Interface/scan/human factors
This discussion has been largely focussed on the central and tragic sustained pitch-up; may I request attention to a possible clue which not all have picked up? Please take another look at the roll axis traces (bottom of pp 29, 30 of report 3). (I am grateful to xcitation et al. for their posts yesterday which made me revisit in more detail my bad gut feelings on first viewing these traces).
1. I'd expect anyone having the aircraft unexpectedly dumped in his lap on AP disconnect to take a few moments to 'get hold of her'. It took the PF ~ 35 seconds to control the roll, and in that period I see what looks to me like overcontol and PIO; the oscillations are too regular for this to be attributable just to (midocean) turbulence. In this period he seems to be struggling more that I would expect of a fully fit and current-on-type pilot; am I being unfair?
2. During this 35 second period I surmise that the PF's attention was focussed strongly or exclusively on his display re. roll - which was evidently working perfectly well - to the possible exclusion of a better scan.
The PNF starts very rightly to 'nag' towards the end of this period, and what happens afterwards seems to me uncannily like what instructors recognise as 'freeze' when a student is unintentionally overloaded. (My compliments to BEA for their very helpful combination of CVR and FDR events).
I am neither willing nor able to seek to apportion any blame - I have never flown a sidestick aircraft, nor indeed any without the 'luxury' of tactile feedback on the controls - can Airbus drivers please comment on the significance or otherwise of the early roll-axis traces?