I guess the rest of the process is to very carefully drop the chip package where you want it to go, and convince the solder to melt. How do you do that? You mentioned a "reflow oven"? I Googled that, and there are lots of different ovens out there.
If you're using old-fashioned lead/tin solder you don't have to be all that careful about the alignment: if you have 10% overlap between the ball and the solderpad on most all the pads, the capillary action of the solder will cause the bga to self-align. (I do this on a regular basis with smaller but similar chips.) Now, if you're using modern no-lead solder it's more difficult. With a BGA you can do an okay job of alignment with a boom microscope set at an angle so you can see under the edge of the chip to the first row of balls: if you have opposite corners on the pads you're good.
And I use a hotplate and a thermocouple taped adjacent to the chip with the tip touching one of the pads. You don't want to overheat the chip itself -- no higher than 260C if you can possibly manage it. A lab hotplate is faster than a toaster oven, and you need to get the heat up fast enough the flux doesn't boil off before the solder reflows. Ideally you'd like it to reflow and hold above the liquidus of the solder (which is about 220C for lead and about 240C for no-lead) for 30 seconds, but managing that without going over 260C is pretty difficult. Then you cool it down pretty quickly. I use a hairdryer with the heat turned off so it's just blowing scads of cool air.
Likewise, by the way, I use a hotplate for adhering the balls to the BGA, rather than a hot air gun -- it's more predictable, temperature-wise.
Well, okay, all the above should be put in the past tense since as of a week ago we now have an OKI bga rework station, with a vision system that aligns the chip/board and places it, and a controllable temp profiler, and it's awesome. But I do have three years' experience hand-placing leadless chips and micro-SMD's, so I know a thing or two about it.
Doesn't the Roomba automatically go back to its charging station after it's done cleaning?
The only thing missing, AFAIK, it's the "empty itself" part. I'm puzzled over that one because they could simply add a "garbage can" part to the charging station...
It depends on the level of roomba -- both my roomba and my scooba have standard wall-wart chargers that I have to plug in by hand. But really that isn't the limiting factor: the brushes get stuff in them, particularly hair, and that more than anything else is the most difficult maintenance task.
They also have a paper filter in front of the main debris container, that has to be manually removed; automating that into a 'garbage can' sort of setup, would be tricky. Old-fashioned vacuums used a disposable paper bag to handle both tasks, dust removal and debris, but thankfully the roomba doesn't work that way or we'd be bankrupt since we run ours pretty much every day in one room or another. That's one of the good design decisions they made, in my opinion: a bit more work but a *lot* less of an ongoing money-sink, and since you have to clean the brush anyway, which requires opening the debris container, you wouldn't save any time by having a bag-type debris collector.
So, maybe you can tell me, what did this first chicken mate with? And after it mated with the non-a-chicken, was the offspring a sterile hybrid like a mule?
It's possible it was like lions and tigers: able to interbreed and produce fertile offspring. But it's more likely that it was simply geography: this group of not-chickens and that other group of not-chickens stopped hanging out because it was too long a walk, and 10,000 years later that group of not-chickens wasn't really not-chickens anymore, but something rather like chickens, and a million years later they were definitely chickens. The concept of species is entirely a human invention, and nature has no respect for it whatsoever: stuff that's living close together and is closely related, interbreeds way more than stuff that is distant or distantly related. Stretch that out a few hundred thousand generations, and you have what humans call species.
I do test circuit hardware design and we use standard relays all over the board, for switching bits of circuitry into and out of contact with an integrated circuit we're testing. We use mechanical relays because of the same reasons they say: zero leakage current when they're open, and extremely low resistance when they're closed, which semiconductor switches just can't equal. The problem is the lifetime of the relays, so we have to socket them all (which, when you're building a board with 500 relays on it, is a significant time and money sink) and replace them pretty often on high-running parts (some of our parts have been in high-volume production for 20 years.) Plus they're big and take up the majority of the board. Having a device that's tiny and can last a billion cycles would be completely awesome.
