EMIT

The problem that I see is in regarding the Flight Director as a COMMAND (as in a command that has to be followed blindly).
I command the aircraft to go where I want it to go, and if I have properly managed my autoflight system, that system will suggest attitudes to me that will be pretty close to, and possibly a refinement of, what I am already doing.

Any other approach to flight guidance is rubbish.

Vessbot

1. You're substituting your desire for how the FD should be followed, for the reality of how it is in far too many cases (including, almost certainly, this one).

2. With any level of intensity of following the FD, ranging from blind slavishness to skeptical dismissal in favor of the raw data behind it, is there any benefit to it coming on at VS 6.0 over staying off? If yes, what is it?

Concours77

EMIT:
"True, the speed indication was weird for a short while, the altitude indication dipped a bit at the very beginning, but the ATTITUDE INDICATION, that is THE MOST IMPORTANT CONTROL INSTRUMENT in INSTRUMENT FLYING, was indicating with perfect precision what was happening."

The speeds were lost for sixty seconds. ATTITUDE? How would you know? From the report? PF PFD was not recorded, and he was flying. "We have NO INDICATIONS..." Until that comment is understood, making statements about panel are not reliable.

I think PF was relying on his FD, which was int. I also believe, and from early on, believed his inputs were mimics of what he saw on FD.

Without CVR data about PITCH discussions, we rely on the results of same. It is likely this was done because it is rude to point out what a fellow pilot is doing with his stick, rather the result can be discussed.... culture.

At Captain's re entrance, he had just clambered up a deck that was at fifteen degrees inclined Nose Up. A check for your "explicit ATTITUDE data" would have informed him they were PITCH/in the WEEDS.

"...Note, that by zooming up untill all speed was bled off, the climb automatically was turned into a descent (GO DOWN), but the way in which they went down was not a wise one..."

The beginning of descent could not have been associated with STALL. They transitioned from ascent to descent with no appreciable change in ATTITUDE?

Having no discernible aero cues to show STALL, PF assumed instead they had started accelerating. With PITCH at ten degrees, that must have been very confusing....

He became obsessed with noise, and a descent that made no sense. The other two did not ever consider STALL

If as you say the ATTITUDE was accurate, and should be employed in flight, you have isolated a disturbing circumstance where the Airplane was reinforcing there was no STALL, and arresting descent should be done by commanding NOSE UP....instead of a workable plan to regain aerodynamic flight....no?

Vessbot

You're not reading him right. His first sentence is "The problem that I see is in [what the problem is]"

Concours77

Apologies. I thought you were promoting the Use of the FD in their situation?
There is a way to compute AoA with the FD? Also, do you think he may have been trying to project the Flight Path Vector for some help?

Vessbot

You deleted the post you're apologizing over, so now mine looks like nonsense to those who didn't read yours.

_Phoenix

I think this is the only rational explication for the zoom climb, PF was following the FD cross bars.
Also, maybe the seat was not properly adjusted, at the moment of AP disconnect. The arm rest was recovered in the up position. In the simulation below, the perceived neutral position of the sidestick seems to be at 2/3 of the total travel, instead of 1/2, visible in several occasions, where the lateral input goes from full left to full right and reverse.

The FD indication was exactly as per PF preference expressed (earlier in CVR), to climb above the weather.
The stall warning was disregarded since it was in conflict with two indications. FD -> climb and variometer showed steady climb of 1000-2000 ft/min (a stalled wing doesn't sustain prolong climb).

That fast they got into coffin corner with THS at max NoseUp. At that point they had 3 puzzles to solve:
1 - to recognize the deep stall and the aerodynamic noise of the vertical fall.
2 - to recognize the excessive AOA and THS at max NU.
3 - to reduce the AOA by all means: prolong input of the sidestick, NoseDown, THR at idle, manual trim for ND.
They solved only the puzzle no. 1, partially.
https://www.youtube.com/watch?v=BR5kFOHVnUU

Concours77

So if they had the flight path vector alive, (the bird) does it point down to the sea? Which cue to follow, avoid?

With GPWS ordering "PULLUP", does the computer reject the actual attitude?

