Originally Posted by
rog747
Not sure I buy the bird taking out the AoA vane shortly after lift off...did they find a dead bird or the vane on or near the runway at ADD?- in this day and age of forensics can these items be found?
I agree that loss of a vane is questionable for 2 reasons:-
1) The idea that there are 2 (or 3) different AoA sensor faults were experienced on the 2 Lion Air flights and ET302 flight seems to me to be very unlikely from a probabilistic point of view, especially so as the AoA sensor is apparently generally very reliable.
2) The lack of noise on the high value of AoA of a flat-line value of 74.5deg. Noise is seen on both R & L AoA right up to take off, and stays at about the same amplitude on R AoA throughout the flight, suggesting it is not caused just by air speed buffeting of the vane. Instead the noise seems to start just as the engines reached 94%, so I suspect the noise is more air-frame vibration related. I find it hard to believe that a counter weight (without a vane) would stay exactly on 74.5 for minutes and not now be affected by vibration plus all of the other changes to flight dynamics. A fixed offset, sometimes tracking the R AoA value, sometimes not, looks to me like a software generated value.
I think the “lost vane” idea came from an observation on ET 302’s FDR Data chart that Vertical Acceleration lines up with a jump in AoA. It is possible that “cause and effect” may have been misinterpreted here because Vertical Acceleration is a consequence of, and directly related to Pitch Rate. This would fit in with the fact that AoA L was disrupted on the 2 Lion Air flights (as well as on ET 302) just around take off when there would have been a large value for pitch rate.
This could well be relevant as one of the AoA correction factors is Pitch Rate, according to Fig 9 in Boeing’s aero_12 magazine (
Figure 9. AOA Measurement Errors)
As I mentioned in an earlier post (#2744), I think an error in the correction calculating software may be the source of the AoA corruption problem. Earlier discussions have suggested that the AoA sensor itself must be the single source of a bad analogue value as both the SMYD (for Stick Shaker on/off determination) and the ADIRU receive bad data. However, either the SMYD uses the corrected AoA from the FCC (via ARINC 429) or it computes the same correction (possibly using an exact copy of the software) as the ADIRU. If the SMYD does its own computation of AOA, then both SMYD and the ADIRU took take a good analogue signal, apply an incorrect correction to produce garbage for AoA. Obviously there has to be a common failure that triggers both sets of computations to produce garbage for AoA. As the fault is always on the Left, the common failure could well be the loss of the signal that tells the L SMYD and the L ADIRU that they are Left, resulting in both of them using the correction algorithm for a Right AoA sensor on data from the Left AoA sensor and computing garbage values.