Originally Posted by
Smilin_Ed
Quote:
It seems that a number 1 priority is that AB recognise that their aircraft can stall....
BOAC:Fully agree!
If they didn't think it could stall, why did they build in a stall warning and stick shaker ? Why did they document it ? Why did they publish info on stall recovery procedures ?
Airbus knows, and always has, that their aircraft
can stall. Airbus state, and always have, that their aircraft
can stall. So where does this "cannot stall" come from, who is saying it ? Are there really actual airbus pilots (as opposed to clueless journalists etc) who believe their a/c cannot stall ? If so, the training is way more broken than even the most extreem comments on this thread have suggested.
And number 2 is that when airspeed information, something that the flight control system relies on heavily, becomes unreliable, that autotrim of the THS should drop out with the autopilot.
737 (for one) does that (and the bus in direct law - see perpignan). Works really well - not. A/P trims a/c up (eg trying to follow glide slope with too little thrust), drops out at stick shaker and leaves pilot approaching stall fully trimmed up. How many then remember to re-trim ? Schipol ? Perpignan ? Bournemouth ?
I am starting to believe that from a user interface point of view, autotrim should either be
always on (extension of elevator control,
always) or
always off. Failure to make the transition from auto to manual trim is becoming a pattern in an lengthening list of accidents and incidents.
We won't know until we can see the plot of stick motion but it seems to me that the PF didn't know, or forgot, that when he pulled back on the stick, he was trimming into the stall. Do we know that the initial pitch excursion was the result of some action of the flight control system and/or the autopilot, or was it the result of PF action in response to erroneous indications?
Autopilot was out before pitch up. As to FCS, we do not know yet, and even the BEA may still be working through the data to determine whether control surfaces responded correctly to pilot inputs (and only pilot inputs) through the event.