In some defense to the PF, the PNF and the Captain apparently had SA issues as well.

PNF briefing to the Captain when he returns:

"What's happening? I don't know I don't know what's happening"

Later:

PNF: What do you think ?what do you think ? what do we have to do?

Captain: I don't know we're going down

Later:

Captain: It's not possible

Given that you're now using phrases like "the aircraft 'died'" I suspect that would be a "yes", then. I still think it incredibly unlikely, and hopefully the final report will include a metallurgical analysis of the THS jackscrew assembly that will determine whether any fractures are consistent with ground impact or otherwise (I suspect that is the reason the assembly is one of the things they raised).

I also suspect the only way it could be proven to your satisfaction with the information we have at this point would have been if the PF had made an input sufficient to move the THS in the opposite direction and not got a response, which he did not do. The traces of what you describe as "making mayonaise" prove it.

@thermalsniffer - Sure thing. I was not intending to single out the PF for censure over and above the other crew members, just stating what could be a possibility based on the information we have. That said, the PNF queries the PF's actions several times ("Why are you going up?" "Descend, descend descend!"), but never seems to reach the point at which he has had enough and takes control. By the time the captain reaches the flight deck, they are past the pont of no return in any case, but it does appear that he did work out what had happened seconds before impact ("No, no, no don't pull back up").

Not the mayonnaise end of the traces, the other end. Distinctly show the elevators going NOSE DOWN, The THS trace remaining flat.

The "aircraft died" I used after seeing "PULLED UP like a madman"....



I'll be more careful. Hyperbole is my enemy. Especially amongst the true tracking adherents!

I admit to being a merely SLF, but I think you guys are making a meal out of this. I do not see how it is so difficult, given the evidence now presented, to come to some fairly straightforward conculsions.

First, it is clear that the investigation has been well handled and no expense has been spared in gathering evidence in order to understand what happened.

There is no evidence that BEA or any other party is manipulating the evidence to promote, or protect, any other party.

It is clear that in a modern cockpit, when things go wrong, the independent software streams are capable of presenting the human beings with an cachophony of instructions, advice and warnings.

There is no evidence of unwarranted presence on the flight deeck.

Training, and modern concepts of technology, give a basic belief in the security of systems.

It is easier to explain, by way of repetition, the capabilities of a system designed by humans than it is to explain the workings of the physical universe.

If circumstances project you into the gap between your training and reality, you need a more fundamental understanding than your training provided.

A reflex action
 

1/ " The statistical data shows that, when confronted by a stall, in 80% of cases, pilots pull back the control column, in a sort of reflex movement, which continues the loss of control."

you can read this in the REPORT on the incident on 24 September 1994 during approach to Orly (94) to the Airbus TAROM

http://www.skybrary.aero/bookshelf/books/680.pdf

It means that 80% of the crews in the sky couldn't recover from a stall except by chance.

The reaction to a stall warning should become a reflex => weekly practice.

Therefore, for the industry, it is safer to prevent the pilot from being out of the flight enveloppe than to train the pilot to recover, what would need a continuous training as some have proposed here.

This accident is only within the statistic.:{

2/ Don't forget that neither the PNF nor the Capt. ever knew that the PF made an initial huge nose up => how many of us could have imagine that ?

In the TAROM incident, the crew could identify that a human action made the stall.

In AF447, the PF never confess his mistake and it is logical that the PNF and the Captain were looking for any cause except a human mistake

Am I right in saying that on this type if inputs are made on both sidesticks the system will take the algebraic sum of those inputs?

In which case if one pilot was maintaining full and up and the other full down this would equal neutral - not what you need for stall recovery.

At which time exactly go the elevators ND? I see two instances where they reach -15 (2:12:45 and just before 2:14), but never 0 (after 2:11:00) or ND (>0).

