683lm/W is the maximum luminous efficacy for light, yes, but that's green light.
To reproduce in full the solar spectrum so that it is indistinguishable from white light requires you to produce a 'white' that produces light from about 400-700nm (UV to IR borders). If you take into account flourescence and its effect on colour, perhaps 350nm is the top end. This would take perhaps 180lm/W.
As you move from near-solar (or tungsten) identical bulbs to more limited 'whites' - you get about 250-400lm/W being the maximum. This varies from pretty good white that you won't notice being different from actual white to something rather more limited, with just blue at 430nm or so, and greenish yellow at 560nm. This will to a cursory glance look right, but will have truly wretched colour reproduction.
Which is not completely, though largely irrelevant to the question of using them as small PCs or embedded controllers. The Ouya has - on a quick glance - no onboard accessable peripheral interfaces.
Then there is the elephant in the room. The Pi is deeply unexceptional, and rather boring hardware. Even the price isn't that special. The exceptional bit is that there are a sizeable slice of half a million of them. This means that even if 99% of them are sitting on a shelf, you have many thousands of people banging on the hardware, and bugs are at least likely to be found in many cases.
With most of the alternative boards, you're going to be the only one of a relative few with them.
Because they're not innovations, in most cases. Being the first to come up with a problem, too often these days leads to a patent. Take multitouch gesture patents. These are essentially all non-innovative. You start out with one perhaps innovative invention - the touchscreen over a computer display. Many subsequent developments were not really innovations, but outgrowths of what is made technically possible by the engineering. Once you have the concept of a computer display with a touchscreen on top of it, the initial concept - probably of a screen you can't lift, with a coarse resolution grid on top just evolves as technology makes it possible to form factors of phones and tablets. Putting a touchscreen on a phone is not a novel or useful idea which should be patentable. Humans have a limited humber of fingers, but it's more than one. The concept of tracking them all isn't really novel, it's just that it became an engineering possibility. Computers got fast enough that they could meaningfully track gestures - again this is not really innovative.
If you are the first person with a multitouch device - pinch and zoom is _obvious_ - it's the most sensible way to implement zoom. This should never be a patentable innovation. Patenting it does nothing but _slow_ innovation, as clearly obvious techniques like this have to be worked around.
Essentially all the parts of a microsatellite are deeply involved with electronics. The smaller you can make them, and the lower power, the better. Lower power smaller electronics mean simpler cases, lighter batteries, solar cells,... Cleverer electronics can remove the need for mechanical parts - if you can trigger the shutter of a camera when pointing in the right direction, versus having a fully stabilised satellite, for example.
The hard part is to ensure it's powering 'your own home'.
Or is it powering your own home, plus the dangling wire that the power company guy is going to reconnect, and doesn't bother testing, as he knows it's on the 'dead' side of the break.
This is why all grid tie inverters have robust 'islanding' protection.
I do hope you've not been using that sig for long.
"The use of the term "patent pending" or "patent applied for" is permitted so long as a patent application has actually been filed. If these terms are used when no patent application has been filed it is deemed as a deceptive act and a fine of up to $500 may be imposed for every such offense.[6] Under the current interpretation of "offense", each mis-marked article constitutes an offense, which permits theoretical damages in the hundreds of millions of dollars for high-volume consumer goods.[7]"
You're assuming a whole bunch of things in that. Primarily that the number of people needed to service the machine is equal to the number of staff replaced. This seems at best extremely questionable.
Secondly - half of people are not as smart as the average. They are unlikely to be able to get employment designing robots, or...
The noise floor is not that simple. If you analyse the noise, you find it's broadly spread over a wide variety of frequencies. If you analyse 10s or so slices of signal, over the range 47-53Hz, things get considerably easier. 10 seconds means that your effective number of signal levels is not 2^16 (65536) It's sqrt(44100*10)*65535, about 25 bits. Throwing out the out of band noise means you lose >99.5% or so of it.
As a practical measure. Go to pretty much any CD you have, and do a FFT in the range 47-53 or 57-63hz. In the vast majority of cases, you will find a signal.