Gonna have to disagree with you here: there were eggs long before there were chickens. They weren't chicken eggs, but they were eggs nonetheless. Something that wasn't a chicken laid an egg, and from that egg hatched a chicken, to oversimplify a million years of evolution.
IANAL, but in general courts are very reluctant to find an entity liable for negligence for failing to anticipate someone else's negligence. That's sensible: "he should have anticipated that someone might be driving the wrong way down the one way street, so the accident is his fault!" would lead to complete chaos. In this case, the fire department could have anticipated his negligence, and did -- by trying to contact him to get the bill paid, they made a good faith attempt to anticipate his negligence. That's not to say that he deserved this -- he, and his pets particularly, deserved much better than the outcome.
But with that said, even if you could show that the fire department had some contributory negligence, if they're a part of the government in any capacity they have immunity against suits for negligence unless they specifically and explicitly allow themselves to be sued for negligence.
His machines could have been built in his time with his technologies. The problem was just that doing so would have taken a *lot* of time -- dozens of man-years of work. He was coming up with additions and modifications ten times as fast as the manufacturing capability of the time could produce.
The bearings of the time were primitive: a lot of machinery was still using greased leather in a compressable joint, and only the nicest equipment used plain bronze sleeve bearings. The gears were not precisely formed so there was a lot of friction between them. A major problem would have been the heat from friction, and the amount of power required to drive the machine against its internal friction, exceeding the mechanical strength of the components. You can to some extent solve the former by just running it slowly. You can to some extent solve the latter by lapping gears against each other with very fine grinding compound between the teeth. They knew these things -- they were custom-lapping threaded shafts to excellent tolerances at the time, for use as master threads in thread-cutting machines -- but the designs Babbage was coming up with would have required a very significant manufacturing effort, that would've required the output of dozens of factories, and there wasn't any obvious reason to make that effort.
I have a watchmaker's lathe built roughly in 1840, and I'm amazed how precisely they could make things back then almost entirely by hand. Ramsden and Maudsley were building stuff even more accurate before 1800.
So next, Google will get a patent on a super-ad that advertisers can pay a surcharge for, that will be played by broadcasters despite the $2 charge consumers have paid. And so the race continues...
Which makes me think of a question maybe someone here at/. can answer: Can miners/drillers set off enough "little ones" to cause a big one?
Yes. Doesn't even have to be mining. Deep well injection set off a series of earthquakes in Denver in the 1960's -- when what they were injecting was millions of liters of nerve gas from Rocky Mountain Arsenal. Yes, indeed, your government funds at work: make hundreds of thousands of pounds of wildly toxic weapons of mass destruction in contradiction to signed treaties, and then when you have them all and don't know what to do with them, pump them into a 25,000-foot-deep hole. They lubricated an old slip fault and caused a half a dozen earthquakes.
The thing is: it's not easy to tell, prima facie, whether you're going to cause a bunch of little earthquakes, or one big one, by doing this. It has been proposed that we should try to set off small earthquakes on purpose, to reduce the strain on tectonic plate boundaries and reduce the chances of a much larger quake, but there's no way to ensure that it would do that rather than just setting off the larger quake right now. If we do that, even if it just sets off small quakes, the situation is no longer an earthquake that's an act of God, but an earthquake that's an act of man, meaning even if it's much smaller than it would have been, specific people are now liable for causing an earthquake, and that's a legal minefield.
I sometimes wrestle with th question of whether or not dangerous behaviors should be punishable at whatever level it takes to make people stop doing them.
I think one problem is that for many people there isn't any deterrent effect to punishments because they don't think they'll get caught. Young children are nearly incapable of long-term planning. Teenagers are better at it, but still aren't always good at imagining how their actions right now could affect the entire rest of their lives, as a car crash could.
Another problem is that people, in general, are really bad at learning something when there isn't a consistent behavior/response system. You text something really important to someone while stopped at a stoplight, and nothing bad happens. So you do it again, and then while you're out on a country road, and still, nothing bad happens, so it seems okay and you keep doing it.
Architects are rarely versed in function, and are almost always about the form.
