Details emerging on AF447 crash ahead of official report

[Rob Silliman]

I don't know how Air France and possibly Airbus dodge the blame for this one.  It seems pilot error is the primary cause.

http://www.telegraph.co.uk/technology/9231855/Air-France-Flight-447-Damn-it-were-going-to-crash.html

[Bill]

It's an interesting article but I wonder if it goes far enough in questioning the Airbus design. I don't thing the physical design and location of the side stick is as big a problem as the software design. At one point Robert did try to take control from Bonin and push the nose down and the aircraft did not respond because Bonin was still pulling up on his side stick. When the Airbus flight control computer receives conflicting inputs from the two side sticks it averages the inputs. Conflicting control inputs should result in a voice alarm warning the flight crew of the conflicting inputs. Had such an alarm sounded Robert would have known immediately why he could not pitch down.

Another problem is that the stall warning is disabled, apparently without warning, if the computer considers any of the necessary inputs invalid, for example if indicated airspeed is less than 60 knots. The flight crew should have known this, however, again, an audible warning would have helped the crew avoid the confusion that occurred later in the event when the stall warning came on when they pitched down and went off when the pitched up. This happened because they were already stalled and airspeed was so low the stall warning was disabled. When they pitched down and airspeed increased enough for the computer to consider it valid and the stall warning sounded again.

Another problem with the Airbus design is that it will not allow the pilot to fly the aircraft outside what the computer considers the normal envelope for the circumstances. Capt. Chesley Sullenberger pointed out that his landing on the Hudson was harder than it needed to be because the Airbus flight control software would not allow him to pitch up as much as he needed to just before impact. A Boeing aircraft would not have limited what the pilot could do because Boeing's fly-by-wire system uses push-back servos to give the pilot tactile feedback that he is trying to fly outside the envelope but the pilot can simply push harder and do what he/she needs to do. Computerized flight control systems have a fundamental limitation in that they can only cope with situations anticipated by their designers. The Airbus engineers did not anticipate landing an A320 in the Hudson River.

If you read what has been published so far, and in particularly the leaked transcript of the cockpit voice recorder (see Air France flight 447 on Wikipedia), it is clear that the primary responsibility belongs with the flight crew but they were hampered more than helped by some of the Airbus flight control system "features".

Bill