That could equally easily be explained by the lack of time for the nose-down inputs to take effect. Looking at the traces (if I understand them correctly), the elevators never return to neutral completely, but begin to move towards "nose-down" twice - once in response to the PF's brief nose-down input at about 2:13:32, and again when the PNF begins pushing nose-down at about 2:13:45 (which correlates with the CVR). This begins with a full-authority push, but gradually tends back towards neutral over the next 15 seconds ("I'm pitching up?") - not enough to move the THS.

During this time the PF returns his stick to neutral, but then begins pulling again at 2:13:52 or thereabouts. At 2:14:05 he makes the statement "We should be pulling, we're at 4,000 ft", at which point the PNF gives up pushing and starts pulling - by this point they are seconds from sea level.

At no point is there a commeasurate nose-down input capable of counteracting the full-authority nose-up command made by the PF at around 2:11:40 (or thereabouts) and subsequently held for over a minute.

Thanks, Doze. I sense some hesitancy there, just a bit. I see the traces as elevators commanding NOSE DOWN, and the THS not moving even a bump off the dead bottom (Closed, NU). Now you say that this "relaxation of NOSE UP" is insufficient to pry the slab off its perch. How about this. The Pilot does not know the THS is NU max. His "Nose Down", when THS is planted, may actually cause the THS to nestle even further into its stop.

Aerodynamically. Now the sensor does not do "Aerodynamically", so shouldn't the slab start to inch its way back ND? Do the elevators have to reach past 0 degrees (mechanically) to encourage the THS to migrate? Because moving the elevators "Less NOSE UP" does not move it?

If this is so, I can see where the case can be made that the THS caused the crash, with some help from a PF who was unaware of its importance in its position working against him. His "Feedback", via his instruments, would tell him "Up is working fine". "Nose Down is not working all that well", with overspeed on his mind, he's thinking (wrongly) the aircraft is aiding the "recovery from OS". Nose UP is what he wants, and the a/c seems to agree, and he is taught to trust the a/c. (Not that there is anything wrong with that).

BTW, and I'm sure you have thought of this: In a too fast condition, one instrument will tell the truth the same way as in a deep STALL, the VSI.

High rate of descent, on the panel, in his belly, ("We have crazy speed, Non?" etc.). Just another hole in the stinky cheese? And another reason to hold "Back stick for a long while".

The points I am trying to make are not substantial, they are made in an inquisitive way, and not to irritate. We may be left forever with a non registration of #2 Panel, and the pilot making use of it locked in a seemingly personal battle with demons he imagines, or may actually see in evidence on his screen. I have suspicioned Overspeed, or its mimic, or its position in PF's thought process, for some time. His initial input, repeated at the last with one identical, is troubling. For you?

I am exceedingly curious about these three Pilots. There is much complication on this flight deck, and soon the engineers will be satisfied.

The upshot of this accident is neither simple, nor is it in any way suggestive of progress. Yet.

Hi,

Maybe for the sake of safety and ease the investigations and find origin of faults .. it will be better to install cameras and OVR in the offices of the decisions makers and beans counters ...... :8

@DW
 

@DW

I have not held in any post in this thread that the stall warning caused this accident. I have held that (1) the logic of the stall warning system is flawed (2) that this accident illustrates the nature of those flaws (3) that these flaws are one possible explanation for the pilots behavior doing one specific phase of the accident (4) that the professionals who designed the system should be held accountable for those flaws to the extent they played any role in this accident (a point which I believe has yet to be conclusively determined because the final report is not out yet).

That is all I have held and anyone who asserts anything else fails at reading comprehension.

Re stall warning inhibit. You do not know. I disagree.

Look very closely at the THS trace (as in zoom in to at least 300%). Unfortunately the PDF render is in too low a resolution to see much more than a couple of pixels difference, but there is movement from the THS actuator trace towards the end of those nose-down inputs. Unfortunately in both cases, the nose-down input was not applied long enough to make a significant difference.