And it's not quite a timestamp, unless the recording is quite long. http://www.nationalgrid.com/uk/Electricity/Data/Realtime/Frequency/Freq60.htm - for example - this is the UKs last hours frequency. The graph is clearly enough to show that even if you can 'only' resolve in 10 second intervals the frequency, it's quite plausible to say if a 10 minute video can be one specific timeframe, or not.
GPS was essential in 1989. Today, your average smartphone (without GPS) just using the camera, onboard sensors, and a few tens of gigs of stored imagery can get really accurate position tracking, at least in good weather. Less ideal in broken cloud.
Robot warehouses, trucks, inventory, already exist. Many of those jobs are temporary. Retooling and programming the robots, sure, those are high skilled jobs. That 99.9% of people won't be able to do through not having the skills or intelligence.
Which you can't do without ripping out the library that talks to the original host, as you won't have a licence for it. So, you now need to reimplement both client libraries, and servers, at a time when you can't test your mods against the original, because they've shut down the servers. This is not going to be cheap.
The above includes air and water as consumables. 4.5Kg is about a minimum for no recycling, just dumping the waste, and inputting food, water, and oxygen.
Well, yes, and no. Fetch all pages to a depth of three clicks on a fetch, and you've gotten to 14 min a page. Preload sites you regularly visit. There will be some commonality between the requested pages for the Aeronauts, so caching helps. Randomly clicking around will be tedious at best.
But, you can have slashdot, xkcd, wikipedia all effectively 'live'. Youtube would be challenging, for more than your subscribed channels. netflix et al would be fine.
As some context on supplies. Completely open loop - no recycling - is 4.5Kg/day. If you recycle the dehumidifier water and urine, that comes down to 1.6kg/day.
It's plausible to get it down to about a kilo a day, without having to do really hard things - close the carbon/nitrogen loops. This means ten tons gets you 30 years of food.
DRM, in some aspects - trusted computing - can be a positive thing. My ideal system would have a root key I can set, that without software signed by it, it is a rock.
To decode this somewhat. "One independent claim requires that "the device [to be charged] tunes the resonance circuit to at least one of the resonance frequencies of the NFMR power supply and subsequently de-tunes the resonance circuit to provide a device identification to the NFMR power supply using a change in a resonance circuit load factor".
Sounds at least basically clever to me."
This is how essentially 90% of existing RFID tags work.
"3. A battery charging circuit, comprising: a first node arranged to receive wirelessly provided power;..." - they connect the antenna through a rectifier to a capacitor, and thence to a battery. This is the most trivially obvious way to do it. I built one of these in around 1986, and every wireless toothbrush in the world does this.
"5. A method of wirelessly transmitting power," and this is a wireless wireless charger. This is rather more novel, and relies on the fact that the wireless wireless charger can be larger than the final device to be charged, so have a greater range from the original power source than the charger.
This would require a fairly large 'transmit' antenna. The problems with near field charging are well known. The field cannot be focussed effectively, due to the nature of the magnetic field. the field production efficiency is limited by the size and weight of the transmitter. Every metal object in the field (which may extend behind/underneath the charger) will 'short out' the field, and take some energy. Resonance is not magic in any way at all.
For every arrangement of coils, and random debris (keys,...) in the way, there will be an optimum frequency that results in smallest loss.
As an analogy - you're trying to power a device by jumping up and down on a trampoline, and having a reciever of that motion somewhere on the trampoline.
Other large masses will disrupt the bouncing, and reduce the fraction of power going into your device.
The average worker, doing an average job, unless working on the concept of the patent takes a significant time, should never ever be able to be granted a patent.
At the moment, you can get a patent for being the first to think of a problem (not a solution) in an emerging field, and coming up with the obvious solution.
I'm reminded of a patent on how to make a gasket that wouldn't absorb soundwaves. It was about the second or third thing any competent engineer would try, given even a poorly equipped engineering shop.
If an engineer in the future that's never heard of your patent can reinvent it trivially, you should never have been granted it. The point of patents is to protect truly novel and important ideas.
Slashdot Mirror
It's way, way, way more complex than this.
683lm/W is the maximum luminous efficacy for light, yes, but that's green light.
To reproduce in full the solar spectrum so that it is indistinguishable from white light requires you to produce a 'white' that produces light from about 400-700nm (UV to IR borders).
