6 ms·

Here's what I don't understand, and admittedly I have very little knowledge of the industry. There seems to be some sort of approval process for electronics to

by arn 15y ago

Here's what I don't understand, and admittedly I have very little knowledge of the industry. There seems to be some sort of approval process for electronics to be vetted during takeoff and landing. Several in-chair entertainment systems seem to have this rating/designation and are allowed during takeoff/landing. So, is there no dollar amount that could be paid to properly test a particular product? Let's say the Kindle or iPod. It may be millions of dollars, but wouldn't there be a huge word-of-mouth and marketing upside if you could get approved for such a thing?

lucasjung 15y ago
I currently work in developmental flight test. Any time we add a new piece of electronics to an aircraft (or replace an existing piece with a new version), be it a radio, a computer, or test instrumentation, we have to conduct an Electromagnetic Compatibility (EMC) Safety of Flight Test (SOFT) to make sure that the new equipment does not interfere in any way with avionics that are critical to safety of flight. Even if it's something that's not being installed per se (e.g. a laptop for collecting test data), it has to be tested. Sometimes there are exceptions, such as very minor upgrades, especially minor software updates. Generally everything is designed so as to not interfere, but every once in a while something causes unexpected interference. Sometimes it's because the designers screwed up, sometimes it's because a bespoke item wasn't manufactured to the correct spec, and sometimes it's because of something nobody anticipated. The problem with consumer electronics is that they haven't gone through this process. None of them should cause interference, but we can't be sure because they were never subjected to comprehensive EMC SOFT. Nor would it really make sense to spend that kind of time and money when we can just mitigate the risk of interference by having everyone turn everything off during critical phases of flight.
- RandallBrown 15y ago
  but wouldn't it make sense for Apple to pay to do the testing? Being able to say "The iPhone is the only phone you can use during your whole flight" would be worth quite a bit of money to them.
  - arn 15y ago
    right, I was suggesting that the vendor (Apple, Amazon) of said products would undertake the testing cost.
  - lucasjung 15y ago
    They would have to test every model of iphone against not only every model of commercial airliner in operation in the U.S. (including puddle-jumpers), but against every major variation in avionics suite among each model. This testing is not cheap. Even for Apple, it would be prohibitively expensive.
    - arn 15y ago
      one airline should go halvsies with Apple. Delta... the only airline you can use your iPhone 4S from takeoff to landing...
    - ghshephard 15y ago
      There's probably only 80 different commercial airplane models. This testing probably costs around $1000/hr - figure 5 days testing @ 8 hours a day = 3200 hours = 3.2 million dollars. Seems to be a reasonable price to pay.
      - lucasjung 15y ago
        As I said, it's not just the different models, it's the different avionics suites within each model (e.g. two aircraft of the same model might have different radios or different navigation equipment, etc.). For each plane, you would need a few hours of testing, which probably costs more than $1000/hour: you have to pay for the plane, the crew (who have to be qualified to perform this type of testing), the ground support equipment, maintenance support, and any fuel burned during testing (e.g. running the APU to provide power during testing will consume a small but not insignificant quantity of fuel). EDIT: Almost forgot: After the testing is complete, you have to pay to have a test report written up and published, and then you have to pay for the FAA certification process. Probably not cheap.
- politician 15y ago
  Could you explain to what degree incidents of unexpected interference result in dangerous failure modes? Is that sort of follow-up analysis done?
  - lucasjung 15y ago
