It's not just irradiance (W/m2) that counts for retinal damage for two reasons. 1. Removal of heat is much more efficient from a small spot than from a big spot (3D versus 2D heat transfer). 2. Involuntary drift of the eye spreads out the dose if the spot size is small. (Try fixing your gaze at a spot for 10 seconds - you can't).
Some of the camera damage was in the aperure blades. Those were not in the image plane of the lens (similar to your irises). Those get quite a bit more dose if there is a big-diameter lens in front of them.
Disclosure: years ago, I reasoned that you wouldn't get blind from looking into the sun for 0.3 seconds, with binoculars, based on your irradiance argument. And tested it. Well, I didn't get permanent eye damage, but the after-image was 8x bigger in diameter than that of the sun with the naked eye and lasted for a day - rather disturbing. That was before I learned about the mechanisms of laser-induced damage.
I wonder whether unadvertised SDXC capabilities exist in devices, where the electric signals all work, but without the OS support for the patented ExFAT file system. You can't advertise SDXC without ExFAT support, but surely you are allowed to make an electrical connection?
Why would there be much labor? Cameras aren't mounted in the dashboard
The monitors are (I quote you: "and that's including the monitor!"). You want the camera without a monitor to hook into the already-present dashboard monitor syste? Now you need to ensure that the camera and the dashboard hardware talk to each other reliably even though they are made by different companies. I've had a phone where the camera would occasionally stop working until a reboot.
The automaker is buying *millions* of the same camera. You've never heard of price breaks?
The company I work for sells expensive industrial equipment that costs our customer tons of money if it's malfunctioning due to a faulty component (and we are on the hook, too, if the yearly downtime hours exceed a certain threshold). The components that we buy from our suppliers need to comply with many pages of requirements and the suppliers need to provide a paper trail of all the quality checks and test results. Believe me, that does not make things cheaper. I suppose that the car industry is more price-conscious than we are, though.
You say: "it's cheaper because they buy them by the million". I say: "it's probably not cheaper because of (1) quality differences, (2) labor/logistics costs, and (3) need to have the same camera model on stock for 10 years." But neither of us really knows the numbers.
You can get Chinese backup camera systems on Ebay for under $100 now, and that's including the monitor! The cost to an auto OEM for just a backup camera is going to be a fraction of that.
I wouldn't be so sure about that. Car manufacturers have to factor in the serviceability. If a camera system breaks during the warranty period, they are on the hook for the cost of the replacement, which may involve disassembling the dashboard of the car (i.e. lots of labor hours.), not to mention the reputation damage if it turns out that the camera system of car X tends to break within 3 years. Hence, they have a quite different requirement on reliability than a private person who orders a gadget from Ebay. And they need to ensure that spare parts will still be available for the next 10 years.
If you look at the front-page figure and Fig. 1 in the patent application, it will become apparent that the device only works for light rays entering at one particular angle. Essentially, it's like looking through louvers with an aspect ratio of 1:4 (fig. 1) or 1:6 (front page), which means that you won't be able to see through at all if your eyes are off by 25 cm horizontal at 1 m distance (i.e., passenger-side pillars) and whatever you see is substantially obscured for smaller angles. This is roughly how those 3M privacy screens for laptops work. That might be barely acceptable for the passenger-side pillars, but would be completely unworkable for the driver-side pillars.
Moreover, the surfaces 126 and 146 in Fig. 2 will need to be polarizing filters or opaque black surfaces so that you don't get to see spurious overlaid images. If you make them black, you will have replaced the obscuration of the pillar by two big black sheets that are only invisible if you look at them from one particular angle. If you make them polarizing absorbers, good luck in manufacturing those such that they don't reflect at grazing-incidence angles. (Those surfaces are mentioned in paragraph 37, without further reasoning about the benefits or tradeoffs, suggesting that the inventors don't know yet how to deal with these issues.)
By the way, the inventors have the polarizations the wrong way around in the figures. Although the correctly mention that p-polarization is transmitted and s-polarization is reflected, they have the arrows indicating the light polarization the wrong way around...
