I'm not saying that there isn't a problem. The article reads more like an advertisement in spots, but they do give a modest amount of technical info -- enough that I'm willing to believe the problem is real. It appears that the spread-spectrum controller is interfering with the WiFi signal. That's not overly surprising, but it has absolutely *nothing* to do with the fact that the data is digital. It has everything to do with the fact that these two devices are using each other's bandwidth and not handling the interference well, which is unsurprising given the relatively unregulated nature of the 2.4GHz band. The intereference could just as easily be caused by an analog source as a digital one.
Also, unless you're really experienced at it, you can't tell a clean, bandwidth-limited signal by looking at it in the time domain -- you need a spectrum analyzer. (If you're really experienced, you'll do ok, but the spectrum analyzer is still important.) Furthermore, "spread spectrum" is not the same thing as "not clean" -- not in the slightest. From the perspective of the other device, though, they may produce similar results (undesired interference).
The lesson here is not that the radio engineers are screwing up. (They might be, but there is no evidence presented to that effect.) Rather, it is that using multiple different transmission schemes in the same band without any coordination is likely to cause problems. And really, that's not exactly a surprising result. If you want someone to complain at, complain at the regulators for not providing more bandwidth with better negotiation protocols mandated.
It's bad because 2.4 Ghz is radio, carrying digital info, which due to the nature of the produced sign wave results in a signal distortion more commonly known as "bleed over". Without the ability to separate the signals by a large frequency, digital over analog bleeds all over the place.
The hell? There is nothing magic about digital data that means you can't bandwidth-limit the outgoing transmission. There are plenty of digital radio protocols that use a very well defined slice of bandwidth, without any more bleed over than traditional AM or FM radio analog broadcasts. Just because the signal represents digital data doesn't mean you have to use square waves or something.
I suppose we should all be thankful that radio engineers are better educated than the average Slashdot poster...
(Of course, it's entirely possible there's something broken about the XBOX radio. It's also entirely possible it's just a spread-spectrum transmitter doing exactly what it's supposed to do in a largely unregulated piece of spectrum.)
How is a power-brick sized docking station any harder to carry around than a power brick? So it has a connector on top that plugs directly into the laptop instead of having a cable. I don't think that makes it harder to carry...
A docking station need not be all that much bigger than the larger laptop power bricks out there today -- eg, the IBM one I have here is approaching half the size of my desktop PSU in volume. By docking station, I was referring to form factor rather than volume. It need not be the whole length of the laptop, just something that has a direct connector instead of a cable. Imagine a modestly enlarged power brick that sits under your laptop instead of next to it.
In reverse order... My (very limited) understanding is that waste heat from charging is one of the main limitations to charge rate for normal batteries. As such, high charge rate capabilities in battery packs without special cooling implies higher charging efficiency. 90% charging efficiency therefore seems a conservative estimate to me.
That same Google paper is the one I was basing my 95% number on, in part. You'll note the part about "our servers' power supplies now run at 90% efficiency or better." I'm not invoking any sort of "green" motivation here, btw -- efficiency is mandated by waste heat concerns, even if there isn't any particular consumer demand for it. I go from 90% to 95% based on several things. First, their paper suggests "better than 90%," and power electronics are generally improving. Second, I've seen commercially viable designs for power supplies of this class at that efficiency level. And third, there's an obvious difference that would provide improved efficiency: laptop battery packs can be made to charge at higher voltage (eg the 24V suggested in the article) than the 12V Google uses as their power supply output, and given comparable power transistors a higher output voltage from a buck converter will tend to be more efficient (mosfet on resistances have a lower impact, as do things like diode drops and coil series resistance).
I think 90% is a good lower bound for a high-output laptop pack (though your 80% number is quite reasonable for a more modest output pack like those used today), with 95% being not at all unreasonable.
The power brick would be about half the size of your desktop's PSU, or a bit less by the time you account for more care toward packing all the components. It'll be more efficient, too, because it only has to output one voltage. And it'll be power factor corrected while they're at it, if they think there's any demand or need for that. 500W worth of power supply circuitry just isn't as big a pain as it used to be. (500W is what it would take to charge a quasi-typical laptop battery from 10% to 90% in 5 minutes.)
