I would state that it is pure inference, rather than speculation.

At 05:43:11, about 32 seconds before the end of the recording, at approximately 13,4002 ft, two momentary manual electric trim inputs are recorded in the ANU direction. The stabilizer moved in the ANU direction from 2.1 units to 2.3 units.

That could not have happened if the cutout switches had not been turned back on.

Oh! :ouch: I do apologize if I have posted wrong information. I had thought it was established fact.

Couldn't agree more.

Memory Items for Airspeed Unreliable?
Memory Items for Unscheduled Stab trim?

It appears the aircraft was never put back in trim after the initial MCAS input and although an attempt was made with the cutouts the aircraft was out of trim. Turned back on again contrary to advice (accept this is speculation at present) and the system was allowed to trim forward again.

Unfortunately these points will be lost to the majority on here who simply want to flame Boeing.

Thoughts.

This aircraft had multiple pitot static malfunctions in the days prior to the accident as per the tech log quoted in the report. Altimeter and VSI erratic.

If they had had a bird strike on the vane, they would have noticed the thump and said something. I think we can rule a bird strike out. And frozen things. It was 17 degrees C on the ground.

Just after takeoff the crew got an almost immediate
Stick shaker
Airspeed disagree
Altitude disagree
F/D disagree
Which constitutes an unreliable airspeed non normal
(Together with an anti ice caution)
And did not apply the memory items for that , and seemed to be prepared to continue to FL320.
The trim problems started once the flaps were retracted , however the flaps seem to have been retracted despite the left stick shaker being continually active and the aircraft in an unreliable airpseed state which calls for a pitch and power setting with the flaps down (if I recall, 10deg NU, 85%N1) but the power wasn't reduced from 94% either.
The stabtrim cutout switches seem to have stopped the MCAS from moving the stab, but then the switches were re-engaged (2 ANU trims were recorded 2,5 min after the stab trim switches were set to cutout) and this then allowed the MCAS to pitch it down again.
As regards thrust and pitching moments I personally would prefer to take my chances with speedbrake, idle thrust and a bit of a dip than two engines at 94% when I'm pointing downhill.
And I don't think I'd be getting ATC involved until I had the beast back under some sort of control.
This seems to be more than an MCAS thing, it may be a pitotstatic adiru thing.
If one sensor gets extreme readings or if there's a sensor disagree, then no system should rely upon that sensor data.
Systems shouldn't rely on one sensor only.

The 2 blips occurred before the final MCAS action, in fact likely triggered it so the extreme final G load is not a factor. They were flying but with (great?) difficulty before then.

Not sure 3 units AND (from trace) while 'stable' is not correct. From the report it appears they tried and failed to use manual trim wheel after cutout and then (last ditch option) re-enabled it to attempt electrical trim.

Gets back to the question of why 2 short blips instead of continuous until collum force approaches neutral?

AoA back to normal
 

Any ideas on why the faulty left AoA goes to values even lower than the assumed correctly working right AoA at the same time the terminal descent starts? MCAS works for a while, left AOA goes down, stick shaker is happy, MCAS stops.

I'm not seeing it that way. With the AP engaged the 3x AND auto trim events kill the pitch attitude - from 8 deg positive decreasing all the way to zero by the time AP tripped-out. At that stage with the AP engaged, FD commanding a climb and with throttles parked forward with excess speed in hand the aircraft should have been going up like the proverbial. Indeed, the pitch attitude only becomes positive again when PF takes control and applies a generous amount of aft stick and ANU blips. Even then he does not achieve anything like the 8 deg positive achieved earlier.

Clearly I know what the aircraft 'should' be doing, but I'm just not finding it in the data provided.

If you see it differently please shout out as even after years of flight testing this has left me scratching my head.

Yeah, from 2.1 to 2.3 units, I know.

It could very well be that the initial blip moved the trim slightly and then stopped, they released and try again but no avail. Each square in the FDR is 3 seconds, so the second "short" blip is 2s or so. Between the two, looks to me in the 3s area. (one full square) Let's say 2.5 to be safe.

