Lion Air 737-Max missing, presumed down in the sea near CGK (Jakarta)

[Gabriel]

EVERY? To how many of them have you spoken?

Isn't runaway trim part of the very basic standard curriculum of what is done in the sim as part of any type rating?

[Gabriel]

Ok, I see. You are saying the trim would get stuck at full ANU if it got there.

Not exactly, I am saying that there is a scenario where the autotrim may stealthly trim all the way up (or close to that, doesn't matter) as response of pilot's sidestick or autoflight inputs and then something else happens (maybe it was the plane coming in direct law, I am not sure) that makes the autotrim stealthly disconnect (except for an EICAS message) so pushing nose down would not make the stabilizer move from the (almost) full nose up.

Again, I don't remember the scenario exactly, but I do remember we discussed it a length in the AF thread before we had the CVR and FDR info, but it was already quite clear from the ACARS and the forensic analysis if the wreckage that the plane had a UAS event and stalled off the sky.

Some were speculating before the wreckage was found that the crew might have not remembered to move the throttle, causing the stall. But then it became apparent the pilots were using the throttles.

.

Yes, when the plane stalled at the top of the impossible climb. Only then they touched the throttles and added full thrust (by the way, all three airspeed indicators were working ok and giving correct and consistent readings since a few seconds by then). They did a 1.5, 7000+ fpm, 2500ft climb from the maximum altitude, without adding thrust. Duh! (but that's for another thread)

[ATLcrew]

Every pilot who is properly trained on the 737NG/MAX has practiced the runaway pitch procedure very well.

EVERY? To how many of them have you spoken?

[Evan]

How many incidents/accidents do we know of involving faulty AoA data? I can only recall the B-2 Spirit at Guam (though I'm not sure whether AoA vanes are even involved, or if it was derived AoA data), and of course, XL 888T. Two of the three AoA vanes had frozen in a neutral position, while the correct vane was deemed invalid by the system and was rejected. The plane didn't make any uncommanded inputs, though they still crashed. But hypothetically, what if the faulty vanes were frozen in a much higher AoA? I figure that's a lot less likely to happen, but still makes me wonder how the plane would have behaved.
But only two AoA vanes on the Max with increased auto trim authority - they now say this only applies to the MAX and the NG would not behave like that in manual flight - does not seem redundant enough to me.

Does not seem redundant at all.

[Black Ram]

• The Airbus system does not act on any invalid air data. When the airspeed is below 60kts (and why would that ever happen?!) the AoA sensors cannot work properly. The Airbus philosophy is therefore to not use them for any critical reason. That's intelligent.
• The Boeing philosophy seems to be quite different. Boeing, apparently, allows a single, un-verified, potentially invalid source of AoA data to command a flight control surface. That's bonkers.

How many incidents/accidents do we know of involving faulty AoA data? I can only recall the B-2 Spirit at Guam (though I'm not sure whether AoA vanes are even involved, or if it was derived AoA data), and of course, XL 888T. Two of the three AoA vanes had frozen in a neutral position, while the correct vane was deemed invalid by the system and was rejected. The plane didn't make any uncommanded inputs, though they still crashed. But hypothetically, what if the faulty vanes were frozen in a much higher AoA? I figure that's a lot less likely to happen, but still makes me wonder how the plane would have behaved.
But only two AoA vanes on the Max with increased auto trim authority - they now say this only applies to the MAX and the NG would not behave like that in manual flight - does not seem redundant enough to me.

[Black Ram]

Right, but you also quoted me saying "In certain conditions it might have reached the full-nose up position and deactivate there and the only clue for the pilots would be a "manual trim" message in the EICAS.....

Ok, I see. You are saying the trim would get stuck at full ANU if it got there. That's also weird to me, and I don't see the logic in having that (not that I am the ultimate authority here). But since the trim on AF447 didn't reach full ANU, wouldn't it seem the trim is not designed to go there and get stuck?

And I do remember the "dead" throttles after autothrottle disengage. Some were speculating before the wreckage was found that the crew might have not remembered to move the throttle, causing the stall. But then it became apparent the pilots were using the throttles.

