Originally Posted by
WideScreen
It's always easy to complain about insufficient URT, though when checking the diagram figures, I don't think, their UR is that bad:
- Within 2 seconds after the AP-off alarm fired, they did have the control wheel back in the position, corresponding to the pre-AP-off position (though that position was already granny compared to what was needed).
- Within subsequent 3 seconds, they were already upside-down in a 360 roll, so, control wheel full to the other side to complete, iso trying to stop the roll.
- Another 3 seconds later, they had wings "somewhat" level.
- Catching the Nr1 still at full throttle AND correcting was 12 seconds after AP-off alarm fired. Late, though given the roll going on and bringing wings somewhat level, let's give them the benefit of the doubt. We should not forget, it's probably not a coordinated roll, so gravity and wild 3D accelerations will throw around people and stuff.
- Looking at the last 1.5 seconds, the Ground speed and Computed Airspeed largely overlap, implying a far from vertical speed-vector, or better, nearly near horizontal.
- Looking at the VS vs. PA, the VS reached nearly 0 at the end, so they nearly made it to horizontal flight, before hitting the water.
- There were 24 seconds between AP-off alarm and hitting the water......
All in all, it's killing they didn't notice the throttle position discrepancy, though the recovery wasn't that bad, I'd say. They just did have the bad luck of just running out of height.
Really shocking, that an AP-off can in just 2 seconds develop into a such an irrecoverable and catastrophic situation.
Not to say, this accident clearly proves, the AP should not "let go" when reaching the end of its control wheel authority, but "freeze" (of course, with alarm), until manually turned off, when the pilots are hands-on. THIS "AP let go" is what caused the Upset to develop (rapidly), not the pilots themselves. Shocking.
Feel free to correct me.
The crew had asymmetric thrust for approximately 90 seconds, not because an engine had sneakily failed, but because the lever for that engine was not in an eye pleasing location. That places a yaw on the aircraft that the B737 AP can only counter with aileron. The Thrust lever is in the view of a crew that are not asleep. The control column is in view of the crew assuming they aren't doing a crossword puzzle or watching movies on an iPad. The left yaw puts a force on the drivers... pushing the pilot sitting in the left seat (as surely in command is a poor description) head to the right, towards the thrust levers. The co-seat occupier has his right ear being forced towards the #3 window. and there is no interest in the proceedings?
The AP gives up when it gets fed up with the lack of supervision. It cannot and must not freeze in position, it must decouple. Indeed, that gives a removal of the sneaky perfidious autopilots roll countering input, and that results in a rapid roll onset due to the asymmetry that has sat there for... a minute and a half.
A minute and a half, that's about a 5th of a Classic Sudoku, its about 90 seconds of video... at 100m/sec, its 9Km...
Thrust did come off eventually, at VNE. Its supposed to.
The aircraft impacts at around 30 degrees nose down, relatively wings level and about 100 kts faster than VNE. ( the data traces here are lousy scaling of data, the raw data would be interesting, but only from a morbid curiosity POV). At least it was nearly wings level. The crew would not have seen a horizon, apart from it being dark...
Did the crew respond? Lets see, thrust came off at around VNE, and inverted, in a steep dive, with the plane continuing to roll. Aileron inputs were made in both directions and yes, better data would help determine if they actually knew which way was up.
Once wings level, and about 250kts above the trimmed speed... the plane is going to be pitching up due to Static Longitudinal Stability [§25.173(b)] so we want to see the control column input or elevator position to determine what the crew were doing, as the plane would be reducing the dive angle naturally anyway.
From reaching 90 degree bank, which should have got some attention, it takes.... ? 8 seconds, 9 seconds? to reach maximum dive angle, and the plane slowly continued the roll through inverted to upright. As said, a half descent barrel roll, other than going down hill, maybe it was a descending scissor. What it isn't is a timely recovery from what should have been obvious for a minute of a half.
I credit the plane with more effective recovery as part of the §25 certification requirements. The crew were not In command of the plane.
If pilots want to dress up in fancy uniforms, I for one would like them to actually care about their profession. When I fly my own jets today, I no longer dress up, after 27,000 hours I am embarrassed to call myself a member of our profession. When asked what I do for a living, I prefer to tell the curious "I am a honky tonk piano player at a brothel..."
This crew at least should have died relaxed. The passengers were probably less so.
The flight path of this was obvious at the time of the event. The precipitating factor was evident about a week later. My frustration is that the data, such that it is, (please get expanded scale of the performance parameters, and add elevator and stab... ) is that it took so long and was so blindingly evident that it was amiss, and the crew snoozed through it all. 90 seconds of SA Type I error... with multiple clues, visual, physical...