If you take into account flourescence and its effect on colour, perhaps 350nm is the top end.
This would take perhaps 180lm/W.
As you move from near-solar (or tungsten) identical bulbs to more limited 'whites' - you get about 250-400lm/W being the maximum.
This varies from pretty good white that you won't notice being different from actual white to something rather more limited, with just blue at 430nm or so, and greenish yellow at 560nm.
This will to a cursory glance look right, but will have truly wretched colour reproduction.
Which is not completely, though largely irrelevant to the question of using them as small PCs or embedded controllers. The Ouya has - on a quick glance - no onboard accessable peripheral interfaces.
Then there is the elephant in the room.
The Pi is deeply unexceptional, and rather boring hardware.
Even the price isn't that special.
The exceptional bit is that there are a sizeable slice of half a million of them.
This means that even if 99% of them are sitting on a shelf, you have many thousands of people banging on the hardware, and bugs are at least likely to be found in many cases.
With most of the alternative boards, you're going to be the only one of a relative few with them.
Not worksafe.
Because they're not innovations, in most cases.
Being the first to come up with a problem, too often these days leads to a patent.
Take multitouch gesture patents.
These are essentially all non-innovative.
You start out with one perhaps innovative invention - the touchscreen over a computer display.
Many subsequent developments were not really innovations, but outgrowths of what is made technically possible by the engineering.
Once you have the concept of a computer display with a touchscreen on top of it, the initial concept - probably of a screen you can't lift, with a coarse resolution grid on top just evolves as technology makes it possible to form factors of phones and tablets.
Putting a touchscreen on a phone is not a novel or useful idea which should be patentable.
Humans have a limited humber of fingers, but it's more than one.
The concept of tracking them all isn't really novel, it's just that it became an engineering possibility.
Computers got fast enough that they could meaningfully track gestures - again this is not really innovative.
If you are the first person with a multitouch device - pinch and zoom is _obvious_ - it's the most sensible way to implement zoom.
This should never be a patentable innovation.
Patenting it does nothing but _slow_ innovation, as clearly obvious techniques like this have to be worked around.
Before you have to make something that works in space, you first have to be able to make something equivalent that does not work in space.
http://www.youtube.com/watch?v=JHHI2Lk79cY
Essentially all the parts of a microsatellite are deeply involved with electronics. ...
The smaller you can make them, and the lower power, the better.
Lower power smaller electronics mean simpler cases, lighter batteries, solar cells,
Cleverer electronics can remove the need for mechanical parts - if you can trigger the shutter of a camera when pointing in the right direction, versus having a fully stabilised satellite, for example.
You install or do not install.
The hard part is to ensure it's powering 'your own home'.
Or is it powering your own home, plus the dangling wire that the power company guy is going to reconnect, and doesn't bother testing, as he knows it's on the 'dead' side of the break.
This is why all grid tie inverters have robust 'islanding' protection.
I do hope you've not been using that sig for long.
"The use of the term "patent pending" or "patent applied for" is permitted so long as a patent application has actually been filed. If these terms are used when no patent application has been filed it is deemed as a deceptive act and a fine of up to $500 may be imposed for every such offense.[6] Under the current interpretation of "offense", each mis-marked article constitutes an offense, which permits theoretical damages in the hundreds of millions of dollars for high-volume consumer goods.[7]"
.2%
You're assuming a whole bunch of things in that.
Primarily that the number of people needed to service the machine is equal to the number of staff replaced.
This seems at best extremely questionable.
Secondly - half of people are not as smart as the average. ...
They are unlikely to be able to get employment designing robots, or
The noise floor is not that simple.
If you analyse the noise, you find it's broadly spread over a wide variety of frequencies.
If you analyse 10s or so slices of signal, over the range 47-53Hz, things get considerably easier.
10 seconds means that your effective number of signal levels is not 2^16 (65536)
It's sqrt(44100*10)*65535, about 25 bits.
Throwing out the out of band noise means you lose >99.5% or so of it.
As a practical measure.
Go to pretty much any CD you have, and do a FFT in the range 47-53 or 57-63hz.
In the vast majority of cases, you will find a signal.