Anyway, looks like the Lion Air crew may have been overwhelmed by multiple things happening. Many have said the flight data shows they were controlling it for some time, then suddenly lost it. Hopefully the investigation will get to the bottom of it.

Some worrying allegations regarding maintenance and air-worthiness at Lion:

https://www.nytimes.com/2018/11/08/w...t-moments.html

[Evan]

Right, but you also quoted me saying "In certain conditions it might have reached the full-nose up position and deactivate there and the only clue for the pilots would be a "manual trim" message in the EICAS."

I don't remember well what these conditions were. Probably Evan can help here. I think it didn't reach the full up limit but was very close to it, and that if it had reached the limit the autotrim (that as you said went almost full up in response to the pilot's nose-up commands on the sidestick) would have self-disconnected (stealthy except by the EICAS message) and then it would not have gone down again in response to eventual nose-down commands. Again this is not what happened in AF (the trim never reached the self-disconnect condition) and, actually, when the other pilot briefly requested control and pushed down, the trim started to move down and the AoA started to diminish. In another example of a confusing effect, when the AoA diminished the stall warning (that had gone silent) started to sound again, which I hypothesize caused confusion (I am pulling up and the stall warning is silent, when I push down the stall warning activates? Better I don't push down!!!). The reason why that happened is that the software logic has a minimum airspeed blow which the AoA is considered not reliable (and hence the stall warning will not consider a high AoA reading as a sign of a stall). When the AoA was too high, the indicated speed was too low due to the high angle of attack of the pitot tubes (that were not frozen anymore). When the AoA started to go down (but was still too high), the real airspeed increased and the indicated airspeed increased even further due to the pitot tubes being not so much misaligned to the relative wind, so with the new airspeed the lower but still to high AoA reading was considered valid and since it was still beyond the stall AoA, there it went the stall warning again. Or how about the fact that, after the AP and AT disengaged, the UAS memory items called for climb thrust and 5 deg ANU, and the pilot were required to move the thrust levers OUT of the CLIMB thrust detent (where they were) to be able to establish climb thrust? (that is because of the thrust lock that comes with the autothrust disengagement and the fact that the thrust was below climb -even with the throttles in the climb detent- when the thrust lock kicked in.

Aircraft do have "stealth" factors (especially for pilots that are not really experts in the system). That said, click click, clack clack, grab a handful of yoke with one hand and a handful of yoke with the other, and make the plane do what you want it to do (and you better want it to do something that is within the laws of Physics) (and remember that the trim, whose control is located in the aforementioned yoke is an integral part of the pitch control system. In the scenario we are imagining for the Lyon crash, even without using the trim cutoff (as pilots were required to do by procedure and experts in the system would have done) the pilots could have still flown away with manual trim inputs in the trim switch located in the yoke.

Why are we back to the scarebus narrative? The BEA AF447 investigation concluded in its final report:

Throughout the flight, the movements of the elevator and the THS were consistent with the pilot’s inputs.

The PF made sustained upward pitch commands, included one lasting 30 seconds. The THS went to 13 ANU and remained there for the duration of the flight. Downward pitch commands were too brief and too transient to have any effect.

But the AoA invalidity issue is at the core of what we are discussing here.

• The Airbus system does not act on any invalid air data. When the airspeed is below 60kts (and why would that ever happen?!) the AoA sensors cannot work properly. The Airbus philosophy is therefore to not use them for any critical reason. That's intelligent.
• The Boeing philosophy seems to be quite different. Boeing, apparently, allows a single, un-verified, potentially invalid source of AoA data to command a flight control surface. That's bonkers.

Regarding the transient stall warnings on AF-447, the stall warning is an approach-to-stall warning. There is no scenario in which an A330 is going to be in that regime below 60kts. It is not intended as a stall-recovery aid. Stall recovery procedure is intended as a stall recovery aid.

However, I completely agree with you on stressing the role and importance of pitch trim. Any pilot who does not instinctively go to pitch trim when elevator authority is diminished shouldn't be allowed into a commercial cockpit. Yet this happens, again and again...

[3WE]

The trim blah blah blah...