The GPWS commands the wrong move.

"DO NOT PULL UP! ASCEND" What's a computer to do?

GPWS: One trick pony. Not that it mattered.

Why do we suppose the Captain declared "PITCH ten Degrees..." 1.6 seconds before impact?

We have been told the attitude of the aircraft was accurate and available to the pilots.
From the FD?

If, after sixty seconds the speeds were accurate, (only thirty seconds prior to forever STALL), and the attitude was displayed, the FDs were active and correct, what is the predicament of the Pilot Flying?

250 knots, high (10-15 degrees NU) Pitch, heading 1 1/2 degrees per second turning right, And conceivably the FPV pointing to the floor, and high noise level.

What would you do? We determine that the pilots do not accept the STALLWARN as valid, so perhaps for the first time, can we forego the "...Get the GD Nose Down, and hold it down!..."

The 330 needed handling at loss of AP, fact. The Roll Axis was a surprise and a challenge to PF. fact.
The neutral point of the stick may have been displaced aft. fact. Max THS NU. fact. The roll remained split, throughout, fact.

With respect, what do you do?

EMIT

Vesbot,

Answer to your post 1520

You asked whether it is useful that a Flight Director comes back into view (is displayed again) after it has been removed from view because of unreliable data?

You dismissed my post 1522 as being not an appropriate answer to your 1520.

You have to realize that no amount of clever engineering can protect an aircraft against pilot actions like Mr. Bonin's.

If an F/D would remain removed from display, even though valid data would be available again, and would require the F/D switch to be moved from ON to OFF to ON in order to get the F/D back on display, then NO DOUBT in the future you will get a crash because another F/D slave will not be able to fly the aircraft BECAUSE HE HAS NO F/D TO FOLLOW (HE will be under the impression that when an F/D switch is ON, there SHOULD BE an F/D in view and he will never reach the point that he cycles the ON/OFF switch). After that future crash, PPRUNE will be full with posts why the hell those stupid aircraft builders don't let the F/D return to display (in accordance with the ON status of the switch) once the applicable data have become valid again.

So, if the AF447 Flight Director did come back ON in the V/S mode at the momentary vertical speed of 6.000 ft/min, that was in itself not particularly useful, but the proper way of using the system would have been, to announce (and so acknowledge) the flight mode annunciation of VERTICAL SPEED, to check the present rate as displayed in the V/S window on the Flight Control Unit, to conclude that a 6.000 ft/min climb was not what was desired, to adjust the value in the FCU V/S window to a more sensible value and to (only then) follow the flight director. In that way, the return of the F/D display could have been used in a useful way. Of course, pilots that would have been that clever, would never have pulled up the nose to 15 degrees above the horizon to begin with.

So, in reality, the answer to your question is not a simple Yes, useful or No, not useful. The usefulness has to be embedded in a proper way of using the system, and, sorry to say, the AF447 crew did not use their A-330 aircraft in a proper way, they f****d up big time.

For that crew, the most useful piece of equipment on the flight deck would have been a newspaper, in which they should have kept their noses deeply dug-in, until the 1 minute speed unreliability had passed. After that, all they had needed to do was to switch the autopilot back on.

EMIT

[QUOTE=Concours77;9836181]EMIT:
"The speeds were lost for sixty seconds. ATTITUDE? How would you know? From the report? PF PFD was not recorded, and he was flying. "We have NO INDICATIONS..."

I think PF was relying on his FD, which was int. I also believe, and from early on, believed his inputs were mimics of what he saw on FD."