My sample size is small, but all three architects I know were required to take a statics class and do load calculations before they got their degrees. This wasn't sufficient for doing engineering for building design, and wasn't intended as such, but they were required to know enough to not hand engineers unbuildable designs.
The world record for preemies is about 18 weeks. My girlfriend was about two months premature, and her younger sister was about 16 weeks premature -- and is doing surprisingly well, after a whole lot of surgery in her first six months. She had her twenty-first birthday last month and we were joking that she was actually only 20 1/2. But my girlfriend said that when her sister was born, she easily fit in a person's hand, and was covered in dark brown hair all over her body, so for years they called her 'mouse' or 'rat'.
A few thoughts: the ADE7763 I mentioned is nice because it's so cheap, and a board built around it would cost about $5, with roughly $8 of components (chip, crystal, screw terminals, header for the SPI bus) so you could afford an Arduino and a half-dozen of these, and put them in the breaker box to measure multiple circuits simultaneously. But if you want to go appliance-by-appliance, the bus length and the per-unit cost are going to get very painful, and it seems like the tweet-a-watt kit, using a cut-up kill-a-watt meter with networking added, would be a better idea. Of course, that doesn't work with 220V appliances, and you have to get something like the SmartWatt, which isn't as user-friendly, although it does come with networking built in. It's also $20 over the article's price budget, which is a shame because it's pretty cool.
But if you're a serious DIY person, I think that analog chip is going to be hard to beat.
At work I use a Yokogawa WT210, and it's fantastic -- distortion out to the 50th harmonic, scads of power factor and real power measuring abilities -- but it's also like $7K, and a bear to learn to use. Le sigh.
With all THAT said, I don't know of any accurate, networkable, completely non-intrusive load monitoring hardware, which is a shame.
And Newark has them in stock. We've been having a horrible time in the last year trying to find distributors that actually have stuff in stock, and end up buying small lots at high prices from frightening places, or pleading for samples directly from the manufacturers. Parts we've been using for years suddenly have 10 week lead times. I'm trying to source one part the manufacturer swears is in production and rolling off the lines but "nobody's buying" and distributors are showing a 30 *week* lead time. Gah. No wonder nobody's buying.
You joke, but I've done that: put a handheld digital multimeter face-down on a flatbed scanner, and used gocr to get data. It was *very* awkward, and once I got gpib running, I've never looked back at that without shivering.
The Analog Designs ADE7763 is a pretty awesome chip for doing this sort of stuff. Here's the appnote in a pdf, and here's the chip itself. It's quite easily interfaced to an Arduino using SPI. I just laid out a board interfacing this to an ATMEGA1284 for doing power quality monitoring and logging, but it's for an internal project so I can't just hand out the code or layout, but it was a dead simple chip to work with: one crystal and two caps were all it required for support, and if it were interfaced to an Arduino, that could handle all the I/O to a computer or write to an SD card.
In addition to trying to deal with someone who doesn't remember you, there can be major inter-family problem because at least some Alzheimers patients can no longer tell what happened from what they think happened, so you run into situations where the person says "why doesn't my son ever come visit me?" when the son in question is sitting in the room, and other family members, not realizing that the person is flat-out wrong, start resenting one another based on the testimony of, well, a crazy person.
And then the patient starts saying the nursing staff is abusive, and that the doctor hurts her when he checks her out -- and who do you believe? because while you know your grandmother is now basically totally unreliable, you also know that the staff knows that too and if they *were* abusive, who would know? and so you run into situations where some family members are trying to keep the patient happy and others are saying the patient needs to be moved to another facility, and others are so upset about the whole situation that they've stopped even visiting...
When my grandmother finally died, the sense of relief was palpable, and several years later, most all her relatives are friendly to one another again, and that was with a good nursing home, everyone living locally and visiting every week, and all of us talking regularly about the problems with each other to try to keep exactly this from happening.
I'm not a lawyer. However, "knowingly" is usually done through discovery -- going through email, like you say -- but can be shown by demonstrating that a "reasonable person" would not have behaved that way.
