bearfoil

Hi RBF, thanks.

Did you notice when PF mentioned "9000"? One second later he remarks:
"8000". That is 15,000 feet per minute! If he is accurate, (and most pilots are, at rating VSI with altitude) that is ~150 KNOTS vertical speed! Later, we hear "one would hope, as we are at 4000".

I am at a loss to understand their "nonchalance"?

speedbirdconcorde. Thanks for your reply Sir.

RetiredF4

DozyWannabe

could you point me to the indication, that the THS is moving? THS is the blue straight line on the bottom?

But by chance, how should THS move when after 15 sec. of SS ND the elevator peaks max at -15°NU (the magenta line)? Would the THS start moving prior elevators move to the ND range? How long is SS ND and in what magnitude required to get the elevator in a ND position?

ChristiaanJ

Let's not confuse the issue... they did not "stop above a certain AoA", but below a certain measured airspeed (< 60kts), because of the UAS.

What personally bothers me, is that the report does not stress sufficiently that the quoted "airspeeds" are those measured by the ADS and displayed in the cockpit and recorded on the FDR, and NOT the real airspeeds.

RetiredF4

It is the revised procedure (ISSUE DATE: 15 JUN 10). The old procedure was different with emphasize of gaining speed and maintaining altitude.
No sense to repaeat the new revised procedure again, everybody should know that one now, at least i hope so.

DozyWannabe

I'm not saying that the behaviour isn't problematic, I'm saying that I don't think that the return of the stall warnings is unambiguously triggered by the nose-down inputs. The only source that suggests that there was internal disagreement at the BEA is that one "La Tribune" article, and as such until I see some corroboration I'm going to be sceptical.

And if you think I'm being stubborn about press accuracy, you wait until you've spent a little more time on this board!

Again, I didn't say that it doesn't matter (I'm an engineer, I like to find things to fix!), I said I'm not 100% convinced it played a causal factor in this accident

Based on the evidence of a single press article that almost certainly comes from within Air France (which as an entity would benefit financially and in PR terms from Airbus having to shoulder a larger percentage of the responsibility), I'm not buying that until I see some better traces - right now it looks ambiguous to me.

That's fair - however as I said before, this is the only airliner to my knowledge that has been that far outside the envelope for that long, falling from that high - so at present it's not clear whether there is a deficiency in the stall warning design specific to the A330 (and by extension the entire Airbus FBW range), or whether this is something that needs to be examined on an industry-wide scale.

Well, not quite - it was still stalled. If the nose-down inputs had been maintained before passing, say, 15,000ft on the way down then it might have stood a chance of coming out of the stall, but that's not a given. Remember that a stall warning is designed to activate before reaching the stall itself, so once it had picked up speed and the wings were unstalled, the warning would continue for a few seconds until it was out of the stall warning regime. That's not bogus, it's just a factor of the design.

Hypothetically, if they had successfully unstalled the wings, started bringing it back under control, but the extra couple of seconds of stall warning meant the difference between successfully pulling out of the recovery dive and crashing - would you be arguing deficient design on the part of the aircraft? Do you think Air France would?

It matters if you want to think about how to fix it. The guys who designed these systems didn't just slap it all together - everything they did had to fit a very strict set of criteria. There was a reason the logic was designed as it was and as such, any change is going to have to be considered very carefully.

Well, this is the thing - we've had two explanations doing the rounds - one, that the stall warning is inhibited by a software setting and the other is that once the airspeed has fallen that low that the AoA vane no longer functions. No doubt we'll find out if one, the other, both or neither is true in the coming weeks and months.

I don't know if you've noticed yet, but there is a small cabal of pilots, reflected in a number on here who have a kneejerk hatred of anything Airbus. This is in part based on press assertions made since the late '80s that Airbus was a prime mover in "taking pilots out of the loop". Some believe this was a deliberate attempt to deskill the job of airline pilot and undermine their livelihood. This is complete nonsense of course, but the press fed that rumour mill and made it what it is today, which is kind of ironic when one considers the short shrift most pilots give the majority of mainstream media articles on aviation.

I'm not that bothered about people who take the time to understand the issues, but when the press get it wrong, or chase a certain angle to get a juicier story, it pollutes the general public's understanding of the issues at hand, which is why you've still got a lot of people today thinking that the infamous 1988 crash of the A320 was caused because "the computer thought it was landing and overrode the pilot", or that "Airbus makes planes that use computers, Boeing still use good old-fashioned cables and hydraulics in their new designs". This bothers me because regardless of whether related to aviation or not, I can't stand people who should know better spreading misinformation.

Interestingly at the moment it's Air France that's making all the noise - Airbus seems to be keeping their cards very close to the chest at the moment.

It's helpful to stop the "If it ain't Boeing, I ain't going" crowd from crowing.

But ultimately the subject of Stall Warning is a prickly one - as far back as 1972 you have crashes caused in part because the pilots thought they were getting a false stick pusher activation, dumping the stick push and in doing so sealing the aircraft's fate.

And there are equally plenty of major reasons not to, not least of which being what would happen if a back-driven sidestick was wired up in reverse during a critical stage of flight (which has happened). Others include the assistance in enforcing SOPs, crew roles and CRM given by having one person with their hand on the stick at any one time, the extra systems complexity and weight that setting up back-drive entails (the 777's backdrive system by its very nature has more potential points of failure as a result).

It has become a matter of personal preference really.

In Normal Law you are commanding *rate* of movement in the axis rather than deflection. The FCU will do whatever it deems necessary with flight controls and thrust to get you the rate you're commanding. As you progress down through the laws, channels move from rate to deflection. In that case, I don't think the amount of time you hold the stick in a given position changes the deflection, but obviously the longer you hold the stick in position, the more time that deflected surface has to act, same as any other aircraft.