A bit of apples and oranges. Wifi is restricted to max 20 mW or so with an omnidirectional antenna, less if you use a directional antenna. Telecom uses higher power (250 mW on the phone from a quick Google) and uses high-gain directional antennas at the cell tower. You would't be allowed to use such high powers in the unlicensed wifi band.
It's difficult to do because Google provides this root detection service (Google Safety Net). That 'snet' program is self-updating and probably runs with elevated permissions. Whenever a developer releases an easy-to-use and somewhat popular root-hiding tool, Google will implement a tailored detection scheme for that particular tool and silently roll it out.
Printer dots are binary; screen pixels have 256 shades of gray or 256^3 colors. When printing full color, the effective resolution of a laser printer is way lower than 1200 ppi.
Adiabatic compressed air energy moves the heat from compression into an insulated thermal mass chamber, and uses that to heat the expansion vessel. It recouperates that loss and has 70% total effective energy storage--higher is possible, up to 90%.
The way you describe it sounds not at all like an adiabatic (no heat exchange) process but rather like a thermodynamically irreversible process. Maybe you mean an isothermal process?
And from https://en.wikipedia.org/wiki/... : "real compressors and turbines are not isentropic, but instead have an isentropic efficiency of around 85%, with the result that round-trip storage efficiency for adiabatic systems is also considerably less than perfect."
Wikipedia doesn't say so much about isothermal compression in practice, but it sounds to me like practically infeasible to reach your 70-90% roundtrip efficiency. You'd need an enormous thermal mass that doesn't change much in temperature despite 1/3 or so of the energy being stored as heat, you need enormous heat exchangers to transfer the heat with negligible temperature differences, and your compressors and decompressors need to have a very high efficiency.
IMHO: A USB device that depends on its power source to limit its input current, and can be damaged by a host that is willing to deliver more current that it requested, is defective by design.
Current limiting is to protect the supplier of the current. Bad current negotiation can damage the power supply, so of course the power supply should limit the current. A bad power supply may break, though. I think the main problem is that USB-C can use a range of voltages and a 5 V device plugged into a 20 V power supply will blow up the device unless the 20 V supply is signaled to throttle back to 5 V.
But maybe I misunderstand. Unfortunately, the reporting about this topic (Leung's findings) is very fuzzy about what happens exactly. But I don't see any scenario where connecting a good quality 5V, 1A or 2A charger to any USB-C device can lead to damage.
"your uncommon hardware may get burned out early (I've lost many wifi / gps / bluetooth / 3g-4g chipsets that way)."
This is the first time I hear smt like this. Do you have evidence?
The gps radio in my nexus 5 (with CM) tends to deteriorate over the months, but it turned out that tightening the screws of the rf shield restores the reception. Not a software issue.
Moreover, I think that Google's three support years count from the date the model is discontinued from the Play store, which may well be 2 years after the model is released.
A modern dac is supposed to digitally filter at 20 kHz (very hard cutoff), in exchange for lots of noise above 100 kHz. A soft roll-off analog filter takes care of the content above 100 kHz.
I suspect that ultrasound in this context really means 16 kHz or so, at volumes that are too low for the ear, but easily picked up by a microohone and some signal processing.
What kind of audio source did you use to find that you can hear 22 kHz? Unless you have an ultra low noise analog sine wave generator and amplifier, you are likely to hear noise, artifacts of the DA converter, and effects of clipping when you crank up the volume. The stated limit of 20 kHz for the human ear is the frequency where the pain threshold and the hearing threshold coincide for an average young person, so it is likely that you need >110 dB SPL to have any chance of hearing above 20 kHz. A device with 90 dB S/N ratio will likely blast you with lower frequencies, which you could easily mistake for "I can hear 20+ kHz".
I did once fry the tweeters of my (60 watt rated) loudspeakers while trying to test the limits of my hearing as I was sending all amplifier output power to the tweeters.
Although the general point is probably correct that min/max charge levels affect Li-ion lifespan, I would be careful with using batteryuniversity as a source of information. It's really a weblog of a small manufacturer of battery chargers that presents a couple of small-scale test results as if they represent the absolute truth for all brands and types of Li-ion batteties.
Until a music-streaming provider complains about being discriminated against, I don't think the review of the whitelist is the problem. The legal issue seems to be about the question whether discrimination by type of content (music, in this case), regardless of the company that provides the content, is allowed or not.