The final AND input by MCAS was 5s and moved 1.3 units. The last pilot input ANU was 2.5s (combined) and moved 0.2 units. Something odd is there.

It is a theory, I could be mistaken. But it's plausible to my eyes.

I am struggling to see any effect, and I didn't see any other indication in the report that the crew re-engaged.

I suspect that cutout switch position is not directly recorded and has to be inferred, possibly this is where the rumours of disagreements and CVR reading problems originate from - maybe there are sounds or speech on the CVR that could indicate re-engagement but it isn't definite, or agreed. Any reference to the disputed sounds may then have been excised from the report.

This MCAS signal got through, which some will take as proving the switch position was changed, there is so much else going wrong not sure I would agree (that it is proven). Also, this still doesn't look right per the descriptions we've had:
- MCAS is supposed to have less authority at higher speed (maybe that's it, it is shorter input?)
- If AOA reduces then MCAS is apparently supposed to put the trim back up, if no pilot trim input - yet there isn't and it doesn't

Re: The FDR plot

Note the call-out box which states "Automatic Trim Command With No Change In Pitch Trim". However, there is another downwards blip of automatic trim command shortly before the crash, which clearly causes more nose-down trim. This must be the result of the MCAS being re-enabled. For some reason, the CVR transcript does not say so, but the pilots must have re-enabled the Cutout Switch.

Could be a DFDR artefact. On another aircraft type I am more familiar with a continuous blip when the trim motor does not run just poles the DFDR data flag on and off again. This single pole will appear as a spike when graphed.

In the accident graph the 2 final electric trim blips do not appear to move the pitch trim, either due to cutout or aerodynamics overcoming the drive.

Unfortunately the plot has artificial slopes on binary events so without the raw data is hard to tell the actual length of short inputs so the total 'blip' could be significantly less than 2.5s.

Interesting idea that when they tried the electrical trim the system stalled so they tried again, hard to tell without raw data and CVR though.

AFAIK MCAS uses the same rate of nose down trim of 0.27 units per second, it is the duration that varies, with a maximum of 9 seconds.
MCAS will 'unwind' the trim only when the AOA reduces below the 10 degree threshold, not when AOA reduces relative to its previous value.

Unless I misread the preliminary report, it appears that the "advice" that was followed was that of the FO, who advised that manual operation of the trim wheels was not possible. It's fairly easy to understand why the crew decided to flip those switches back on (if they did).

Questions after reading the report
 

The preliminary report leaves me with several questions:

Shortly after Takeoff the Captain who was pilot flying had stick shaker on his side only. Why not transfer control to the first officer at this point or at least a little while later? That is actually what the Captain of the Lion Air flight previous to Lion Air 610 flight did. If you see that there is stick shaker on your side but not on the other wouldn't it make sense to rely on the FCC B side in this case? Why engage the autopilot on the A side that has the stick shaker? The captain tried to engage Autopilot A two times unsuccessfully at 400 Ft and 630 Ft. That would be another indication to either continue manually or at least try the other side. Finally he succeeded in turning the autopilot A on at 1000 Ft (for 33 seconds). Possibly Autopilot B would have worked since it seems like there was no false sensor indications on that side an MCAS would not have activated (speculation).

Based on the report the situation for ET302 was basically exactly the same as for Lion Air 610. Stick shaker right after Takeoff and then as the flaps were retracted MCAS started trimming the nose down.

At 05:40:00 shortly after the autopilot disengaged, the FDR recorded an automatic aircraft nose down (AND) activated for 9.0 seconds and pitch trim moved from 4.60 to 2.1 units. The climb was arrested and the aircraft descended slightly.

Knowing about the Lion Air Accident as the ET302 crew must have, were they expecting the nose down trim as they retracted the flaps? Based on their actions it seems like they did not even though the circumstances were the same. Would this not have been the point to carry out the runaway stab trim memory items?

At 05:40:12, approximately three seconds after AND stabilizer motion ends, electric trim (from pilot activated switches on the yoke) in the Aircraft nose up (ANU) direction is recorded on the DFDR and the stabilizer moved in the ANU direction to 2.4 units. The Aircraft pitch attitude remained about the same as the back pressure on the column increased.

