Most active commenters

    ←back to thread

    279 points Michelangelo11 | 23 comments | | HN request time: 0.686s | source | bottom
    1. freefaler ◴[] No.45038516[source]
    So as a pilot you can't override the software to stop it from "thinking that the plane is on the ground" mode?

    Something similar happened recently with A320 when it didn't want to land on an airfield during emergency unless it was flown in a special mode. But F-35 doesn't have that?

    replies(6): >>45038623 #>>45039289 #>>45039382 #>>45040610 #>>45041772 #>>45045256 #
    2. ◴[] No.45038623[source]
    3. netsharc ◴[] No.45039289[source]
    > unless it was flown in a special mode.

    What fresh hell is that... reboot, jam F8 just as the "Airbus" logo shows up, and then select "Boot in safe mode"?

    replies(4): >>45039408 #>>45039526 #>>45040492 #>>45041141 #
    4. seethishat ◴[] No.45039382[source]
    "On the ground" = WoW sensors. WoW sensors have been around a long time (see link). And, humans probably should not have any say about that. If humans could override WoW, then the landing gear could be deployed or retracted when it should not and cause a lot of damage due to human error.

    https://en.wikipedia.org/wiki/ACARS

    replies(3): >>45039495 #>>45039726 #>>45040325 #
    5. xattt ◴[] No.45039408[source]
    Fly-by-wire aircraft have changeable “flight laws” that correspond to different levels of computer intervention to mitigate situations incompatible with controlled flight.

    Think of it as various stability control modes in a modern car. Likely the aircraft needed to be put in the least restrictive flight law mode as a workaround.

    replies(1): >>45040078 #
    6. jcalvinowens ◴[] No.45039495[source]
    I know the 737 allows the pilot to override that (force the wheels to raise even if the airplane thinks it's on the ground). I think most airliners do. I can't find a good succinct reference though.

    EDIT: Remembered Airbus exists

    7. crote ◴[] No.45039526[source]
    Airbus is fully fly-by-wire. Without some kind of computer intervention, nothing would be stopping an accidental bang against the flight stick from causing a maneuver violent enough to rip the wings off.

    An Airbus can operate in three modes. With Normal Law, the airplane will refuse to do anything which will stop it from flying. This means the pilot cannot stall the airplane, for example: the computer will automatically correct for it.

    With Alternate Law the pilot loses most protections, but the plane will still try to protect against self-destruction. The plane no longer protects against being stalled, but it won't let you rip the wings off.

    With Direct Law all bets are off. Controls now map one-to-one to control surfaces, the plane will make no attempt to correct you. All kinds of automatic trimming are lost, you are now essentially flying a Cessna again. The upside is that it no longer relies on potentially broken sensors either: raising the gear while on the ground is usually a really stupid idea - until the "is the plane on the ground" sensors break.

    So no, a "Boot in safe mode" isn't as strange as it might sound at first glance. It significantly improves safety during day-to-day operations, while still providing a fallback mechanism during emergencies.

    replies(2): >>45040099 #>>45042342 #
    8. agos ◴[] No.45039726[source]
    this makes sense but why is the decision based only on the state of the landing gear? Is it dumb to expect altitude and speed to be considered?
    replies(1): >>45042434 #
    9. lazide ◴[] No.45040078{3}[source]
    ‘Incompatible with controlled flight’ is my new ‘rapid unscheduled disassembly’.

    Notably, most drones have similar levels of control. Everything has to go through the IMU of course (nobody is manually going to be managing 4 separate motor controllers at once), but depending on the modes, the type of control is wildly different.

    ‘Consumer’/‘idiot’ mode - you tell it which direction to go, and how high/low you want it, and it’ll do that safely. Usually with some sort of object detection/avoidance, auto GPS input, so you won’t accidentally wander into something or hit something. Goal is stable, level flight.

    ‘Sport’ mode - go fast, usually disables all but the most simple collision avoidance. Sometimes even that. Still provides stable, level flight, but you can easily crash it. Usually goes 2-3x faster than ‘idiot’ mode.

    ‘Attitude’ or ‘acrobatic’ mode - you’re directly commanding the target 3D pitch/yaw, and aggregate power output. No provision is given to automatically maintaining level flight (won’t auto level), generally no regard is given to airframe integrity, collision avoidance, or engine life, and boy is it fun.

    It’s really common to crash in this mode, because people are also doing flips, acrobatic maneuvers, running courses, etc.

    Drones in even ‘sport’ mode can’t do flips because it’s fundamentally at odds with auto maintaining level flight, etc.