And as for lawyers shielding their clients, in the Hewlett Packard pretexting scandal, CEO Patricia Dunn repeatedly asked her lawyers, in writing, if the methods involved in finding the leak were legal, and they repeatedly assured her, in writing, that they were. Despite that, several jurisdictions filed criminal charges and arrest warrants against her, although the charges were eventually dismissed.
I'm going to guess they'd launch it from one of the big strips in the Mojave desert and then set it over the US/Mexico or US/Canada border as very-long-term surveillance.
Plus if it runs for 5 years, why not just make 40 of them, launch them all, and leave them all over the world, like the GPS satellite network? If you've got 5 year misson plans, why bother hurrying when you can just have one local?
Both piston aircraft and commercial jet aircraft are propeller-based. Piston aircraft and turboprop aircraft have a big exposed propeller, powered by pistons or a geared jet, while commercial turbofan engines, as used on 90% of airline transport aircraft, have a set of propellers enclosed in a housing, that are powered by a jet.
Piston aircraft are *much* less reliable and have *much* lower time between overhauls than jets, because jets basically have a very small number of non-contact bearings, while piston aircraft have a quite large number of physically contacting bearings and sliding surfaces. Jet engines are much more expensive to build, but they last for tens of thousands of hours.
Jet engines are attractive to aircraft designers because they have high reliability, low vibration, have a much higher power-per-unit-weight and power-per-unit-volume, and are much less affected by the lack of air pressure at high altitude. They aren't as efficient as piston engines, and high bypass ratio jets, as seen on most transport jets, are not as efficient as turboprops. But their size, reliability, and power overcome the very slight efficiency loss.
It is possible you could run a turbofan with electric motors rather than a jet, but that would be somewhat pointless for this kind of design requirement.
I wish we could have made it less complicated but as the first off-line triac-decoding LED driver, it sort of needed to do everything. The eval boards for it aren't *that* expensive and do a good job of getting an experimental setup running... but it costs a lot less to buy one of the ecosmart bulbs from home depot that says "lighting science group" on the box -- this one, I believe -- because that already has the chip, the electronics, and blah blah blah, and it's a good layout with high-quality capacitors. One nice thing about modern drivers is less reliance on electrolytic caps, and there are some ref designs for this chip that use 100% ceramic caps (a lot of them, granted...) which means they should last tens of years with no problems.
One of the several tricky things about building a driver that runs off standard triac-based dimmers is that the way triac dimmers work is by either staying at zero volts until some point after the AC input starts to rise, then snapping up to that point, or calculating where in the AC cycle they are and shutting off before the AC goes back to zero. The issue at hand is that you have to run your whole lighting system based on having zero voltage running to your driver for at least part of the time, so there are all these fancy things you need to do to make sure you have power locally even if none is coming in. Most of those fancy things end up making a really messy load, with voltage and current flows that are very poorly aligned with each other, so then there's all this power factor correction stuff you have to run, in addition, and that soaks up a bit of power, too. Triacs are painful to work with, especially the really cheap dimmers that are, unsurprisingly, the most common ones around.
So, getting energy savings using standard dimmers is *difficult*, but it is possible. I've yet to see a CLF that dimmed nicely (or without putting a lot of noise on the line, or both) and at least we've managed to handle both those issues.
LED's fade but it should take 50,000 hours before the fade is even 25%, unless the light is designed terribly and the LED's are running hot.
Within 5 years most good LED lights will include closed-loop feedback with sensors that check both the light intensity and color and correct for LED aging by changing how the LED's are driven (underdriving them at first, and including about 15% red LED's in with the whites) so soon we should have lights that provide arbitrarily perfect light until they fail, which should be after about 75,000 hours. And in any case, they're *way* more efficient than incandescents and in my opinion look better than CFL's.
However, also in my opinion, the worry about mercury in CFL's is wildly overrated.
Dimmers are not suitable for fluorescent or LED bulbs, each should really be dimmed with a control signal instead of a rheostat. Hopefully this kind of technology will be common in homes soon so we can get rid of the buzzing from dimming fluorescent and LEDs.
