The NTBS came out with a report recently: “Assumptions Used in the Safety Assessment Process and the Effects of Multiple Alerts and Indications on Pilot Performance”
Basically:
The NTSB cite is a good read. The end result of the investigation is going to affect how certifications are done on all aircraft.
But the synopsis of the 3 flights show all 3 crews reacted as expected to the errant computer trim and re-adjusted for it in real time.
the preceding Lion Air flight had the same mechanical issues. They didn’t explain what was done to fix or test it. The stick shaker on the Captain’s side was active the entire flight and the computer activated trim for 10 seconds after rotation. They countered with pilot electric trim and eventually shut it down and completed the flight with manual trim.
There are detailed graphs of dozens of relevant items in the ET302 preliminary report [PDF], pages 26 and 27. I don’t know how anyone can even comment on the crashes without reading this report.
It’s not that the dive is too steep, but the stabiliser, the tail-plane, is so far out of trim that the elevator, a much smaller surface, is not powerful enough to counteract it. There is also the problem that they were going very fast which makes all of the control surfaces more powerful, once again the stabiliser beats the elevator due to its size. A further problem is that at high speed, and with the stabiliser trimmed all the way nose down, the manual trim wheel did not have the ability to move the stabiliser against the air loads. The electric trim should still have worked though. It is a mystery why, after reinstating the electric trim, the pilot’s only appear to have made a few blips of trim input instead of the 20 seconds or so it would take to get the stab back to a reasonable position. They seemed to know what had to be done, but they either didn’t or couldn’t do it for some reason.
It’s interesting that there is no change in pitch trim corresponding to those two blips of electric trim commands. It looks like they had no effect on the trim.
In the earlier electric pitch trim commands, just after the stick shaker started, you can see immediate changes to the pitch trim. Those last two inputs seem to have done nothing.
Also, corresponding exactly with the second blip a warning is triggered, ‘AP WARN CAPT’. I’m not sure what that autopilot warning sounds like, but it might have been distracting.
Thank you very much for the clear answer. I guess, perhaps due to the inevitable panic induced by flying towards the ground at several hundred mph (which no amount of training is going to completely eliminate), they didn’t disconnect in their minds the automatic trim (which was causing the problem) and electric trim (which might have saved them), instead going straight from “automatic trim” to “manual trim”, neither of which could help them.
ETA: the above in reply to Richard Pearse; I see now from GreenWyvern’s intervening post that another possibility is the electric trim wasn’t working (or at least they thought it wasn’t).
I was looking for a link to a synopsis of the graphs that was easier to read.
It would be nice if you understood the subject matter and made an opinion based on facts instead of forming an opinion and using other opinions to back it up. I invite you to parse out your own link with any specific fact if you disagree with what I’ve said.
The previous crew that flew the Lion Air plane went through the same problems as the fatal flight that followed. They maintained control of the plane while assessing the situation and simply trimmed the plane to a desired setting and then shut off the electric trim. Problem solved.
The pilots of the fatal crashes had sufficient control of the planes to bypass the MCAS system. They also had enough information to understand the AoA mis-match and determine which of the 2 was giving good information. They also had enough information from other gauges to ignore the AoA data altogether. they allowed the trim system to extend to a range that was difficult to handle.
The Ethiopian flight was the result of the same problem and required the same resolution.
This is ironic. For starters, there is no AoA data presented to the pilots. They can’t see a mismatch. They had airspeed discrepancies, but those are more often caused by pitot/static problems rather than AoA data issues.
There is no doubt that there was a sequence of actions the pilots could have followed that would have prevented the crashes, but those actions had not been clearly set out by the manufacturer. The importance of using electric trim to get the aircraft in trim prior to disconnecting the electric trim was severely understated. It should have been in big bold you’ll-die-if-you-don’t-do-this font at the start of the procedure, instead it was almost a footnote at the bottom, a mere suggestion, and only added once one 737 had already crashed.
An MCAS runaway was different to how a trim runaway is typically presented. Because the MCAS could run the trim much faster than the normal electric trim, it took very little time for the MCAS to make the aircraft unflyable unless the crew did everything right, but no one knew what “everything right” was until after the accidents. The trim runaway procedure didn’t have “everything right”. Sure, the information was there, but it needed to be interpreted and practiced.
The fact is that with abnormal procedures you really need to have a practice so that you can see and feel how to do it right and what happens if you do it wrong. Boeing didn’t want anyone to practice it though, that would suggest that additional sim training was required and they were under immense pressure to provide the 737 MAX with no sim training requirements. Southwest Airlines had a one million dollar penalty clause for every jet they’d ordered if sim training was required.