My feeling is that T-mobile is in a weak position here. Because if T-mobile is right, they would also be allowed to discriminate against other types of content; for example by making VoIP traffic more expensive. In fact, this is already happening: the cost of any data traffic is increased to subsidize the bandwidth for music data.
Clearly, IANAL, and the difference between "law" and "regulation" is fuzzy for me. But I think it happens every now and then that a local law/regulation is overturned because it conflicts with an EU regulation. Couldn't that be the case here?
Spotify, mentioned in the summary, is a bad example, since it is one of the many streaming services that is whitelisted. T-mobile allows any streaming provider such as spotify to sign up, without restriction or charges. According to T-mobile, this is allowed by the European law, which takes precedence over Dutch law. So, they are appealing in court and continuing their service for the time being.
What is a typical cost of a car, per distance, in the US? In the Netherlands, about 0.19 EUR/km is the bare minimum in marginal costs for a small car with good fuel economy.
"I've yet to hear a reason why removable batteries are a bad thing for consumers with the sole exception of anorexia."
I used to have an HTC Desire S, with removable battery. A drop from 5 cm with the wrong angle would temporarily disconnect the battery, leading to a power-off. I had to put in a spacer to push the battery harder against the spring contact. (it took quite a while to figure out what the problem was)
I have a low-end LG L40 with removable battery. Every time I drop the phone, the back cover will come off and I end up with a separate battery, cover, and front. At least it immediately removed its own battery when I dropped it in the toilet. (phone is still working after a rinse in distilled water and thorough drying)
And I have a Nexus 5 with "non-removable" battery. The back cover can still come off with some effort; the battery can be replaced although it's a bit more work. And I don't have the above problems with self-disassembly and flaky contacts.
Norwegian 100% hydro power is subject to double book keeping. Other European countries try to meet their CO2 goals by buying green certificates from Norway, so that on paper lots of the Norwegian electrons become "coal" and lots of other European electrons become "green".
Where did you get "It only became a serious flaw recently"? It happened somewhere between 11 years ago and present, but no date is given for commit f33ea7f404e5 (I did try Google).
Slashdot Mirror
It's not just irradiance (W/m2) that counts for retinal damage for two reasons. 1. Removal of heat is much more efficient from a small spot than from a big spot (3D versus 2D heat transfer). 2. Involuntary drift of the eye spreads out the dose if the spot size is small. (Try fixing your gaze at a spot for 10 seconds - you can't).
Some of the camera damage was in the aperure blades. Those were not in the image plane of the lens (similar to your irises). Those get quite a bit more dose if there is a big-diameter lens in front of them.
Disclosure: years ago, I reasoned that you wouldn't get blind from looking into the sun for 0.3 seconds, with binoculars, based on your irradiance argument. And tested it. Well, I didn't get permanent eye damage, but the after-image was 8x bigger in diameter than that of the sun with the naked eye and lasted for a day - rather disturbing. That was before I learned about the mechanisms of laser-induced damage.
I wonder whether unadvertised SDXC capabilities exist in devices, where the electric signals all work, but without the OS support for the patented ExFAT file system. You can't advertise SDXC without ExFAT support, but surely you are allowed to make an electrical connection?
The monitors are (I quote you: "and that's including the monitor!"). You want the camera without a monitor to hook into the already-present dashboard monitor syste? Now you need to ensure that the camera and the dashboard hardware talk to each other reliably even though they are made by different companies. I've had a phone where the camera would occasionally stop working until a reboot.
The company I work for sells expensive industrial equipment that costs our customer tons of money if it's malfunctioning due to a faulty component (and we are on the hook, too, if the yearly downtime hours exceed a certain threshold). The components that we buy from our suppliers need to comply with many pages of requirements and the suppliers need to provide a paper trail of all the quality checks and test results. Believe me, that does not make things cheaper. I suppose that the car industry is more price-conscious than we are, though.
You say: "it's cheaper because they buy them by the million". I say: "it's probably not cheaper because of (1) quality differences, (2) labor/logistics costs, and (3) need to have the same camera model on stock for 10 years." But neither of us really knows the numbers.
