should maintain it's exponential speed curve.
This is why the spectrum is underutilized.
You do realize that as sampling rate goes up, spectrum use (bandwidth) goes up, right?
Any given region of spectrum can only carry so much data, any way you slice it. Power, noise/clutter, and bandwidth combined determine (and limit) data rate.
For AM transmissions, theoretically a single, exact frequency can suffice. Assuming the transmitter is truly on the expected frequency, all you need is a very narrow bandpass filter.
If you try to send an AM signal across a 1 Hz band, you will get a 1 Hz bandwidth signal out at the other end. Not very useful if you were trying to play music. Definitely not useful if you were trying to transmit data.
The number people are interest in is data rate. Data rate is bandwidth times the log to base 2 of the number of levels you can distinguish. Different encoding schemes (FM, wide-spectrum coding) express the relation differently, but the same limit applies.
You can narrow the bandwidth, but as soon as you hit noise limits, your data rate starts going down too. *That's* the problem. Low-noise electronics doesn't help if the noise is from other users.
The only way to avoid user clutter is to switch to something other than a broadcast system, which involves either large dishes or short-range transceivers and hubs connected to a _wired_ backbone.
But at present making longer fibers is literaly "only a matter of time;" with a continious deposition process, the longer you let it run, the longer the fibers are.
My understanding is that this is not correct. The nanotube fibers cap themselves after a certain average length (consistent with this being a random process with low but significant probability). Leave it on, and you get more fibers, not longer. Last I heard they were experimenting with different additives, as various metals inhibit capping for poorly understood reasons (which is how they produced long tubes in the first place, as opposed to spheres and short tubes).
If you have a citation that demonstrates otherwise, though, I'd be glad to look it over.
if they are too well aligned they either act as fracture features or "pull out"
My understanding was that this was the failure mode of concern. For a space elevator, you want aligned fibers for maximum tensile strength (you don't care as much about other stresses).
The present problem is the old "how do they stick the teflon to the pan" problem; getting them to play nicely in composites, and finding ways to manufacture the composites.
When I asked a polymer chemist friend about this, he said that longer fibers would solve the problem. This suggests that we cannot yet manufacture fibers long enough to be useful in composites.
You're missing something here: the difference between science and engineering. Space elevator advocates often point out that most of the remaining problems are not in the realm of science but instead tech and financing. So progress is not dependent on some long haired genius in a basement lab having a brainwave. You can make confident predictions that technology will improve and that the material with the required tensile strength will be constructed soon in the future.
I agree except for the "soon" part.
You can make exactly the same arguments about whisker fibers, which have been around for quite some time and would be a wonderful material if we could solve the degradation and production problems.
You can also make the same argument about controlled fusion. Few doubt that it's possible; there's just been a history of very optimistic estimates for when we'll finally have all of the engineering problems solved.
If we have nanotubes in quantity tomorrow, I'll be the first to cheer, because you can do many interesting things with them on a smaller scale than building space elevators. However, I'm not holding my breath.
How do LEDs hold up to high frequency PWM? That sounds like a really cool idea.
Frequency limits are not a problem for a device like this (a friend had fun making an RF link between an LED and a photodetector).
As long as you stay within the rated current limits, you're fine.
You still need to get rid of the heat somehow.
Conduction from the much larger surface area of the oil resovoir to the surrounding environment should work. Think of this a better way of coupling the cpu to the case heat-wise (use the whole case/resovoir basin as a heat sink to dump heat into the environment).
I've thought about doing something like this off and on for years.
The reason: I hate CPU fans.
They're loud. They die with distressing regularity. They're louder *as* they die - the death rattle can last for a year or more.
Put the motherboard in a bin of vegetable oil, keep the drives and power supply out of it (or even put the power supply into it), and you get convection cooling with heat sinks and no fans at all.
The only catch is that you're going to have to either filter the oil or change it regularly even if it _is_ in a sealed container, and have a working procedure for draining the box and cleaning the components if you ever want to swap out a card or perform other reconfigurations.
Still awfully tempting. My second fan is on its last legs at the moment. And don't get me started about chipset fans.
It's also a gun easily disabled by an electro-magnetic pulse, which is especially relevant since the military now has EMP bombs.
The military will always have their own weapons; they won't be stuck using something so failure-prone. This is great for civilian use. It doesn't hamper legitimate uses while making illegitimate ones more difficult to execute and get away with.
Sorry - I don't buy that. You're talking about an industry that has prospered over the past 20 years largely by creating the perception that there is a continual need to upgrade. They'd probably be quite cheerful as they sold 0.001% of their market a special, expensive, backwards-compatible chip for legacy systems.
Remember the Cyrix "486". A chip that doesn't support _everything_, is broken as far as the customer is concerned. Selling a "mostly" compatible chip would be a great way to torpedo their own sales.