The hypocrisy is that some people suggest that the accident pilots weren’t trained sufficiently to deal with the problem, but the lack of training stems directly from Boeing who in the first instance didn’t tell anyone about the new system, and, once there was a crash, tried to say that an existing procedure was all that was required without stressing the important differences between a traditional trim runaway and a rogue MCAS.
they see a mismatch of data in the warning. If my assumption that there is a displayed number associated with it is wrong then I stand corrected. but it was a substantial disagreement between PIC and FO AoA input. Something like 50 degrees. there is a stick shaker on one side and nothing on the other. One of the systems is messed up.
You don’t need pilot electric trim to counter the computer. You need to arrest the downward trim before it gets out of hand. The procedure is effectively that of runway trim. Having pilot electric trim to counter computer electric trim is simply a bonus and all three crews demonstrated that pilot electric trim worked.
yes but you see and hear the trim wheel as well as see a change in pitch. since they reacted to the change in pitch to counter the computer it seems to me they knew what was going on. there are clearly the distractions of the alarm and stick shaker but the PIC should be focused on the flight envelope.
Yes, there should have been better training. Boeing should have jumped through hoops after the Lion Air crash. and then jumped through some more.
Looking at the altitude the F/O called 8,400 feet. That would be less than 800 feet AGL from ADD. I just don’t understand why a nose down trim input wasn’t taken more seriously at such a low level. There is the repeated up-trim by the pilots so they’re clearly aware of the nose dropping as well as the source which is the computer generated down trim. They knew the electrical up trim works and they knew how to shut off the trim. There is nothing subtle about what is going on.
Boeing owns this but it was well within the skill of the pilots to correct it. At 800 ft the reason behind the faulty trim input is overshadowed by the need to correct it.
There is no number or anything. There is no AoA indication of any sort on the vast majority of airliners. The B737 MAX had AoA indicators as an option, for a fee. Only Southwest and American had taken up that option. There was supposed to be an AoA disagree alert that was fitted to all B737 MAX aircraft, but due to an error, it was only enabled on the aircraft that had AoA indicators.
It is pretty clear that they did need to do this. Maybe not if they were expecting the MCAS problem and were primed to take action (as in sitting in a simulator and being told they were going to do an MCAS exercise). It takes time to comprehend what’s going on and recognise what actions needed to be taken.
Remember that a trim input stops the MCAS for 5 seconds, so each time they applied nose up trim it seemed like the problem was fixed. It may have taken a few cycles to realise that the problem is not going away and by that time, if they had not countered the MCAS sufficiently with electric trim they then find that the MCAS has put the stab so far out of trim that they can’t trim with the manual trim wheel. Then they reactivate the electric trim in a desperate attempt to use it but it just sets MCAS off again, puts the plane into a -2G bunt, lifts the pilots hard up against the flight deck ceiling because they don’t have negative G straps like a hardcore aerobatic plane does, and they can no longer adequately reach most of the controls to do anything more about it.
What is a stick shaker if not a flight envelope warning? In hindsight it is easy to see that the trim was the problem and the stick shaker and other alarms were distractions, but when you are there in the moment there is no bright beacon pointing out the core problem, there is just a bunch of contradictory stuff happening and any part of it might be the core problem. What if the stick shaker was genuine? Then the MCAS activation would also be genuine and turning it off would be the exact wrong thing to do. They have to work out what’s what for themselves and this takes time
I work with a flight attendant who likes to give word puzzles written on our coffee cup lids. The other day it was “GESG”. Now that is obviously “scrambled eggs” right? Well no, not obviously. It’s obvious to me because I have seen the puzzle before, it may be obvious to you if it just “clicked” or you do a lot of similar puzzles, but it’s not objectively obvious.
The MCAS “problem” has been solved by one crew before and not solved by two. The first and second crews had no prior knowledge, the third crew had been given an incomplete solution along with some hints. Of course it is obvious to us now, but that is because we have been given the complete answer, and even then, there is no guarantee that you or I would act appropriately if faced with the MCAS problem now.
They were well aware that the nose down trim was a big problem, the captain was calling for the FO to help trim with him and pull up with him, but despite how easy it seems to us who weren’t there, they couldn’t fix it. They didn’t fail to fix it on purpose. They didn’t set out to crash. They were fighting for their lives and it wasn’t enough.
Correct that it was well within the skills of the pilots to correct it, provided they know what the problem is and what the solution is and what the repercussions were for mistiming of the necessary steps. Where Boeing failed miserably is in not clearly identifying the problem. It is not a classic trim runaway, continuous uncommanded movement of the trim. MCAS isn’t continuous. In a normal trim runaway pulling the control column aft activates cutout switches that disable electric trim. In a normal trim runaway, all you need to do to not die is pull the stick back and the trim stops. Those cutout switches were disabled with MCAS.