I wouldn't be so sure about that. Car manufacturers have to factor in the serviceability. If a camera system breaks during the warranty period, they are on the hook for the cost of the replacement, which may involve disassembling the dashboard of the car (i.e. lots of labor hours.), not to mention the reputation damage if it turns out that the camera system of car X tends to break within 3 years. Hence, they have a quite different requirement on reliability than a private person who orders a gadget from Ebay. And they need to ensure that spare parts will still be available for the next 10 years.
If you look at the front-page figure and Fig. 1 in the patent application, it will become apparent that the device only works for light rays entering at one particular angle. Essentially, it's like looking through louvers with an aspect ratio of 1:4 (fig. 1) or 1:6 (front page), which means that you won't be able to see through at all if your eyes are off by 25 cm horizontal at 1 m distance (i.e., passenger-side pillars) and whatever you see is substantially obscured for smaller angles. This is roughly how those 3M privacy screens for laptops work. That might be barely acceptable for the passenger-side pillars, but would be completely unworkable for the driver-side pillars.
Moreover, the surfaces 126 and 146 in Fig. 2 will need to be polarizing filters or opaque black surfaces so that you don't get to see spurious overlaid images. If you make them black, you will have replaced the obscuration of the pillar by two big black sheets that are only invisible if you look at them from one particular angle. If you make them polarizing absorbers, good luck in manufacturing those such that they don't reflect at grazing-incidence angles. (Those surfaces are mentioned in paragraph 37, without further reasoning about the benefits or tradeoffs, suggesting that the inventors don't know yet how to deal with these issues.)
By the way, the inventors have the polarizations the wrong way around in the figures. Although the correctly mention that p-polarization is transmitted and s-polarization is reflected, they have the arrows indicating the light polarization the wrong way around...
A bit of apples and oranges. Wifi is restricted to max 20 mW or so with an omnidirectional antenna, less if you use a directional antenna. Telecom uses higher power (250 mW on the phone from a quick Google) and uses high-gain directional antennas at the cell tower. You would't be allowed to use such high powers in the unlicensed wifi band.
It's difficult to do because Google provides this root detection service (Google Safety Net). That 'snet' program is self-updating and probably runs with elevated permissions. Whenever a developer releases an easy-to-use and somewhat popular root-hiding tool, Google will implement a tailored detection scheme for that particular tool and silently roll it out.
Printer dots are binary; screen pixels have 256 shades of gray or 256^3 colors. When printing full color, the effective resolution of a laser printer is way lower than 1200 ppi.
What did Apple do to you personally that makes you breach your NDA so blatantly?
Hum, I guess I should have read the article front to end rather than skim it...
Adiabatic compressed air energy moves the heat from compression into an insulated thermal mass chamber, and uses that to heat the expansion vessel. It recouperates that loss and has 70% total effective energy storage--higher is possible, up to 90%.
The way you describe it sounds not at all like an adiabatic (no heat exchange) process but rather like a thermodynamically irreversible process. Maybe you mean an isothermal process?
And from https://en.wikipedia.org/wiki/... :
"real compressors and turbines are not isentropic, but instead have an isentropic efficiency of around 85%, with the result that round-trip storage efficiency for adiabatic systems is also considerably less than perfect."
Wikipedia doesn't say so much about isothermal compression in practice, but it sounds to me like practically infeasible to reach your 70-90% roundtrip efficiency. You'd need an enormous thermal mass that doesn't change much in temperature despite 1/3 or so of the energy being stored as heat, you need enormous heat exchangers to transfer the heat with negligible temperature differences, and your compressors and decompressors need to have a very high efficiency.
IMHO: A USB device that depends on its power source to limit its input current, and can be damaged by a host that is willing to deliver more current that it requested, is defective by design.
Current limiting is to protect the supplier of the current. Bad current negotiation can damage the power supply, so of course the power supply should limit the current. A bad power supply may break, though. I think the main problem is that USB-C can use a range of voltages and a 5 V device plugged into a 20 V power supply will blow up the device unless the 20 V supply is signaled to throttle back to 5 V.