Except that Lithium Ion batteries don't actually contain metallic lithium. They contain lithium ions -- ie, the lithium is already oxidized. That's true for both the charged and discharged state. Some other metal (cobalt traditionally, I think iron and a couple others are used in newer experimental chemistries) is being oxidized and reduced. Wikipedia has more about the relevant electrochemistry.
Non-rechargable lithium cells (most 3V coin type cells) have metallic lithium. The rechargable chemistries don't, though -- hence the name lithium ion.
So the random laptop battery I have handy is rated 10.8V, 4.8Ah -- 52Wh. 5 minutes for 80% charge (from 10% to 90%, you're unlikely to let it go all the way to zero) is just shy of 500 watts. Your average wall outlet is easily capable of that (12A at 115V is a nice, conservative estimate). The power brick to handle that won't be huge -- think about a 500W computer power supply, and then remember that this will be noticeably smaller and more efficient because it only has to provide one output voltage instead of the mess your average computer wants. It'll need some cooling (even at a mildly aggressive but reasonable 95% efficiency, that's 25W of waste heat), but the fan will still be reasonable.
At first glance it would appear that the cable from power brick to laptop would be huge and awkward, but that can be solved fairly easily by having the connection be more like a docking station cradle. That would also let the charger supply additional airflow for the battery with a larger fan that you'd find on the laptop itself -- the battery will get rather warm during this process, and battery heating is probably one of the limiting factors on charge rates for something like this.
The robotic mission would be a needed precursor to manned trips to the red planet.
No, it wouldn't. We know enough about Mars to send a human or three there on a mission now, especially with a plan like Mars Direct. (Short version of plan: send an automated small chemical plant there with a hydrogen cargo. Turn the hydrogen plus martian CO2 into methane + oxygen. When the return vehicle is fully fuelled, send the human crew along on the next ship. They don't launch until they have a confirmed return ship ready, so if the hard part doesn't work they don't go. It needs two launches with payload capability roughly on par with either a Saturn V or a Shuttle stack converted to cargo use instead of flying the orbiter -- or an Ares V, roughly.)
Of course, sample return missions are interesting and useful in their own right. But don't confuse the issue; we went to the Moon without a sample return mission, we can do the same for Mars.
No, he basically has a point. It only gets damaged if the peak power of the laser is high enough. The problem is, it only works once -- you can force a 10x increase in laser size, but you really can't force a *second* 10x increase. 90% is a very good mirror in a battlefield, but not entirely unreasonable. For a nuclear warhead reentry vehicle, you can likely do significantly better (the launch vehicle is another matter, and laser system proposals frequently discuss boost-phase targetting for a variety of reasons).
How easy do you think it is to keep a mirror clean on a battlefield?
OK, now how easy do you think it is to keep a precision optics grade mirror at clean-room standards on a battlefield? Because I'm pretty sure that's what it takes to stop a potent laser -- if you're mirror is absorbing more than perhaps 0.1% of the incoming light, it will get hot and melt rather quickly, which makes it absorb a lot more light...
No, it shouldn't. There's two pieces here -- the executable file on disk (which is also cached in memory for speed reasons), and the memory the program is using. It's almost exactly like you took a few hundred MB of ram, decided to use it as your hard disk, and ran programs off that. Those programs still use memory normally, so when you run them ram usage jumps. In practice, of course, the amount of memory used for this varies (mostly based on how much memory you're actually using for runnings programs), and it tries to do a good job of only keeping the things that will be helpful copied into memory.
Doesn't everyone here have a hobby or two they spend a fair bit of money on? Perhaps it's your computer gear, maybe it's model airplanes, maybe it's your car or your audio system. Last I checked, an awful lot of geeks had a particular hobby they enjoyed and spent money on, and they don't have to be 'rich bastards' to do so. They just have to value enjoying themselves over... What? Hording money? So this man's hobby is reading and his library, and he enjoys organizing it in a creative way.