I think an overspeed warning would require agreement from all speed sources, and therefore would have been annunciated to the PNF at the same time. The use of emotive words like "madman" and "demons" has come from you, not me. I'm trying to remain in the realms of dry language and logic. :)

[EDIT : I haven't backtracked at all - I merely gratefully accepted a correction on the part of why the initial stall warning events happened. I also know that "madman" originated from a post on the other thread, but I certainly never used it. ]

@MountainBear - I agree that the logic needs revisiting, but to say that the stall warning inhibit may have contributed to the pilots actions is to go directly in the face of the FDR traces. Stall warning starts at the apogee of the climb and persists for almost a minute. The correct response to stall warning is to lower the nose until the stall warning goes away and then level out once stable. The *trained* response to stall warning was to maintain a level attitude and increase thrust. The PF's response to the stall warning was to continue holding the stick halfway back (with two blips of nose-down lasting less than a second each), and then, 30 seconds into the stall warning holds the stick in the full nose-up position for more than 30 seconds, then returns to hovering around the halfway back position. At *no* point is there any sidestick input which correlates to the end of the stall warning at 2:12:57.

NO. Madman is not mine, and demon is in my view a descriptive worthwhile. I explained this before, you missed it?

It goes to style, and I am working against an old habit of vocabulary.

Can you relent ever?

As above, and before, you have had to backtrack. The STALL WARN inhibit may have played a crucial role in the accident.

Even with the VS, I did not use Bold! No grades, no promotions, no degrees. This is conjecture and informed discussion, in good humour. I take my lumps with the rest, as I get carried away also.

From a legitimate investigatory standpoint, there is not one creature present who exhibits credentials sufficient to convince me that this thread is take it to the BANK........ I try not to take my self too seriously. Forget my threadbare credentials. They are evident, I am sure. This is how I do not take personally the insults and derision. Water off a duck's back.

Why did the autopilot disconnected ?
 

As a result, (TAROM accident see my post above) the Bureau Enquêtes-Accidents recommends:

that a study be launched so that the pilot’s priority over all Automatic Flight Systems is
maintained in all circumstances.

This could be done :

a) by the disconnection of Automatic Flight Systems (automatic pilot and auto-throttle
lever or auto thrust) in the event of conflict between the pilot’s actions and those of the
Automatic Flight System or Flight Director.

b) and/or by clear information in the cockpit (possibly an alarm) warning the flight crew of
such a conflict.

The stall started with the nose up by the PF.

Did the PF started this nose up BEFORE the autopilot disconnected ?

reflex
 

JJFFC

Ok I understand an incorrect reflex action by 80%.
However this incident went way beyond reflex with 4 minutes of stick back, with PF saying he has been sticking back is not a reflex. It is some evidence of a conscious action i.e. observe, analyse, take action and communicate. Also 2 others in the cockpit also agree with the nose up attitude and resort to focussing on keeping the wings level.
Under your hypthesis we say they correctly identified stall, did a bad reflex for 4 minutes. All 3 failed to communicate stall. In fact they all appear to say they have no idea what is happening.
IMHO the facts as they stand do not back this hypothesis up.

Or was it the "lurching elbow" hitting the ss when he turned back to the front after chattin his wife? Do not say this is impossible.

This would be typical of one of the TV network's daytime dramas,e.g., "Soaps".:suspect:

Best to forget the drama and stick to the technical stuff.

The THS didn't break...;)

@xcitation : It was at a much lower level, but remember ColganAir, where the reaction to a Stall Warning was a pull-up to the stops that was held for the duration of the subsequent stall, spiral dive and crash.

OK, so here's the best I can manage given the limited PDF resolution :

http://i1088.photobucket.com/albums/...dr-munge-2.png

The only point at which I see a possible positive correlation between stall warning stopping and relaxing of nose-down input is at approximately 2:12:35, but this doesn't then explain why he holds the stick around neutral when the stall warning returns a 2:12:40 and holds it *back* when it returns at 2:12:50.

One of the things I noticed as I was munging this graphic is that the traces on pages 110 and 111 are not scaled t oprecisely the same size, so I had to re-scale the graphics to match up. Be very careful when flicking between the pages and trying to draw correlations.