Instead the crew trimmed the aircraft nose up from 2.1 units back to 2.4 units. Why not trim back to the original 4.6 units before MCAS engaged so that it isn't necessary to increase back pressure on the column? Some people on this forum suggest the electric trim switch will not work sufficiently against MCAS. The preliminary report on Lion Air 610 shows otherwise as they countered MCAS up to 30 times.

By now we know that MCAS activates again 5 seconds after the electronic trim switches have been used. They were used by the crew but apparently only to offset the MCAS trim input to a small degree and now MCAS starts trimming the nose down again:

At 05:40:20, approximately five seconds after the end of the ANU stabilizer motion, a second instance of automatic AND stabilizer trim occurred and the stabilizer moved down and reached 0.4 units.

This time the crew did trim against MCAS from 0.4 units back to 2.3 units therefore basically reversing the input from MCAS and showing that this is possible.

Now the First Officer suggested to set the stab trim cut out switches to cut out and MCAS was stopped.

Would it have been possible at this point to reduce the power and then trim the aircraft manually?
At the time the crew did try to trim the aircraft manually the aircraft was flying at the edge of it's envelope at 340 Knots according to the left side or outside of it already at 365 according to the right side. I am not surprised that at this point it was not possible to trim manually anymore.

At 05:43:11, about 32 seconds before the end of the recording, at approximately 13,4002 ft, two momentary manual electric trim inputs are recorded in the ANU direction. The stabilizer moved in the ANU direction from 2.1 units to 2.3 units. At 05:43:20, approximately five seconds after the last manual electric trim input, an AND automatic trim command occurred and the stabilizer moved in the AND direction from 2.3 to 1.0 unit in approximately 5 seconds.

Since the crew said at 05:41:46 that electric trim did not work and now it was possible for them to trim electrically again it is likely that the stab trim switches were set back to their original position before being set to cut out. Aside from not reducing the power even when there was the overspeed warning, why did the crew not trim continuously as they did before in order really raise the nose? Also why did they then not set the stab trim switches back to cut out since it was to be expected that MCAS would engage again 5 seconds after release as it did before?

I am not judging the pilots or saying the same thing could not have happened to me. I am just asking myself these questions based on what I have read.

Step 2 - Control airplane pitch attitude manually with control column and main electric trim as needed.

Was this step done? 2 units is nowhere near in trim and significant control column input would've been needed. Full stabilizer authority was available to the pilots using the electric trim.

If runaway continues... I'd argue people are getting hooked up on semantics here. You can't put a long winded paragraph in covering every eventuality. The system is still not performing as expected so you cut out the stabs, ensuring step 2 is complete and the aircraft is in trim.

The bulletin Boeing sent to all operators and should have been read by all pilots was very clear on this point in 'Operating Instructions'.

737mgm

There is a subtle but perhaps crucial difference between the flights. With Lion Air 610 the autopilot was never engaged, and MCAS activated as soon as the flaps retracted. With ET302 the autopilot was engaged before flaps were retracted, and remained engaged for a few seconds longer, during which time MCAS was inhibited. The trigger for MCAS activation was autopilot disengage, not flaps retraction.

Cognitively this is a very different situation, and focusing on why the autopilot disengaged may have been a distraction from the MCAS activation. Conversely flaps retraction is a non-event cognitively, and more attention could have been given to MCAS activation. I do not know whether any of this is relevant, but it is perhaps important to see the whole sequence of events.

I don't disagree and I'd have probably done the same. Surely the moment they were back on though you'd be trimming hard?

From the narrative it sounds as if the Capt hasn't grasped what is happening and it's a suggestion from the FO to cut the switches out in the first place. So when they turn them back on is he 'in the loop'?

There are multiple reports suggesting that the deadheading pilot was a Batik Air captain. I haven't seen any that mentioned their nationality, and I note that the poster who suggested it was a Brit hasn't come back with any confirmation of that.