For Lithium Ion batteries, most of the lifetime of the battery is determined by time since manufacture (with modifiers for how charged it is -- 40% or so is best, iirc -- and temperature and such), with charge cycles being a second-order effect. Of course, that assumes you take good care of it, but the charge controller in the car should be able to handle that anyway.
Of course, as you say, supercapacitors are the interesting technology. AIUI, all the pieces exist in the lab to make supercaps that beat LiIon for storage density. Now it's just a small matter of engineering (tm) before we can use them in cars. They're getting better quickly, in terms of what you can actually buy.
Well, if it's not economical to buy the battery for the purpose, but you're going to buy the battery anyway to use it in the car... A lot of batteries (especially LiIon) have a significant component of their lifetime measured in years of service, not charge cycles. So if it's not costing you anything to use the battery like this, and you already own it...
You do realize that this already happens, and the electric companies do pay you for it? Industrially, power compaines give large consumers a break on rates if they get a say in when the power gets used, for exactly this reason. Some consumers need fairly large amounts of power, but don't care when they use it. Think refrigerated warehouses -- you can turn off the refrigeration for hours to reduce load without trouble, but then they have to use more later. In exchange for doing this, they get reduced rates. In some areas, you can also buy time of day metering -- handy if you have grid-tie solar panels, as you get to run the meter backward at day rates, then come home and use power at night rates.
I imagine they would be happy to extend the same basic deals to your car. And as you point out, you're not required to do so, so if they want you too, they'll have to offer such things.
Slashdot Mirror
I'm not saying that there isn't a problem. The article reads more like an advertisement in spots, but they do give a modest amount of technical info -- enough that I'm willing to believe the problem is real. It appears that the spread-spectrum controller is interfering with the WiFi signal. That's not overly surprising, but it has absolutely *nothing* to do with the fact that the data is digital. It has everything to do with the fact that these two devices are using each other's bandwidth and not handling the interference well, which is unsurprising given the relatively unregulated nature of the 2.4GHz band. The intereference could just as easily be caused by an analog source as a digital one.
Also, unless you're really experienced at it, you can't tell a clean, bandwidth-limited signal by looking at it in the time domain -- you need a spectrum analyzer. (If you're really experienced, you'll do ok, but the spectrum analyzer is still important.) Furthermore, "spread spectrum" is not the same thing as "not clean" -- not in the slightest. From the perspective of the other device, though, they may produce similar results (undesired interference).
The lesson here is not that the radio engineers are screwing up. (They might be, but there is no evidence presented to that effect.) Rather, it is that using multiple different transmission schemes in the same band without any coordination is likely to cause problems. And really, that's not exactly a surprising result. If you want someone to complain at, complain at the regulators for not providing more bandwidth with better negotiation protocols mandated.
It's bad because 2.4 Ghz is radio, carrying digital info, which due to the nature of the produced sign wave results in a signal distortion more commonly known as "bleed over". Without the ability to separate the signals by a large frequency, digital over analog bleeds all over the place.
The hell? There is nothing magic about digital data that means you can't bandwidth-limit the outgoing transmission. There are plenty of digital radio protocols that use a very well defined slice of bandwidth, without any more bleed over than traditional AM or FM radio analog broadcasts. Just because the signal represents digital data doesn't mean you have to use square waves or something.
I suppose we should all be thankful that radio engineers are better educated than the average Slashdot poster...
(Of course, it's entirely possible there's something broken about the XBOX radio. It's also entirely possible it's just a spread-spectrum transmitter doing exactly what it's supposed to do in a largely unregulated piece of spectrum.)
How is a power-brick sized docking station any harder to carry around than a power brick? So it has a connector on top that plugs directly into the laptop instead of having a cable. I don't think that makes it harder to carry...