The National Semiconductor LM3445 -- which I helped design -- is a fantastic LED driver specifically designed to decode standard wall triac dimmers. It works better than an incandescent light does: I haven't seen a design yet where it couldn't manage 1000:1 dim ratio from full on to as dim as possible. (Which means the light is emitting photons too dimly for you to see except in an extremely dark room.) There are many others coming onto the market right now that do the same thing, but I think we have a first-to-market advantage. Lightbulbs based on our chip are showing up in Home Depot and the like.
My guess is that within a couple years only the cheapest LED lights won't be dimmable because it's such a common expectation.
1. If the copyright mode failing means no more Twilight and Star Wars, I'm all for it.
2. I'm quite familiar with every one of those authors. I think it was unfortunate that Cervantes was so upset by an "unauthorized sequel" -- but, hey, it got him off his duff to write the authorized sequel. So we got *two* great books out of him. Compare to Edmond Rostand, who spent his whole life writing one (really excellent) play and some other also-rans. And while I'm at it -- how is writing an unauthorized sequel going to be solved by copyright? That's not copying, that's impersonation.
3. There probably wasn't a Homer -- The Odyssey and The Illiad were likely written by different people, and it's not even clear that what we have was anything like the originals. It's more likely they were community accretions of the works of many poets and playwrights, who were all, y'know, violating someone else's copyright.
People have *always* written stuff for fun. If we had no copyright, it wouldn't stop people writing stuff for fun, it'd just reduce the number of people doing it for money. I, personally, don't have a problem with that.
Slashdot Mirror
I guess the rest of the process is to very carefully drop the chip package where you want it to go, and convince the solder to melt. How do you do that? You mentioned a "reflow oven"? I Googled that, and there are lots of different ovens out there.
If you're using old-fashioned lead/tin solder you don't have to be all that careful about the alignment: if you have 10% overlap between the ball and the solderpad on most all the pads, the capillary action of the solder will cause the bga to self-align. (I do this on a regular basis with smaller but similar chips.) Now, if you're using modern no-lead solder it's more difficult. With a BGA you can do an okay job of alignment with a boom microscope set at an angle so you can see under the edge of the chip to the first row of balls: if you have opposite corners on the pads you're good.
And I use a hotplate and a thermocouple taped adjacent to the chip with the tip touching one of the pads. You don't want to overheat the chip itself -- no higher than 260C if you can possibly manage it. A lab hotplate is faster than a toaster oven, and you need to get the heat up fast enough the flux doesn't boil off before the solder reflows. Ideally you'd like it to reflow and hold above the liquidus of the solder (which is about 220C for lead and about 240C for no-lead) for 30 seconds, but managing that without going over 260C is pretty difficult. Then you cool it down pretty quickly. I use a hairdryer with the heat turned off so it's just blowing scads of cool air.
Likewise, by the way, I use a hotplate for adhering the balls to the BGA, rather than a hot air gun -- it's more predictable, temperature-wise.
Well, okay, all the above should be put in the past tense since as of a week ago we now have an OKI bga rework station, with a vision system that aligns the chip/board and places it, and a controllable temp profiler, and it's awesome. But I do have three years' experience hand-placing leadless chips and micro-SMD's, so I know a thing or two about it.
Doesn't the Roomba automatically go back to its charging station after it's done cleaning?
The only thing missing, AFAIK, it's the "empty itself" part. I'm puzzled over that one because they could simply add a "garbage can" part to the charging station...
It depends on the level of roomba -- both my roomba and my scooba have standard wall-wart chargers that I have to plug in by hand. But really that isn't the limiting factor: the brushes get stuff in them, particularly hair, and that more than anything else is the most difficult maintenance task.
They also have a paper filter in front of the main debris container, that has to be manually removed; automating that into a 'garbage can' sort of setup, would be tricky. Old-fashioned vacuums used a disposable paper bag to handle both tasks, dust removal and debris, but thankfully the roomba doesn't work that way or we'd be bankrupt since we run ours pretty much every day in one room or another. That's one of the good design decisions they made, in my opinion: a bit more work but a *lot* less of an ongoing money-sink, and since you have to clean the brush anyway, which requires opening the debris container, you wouldn't save any time by having a bag-type debris collector.