    I would hesitate to use the term "dangerous failure modes." The danger generally stems from interference with key avionics during critical phases of flight. For example, if your ILS starts acting wacky right in the middle of an approach in actual IMC, you've got a problem. Likewise, if your radios become unusable due to noise anywhere in the terminal area, you've got a problem. Generally, this sort of thing is caught during the EMC SOFT, which is conducted on deck so that there is no immediate danger. Once the interference has been discovered it's basically a debugging process to figure out what the hell is going in (where is the interference originating, and how is it causing the observed effect). The plane doesn't fly again until it can pass an EMC SOFT. Usually dangerous interference comes through hard-wired connections: a device that feeds back frequencies into the power system, which then disrupts other devices; or a device that puts noise on an avionics bus, disrupting communications on that bus. However, sometimes a device, even a device not designed to transmit, can create significant RF interference. For example, a device that has an internal power supply or transformer of some sort. If that power supply is not working right, it might generate an oscillating signal. Maybe even the device itself is tolerant of a certain amount of oscillation in its power and so is not affected by this. Then, there just happens to be a wire inside the device, connected to the power supply that's just the right length to act as an antenna at that frequency, and now you've got an RF transmitter on your aircraft that you didn't expect or want. There are bazzillions of wires in most medium-to-large modern aircraft, and some of them are bound to be the right length to act as an antenna and receive that signal, passing interference into who-knows-what. That example may sound highly unlikely, but it does happen.
    - politician 15y ago
      "Once the interference has been discovered ... the plane doesn't fly again until it can pass an EMC SOFT." Yet, turning off consumer electronics is enforced by the honor system? Is there something else going on that makes consumer electronics unlikely sources of interference that would justify this lazy approach to safety? Why aren't all of these devices confiscated and locked in Faraday cages? It just strikes me as odd that failure to pass an EMC SOFT grounds the plane until the issue is debugged, but then they just allow an random passenger to bring any random radio aboard just so long as they promise/remember to turn it off during critical phases of the flight. That example may sound highly unlikely, but it does happen. So highly unlikely that an honor system approach is justified?
- tomkarlo 15y ago
  From a safety standpoint, isn't the right thing to do to test flight equipment against interference from consumer devices? Simply assuming that everyone on the plane will turn off their cell phone seems like a cop-out. Having flight safety depend on the (unverified) compliance of consumers is ridiculous. There are only two cases here: 1. Consumer electronics represent a real threat to air safety. They should be banned from planes or disabled (i.e. the battery removed and turned over until the flight ends.) There should be real penalties for non-compliance. 2. Consumer electronics are not really a threat to air safety. We should stop wasting everyone's time and adding unnecessary stress to the travel process.
  - gnaffle 15y ago
    3. Consumer electronics are not a big threat to air safety, but they can (and have, see Wikipedia article / NASA study) interfered with airplane navigation instruments. Since it's much more likely that 1 out of 100 or 300 active devices will cause interference during takeoffs and landings than 1 out of 4-5 that are turned on by accident or ignorance, it's better to err on the side of safety and tell everyone to turn them off.
    - tomkarlo 15y ago
      I'm not sure that's valid case... the number of devices per person, and therefore per plane, has been rising rapidly over time (I carried at least 5 "electronic devices" (kindle, laptop, 2 phones, camera) onto my flight home last night.) A plane with 200+ people could soon have 400-500 devices versus 50 a decade ago - at one end or the other of that range they're either not a threat (and we needn't turn them off) or they're a real threat and need more than voluntary shutdown. (BTW, I watched the folks in my row on the plane, and not one person besides me actually shut down their phone or Kindle. They all just either put them in airplane mode or standby by tapping the off key. So I'm guessing that the reality is that most devices are not actually shut off by non-technical users. Even typing this I realize my Kindle wasn't actually shut off, its 3G was still active.) I find it hard to believe we exist exactly in the statistical sweet spot where voluntary compliance results in a materially safer outcome than broad non-compliance.
magic_haze 15y ago
DO-254 [http://en.wikipedia.org/wiki/DO-254 http://en.wikipedia.org/wiki/DO-254] seems like what you're looking for. I'm not familiar with this particular document, but the equivalent certification process for software (DO-178B) is not something that can be affixed lightly on any random product - the overhead it imposes on the development is simply way too high to be viable on commercial software.