A docking station need not be all that much bigger than the larger laptop power bricks out there today -- eg, the IBM one I have here is approaching half the size of my desktop PSU in volume. By docking station, I was referring to form factor rather than volume. It need not be the whole length of the laptop, just something that has a direct connector instead of a cable. Imagine a modestly enlarged power brick that sits under your laptop instead of next to it.
In reverse order... My (very limited) understanding is that waste heat from charging is one of the main limitations to charge rate for normal batteries. As such, high charge rate capabilities in battery packs without special cooling implies higher charging efficiency. 90% charging efficiency therefore seems a conservative estimate to me.
That same Google paper is the one I was basing my 95% number on, in part. You'll note the part about "our servers' power supplies now run at 90% efficiency or better." I'm not invoking any sort of "green" motivation here, btw -- efficiency is mandated by waste heat concerns, even if there isn't any particular consumer demand for it. I go from 90% to 95% based on several things. First, their paper suggests "better than 90%," and power electronics are generally improving. Second, I've seen commercially viable designs for power supplies of this class at that efficiency level. And third, there's an obvious difference that would provide improved efficiency: laptop battery packs can be made to charge at higher voltage (eg the 24V suggested in the article) than the 12V Google uses as their power supply output, and given comparable power transistors a higher output voltage from a buck converter will tend to be more efficient (mosfet on resistances have a lower impact, as do things like diode drops and coil series resistance).
I think 90% is a good lower bound for a high-output laptop pack (though your 80% number is quite reasonable for a more modest output pack like those used today), with 95% being not at all unreasonable.
There may be other obstacles to us finding another civilization, you know. The heat death of the universe isn't the one I'd worry about.
You didn't happen to read the remainder of my comment, did you? I offered both a size estimate and a possible solution to the big thick cable problem.
The power brick would be about half the size of your desktop's PSU, or a bit less by the time you account for more care toward packing all the components. It'll be more efficient, too, because it only has to output one voltage. And it'll be power factor corrected while they're at it, if they think there's any demand or need for that. 500W worth of power supply circuitry just isn't as big a pain as it used to be. (500W is what it would take to charge a quasi-typical laptop battery from 10% to 90% in 5 minutes.)
Except that Lithium Ion batteries don't actually contain metallic lithium. They contain lithium ions -- ie, the lithium is already oxidized. That's true for both the charged and discharged state. Some other metal (cobalt traditionally, I think iron and a couple others are used in newer experimental chemistries) is being oxidized and reduced. Wikipedia has more about the relevant electrochemistry.
Non-rechargable lithium cells (most 3V coin type cells) have metallic lithium. The rechargable chemistries don't, though -- hence the name lithium ion.
So the random laptop battery I have handy is rated 10.8V, 4.8Ah -- 52Wh. 5 minutes for 80% charge (from 10% to 90%, you're unlikely to let it go all the way to zero) is just shy of 500 watts. Your average wall outlet is easily capable of that (12A at 115V is a nice, conservative estimate). The power brick to handle that won't be huge -- think about a 500W computer power supply, and then remember that this will be noticeably smaller and more efficient because it only has to provide one output voltage instead of the mess your average computer wants. It'll need some cooling (even at a mildly aggressive but reasonable 95% efficiency, that's 25W of waste heat), but the fan will still be reasonable.
At first glance it would appear that the cable from power brick to laptop would be huge and awkward, but that can be solved fairly easily by having the connection be more like a docking station cradle. That would also let the charger supply additional airflow for the battery with a larger fan that you'd find on the laptop itself -- the battery will get rather warm during this process, and battery heating is probably one of the limiting factors on charge rates for something like this.
and my kids (14, 16) have zero curiosity about either.
Perhaps this is why OLPC intends to give laptops to younger children than yours.
Yes. They're ill-tempered, too.
The robotic mission would be a needed precursor to manned trips to the red planet.
No, it wouldn't. We know enough about Mars to send a human or three there on a mission now, especially with a plan like Mars Direct. (Short version of plan: send an automated small chemical plant there with a hydrogen cargo. Turn the hydrogen plus martian CO2 into methane + oxygen. When the return vehicle is fully fuelled, send the human crew along on the next ship. They don't launch until they have a confirmed return ship ready, so if the hard part doesn't work they don't go. It needs two launches with payload capability roughly on par with either a Saturn V or a Shuttle stack converted to cargo use instead of flying the orbiter -- or an Ares V, roughly.)