So, maybe you can tell me, what did this first chicken mate with? And after it mated with the non-a-chicken, was the offspring a sterile hybrid like a mule?
It's possible it was like lions and tigers: able to interbreed and produce fertile offspring. But it's more likely that it was simply geography: this group of not-chickens and that other group of not-chickens stopped hanging out because it was too long a walk, and 10,000 years later that group of not-chickens wasn't really not-chickens anymore, but something rather like chickens, and a million years later they were definitely chickens. The concept of species is entirely a human invention, and nature has no respect for it whatsoever: stuff that's living close together and is closely related, interbreeds way more than stuff that is distant or distantly related. Stretch that out a few hundred thousand generations, and you have what humans call species.
I do test circuit hardware design and we use standard relays all over the board, for switching bits of circuitry into and out of contact with an integrated circuit we're testing. We use mechanical relays because of the same reasons they say: zero leakage current when they're open, and extremely low resistance when they're closed, which semiconductor switches just can't equal. The problem is the lifetime of the relays, so we have to socket them all (which, when you're building a board with 500 relays on it, is a significant time and money sink) and replace them pretty often on high-running parts (some of our parts have been in high-volume production for 20 years.) Plus they're big and take up the majority of the board. Having a device that's tiny and can last a billion cycles would be completely awesome.
Gonna have to disagree with you here: there were eggs long before there were chickens. They weren't chicken eggs, but they were eggs nonetheless. Something that wasn't a chicken laid an egg, and from that egg hatched a chicken, to oversimplify a million years of evolution.
But with that said, even if you could show that the fire department had some contributory negligence, if they're a part of the government in any capacity they have immunity against suits for negligence unless they specifically and explicitly allow themselves to be sued for negligence.
The bearings of the time were primitive: a lot of machinery was still using greased leather in a compressable joint, and only the nicest equipment used plain bronze sleeve bearings. The gears were not precisely formed so there was a lot of friction between them. A major problem would have been the heat from friction, and the amount of power required to drive the machine against its internal friction, exceeding the mechanical strength of the components. You can to some extent solve the former by just running it slowly. You can to some extent solve the latter by lapping gears against each other with very fine grinding compound between the teeth. They knew these things -- they were custom-lapping threaded shafts to excellent tolerances at the time, for use as master threads in thread-cutting machines -- but the designs Babbage was coming up with would have required a very significant manufacturing effort, that would've required the output of dozens of factories, and there wasn't any obvious reason to make that effort.
I have a watchmaker's lathe built roughly in 1840, and I'm amazed how precisely they could make things back then almost entirely by hand. Ramsden and Maudsley were building stuff even more accurate before 1800.
So next, Google will get a patent on a super-ad that advertisers can pay a surcharge for, that will be played by broadcasters despite the $2 charge consumers have paid. And so the race continues...
Which makes me think of a question maybe someone here at /. can answer: Can miners/drillers set off enough "little ones" to cause a big one?
Yes. Doesn't even have to be mining. Deep well injection set off a series of earthquakes in Denver in the 1960's -- when what they were injecting was millions of liters of nerve gas from Rocky Mountain Arsenal. Yes, indeed, your government funds at work: make hundreds of thousands of pounds of wildly toxic weapons of mass destruction in contradiction to signed treaties, and then when you have them all and don't know what to do with them, pump them into a 25,000-foot-deep hole. They lubricated an old slip fault and caused a half a dozen earthquakes.
The thing is: it's not easy to tell, prima facie, whether you're going to cause a bunch of little earthquakes, or one big one, by doing this. It has been proposed that we should try to set off small earthquakes on purpose, to reduce the strain on tectonic plate boundaries and reduce the chances of a much larger quake, but there's no way to ensure that it would do that rather than just setting off the larger quake right now. If we do that, even if it just sets off small quakes, the situation is no longer an earthquake that's an act of God, but an earthquake that's an act of man, meaning even if it's much smaller than it would have been, specific people are now liable for causing an earthquake, and that's a legal minefield.
