Originally posted by Evan
View Post
1) First frame of the video: The plane is in a left bank but rolling right quite fast. The left bank angle was obviously steeper before the start of the video.
2) Bank pass approx wings level, still rolling right fast, indication that we have right-roll command input.
3) Roll spoiler starts to deflect (almost unnoticeable in the picture). However the roll rate had already slowed down a bit by then, indication that the right-roll control input had diminished or been neutralized before this picture.
4) The wings reach the max right-bank, roll rate is hence zero, the spoilers have not reach the max deflection.
5) Wing is coming down, spoilers reach maximum deflection.
6) Bank pass wing level, spoilers still in maximum deflection for left roll.
7) In a left bank and rolling left, spoilers start to retract.
8) Spoilers complete retraction, plane is still rolling left. Cannot be seen in the video but most likely the command has to be for right roll after this frame, however the plane keeps rolling left due to inertia.
9) Pane reaches maximum left bank, roll rate is zero, plane starts to roll right after this frame.
2nd row is the second cycle, which is basically a copy-paste of the previous one, except with a little bit less of amplitude (but then these were the last 2 cycles).
I was thinking that, since the initial roll excursion was uncommanded (an upset rather than an intention), would the AP logic—up to a point— simply rely on the aircraft's lateral stability to restore wings level (no flight control inputs)? And if that point is around 33° bank, would something in the logic THEN command a counter-rolling flight control movement?
Because that is what appears to be happening in the video showing the delay in the spoiler deployment. I wouldn't expect such a delay from the autopilot unless it was intentionally delayed.
The A320 normal law in roll requires a pilot in manual flight to maintain full stick deflection to exceed 33° of bank. It also introduces THS limits at this point. So perhaps there is a similar functional threshold in autoflight at this bank angle.
Is it possible that there is a bit of an oversight in some flight control algorithm that only reveals itself in this rare combination of circumstances? Something that can lead to A/P-induced occillation... until another mode is selected or the A/P is disconnected (something I might expect the pilots to do if this was happening under autoflight)... ?
Comment