Of course, sample return missions are interesting and useful in their own right. But don't confuse the issue; we went to the Moon without a sample return mission, we can do the same for Mars.
No, he basically has a point. It only gets damaged if the peak power of the laser is high enough. The problem is, it only works once -- you can force a 10x increase in laser size, but you really can't force a *second* 10x increase. 90% is a very good mirror in a battlefield, but not entirely unreasonable. For a nuclear warhead reentry vehicle, you can likely do significantly better (the launch vehicle is another matter, and laser system proposals frequently discuss boost-phase targetting for a variety of reasons).
Seconded. Especially if there are more like it...
How easy do you think it is to keep a mirror clean on a battlefield?
OK, now how easy do you think it is to keep a precision optics grade mirror at clean-room standards on a battlefield? Because I'm pretty sure that's what it takes to stop a potent laser -- if you're mirror is absorbing more than perhaps 0.1% of the incoming light, it will get hot and melt rather quickly, which makes it absorb a lot more light...
You misunderstand. C-130H is the designation for the new, genetically engineered, giant sharks.
No, it shouldn't. There's two pieces here -- the executable file on disk (which is also cached in memory for speed reasons), and the memory the program is using. It's almost exactly like you took a few hundred MB of ram, decided to use it as your hard disk, and ran programs off that. Those programs still use memory normally, so when you run them ram usage jumps. In practice, of course, the amount of memory used for this varies (mostly based on how much memory you're actually using for runnings programs), and it tries to do a good job of only keeping the things that will be helpful copied into memory.
Seriously, what the hell?
Doesn't everyone here have a hobby or two they spend a fair bit of money on? Perhaps it's your computer gear, maybe it's model airplanes, maybe it's your car or your audio system. Last I checked, an awful lot of geeks had a particular hobby they enjoyed and spent money on, and they don't have to be 'rich bastards' to do so. They just have to value enjoying themselves over... What? Hording money? So this man's hobby is reading and his library, and he enjoys organizing it in a creative way.
Sheesh.
There are two kinds of people in the world: Those that can extrapolate from incomplete data.
You can get this in many parts of the US. Usually, though, you'll have to pay for the new meter yourself, which makes it not as attractive.
Shall I shed a tear because you have more trouble hiding things from the public?
For Lithium Ion batteries, most of the lifetime of the battery is determined by time since manufacture (with modifiers for how charged it is -- 40% or so is best, iirc -- and temperature and such), with charge cycles being a second-order effect. Of course, that assumes you take good care of it, but the charge controller in the car should be able to handle that anyway.
Of course, as you say, supercapacitors are the interesting technology. AIUI, all the pieces exist in the lab to make supercaps that beat LiIon for storage density. Now it's just a small matter of engineering (tm) before we can use them in cars. They're getting better quickly, in terms of what you can actually buy.
Well, if it's not economical to buy the battery for the purpose, but you're going to buy the battery anyway to use it in the car... A lot of batteries (especially LiIon) have a significant component of their lifetime measured in years of service, not charge cycles. So if it's not costing you anything to use the battery like this, and you already own it...
You do realize that this already happens, and the electric companies do pay you for it? Industrially, power compaines give large consumers a break on rates if they get a say in when the power gets used, for exactly this reason. Some consumers need fairly large amounts of power, but don't care when they use it. Think refrigerated warehouses -- you can turn off the refrigeration for hours to reduce load without trouble, but then they have to use more later. In exchange for doing this, they get reduced rates. In some areas, you can also buy time of day metering -- handy if you have grid-tie solar panels, as you get to run the meter backward at day rates, then come home and use power at night rates.
I imagine they would be happy to extend the same basic deals to your car. And as you point out, you're not required to do so, so if they want you too, they'll have to offer such things.