If as you say the ATTITUDE was accurate, and should be employed in flight, you have isolated a disturbing circumstance where the Airplane was reinforcing there was no STALL, and arresting descent should be done by commanding NOSE UP....instead of a workable plan to regain aerodynamic flight....no?[/UNQUOTE]

Concourse77,

You are contradicting yourself - you assert that, according to an unprofessional, non-informational term about "screens" gone "crazy", that there must have been no attitude indication, but on the other hand, you assert that mr. Bonin must have been following Flight Director commands.
Let me tell you that the F/D is displayed right on the very screen that displays the attitude (the whole display is called the PFD, Pilots Flight Display). That PFD is such an important display, that should the Display Unit (that is, the display screen itself) fail, the PFD (the picture that is, the information) will automatically transfer to the other Display Unit, which normally displays the ND (Navigation Display).
And should, for whatever reason, all Display Units fail, go black, then there is ATTITUDE available on the completely independent, autonoumous standby attitude instrument (I use such a generic term because there are different forms of backup instrumentation available, however, the overriding quality of all those backup systems is that they will function on the most basic form of power supply and are completely independent of all the main systems).
As for your incredulous reaction to my observations about ATTITUDE being the most important control parameter for flying ... sigh ....
I see your join date on PPRUNE is annunciated as of late 2016 - please start a read-up of all previous posts in the existing 12 threads on the subject and also get better informed on how flying is actually done before you spout off with inconsistent theories.

EMIT

Concours77,

A picture is worth a thousand words, so I will try once:

Following pictures of a PFD of a representative airliner in climb at high level and in level flight.
Note that there is only a hear thicknness difference in pitch attitude between the two situations.
The third picture shows the same sort of display of an Airbus.

wiedehopf

@EMIT

there is a difference between a 6000ft/min climb and 500ft/min climb.

additionally you can have a higher angle of attack when you are not flying in the optimal range adding to the attitude.

EMIT

Wiedehopf,

Yes, I know there is a difference between a 500 ft/min climb and a 6.000 ft/min climb.

The most important difference is the fact that a 500 ft/min climb is just about the maximum that an airliner can MAINTAIN at that altitude and that a 6.000 ft/min climb CANNOT be MAINTAINED by an airliner at that altitude.
NOT BE MAINTAINED means that the airliner will loose speed when performing that kind of climb. It will trade speed for climb performance, or speed for altitude, however you will call it, but it can only be a temporary situation. Once the speed has dropped a considerable amount, you do not have enough speed left to sustain 1 g in level flight, you will have to unload, dive to pick up speed, etectera.

The pictures here are presented to show how close the climb ATTITUDE and the level flight ATTITUDE are and to make clear, in a visual way, for laymen, how absolutely idiotic a 15 degree pitch attitude is for an airliner at high level and how unequivocally clear the difference is on the attitude indicating instrument between circa 3 degrees and 15 degrees nose up.

Vessbot

Nothing is an absolute solution, but that doesn't mean that we shouldn't enact mechanisms that have a partial chance of helping. And it certainly does not mean that we should leave in place mechanisms that have a high chance of leading the crew astray (threats). And this is one.

A valid problem, but one easily solved by having that FD on/off control be a pushbutton instead of a switch. Then, there could be no mismatch between switch position and actual system state.

Well yes, lacking a FD to help a task saturated pilot with valid data would be a problem, agreed. But having a FD sneak an unnoticed mode change and lead a task saturated pilot into a zoom climb and stall, is a much bigger problem. In choosing which problem to design the system to account for (and the design solutions are contradictory), we're forced to choose between two evils. And I dare think that even without the hindsight of this accident, I'd say that the latter is the greater evil.

You list a number of actions that a capable crew should have taken instead of mindlessly following the FD. And I agree with all of them - but this all presumes that they are capable, and that's where the mechanism breaks down: they weren't. Instead, they were overwhelmed and tunnel visioned past the point of being able to deal with this.

The point of a safety system is to catch or back up a crew at their last capable. If its design fails to do this, hammering harder on what the crew should have done misses the point. All those things may be true, but that's about the crew, not about the safety system in question. A different slice of the Swiss cheese being addressed.

The FD behaving the way it did is much more than "not useful" as you label it. It's specifically harmful. It needs to be redesigned, unless this need is outweighed by some opposite need. You touched upon such a possible opposing need, but I think it's not strong enough for the reasons I stated.

The FD design needs to be based on the assumption that it will be followed. And on that principle, it follows that a deadly command should not be displayed in the first place, rather than depend on the task saturated crew to do all this extra task sifting and ordering to fix it.