And it's not quite a timestamp, unless the recording is quite long.
http://www.nationalgrid.com/uk/Electricity/Data/Realtime/Frequency/Freq60.htm - for example - this is the UKs last hours frequency.
The graph is clearly enough to show that even if you can 'only' resolve in 10 second intervals the frequency, it's quite plausible to say if a 10 minute video can be one specific timeframe, or not.
GPS was essential in 1989.
Today, your average smartphone (without GPS) just using the camera, onboard sensors, and a few tens of gigs of stored imagery can get really accurate position tracking, at least in good weather. Less ideal in broken cloud.
Robot warehouses, trucks, inventory, already exist.
Many of those jobs are temporary.
Retooling and programming the robots, sure, those are high skilled jobs.
That 99.9% of people won't be able to do through not having the skills or intelligence.
Which you can't do without ripping out the library that talks to the original host, as you won't have a licence for it.
So, you now need to reimplement both client libraries, and servers, at a time when you can't test your mods against the original, because they've shut down the servers.
This is not going to be cheap.
The above includes air and water as consumables.
4.5Kg is about a minimum for no recycling, just dumping the waste, and inputting food, water, and oxygen.
Well, yes, and no.
Fetch all pages to a depth of three clicks on a fetch, and you've gotten to 14 min a page.
Preload sites you regularly visit.
There will be some commonality between the requested pages for the Aeronauts, so caching helps.
Randomly clicking around will be tedious at best.
But, you can have slashdot, xkcd, wikipedia all effectively 'live'.
Youtube would be challenging, for more than your subscribed channels.
netflix et al would be fine.
As some context on supplies.
Completely open loop - no recycling - is 4.5Kg/day.
If you recycle the dehumidifier water and urine, that comes down to 1.6kg/day.
It's plausible to get it down to about a kilo a day, without having to do really hard things - close the carbon/nitrogen loops.
This means ten tons gets you 30 years of food.
DRM, in some aspects - trusted computing - can be a positive thing.
My ideal system would have a root key I can set, that without software signed by it, it is a rock.
To decode this somewhat.
"One independent claim requires that "the device [to be charged] tunes the resonance circuit to at least one of the resonance frequencies of the NFMR power supply and subsequently de-tunes the resonance circuit to provide a device identification to the NFMR power supply using a change in a resonance circuit load factor".
Sounds at least basically clever to me."
This is how essentially 90% of existing RFID tags work.
"3. A battery charging circuit, comprising: a first node arranged to receive wirelessly provided power;..." - they connect the antenna through a rectifier to a capacitor, and thence to a battery.
This is the most trivially obvious way to do it. I built one of these in around 1986, and every wireless toothbrush in the world does this.
"5. A method of wirelessly transmitting power," and this is a wireless wireless charger.
This is rather more novel, and relies on the fact that the wireless wireless charger can be larger than the final device to be charged, so have a greater range from the original power source than the charger.
This would require a fairly large 'transmit' antenna.
The problems with near field charging are well known.
The field cannot be focussed effectively, due to the nature of the magnetic field.
the field production efficiency is limited by the size and weight of the transmitter.
Every metal object in the field (which may extend behind/underneath the charger) will 'short out' the field, and take some energy.
Resonance is not magic in any way at all.
For every arrangement of coils, and random debris (keys, ...) in the way, there will be an optimum frequency that results in smallest loss.
As an analogy - you're trying to power a device by jumping up and down on a trampoline, and having a reciever of that motion somewhere on the trampoline.
Other large masses will disrupt the bouncing, and reduce the fraction of power going into your device.
The average worker, doing an average job, unless working on the concept of the patent takes a significant time, should never ever be able to be granted a patent.
At the moment, you can get a patent for being the first to think of a problem (not a solution) in an emerging field, and coming up with the obvious solution.
I'm reminded of a patent on how to make a gasket that wouldn't absorb soundwaves.
It was about the second or third thing any competent engineer would try, given even a poorly equipped engineering shop.
If an engineer in the future that's never heard of your patent can reinvent it trivially, you should never have been granted it.
The point of patents is to protect truly novel and important ideas.
Pop-quiz.
Will the above take seconds, hours, or a century?