    10. megaloblasto ◴[] No.45040099{3}[source]
    How does a pilot switch between the three modes? Just switches on the dash?
    replies(4): >>45040393 #>>45040549 #>>45040732 #>>45043044 #
    11. tonyarkles ◴[] No.45040325[source]
    Yeah, when I saw that it was related to WoW sensors I was reminded of a chat that I had with an experienced aviation engineer a few years ago when we were considering adding them to the UAS I'm working on. "Every system I've ever worked on that relied on WoW sensors has had some kind of unexpected adverse event due to the sensors either triggering when they weren't supposed to, or not triggering when they should have. If you can figure out a different way to do what you're trying to do, please do."
    12. selectodude ◴[] No.45040393{4}[source]
    It automatically downgrades modes as sensors break. You can force into alternate law by pulling some fuses if you really want to.
    13. voidUpdate ◴[] No.45040492[source]
    The first part is trying to work out which key you need to mash, its always the one you check last
    14. ahartmetz ◴[] No.45040549{4}[source]
    As I understand it, there are no switches because pilots aren't supposed to switch modes, but if necessary, pulling certain circuit breakers will disable subsystems whose failure triggers alternate law. And AFAIU it is documented which breakers are "least unsafe" to use that way.
    15. foxyv ◴[] No.45040610[source]
    That's kind of funny. Most fighter jets have an emergency override for the gear so you can use gravity or maneuvers to drop them to a down and locked position even during hydraulic and electrical failures. I can understand not being able to bring the gear up, but getting the gear down should be very easy.

    I wonder if the F-35 has a similar gear override.

    16. foxyv ◴[] No.45040732{4}[source]
    This is a very "Don't do this" kind of situation but you can force the computer to switch to a different law by pulling circuit breakers for certain instruments. Typically though it will happen automatically when the computer detects sensor failures.

    Changing the law of the aircraft is something you REALLY do not want to do. It's a "The Airplane broke real bad, do that pilot thing!" situation. Especially on a fighter jet with relaxed stability.

    https://www.airliners.net/forum/viewtopic.php?t=748133

    Some fighters have an "EFCS" switch which will switch laws. This can be used during "I'm going to fly into the ground" or "I'm losing this dogfight" situations. Typically this means you are going to be scrapping the airframe soon one way or another.

    replies(1): >>45043418 #
    17. munchlax ◴[] No.45041141[source]
    You've reached the Airbus emergency hotline.

    Have you tried forcing an unexpected reboot?

    18. foldr ◴[] No.45041772[source]
    > Something similar happened recently with A320

    Are you sure? I can’t seem to find any references to any such incident.

    19. libraryofbabel ◴[] No.45042342{3}[source]
    Great explanation. And of course the distinction between Normal Law and Alternate Law or Direct Law was at the heart of one of the most famous crashes of recent years, Air France 447: https://admiralcloudberg.medium.com/the-long-way-down-the-cr...
    20. seplox ◴[] No.45042434{3}[source]
    No, that's not dumb at all. Inputs are filthy and sensors fail. If you're not comparing all available sensor data to confirm your understanding of reality, then a single sensor failure could... oh I dunno... cause your 737 MAX to divebomb.

    The F-35 could compare weight on wheels to airspeed as a simple sanity check.

    21. ssalazar ◴[] No.45043044{4}[source]
    This blog post (and blog in general) has some detailed descriptions of different control laws, how they are activated, and how they contributed to the conditions of this particular crash: https://admiralcloudberg.medium.com/trial-by-fire-the-crash-...

    The above post is focused on a Sukhoi jet, with some comparison to Airbus' design, but they also cover Airbus in another post: https://admiralcloudberg.medium.com/the-long-way-down-the-cr...

    22. megaloblasto ◴[] No.45043418{5}[source]
    Very interesting. Reminds me of 2001 a space odyssey when the pilot unplugs parts of Hals brain in order to regain control over the ship.
    23. nurple ◴[] No.45045256[source]
    I've crashed an RC helicopter because of a similar software issue. Rotorflight is an OSS flight controller, and it has an internal mode for tracking if the vehicle is on the ground or not that isn't always quite accurate at the margins. If you're touching the ground and it's not in ground-handling mode, the I-term in the PID loop winds up really quickly (because the input isn't producing the expected rotation rate) and flips your model on its side.

    Betaflight (flight controller for drones, which rotorflight is based on) has a similar function called "air mode" which is common to either disable or set to a switch for aerobatic drones so that they'll still have full rotation rates at zero throttle.