5 ms·

Did I understand correctly that basically air speed sensor was faulty and because of that the autopilot decided that we need more speed to avoid stall and put t

Did I understand correctly that basically air speed sensor was faulty and because of that the autopilot decided that we need more speed to avoid stall and put the nose down all the way to the ground/sea?

lvs 8y ago
Not only that, but it does so even when the autopilot isn't engaged and the pilots believe they have manual control.
Someone 8y ago
The essential bit is that the autopilot made that decision even if it was switched ‘off’, and Boeing never told anybody the autopilot couldn’t really be switched off. It seems they added a second autopilot that prevents pilots from doing truly stupid things without telling anybody. Problem? Failure modes of that autopilot can easily be lethal. I think they may have been right in saying pilots don’t need additional training for this new feature. If the plane flew itself into the ground, and there’s nothing the pilots could have done to prevent that, they don’t need additional training. Chances are they didn’t tell buyers of the plane because, for years, they have marketed their planes as “if the software fails, pilots can take control, unlike in Airbus planes” (counter-acting Airbus’ message that their planes were more modern)
- VBprogrammer 8y ago
  I can't read the article due to the pay wall but if what you said is accurate then I don't believe it's an accurate reflection of the real complexity of a modern airliner. See http://www.b737.org.uk/flightcontrols.htm http://www.b737.org.uk/flightcontrols.htm The pitch trim can be controlled by a number of different systems for different reasons, see the section on mach trim for an interesting one. I would be shocked if the inputs here weren't disabled by switching thd stab trim - autopilot switch to the off position. I've mentioned in previous comments, this sounds like a trim runaway incident. The mechanism may be new but the underlying fault and the symptoms would have been the same.
  - comex 8y ago
    You seem to be right: http://rgl.faa.gov/Regulatory_and_Guidance_Library/rgad.nsf/0/83ec7f95f3e5bfbd8625833e0070a070/%24FILE/2018-23-51_Emergency.pdf http://rgl.faa.gov/Regulatory_and_Guidance_Library/rgad.nsf/... > Runaway Stabilizer > Disengage autopilot and control airplane pitch attitude with control column and main electric trim as required. If relaxing the column causes the trim to move, set stabilizer trim switches to CUTOUT. If runaway continues, hold the stabilizer trim wheel against rotation and trim the airplane manually.
  - FabHK 8y ago
    Yes, and in particular it seems likely that the pilots could have stopped the automatic down trim by either invoking two trim cutout switches, and/or manually holding/turning back the trim wheel. But it appears they a) had many other things to deal with, and b) did not expect this particular failure mode, so did not recognise it. Terrible.
- Xixi 8y ago
  Airbus planes have four modes: Normal Law, Alternate Law, Abnormal Alternate Law and Direct Law. In Normal Law the computer will basically override anything that would put the plane out of its flight envelope. In Direct Law there are no protections provided (but there are still warnings). (Abnormal) Alternate Law are in between, with some protections active. There's even a mechanical backup in case of complete electric power loss... I don't know if it's available on all models though. So no: in Airbus planes the pilot can and do absolutely control everything. If needed the pilot can always go into Direct Law. When sensors fail (like the pitot tubes in AF447) Airbus planes step down to Alternate Law. “If the software fails, pilots can take control, unlike in Airbus planes” has never been true. There is however a difference in that Boeing pilots are supposed to always have the last word without needing to deactivate protections or anything. The plane would for instance apply physical resistance on the yoke to alert the pilot that something is wrong, but pulling/pushing hard enough is enough to override the protection. Well, that was my understanding until this accident (investigation pending, of course).
- Obi_Juan_Kenobi 8y ago
  This is incorrect. There are two systems here: autopilot and stall/overspeed protection. These are both automated systems, but they are controlled separately. When autopilot is off, you'll still have stall/overspeed protection that will pitch down/up in the corresponding condition. This protection can also be disabled by switching off some flight computer control. This is the case for both Airbus and Boeing, and the specific controls required depend on the particular plane model. The issue at hand is ensuring that training for these situations and knowledge of the flight control computers is adequate.
StreamBright 8y ago
It is hard to believe for me that the stall avoidance system is out of bounds. If somebody would ask me to implement such a system the first thing would be to gather the scenarios when such invasive action would be potentially fatal. Starting with the how far is the ground from the plane following with the condition are we taking off at the moment. If Boeing has a system that does not have bounds and always on than I am going to stop flying with airlines that have these planes in their fleets today.
voiture_rapide 8y ago
No. This is not the autopilot, it is a separate independent system that operates only when AP is off. The issues are a- that MCAS depends on having valid AOA data when there are only two AOA sensors and therefore no way to get a consensus “vote” in the event of a failure but far more importantly b- Boeing did not provide 737 MAX pilots with any information whatsoever about this MCAS system and what to do in the event of an AOA failure triggering invalid MCAS behavior.

lvs 8y ago

Not only that, but it does so even when the autopilot isn't engaged and the pilots believe they have manual control.