But maybe I misunderstand. Unfortunately, the reporting about this topic (Leung's findings) is very fuzzy about what happens exactly. But I don't see any scenario where connecting a good quality 5V, 1A or 2A charger to any USB-C device can lead to damage.
"your uncommon hardware may get burned out early (I've lost many wifi / gps / bluetooth / 3g-4g chipsets that way)."
This is the first time I hear smt like this. Do you have evidence?
The gps radio in my nexus 5 (with CM) tends to deteriorate over the months, but it turned out that tightening the screws of the rf shield restores the reception. Not a software issue.
Moreover, I think that Google's three support years count from the date the model is discontinued from the Play store, which may well be 2 years after the model is released.
A modern dac is supposed to digitally filter at 20 kHz (very hard cutoff), in exchange for lots of noise above 100 kHz. A soft roll-off analog filter takes care of the content above 100 kHz.
I suspect that ultrasound in this context really means 16 kHz or so, at volumes that are too low for the ear, but easily picked up by a microohone and some signal processing.
What kind of audio source did you use to find that you can hear 22 kHz? Unless you have an ultra low noise analog sine wave generator and amplifier, you are likely to hear noise, artifacts of the DA converter, and effects of clipping when you crank up the volume. The stated limit of 20 kHz for the human ear is the frequency where the pain threshold and the hearing threshold coincide for an average young person, so it is likely that you need >110 dB SPL to have any chance of hearing above 20 kHz. A device with 90 dB S/N ratio will likely blast you with lower frequencies, which you could easily mistake for "I can hear 20+ kHz".
I did once fry the tweeters of my (60 watt rated) loudspeakers while trying to test the limits of my hearing as I was sending all amplifier output power to the tweeters.
Although the general point is probably correct that min/max charge levels affect Li-ion lifespan, I would be careful with using batteryuniversity as a source of information. It's really a weblog of a small manufacturer of battery chargers that presents a couple of small-scale test results as if they represent the absolute truth for all brands and types of Li-ion batteties.
Until a music-streaming provider complains about being discriminated against, I don't think the review of the whitelist is the problem. The legal issue seems to be about the question whether discrimination by type of content (music, in this case), regardless of the company that provides the content, is allowed or not.
My feeling is that T-mobile is in a weak position here. Because if T-mobile is right, they would also be allowed to discriminate against other types of content; for example by making VoIP traffic more expensive. In fact, this is already happening: the cost of any data traffic is increased to subsidize the bandwidth for music data.
Clearly, IANAL, and the difference between "law" and "regulation" is fuzzy for me. But I think it happens every now and then that a local law/regulation is overturned because it conflicts with an EU regulation. Couldn't that be the case here?
References (Dutch, you'll have to pass it through Google Translate): https://www.t-mobile.nl/datavr... http://newsroom.t-mobile.nl/t-...
I'm trying to say that the construction that allows easy battery replacement comes with several points of failure that ai'd rather do without.
What is a typical cost of a car, per distance, in the US? In the Netherlands, about 0.19 EUR/km is the bare minimum in marginal costs for a small car with good fuel economy.
"I've yet to hear a reason why removable batteries are a bad thing for consumers with the sole exception of anorexia."
I used to have an HTC Desire S, with removable battery. A drop from 5 cm with the wrong angle would temporarily disconnect the battery, leading to a power-off. I had to put in a spacer to push the battery harder against the spring contact. (it took quite a while to figure out what the problem was)
I have a low-end LG L40 with removable battery. Every time I drop the phone, the back cover will come off and I end up with a separate battery, cover, and front. At least it immediately removed its own battery when I dropped it in the toilet. (phone is still working after a rinse in distilled water and thorough drying)
And I have a Nexus 5 with "non-removable" battery. The back cover can still come off with some effort; the battery can be replaced although it's a bit more work. And I don't have the above problems with self-disassembly and flaky contacts.
I'll take the "nonremovable" battery, thank you.
Norwegian 100% hydro power is subject to double book keeping. Other European countries try to meet their CO2 goals by buying green certificates from Norway, so that on paper lots of the Norwegian electrons become "coal" and lots of other European electrons become "green".
Where did you get "It only became a serious flaw recently"? It happened somewhere between 11 years ago and present, but no date is given for commit f33ea7f404e5 (I did try Google).