Notes on Air France Flight 447: Thoughts on the CVR Facts

George's Point of View

So there you have it: the short version of the investigation’s reading of the Cockpit Voice recorder. If you missed it, we have posted it here in this blog in it’s entirety:

If you don’t like the visual rendition, you can click at the bottom for the .pdf.

The problems seem to begin at 2 h 08 min 07; then at 2 h 10 min 05 autopilot & auto thrust disengages. The pilots note that the speeds do not agree,( which means the speeds are incorrect, and it is an indication that pitot tubes are malfunctioning. Around this time, ACARS sent a PITOT error message, which was not mentioned in the CVR summary.) The PIC (captain) re-enters at 2 h 11 min 40 and it is all downhill from there.

As far as we can tell, everything in the cockpit voice recorder still indicates that the main cause for this crash is Thales defective pitot tubes which froze over and sent incorrect data back. How could anyone make correct decisions without knowing the speed at which the plane was traveling? How would the pilots have discerned when the incoming data was faulty and which of it—if any—was correct?

Based on the pilots’ response to the stall, we can also reiterate points made at the February 24 hearing, where Justice Zimmerman pointed out a lack of training for pilots on how to respond to a catastrophic failure. Shouldn’t pilots (and not just the PIC) be trained in this procedure to the point that the correct corrective response is second nature? The time to try to figure out how to respond is not during the catastrophe, with 228 lives hanging in the balance.

There does not appear to be an emergency procedure from the manufacturer. (This was also noted in the February hearing by Justice Zimmerman.)

It appears that the plane stalled, and that could not be corrected in time to prevent the catastrophe.

So now, all eyes will turn to the DVR, which will hopefully help decode what happened mechanically in the stall.

And I do have questions about the notation, which seems to imply that even if autopilot is not online, some (background?) processes continue to be determined by digital input, which may be faulty.

When the measured speeds are below 60 kt, the measured angle of attack values are considered invalid and are not taken into account by the systems. When they are below 30 kt, the speed values themselves are considered invalid.(Or I am misreading the data and the fact of unrecoverability is due to other system factors. It does appear that the Flight Control System is unwieldy or badly conceived.)

It seems to me as a layman, that this is a fly-by-wire conundrum. If the plane is in crisis, but it is logically disregarding the correct input when it is beyond a “safe or logical” range, then how can it be corrected, if there are no manual controls? (Not to mention no emergency procedures to fall back on.)

Originally posted by George Hatcher on Friday, May 27th, 2011