Originally Posted by
Loose rivets
One of the biggest issues to me was the redatumization of MCAS upon reset. This would allow the inputs to sum and account for the screwjack going to full travel.
I've raised this issue on R&N with no real response. It seems so important.
Another mystery is the clear statement way back of MCAS reestablishing the old setting after actioning. I just don't know how to search that in the vast number of posts.
Reading everything FcEng84 has written is good advise.
One source of confusion is the mixing/confusion of descriptions of MCAS 'normal operartion' and with a failed (positive) AoA sensor.
What I read was that in normal operation absent any pilot trim inputs it would back out it's input once the AoA returned to a lower value.
Any pilot trim inputs cancel entire the sequence (further ND trim and the restore), basically reset to initial state.
The flaw in the design is that in case of a hard AoA fail the action is to trim nose down and then do nothing
-until- it sees a pilot trim input which reset it to initial state.It then waits 5 seconds and begins again.
It it then again detects high AoA and repeats ND with no 'memory' (preserved state) of prior actions. This is the root cause of extreme ND trim.
In ET case you can see this action, including the ~10S attempted ND trim after the trim cutout (no effect).
It then does nothing until after the trim is re-enabled
and 5 seconds after the ineffective short pilot trim inputs.
The Boeing 'fix' description includes a limit on total trim applied as well as using both AoA sensors. It is ambiguous on what constitutes an occurrence though