Replies

I've always been amazed at what careful analysis of the CVR yields. The analysis of the sounds are sometimes as important, if not more important than the actual words. There was once a PBS program on just how much care they take in analyzing every sound on the tape to ensure the entire picture is understood by the investigators.

For example, the AA1420 crash in Little Rock, analysis of the CVR indicates that the speedbrake arm checklist item was never completed.

http://www.ntsb.gov/publictn/2001/AAR0102.pdf (3.1 MB PDF)

From Page 131 and 132 of the report (incredible what investigators can discern from stuff most of us won't even notice):

The CVR transcript identified the sound of two “thuds” similar to the sound of an airplane touching down on a runway along with a “squeak” sound at 2350:20.2, 2 seconds before the first officer stated, “we’re down.” Analysis of the CVR sound spectrum for the accident airplane indicated that the first “thud” was main landing gear touchdown and that the second “thud” was nose gear touchdown. Safety Board investigators determined that the squeak sound, which occurred in between the two “thud” sounds, was the autospoiler actuator operating. Because of the placement of the squeak sound, Board investigators further determined that the actuator was triggered at main landing gear touchdown by the wheel spin-up transducers.
The Safety Board compared the CVR sound spectrum for the accident airplane with those for other CVR recordings obtained during the Palm Springs incident and the flight and ground tests that were conducted as part of the investigations of this accident and the Palm Springs incident.195 According to the CVR sound spectrum study, the sound durations associated with the autospoiler actuator extension for the Little Rock and Palm Springs flights (0.08 and 0.06 second, respectively) were consistent with the ground test in which the spoiler handle was unarmed (0.06 second). The study also indicated that, for the two flight tests and two ground tests involving normal autospoiler operation (that is, when the spoiler handle was armed before touchdown and the spoilers fully deployed after touchdown), the sounds associated with the activation of the autospoiler actuator lasted at least two times longer for the flight and ground test airplane recordings (between 0.16 and 0.19 second) than the sounds for the Little Rock airplane recording.
The ground tests conducted on the Palm Springs airplane also indicated that, when the left throttle was at least 1 3/8 inches above idle and the autospoiler actuator operated, the spoiler handle would fully extend (by means of the autospoiler actuator), be knocked down, and retract about 0.5 second later to the stowed position. Because the two spoiler positions—the left outboard and right inboard—are recorded alternately every 0.5 second with about 0.25 second in between recordings, it is likely that an autoretract event would be captured by an FDR. In fact, the ground test FDR captured all of the autoretract events, and the two spoiler autoretract events reported by American (see section 1.6.2.2) were captured on FDR data. However, no indication of an autoretract event was found on the flight 1420 FDR data.
The CVR sound spectrum study, physical evidence, and testing and teardown results indicate that the autospoiler actuator on the flight 1420 airplane operated properly at touchdown and that the spoiler handle was in the unarmed position when the autospoiler actuator extended. Accordingly, the Safety Board concludes that the autospoiler system operated properly and that the spoilers did not automatically.

According to Le Figaro, first results from the FDR or the CVR analysis shows that Airbus has no responsibility.... The A330 did not fail and is not responsible for the crash.

http://www.lefigaro.fr/flash-actu/20...de-cause-par-les-boites-noires.php

Same article cites sources close to the investigation as saying that new elements will be made public tomorrow regarding Air France and the crew's possible responsibility.

It's impossible for the BEA to have already processed the FDR data and reconstructed a possible scenario. Either it's a fake information leak, or they listened to something "very big" on the CVR.

Of all the scenarii that have been discussed, one IMO sticks out: There was only one pilot at the controls., one who in the noise of the alarms and the gongs going off missed the cockjpit calls and left the other two pilots locked out of the flight deck. Only something that big could have been determined so quickly.

And for the moment, i'm feeling in my stomach pit like I did when the FR report was: "Guys, the shuttle is scheduled to land at this very moment. We have a problem" (they'd not heard from it for almost 1/2 hour by that time and the runway and skies were empty).