I sometimes wrestle with th question of whether or not dangerous behaviors should be punishable at whatever level it takes to make people stop doing them.
I think one problem is that for many people there isn't any deterrent effect to punishments because they don't think they'll get caught. Young children are nearly incapable of long-term planning. Teenagers are better at it, but still aren't always good at imagining how their actions right now could affect the entire rest of their lives, as a car crash could.
Another problem is that people, in general, are really bad at learning something when there isn't a consistent behavior/response system. You text something really important to someone while stopped at a stoplight, and nothing bad happens. So you do it again, and then while you're out on a country road, and still, nothing bad happens, so it seems okay and you keep doing it.
Architects are rarely versed in function, and are almost always about the form.
My sample size is small, but all three architects I know were required to take a statics class and do load calculations before they got their degrees. This wasn't sufficient for doing engineering for building design, and wasn't intended as such, but they were required to know enough to not hand engineers unbuildable designs.
The world record for preemies is about 18 weeks. My girlfriend was about two months premature, and her younger sister was about 16 weeks premature -- and is doing surprisingly well, after a whole lot of surgery in her first six months. She had her twenty-first birthday last month and we were joking that she was actually only 20 1/2. But my girlfriend said that when her sister was born, she easily fit in a person's hand, and was covered in dark brown hair all over her body, so for years they called her 'mouse' or 'rat'.
But if you're a serious DIY person, I think that analog chip is going to be hard to beat.
At work I use a Yokogawa WT210, and it's fantastic -- distortion out to the 50th harmonic, scads of power factor and real power measuring abilities -- but it's also like $7K, and a bear to learn to use. Le sigh.
With all THAT said, I don't know of any accurate, networkable, completely non-intrusive load monitoring hardware, which is a shame.
And Newark has them in stock. We've been having a horrible time in the last year trying to find distributors that actually have stuff in stock, and end up buying small lots at high prices from frightening places, or pleading for samples directly from the manufacturers. Parts we've been using for years suddenly have 10 week lead times. I'm trying to source one part the manufacturer swears is in production and rolling off the lines but "nobody's buying" and distributors are showing a 30 *week* lead time. Gah. No wonder nobody's buying.
You joke, but I've done that: put a handheld digital multimeter face-down on a flatbed scanner, and used gocr to get data. It was *very* awkward, and once I got gpib running, I've never looked back at that without shivering.
The Analog Designs ADE7763 is a pretty awesome chip for doing this sort of stuff. Here's the appnote in a pdf, and here's the chip itself. It's quite easily interfaced to an Arduino using SPI. I just laid out a board interfacing this to an ATMEGA1284 for doing power quality monitoring and logging, but it's for an internal project so I can't just hand out the code or layout, but it was a dead simple chip to work with: one crystal and two caps were all it required for support, and if it were interfaced to an Arduino, that could handle all the I/O to a computer or write to an SD card.
And then the patient starts saying the nursing staff is abusive, and that the doctor hurts her when he checks her out -- and who do you believe? because while you know your grandmother is now basically totally unreliable, you also know that the staff knows that too and if they *were* abusive, who would know? and so you run into situations where some family members are trying to keep the patient happy and others are saying the patient needs to be moved to another facility, and others are so upset about the whole situation that they've stopped even visiting...
When my grandmother finally died, the sense of relief was palpable, and several years later, most all her relatives are friendly to one another again, and that was with a good nursing home, everyone living locally and visiting every week, and all of us talking regularly about the problems with each other to try to keep exactly this from happening.
And as for lawyers shielding their clients, in the Hewlett Packard pretexting scandal, CEO Patricia Dunn repeatedly asked her lawyers, in writing, if the methods involved in finding the leak were legal, and they repeatedly assured her, in writing, that they were. Despite that, several jurisdictions filed criminal charges and arrest warrants against her, although the charges were eventually dismissed.