I deleted a long response to your post because it comes down to a single issue. All 3 crews reacted as expected to errant trim input. All 3 of the crews should have arrested the continued computer input when it was apparent it wasn’t going to stop.
You brought up a good point about what is obvious to one person is not necessarily obvious to another. We just disagree on the degree of it. All 3 crews reacted repeatedly to MCAS by countering it. Had they not reacted this way I would agree with you.
But I think a trim issue calls for focus on the trim system. Unless there’s another mechanical issue not yet revealed it’s up to the crew to ensure the horizontal plane remains within controllable range. The rest of the plane is responding to crew input. They have control of the engines, ailerons, elevator and flaps. The trim issue can be relegated to manual mode. If there’s an alarm that says it’s going to stall that will certainly rattle a crew but they can see they’re well inside the flight envelope.
They did relegate the trim to manual mode, it didn’t work. They likely didn’t appreciate the necessity of having the stab in trim (with man electric trim) prior to using the cutout switches. Why? Because the inability of the manual trim wheel to move the stab in extreme out of trim conditions had been lost to the mists of time. Additionally it has been discovered since the crashes that the B737 simulators didn’t realistically simulate the force required to move the manual trim wheel, so even if they had practiced in the sim they wouldn’t have appreciated that using the cutouts while out of trim could be fatal.
Regardless of “should”, that fact that two fatal accidents happened due to the same issue indicates that a rapid resolution is outside the ability of some percentage of pilots.
And, at the risk of understating the issue, that percentage is higher than we typically prefer.
Your statement seems in line with the NTSB’s assessment. They said Boeing assumed too high a degree of pilot competency in the design and certification of their systems.
I find this disturbing. How does a commercial pilot not understand this? Even putting this aside how does a commercial pilot not understand the problem of full trim?
Great. What else don’t they do right and is this a systemic problem with simulators in general?
Yes. Aircraft need to be designed for the pilots who will actually fly them, not the pilots that we think should be flying them.
Well, if you can hold the force with the elevator then it shouldn’t be a problem disconnecting the electric trim and returning to an in-trim condition manually right? That’s how all the aircraft I’ve flown have worked. The trim runaway scenario kind of assumes that you can’t get it in trim with the electric trim in the first place, otherwise it wouldn’t be a runaway, but MCAS is a special case that doesn’t fit with a traditional trim runaway.
There are some key differences between a trim runaway and rogue MCAS. The control column has automatic cutout switches. In a nose down trim runaway, if all you do is instinctively hold the column aft then the column cutout switches will activate and the runaway will be automatically stopped. Not so with MCAS. MCAS was designed to keep working with the pilots holding back pressure on the column, so those cutout switches do not affect MCAS and the MCAS will continue running the trim nose down.
With a trim runaway, the instinctive action of holding the column back will stop the runaway before the stab is so far out of trim that it can’t be returned using the manual trim wheel. That is a big safety net that is missing from the rogue MCAS scenario.
I’ve never put much stock in them for replicating handling. They are best for procedures.
Yes but the point of arresting a runway trim is to avoid full trim and negative forces on the yoke. Otherwise you’d just shut if off whenever you want and manually run it back.
I agree with your logic puzzle argument but I don’t see this as a series of dots to connect. You avoid runaway trim because of the problems it creates. The continued attempts of the system to retrim for whatever reason becomes a runaway event if left alone. They repeatedly stopped the unknown trim input with counter input. It’s not a huge leap of logic to shut it down before it causes problems.
The plane is 800 feet off the ground and wants to trim down. The solution is to retrim it, shut it off, and fly the plane. Even if the crew ignores the trim gauge they can feel the yoke getting heavier. They can feel the problem getting worse and they know it’s a trim issue. It’s flying at it’s most basic level.
MCAS 2001: "I’m sorry Dave, but I’m afraid I can’t do that "
They DID shut it down, and pretty quickly too.
they eventually shut it down after it exceeded their ability to recover with yoke input.
As it continues to trim down, the yoke force to pull up increases. Everything in front of them showed it was a trim issue and they had electrical control over the trim as well as manual trim. All they had to do was pull the nose up with the yoke, trim to a neutral force, and shut off electric trim.
Interesting new news, unfortunately from a paywalled site:
Previously I had assumed it was a stupid oversight. But instead they knew how to do it correctly–but decided to save a few bucks on each plane.