by arn 15y ago

lucasjung 15y ago

I currently work in developmental flight test. Any time we add a new piece of electronics to an aircraft (or replace an existing piece with a new version), be it a radio, a computer, or test instrumentation, we have to conduct an Electromagnetic Compatibility (EMC) Safety of Flight Test (SOFT) to make sure that the new equipment does not interfere in any way with avionics that are critical to safety of flight. Even if it's something that's not being installed per se (e.g. a laptop for collecting test data), it has to be tested. Sometimes there are exceptions, such as very minor upgrades, especially minor software updates. Generally everything is designed so as to not interfere, but every once in a while something causes unexpected interference. Sometimes it's because the designers screwed up, sometimes it's because a bespoke item wasn't manufactured to the correct spec, and sometimes it's because of something nobody anticipated. The problem with consumer electronics is that they haven't gone through this process. None of them should cause interference, but we can't be sure because they were never subjected to comprehensive EMC SOFT. Nor would it really make sense to spend that kind of time and money when we can just mitigate the risk of interference by having everyone turn everything off during critical phases of flight.

RandallBrown 15y ago
but wouldn't it make sense for Apple to pay to do the testing? Being able to say "The iPhone is the only phone you can use during your whole flight" would be worth quite a bit of money to them.
- arn 15y ago
  right, I was suggesting that the vendor (Apple, Amazon) of said products would undertake the testing cost.
- lucasjung 15y ago
  They would have to test every model of iphone against not only every model of commercial airliner in operation in the U.S. (including puddle-jumpers), but against every major variation in avionics suite among each model. This testing is not cheap. Even for Apple, it would be prohibitively expensive.
  - arn 15y ago
    one airline should go halvsies with Apple. Delta... the only airline you can use your iPhone 4S from takeoff to landing...
  - ghshephard 15y ago
    There's probably only 80 different commercial airplane models. This testing probably costs around $1000/hr - figure 5 days testing @ 8 hours a day = 3200 hours = 3.2 million dollars. Seems to be a reasonable price to pay.
    - lucasjung 15y ago
      As I said, it's not just the different models, it's the different avionics suites within each model (e.g. two aircraft of the same model might have different radios or different navigation equipment, etc.). For each plane, you would need a few hours of testing, which probably costs more than $1000/hour: you have to pay for the plane, the crew (who have to be qualified to perform this type of testing), the ground support equipment, maintenance support, and any fuel burned during testing (e.g. running the APU to provide power during testing will consume a small but not insignificant quantity of fuel). EDIT: Almost forgot: After the testing is complete, you have to pay to have a test report written up and published, and then you have to pay for the FAA certification process. Probably not cheap.
politician 15y ago
Could you explain to what degree incidents of unexpected interference result in dangerous failure modes? Is that sort of follow-up analysis done?
- lucasjung 15y ago
  I would hesitate to use the term "dangerous failure modes." The danger generally stems from interference with key avionics during critical phases of flight. For example, if your ILS starts acting wacky right in the middle of an approach in actual IMC, you've got a problem. Likewise, if your radios become unusable due to noise anywhere in the terminal area, you've got a problem. Generally, this sort of thing is caught during the EMC SOFT, which is conducted on deck so that there is no immediate danger. Once the interference has been discovered it's basically a debugging process to figure out what the hell is going in (where is the interference originating, and how is it causing the observed effect). The plane doesn't fly again until it can pass an EMC SOFT. Usually dangerous interference comes through hard-wired connections: a device that feeds back frequencies into the power system, which then disrupts other devices; or a device that puts noise on an avionics bus, disrupting communications on that bus. However, sometimes a device, even a device not designed to transmit, can create significant RF interference. For example, a device that has an internal power supply or transformer of some sort. If that power supply is not working right, it might generate an oscillating signal. Maybe even the device itself is tolerant of a certain amount of oscillation in its power and so is not affected by this. Then, there just happens to be a wire inside the device, connected to the power supply that's just the right length to act as an antenna at that frequency, and now you've got an RF transmitter on your aircraft that you didn't expect or want. There are bazzillions of wires in most medium-to-large modern aircraft, and some of them are bound to be the right length to act as an antenna and receive that signal, passing interference into who-knows-what. That example may sound highly unlikely, but it does happen.
  - politician 15y ago
    "Once the interference has been discovered ... the plane doesn't fly again until it can pass an EMC SOFT." Yet, turning off consumer electronics is enforced by the honor system? Is there something else going on that makes consumer electronics unlikely sources of interference that would justify this lazy approach to safety? Why aren't all of these devices confiscated and locked in Faraday cages? It just strikes me as odd that failure to pass an EMC SOFT grounds the plane until the issue is debugged, but then they just allow an random passenger to bring any random radio aboard just so long as they promise/remember to turn it off during critical phases of the flight. That example may sound highly unlikely, but it does happen. So highly unlikely that an honor system approach is justified?
tomkarlo 15y ago
From a safety standpoint, isn't the right thing to do to test flight equipment against interference from consumer devices? Simply assuming that everyone on the plane will turn off their cell phone seems like a cop-out. Having flight safety depend on the (unverified) compliance of consumers is ridiculous. There are only two cases here: 1. Consumer electronics represent a real threat to air safety. They should be banned from planes or disabled (i.e. the battery removed and turned over until the flight ends.) There should be real penalties for non-compliance. 2. Consumer electronics are not really a threat to air safety. We should stop wasting everyone's time and adding unnecessary stress to the travel process.
- gnaffle 15y ago
  3. Consumer electronics are not a big threat to air safety, but they can (and have, see Wikipedia article / NASA study) interfered with airplane navigation instruments. Since it's much more likely that 1 out of 100 or 300 active devices will cause interference during takeoffs and landings than 1 out of 4-5 that are turned on by accident or ignorance, it's better to err on the side of safety and tell everyone to turn them off.
  - tomkarlo 15y ago
    I'm not sure that's valid case... the number of devices per person, and therefore per plane, has been rising rapidly over time (I carried at least 5 "electronic devices" (kindle, laptop, 2 phones, camera) onto my flight home last night.) A plane with 200+ people could soon have 400-500 devices versus 50 a decade ago - at one end or the other of that range they're either not a threat (and we needn't turn them off) or they're a real threat and need more than voluntary shutdown. (BTW, I watched the folks in my row on the plane, and not one person besides me actually shut down their phone or Kindle. They all just either put them in airplane mode or standby by tapping the off key. So I'm guessing that the reality is that most devices are not actually shut off by non-technical users. Even typing this I realize my Kindle wasn't actually shut off, its 3G was still active.) I find it hard to believe we exist exactly in the statistical sweet spot where voluntary compliance results in a materially safer outcome than broad non-compliance.

magic_haze 15y ago

DO-254 [http://en.wikipedia.org/wiki/DO-254 http://en.wikipedia.org/wiki/DO-254] seems like what you're looking for. I'm not familiar with this particular document, but the equivalent certification process for software (DO-178B) is not something that can be affixed lightly on any random product - the overhead it imposes on the development is simply way too high to be viable on commercial software.