Someone 8y ago

The essential bit is that the autopilot made that decision even if it was switched ‘off’, and Boeing never told anybody the autopilot couldn’t really be switched off. It seems they added a second autopilot that prevents pilots from doing truly stupid things without telling anybody. Problem? Failure modes of that autopilot can easily be lethal. I think they may have been right in saying pilots don’t need additional training for this new feature. If the plane flew itself into the ground, and there’s nothing the pilots could have done to prevent that, they don’t need additional training. Chances are they didn’t tell buyers of the plane because, for years, they have marketed their planes as “if the software fails, pilots can take control, unlike in Airbus planes” (counter-acting Airbus’ message that their planes were more modern)

VBprogrammer 8y ago
I can't read the article due to the pay wall but if what you said is accurate then I don't believe it's an accurate reflection of the real complexity of a modern airliner. See http://www.b737.org.uk/flightcontrols.htm http://www.b737.org.uk/flightcontrols.htm The pitch trim can be controlled by a number of different systems for different reasons, see the section on mach trim for an interesting one. I would be shocked if the inputs here weren't disabled by switching thd stab trim - autopilot switch to the off position. I've mentioned in previous comments, this sounds like a trim runaway incident. The mechanism may be new but the underlying fault and the symptoms would have been the same.
- comex 8y ago
  You seem to be right: http://rgl.faa.gov/Regulatory_and_Guidance_Library/rgad.nsf/0/83ec7f95f3e5bfbd8625833e0070a070/%24FILE/2018-23-51_Emergency.pdf http://rgl.faa.gov/Regulatory_and_Guidance_Library/rgad.nsf/... > Runaway Stabilizer > Disengage autopilot and control airplane pitch attitude with control column and main electric trim as required. If relaxing the column causes the trim to move, set stabilizer trim switches to CUTOUT. If runaway continues, hold the stabilizer trim wheel against rotation and trim the airplane manually.
- FabHK 8y ago
  Yes, and in particular it seems likely that the pilots could have stopped the automatic down trim by either invoking two trim cutout switches, and/or manually holding/turning back the trim wheel. But it appears they a) had many other things to deal with, and b) did not expect this particular failure mode, so did not recognise it. Terrible.
Xixi 8y ago
Airbus planes have four modes: Normal Law, Alternate Law, Abnormal Alternate Law and Direct Law. In Normal Law the computer will basically override anything that would put the plane out of its flight envelope. In Direct Law there are no protections provided (but there are still warnings). (Abnormal) Alternate Law are in between, with some protections active. There's even a mechanical backup in case of complete electric power loss... I don't know if it's available on all models though. So no: in Airbus planes the pilot can and do absolutely control everything. If needed the pilot can always go into Direct Law. When sensors fail (like the pitot tubes in AF447) Airbus planes step down to Alternate Law. “If the software fails, pilots can take control, unlike in Airbus planes” has never been true. There is however a difference in that Boeing pilots are supposed to always have the last word without needing to deactivate protections or anything. The plane would for instance apply physical resistance on the yoke to alert the pilot that something is wrong, but pulling/pushing hard enough is enough to override the protection. Well, that was my understanding until this accident (investigation pending, of course).
Obi_Juan_Kenobi 8y ago
This is incorrect. There are two systems here: autopilot and stall/overspeed protection. These are both automated systems, but they are controlled separately. When autopilot is off, you'll still have stall/overspeed protection that will pitch down/up in the corresponding condition. This protection can also be disabled by switching off some flight computer control. This is the case for both Airbus and Boeing, and the specific controls required depend on the particular plane model. The issue at hand is ensuring that training for these situations and knowledge of the flight control computers is adequate.

StreamBright 8y ago

It is hard to believe for me that the stall avoidance system is out of bounds. If somebody would ask me to implement such a system the first thing would be to gather the scenarios when such invasive action would be potentially fatal. Starting with the how far is the ground from the plane following with the condition are we taking off at the moment. If Boeing has a system that does not have bounds and always on than I am going to stop flying with airlines that have these planes in their fleets today.

voiture_rapide 8y ago

No. This is not the autopilot, it is a separate independent system that operates only when AP is off. The issues are a- that MCAS depends on having valid AOA data when there are only two AOA sensors and therefore no way to get a consensus “vote” in the event of a failure but far more importantly b- Boeing did not provide 737 MAX pilots with any information whatsoever about this MCAS system and what to do in the event of an AOA failure triggering invalid MCAS behavior.