_Phoenix

EMIT, that’s the imagine portrayed for AF447 accident to mass media, unfortunately it is a convenient finding for some professionals (not only for biased).
AF447 pilots used A-330 aircraft in a proper way, 99.99% of the time, in normal law, which implies a synthetic flying. Flying the automation is one thing, but flying the alternate 2b (pitch normal+lateral in direct) is quite challenging at high altitude, in the night and icy thunderstorm. However, it is manageable for any trained pilot, including PF of AF447. He could fly it properly, if he would have the proper instrumentation and I’m not talking here about airspeed, FD, PFD, etc but about the No 1 instrument installed on the flight deck, that tells you if the wing is flying or not. That stall warning was triggered 20 seconds while the wing was still flying and vice versa was inhibited when they were falling like a stone. I agree with Vessbot, a task saturated pilot might follow the FD only (also a possible a convenient habit). A tunneled chasing of the cross bars might disregard the attitude, until is too late when the PFD is all blue. Yes, 15 degree pitch was idiotic, Bonin put them in the coffin corner with the feet on the wall, but I guess more troubling are the puzzles (listed in my previous post) to be solved in 2 minutes in order to recover an airliner from a stall, with all the sky underneath.

Concours77

"...Yes, 15 degree pitch was idiotic, Bonin put them in the coffin corner with the feet on the wall, but I guess more troubling are the puzzles (listed in my previous post) to be solved in 2 minutes in order to recover an airliner from a stall, with all the sky underneath..."

Then perhaps two "idiots" on the flight deck? Bonin, for consistent up ELEVATOR, and the A330 for full authority AUTOMATIC TRIM FULL NOSE UP?

1. Elevator alone cannot sustain 40 degrees AoA.

2. Elevator alone cannot prevent Nose drop at STALL.

2a STALL with elevator only is standard recovery.

3. Elevator alone cannot make the STALL so stubborn, and flyable in their state.

4. Recovery from inappropriate TRIM is not trained.

5. There are no "Human Data Recorders" available, how convenient, indeed.

"...That stall warning was triggered 20 seconds while the wing was still flying and vice versa was inhibited when they were falling like a stone..."

Perhaps picky, but STALL WARN sounds when wing is still flying. STALL WARN is ALERT of impending STALL. All pilots had ever heard was "Impending Stall", not "In Stall".....?

Without functional AoA indicator, and STALL AoA computed, no definitive conclusion of the wing's status can be had.

No one here posting was present at this accident. Please stop with the "idiot", idiotic", etc.

Turbine D

Nearly, if not all, your questions or poses have been covered in the previous AF447 11 threads, if only you would go back and get educated. What you are doing now is rehashing what has been reviewed extensively...

Concours77

Excuse me. You have something new? The only change possible in the outcome can come from a perspective.

It is likely futile, but there are those here who wish to look differently at the data, and re open what has become closed off. From the outset, there was one poster whose only contribution was limited to endless repetition of narrow minded blame about aft stick.

The conclusion can be amended. Some here are not convinced the cause of aviation is advanced by ignoring the aircraft's design flaws, and focus on the crew.

This one sided and trite approach is in the end costly, and ignorant.

There are no cold cases, only a wish to worship apathy.

Winnerhofer

Round 1: AF 0 Airbus 1
Family tries to go after Airbus but are rejected:
Rio-Paris: une famille déboutée de sa demande en référé contre Airbus - Le Parisien

Centaurus

It should not be forgotten that the original concept behind flight director use was that it was an aid to instrument flying and nothing else.

Unfortunately over the years, that concept has morphed into what many pilots understand the FD to be an absolutely vital ingredient for even basic instrument flying ability. That is despite evidence to show that loss of control has been the occasional result because of pilot blind adherence to following FD indications; even though it was obvious these indications led to over-controlling because pilots were "chasing" the needles.

If mode confusion exists - however temporary, pilots need to be trained to seamlessly revert from FD use to raw data flying,. Until that happens (very unlikely) loss of control will always be possible in todays high automated aircraft. Of course if the pilot lacks raw data instrument flying skills to be able to confidently switch off the FD, then that's another story..