Great post.

Since your bring it up, there's this fundamental that trim ONLY exists to remove control forces, NOT to control the aircraft. (Yes an absolute statement).

AND a statement that is wrong. Trim is required to control airliners through significant speed and flap changes. Trim is required to control airliners through autopilot systems too...I THINK.

So, you aeroengineers deviated from fundamentals and today we are wondering if this deviation drove a plane into the water.

Just food for thought.

[3WE]

...The response from some long standing contributors was to sneer...

Pesagus- please go re-read my response. Also note- the very first comment where I say "ass hat parlour talk"...that is referring to MY comments not yours.

If you read this forum- you know some of us get frustrated when pilots SEEM to forget the basics and crash in seemingly unbelievable ways...But after bitching about it for years, I more and more understand that I'm not sure how to fix it. Most everyone is trying to do the right thing...there's always going to be bad apples and there's always going to be brain farts. And the insiders also like to cite statistics that they do a REALLY great job ALMOST always.

[Gabriel]

The trim is one of the things that are normally overlooked but it is one of the most important things to learn in a new airplane, since it is so different from airplane to airplane and its effect can be overwhelming, literally, since it can exceed the pilot's ability to control pitch with the elevator alone, again, deepening on the design of the trim.

The Piper Cub has a system that is similar to that of the 737 but different from the Piper Tomahawk, which in turn is different that that of a Cessna 152 which is similar to the DC-9, but different than the 737.

THS, spring-loaded, trim tabs.

[Gabriel]

To my understanding, the trim went fully nose up because the PIC put it there with his control inputs. Trying to climb past 38000 feet, burning all their speed, and continuing to keep the nose up....pretty sure it couldn't have happened unless the trim went up, so the system was just doing what the PIC was asking for. What's really ironic - Boeing's "runaway trim" stall prevention could have been more helpful there.

Right, but you also quoted me saying "In certain conditions it might have reached the full-nose up position and deactivate there and the only clue for the pilots would be a "manual trim" message in the EICAS."

I don't remember well what these conditions were. Probably Evan can help here. I think it didn't reach the full up limit but was very close to it, and that if it had reached the limit the autotrim (that as you said went almost full up in response to the pilot's nose-up commands on the sidestick) would have self-disconnected (stealthy except by the EICAS message) and then it would not have gone down again in response to eventual nose-down commands. Again this is not what happened in AF (the trim never reached the self-disconnect condition) and, actually, when the other pilot briefly requested control and pushed down, the trim started to move down and the AoA started to diminish. In another example of a confusing effect, when the AoA diminished the stall warning (that had gone silent) started to sound again, which I hypothesize caused confusion (I am pulling up and the stall warning is silent, when I push down the stall warning activates? Better I don't push down!!!). The reason why that happened is that the software logic has a minimum airspeed blow which the AoA is considered not reliable (and hence the stall warning will not consider a high AoA reading as a sign of a stall). When the AoA was too high, the indicated speed was too low due to the high angle of attack of the pitot tubes (that were not frozen anymore). When the AoA started to go down (but was still too high), the real airspeed increased and the indicated airspeed increased even further due to the pitot tubes being not so much misaligned to the relative wind, so with the new airspeed the lower but still to high AoA reading was considered valid and since it was still beyond the stall AoA, there it went the stall warning again. Or how about the fact that, after the AP and AT disengaged, the UAS memory items called for climb thrust and 5 deg ANU, and the pilot were required to move the thrust levers OUT of the CLIMB thrust detent (where they were) to be able to establish climb thrust? (that is because of the thrust lock that comes with the autothrust disengagement and the fact that the thrust was below climb -even with the throttles in the climb detent- when the thrust lock kicked in.

Aircraft do have "stealth" factors (especially for pilots that are not really experts in the system). That said, click click, clack clack, grab a handful of yoke with one hand and a handful of yoke with the other, and make the plane do what you want it to do (and you better want it to do something that is within the laws of Physics) (and remember that the trim, whose control is located in the aforementioned yoke is an integral part of the pitch control system. In the scenario we are imagining for the Lyon crash, even without using the trim cutoff (as pilots were required to do by procedure and experts in the system would have done) the pilots could have still flown away with manual trim inputs in the trim switch located in the yoke.