Where the hell are you going to launch it from?
I'm going to guess they'd launch it from one of the big strips in the Mojave desert and then set it over the US/Mexico or US/Canada border as very-long-term surveillance.
Plus if it runs for 5 years, why not just make 40 of them, launch them all, and leave them all over the world, like the GPS satellite network? If you've got 5 year misson plans, why bother hurrying when you can just have one local?
Piston aircraft are *much* less reliable and have *much* lower time between overhauls than jets, because jets basically have a very small number of non-contact bearings, while piston aircraft have a quite large number of physically contacting bearings and sliding surfaces. Jet engines are much more expensive to build, but they last for tens of thousands of hours.
Jet engines are attractive to aircraft designers because they have high reliability, low vibration, have a much higher power-per-unit-weight and power-per-unit-volume, and are much less affected by the lack of air pressure at high altitude. They aren't as efficient as piston engines, and high bypass ratio jets, as seen on most transport jets, are not as efficient as turboprops. But their size, reliability, and power overcome the very slight efficiency loss.
It is possible you could run a turbofan with electric motors rather than a jet, but that would be somewhat pointless for this kind of design requirement.
I wish we could have made it less complicated but as the first off-line triac-decoding LED driver, it sort of needed to do everything. The eval boards for it aren't *that* expensive and do a good job of getting an experimental setup running... but it costs a lot less to buy one of the ecosmart bulbs from home depot that says "lighting science group" on the box -- this one, I believe -- because that already has the chip, the electronics, and blah blah blah, and it's a good layout with high-quality capacitors. One nice thing about modern drivers is less reliance on electrolytic caps, and there are some ref designs for this chip that use 100% ceramic caps (a lot of them, granted...) which means they should last tens of years with no problems.
So, getting energy savings using standard dimmers is *difficult*, but it is possible. I've yet to see a CLF that dimmed nicely (or without putting a lot of noise on the line, or both) and at least we've managed to handle both those issues.
Not a chance. "Fading White LEDs"
LED's fade but it should take 50,000 hours before the fade is even 25%, unless the light is designed terribly and the LED's are running hot.
Within 5 years most good LED lights will include closed-loop feedback with sensors that check both the light intensity and color and correct for LED aging by changing how the LED's are driven (underdriving them at first, and including about 15% red LED's in with the whites) so soon we should have lights that provide arbitrarily perfect light until they fail, which should be after about 75,000 hours. And in any case, they're *way* more efficient than incandescents and in my opinion look better than CFL's.
However, also in my opinion, the worry about mercury in CFL's is wildly overrated.
The National Semiconductor LM3445 -- which I helped design -- is a fantastic LED driver specifically designed to decode standard wall triac dimmers. It works better than an incandescent light does: I haven't seen a design yet where it couldn't manage 1000:1 dim ratio from full on to as dim as possible. (Which means the light is emitting photons too dimly for you to see except in an extremely dark room.) There are many others coming onto the market right now that do the same thing, but I think we have a first-to-market advantage. Lightbulbs based on our chip are showing up in Home Depot and the like.
My guess is that within a couple years only the cheapest LED lights won't be dimmable because it's such a common expectation.
2. I'm quite familiar with every one of those authors. I think it was unfortunate that Cervantes was so upset by an "unauthorized sequel" -- but, hey, it got him off his duff to write the authorized sequel. So we got *two* great books out of him. Compare to Edmond Rostand, who spent his whole life writing one (really excellent) play and some other also-rans. And while I'm at it -- how is writing an unauthorized sequel going to be solved by copyright? That's not copying, that's impersonation.
3. There probably wasn't a Homer -- The Odyssey and The Illiad were likely written by different people, and it's not even clear that what we have was anything like the originals. It's more likely they were community accretions of the works of many poets and playwrights, who were all, y'know, violating someone else's copyright.
People have *always* written stuff for fun. If we had no copyright, it wouldn't stop people writing stuff for fun, it'd just reduce the number of people doing it for money. I, personally, don't have a problem with that.