This was quite a revealing bit of the puzzle when it was announced. In short (lay persons interpretation): due to a separate issue they had with the airspeed disagree indicator warning they followed the course of action for that which was to speed up the plane to prevent potential stall. You’re flying, you don’t know for certain the speed, so go a bit faster.
But this meant when they had the AoA/MCAS issue they cut off the MCAS and were going too fast to be able to manually trim the aircraft due to the fact MCAS had driven the stabilizer to full deflection. At that speed and that amount of deflection and pulling back with full elevator to try keep the nose up the forces were too great to manually trim.
The solution to that is to ease up on the elevator allow the nose to go down and manually trim then nose back up. But they were too close to the ground still to do that safely. So they needed the electric trim enabled, but that means having the faulty MCAS enabled which they needed off.
Suggested work arounds were to set flaps to 1 while leaving the drab trim cutoff disabled which would’ve disabled the MCAS and let you still electrically trim. But even if you thought of this you can’t extend flaps at 400kt.
If all this is true then this accident is definitely not as cut and dry as everyone thought initially and the solution of “just cut the Stab Trim cutout switches and trim manually” isn’t a solution in certain failure scenarios, namely this one.
Aren't all these scenarios tested in simulators and in the actual aircraft? If it's not possible to recover the aircraft why was this not know and why was this plane still flying after the first crash?
> Aren't all these scenarios tested in simulators and in the actual aircraft?
In this case the answer is, no:
"During nine to 12 months of MAX flight testing, test pilots injected errors into the flight system that tested stall conditions and runaway stabilizer, among other scenarios, the people said. But no one was aware of a specific test of an MCAS failure mode triggered by erroneous sensor data."
Planes are complicated, and you can't test for all n! different things that can go wrong. I think the case that:
a) an MCAS sensor is giving incorrect data AND
b) the data is leading MCAS to set the trim to the max AND
c) the plane is going so fast that the force of the air over the elevator is too great to manually override AND
d) the plane is too low to temporarily drop altitude to reduce the force on the elevator to allow a manual override
I don't think this is an obvious set of things to foresee happening, especially D.
It's the same way with software. There's just too many things to think of, and errors that slip through the test cases are inevitable. I've had what I've considered to be pretty thorough test cases and had a bug that was "obvious".
I'm not excusing Boeing, but it isn't gross incompetence either.
It makes sense that even with two pilots trying to manually spin the trim wheels from a full nose down configuration would be nearly impossible due to aerodynamic loading. IIRC relatively unloaded manual trim wheel movement is difficult even with both pilots extending the trim wheel folding handles and spinning them with maximum effort. I would like to know if commercial simulators accurately simulate manual trim wheel difficulty under high aerodynamic loading such as full nose down trim and high speed.
I think the electric trim cutout switches should be repurposed as "automation trim cutout switches" so that pilots can continue to use the yoke trim switches under high aerodynamic forces where manual trim reversion isn't possible and electric power is still available... and pilots could still revert to trim wheels during electric power loss conditions. You wouldn't disable the brake booster in a car because cruise control went haywire.
> think the electric trim cutout switches should be repurposed as "automation trim cutout switches" so that pilots can continue to use the yoke trim
Those trim cutout switches were actually first put in place long ago to cutout the electric trim on the yokes for situations when they get stuck or otherwise runaway.
So really what’s needed now is a different set of MCAS disable switches.
Despite the similar title to the WSJ article, this contains heaps of new original information based on actual simulator testing that hints why the pilots might have re-enabled the cut-out switches.
Exactly. This is the original source of an attempt of one real 737 pilot to reproduce the conditions similar to those that happened during the last crashes, using the existing (non-MAX) Boeing 373 flight simulator.
The company he works for requested the takedown of the video, but the video shows a lot that the public shoold know, and the conclusions are still there.
8 comments
[ 2.8 ms ] story [ 30.8 ms ] threadBut this meant when they had the AoA/MCAS issue they cut off the MCAS and were going too fast to be able to manually trim the aircraft due to the fact MCAS had driven the stabilizer to full deflection. At that speed and that amount of deflection and pulling back with full elevator to try keep the nose up the forces were too great to manually trim.
The solution to that is to ease up on the elevator allow the nose to go down and manually trim then nose back up. But they were too close to the ground still to do that safely. So they needed the electric trim enabled, but that means having the faulty MCAS enabled which they needed off.
Suggested work arounds were to set flaps to 1 while leaving the drab trim cutoff disabled which would’ve disabled the MCAS and let you still electrically trim. But even if you thought of this you can’t extend flaps at 400kt.
If all this is true then this accident is definitely not as cut and dry as everyone thought initially and the solution of “just cut the Stab Trim cutout switches and trim manually” isn’t a solution in certain failure scenarios, namely this one.
In this case the answer is, no:
"During nine to 12 months of MAX flight testing, test pilots injected errors into the flight system that tested stall conditions and runaway stabilizer, among other scenarios, the people said. But no one was aware of a specific test of an MCAS failure mode triggered by erroneous sensor data."
March 29, https://www.reuters.com/article/us-ethiopia-airplane-boeing-...
a) an MCAS sensor is giving incorrect data AND
b) the data is leading MCAS to set the trim to the max AND
c) the plane is going so fast that the force of the air over the elevator is too great to manually override AND
d) the plane is too low to temporarily drop altitude to reduce the force on the elevator to allow a manual override
I don't think this is an obvious set of things to foresee happening, especially D.
It's the same way with software. There's just too many things to think of, and errors that slip through the test cases are inevitable. I've had what I've considered to be pretty thorough test cases and had a bug that was "obvious".
I'm not excusing Boeing, but it isn't gross incompetence either.
It makes sense that even with two pilots trying to manually spin the trim wheels from a full nose down configuration would be nearly impossible due to aerodynamic loading. IIRC relatively unloaded manual trim wheel movement is difficult even with both pilots extending the trim wheel folding handles and spinning them with maximum effort. I would like to know if commercial simulators accurately simulate manual trim wheel difficulty under high aerodynamic loading such as full nose down trim and high speed.
I think the electric trim cutout switches should be repurposed as "automation trim cutout switches" so that pilots can continue to use the yoke trim switches under high aerodynamic forces where manual trim reversion isn't possible and electric power is still available... and pilots could still revert to trim wheels during electric power loss conditions. You wouldn't disable the brake booster in a car because cruise control went haywire.
Those trim cutout switches were actually first put in place long ago to cutout the electric trim on the yokes for situations when they get stuck or otherwise runaway.
So really what’s needed now is a different set of MCAS disable switches.
Despite the similar title to the WSJ article, this contains heaps of new original information based on actual simulator testing that hints why the pilots might have re-enabled the cut-out switches.
The company he works for requested the takedown of the video, but the video shows a lot that the public shoold know, and the conclusions are still there.
Moderators maybe please take a new look here.