[Black Ram]

That hypothetical scenario is clearly not what happened here

Yes, obviously. But it is one reason for the system's logic not to consider two very different simultaneous AoA measurements a fault.

And it does. With some not minor caveat, the the pilot would need to be aware of the trim motion (as they should) and react accordingly by acting on the trim controls as needed and required by the procedures.

An uncommanded trim motion in manual flight does not seem like giving the pilots full authority to me. Maybe it's tricky for the system to figure out when AoA measurements are not valid, but if there is a problem with the air data, I think they system shouldn't mess with the trim.
Yes, there is a procedure. There was also a procedure for UAS, but people still attacked Airbus after AF447, even our beloved hero Sully. Also, it seems the Lion Air pilots were dealing with another problem.

This is an issue in Airbus too. AF had the trim automatically and stealthy move almost full nose-up. In certain conditions it might have reached the full-nose up position and deactivate there and the only clue for the pilots would be a "manual trim" message in the EICAS.

To my understanding, the trim went fully nose up because the PIC put it there with his control inputs. Trying to climb past 38000 feet, burning all their speed, and continuing to keep the nose up....pretty sure it couldn't have happened unless the trim went up, so the system was just doing what the PIC was asking for. What's really ironic - Boeing's "runaway trim" stall prevention could have been more helpful there.

I don't think this is just about a runaway trim, because it's an action that's built in.

Btw, though I found AF447's pilots (pretty much all three) to be very incompetent and sometimes unprofessional, and though I didn't think the Airbus cockpit design was at fault, I never said there could still be improvements to help prevent such an event. From what I hear, the BUSS system is slowly becoming standard. Seems it was very useful for MH-134.

[Evan]

I was right to hesitate to join the thread. The response from some long standing contributors was to sneer and having sneered not to respond when I offered some clarification of my background.Notably, they failed to respond to my question as to whether pilots have the ability to go back to basic principles when the systems fail and to establish and maintain the aircraft in a safe and stable attitude.Instead, the response to this crash seems to concentrate on a retrospective forensic examination of how the " systems" failed. When you are sitting at the front and the system has landed you up shit creek it might be more pertinent to remember before anything else what keeps the aircraft flying. Having spent all those hours in the cockpit, don't you have a picture of how the view outside should look or what the artificial horizon might tell you?

Do pilots have the ability to go back to basic principles when the systems fail and to establish and maintain the aircraft in a safe and stable attitude?

Yes.

Except for the one's who aren't very well trained in systems. You seem to be suggesting that, when systems fail, all you have to know is basic airmanship, the skills every glider pilot should excel in. That's a popular belief on this forum. But it isn't true.

Sometimes systems are doing things that might confuse you or trip you up or scramble your situational awareness. If you know the systems as well as the airmanship, you know how to get these system failures under control and out of the way. That is part of stabilizing a modern airliner.

In the hypothetical case we are building here, that a system anomaly led to an uncommanded and unwarranted stabilizer movement, the pilots who know how to stabilize the aircraft would respond, in part, by cutting off the automatic stab trim. Every pilot who is properly trained on the 737NG/MAX has practiced the runaway pitch procedure very well.

And for the ones who haven't, the answer is:

No.

[pegasus]

I was right to hesitate to join the thread. The response from some long standing contributors was to sneer and having sneered not to respond when I offered some clarification of my background.Notably, they failed to respond to my question as to whether pilots have the ability to go back to basic principles when the systems fail and to establish and maintain the aircraft in a safe and stable attitude.Instead, the response to this crash seems to concentrate on a retrospective forensic examination of how the " systems" failed. When you are sitting at the front and the system has landed you up shit creek it might be more pertinent to remember before anything else what keeps the aircraft flying. Having spent all those hours in the cockpit, don't you have a picture of how the view outside should look or what the artificial horizon might tell you?

[Evan]

According to avherald (KNKT) the AOA sensor was replaced prior to the penultimate flight of this aircraft.

This may have been when the seed was planted.