Lonewolf_50

bear, FFS, the transcript isn't parsed to fractions of a second, and he may have noted 8 grand in passing or nine grand in passing and approaching ten grand ... your division presumes a precision in your chosen "data points" that isn't in keeping with the accuracy you wish to assign to your RoD estimate.

If you make the denominator 0.75 seconds or 1.25 seconds to accomodate time to speak and delay between reading and speaking, what numbers do you arrive at? Keep moving your decimal fractions about until you get the numbers you are looking for?

Come on, bear, you know better than that!

Given that the FDR apparently produced some data points and readings around 10,000-11,000 fpm down (closer to the latter), such data is quite enough vertical velocity (or the vector along the negative z axis ) to evoke a grimace on anyone's face who is familiar with flying. There is no need to employ hyperbole and inaccuracy in attempt to craft further drama. There is drama aplenty without resorting to such.

JD-EE

airtren, in precisely what way were the stall warnings when the AoA was absurdly high in any way improper? The cessation of the stall warnings was improper; but, the warnings themselves were very proper. Or am I missing something rather serious?

bearfoil

You misunderstand. Perhaps utterly. My point is that on this flight deck there is just such a disregard for accuracy as you proclaim. So when you are in agreement, but by misunderstanding, it becomes FFS?

Lonewolf it is you who know better. Why announce two impressions of rate, and let it slide as if it matters not? Why announce it so close to the previous one, and let it "slide"? I submit you have a hair trigger when you read my posts. Read it again? In the interests of friendship?

Because we are to assume that the low level and rate are terrifying, and yet remain seemingly unimportant to the crew? Folla?

A33Zab

Triggering a Stall Warning: "STALL STALL" is not the difficult part in the logic.

The challeging part is to trigger a stall warning when it is appropriate to do so
(for the flight condition & A/C Config.) and to prevent the spurious stall warnings.

It requires 25 pages of logic in the FWC ECAM system logic manual.

I'm not here to defend AIB, they can add another ECAM page logic if you like, but would that have made any difference after 54 sec of continuous warning?

This is how the game -to blame- is played.

mm43

CJ;Exactly, ... and if they had done so they would have provided the reason why they are different.

Lonewolf_50

bear, as a friend: please go back to the alleged mathematical operation you presented to us, and your choice of numerator and denominator, and consider how you derived those data points. Compare that sourcing to the FDR which BEA roots its estimates and data from.

Then consider the the FDR analysis came up with a RoD between 10,000 and 11,000 FPM (I suspect it varied early on, and stabilized (per Grity and henra's posts a few threads ago) as the bird descended into denser air and friction came into play, and an approach to terminal velocity. I am within 10% in that error band. I suppose I could go and find the 10,098 or 10,988, or whichever figure is cited as ultimate RoD, and please my pedantic little old self. No drama. henra I thin it was pointed out that the RoD would vary based upon air density, a couple of threads ago.

To arbitrarily add a 40% bonus for dramatic effect on dubious arithmetic grounds doesn't help in understanding.

So, my tone might have been edgy, but I know you've been around long enough to know how number manipulation works. (Heh, you have even complained about such things in various threads ... )

HazelNuts39

Let's put this into perspective. Certification requirements, airplane systems design and airline pilot's training primarily seek to avoid stalling the airplane. Limiting myself to system design, there are first of all the various high-AoA protections in normal law. In the fairly remote failure condition of alternate law the only protection left is stall warning. Its function is to warn the pilot that AoA is exceeding a threshold close to AoA-stall, but sufficiently below it to allow the pilot to avoid stalling. At high Mach it is set to stay below the onset of buffet, which provides additional margin to the stall, and its function is amplified by the occurrence of aerodynamic buffet. In the words of the regulation (which doesn't apply in abnormal systems configurations) stall warning must continue until the AoA is reduced below the threshold at which it is triggered. It would seem that in almost all cases the system will do just that. The only case where it will not function as intended is when the stall warning is ignored, and nose-up commands are maintained, causing the AoA to increase to values (>40 degrees) at which the sensed airspeed is so corrupted that the AoA signal itself is considered unreliable or invalid.

P.S.
Ice detectors detect the formation of ice by freezing of liquid water, which in all probability was not present in the conditions of AF447. They work in cycles: unheated while collecting ice, heated to shed the ice. Ice particles do not adhere to non-heated surfaces, and are therefore not detected.

JD-EE

DozyWannabe - on page 114 look down to the 8th set of graphical elements, dark green titled "ASSIETTE (>D- A cabrer)[DA]". What is it?

When I try to translate it this seems to be an aircraft pitch trace. If so, at many times the PF had full nose up on the elevator the plane had a significant, 10 degrees or more, nose down attitude.

So I'd like to know what that line's title really means.

Please!

A33Zab

If you want to trace and compare, don't forget to take the accelleration into account.
The - accelleration normale* - is the feedback (besides elevator servo deflection) to the FBW pitch channel.

* well not completely sure, awaiting english version.

DozyWannabe

You need to be zoomed right in, but you can see movement in the THS line after an elevator movement of any size. Unfortunately, due to the low resolution of the image in the PDF it only shows as a series of 1-pixel notches in the line, but there is definitely a correlation between elevator activity and those notches at that point. Earlier in the sequence , those notches appear just prior to movement of the THS in reaction to elevator movement. Even a small amont of elevator movement seems to trigger them, but the elevator commands need to be held for a significant amount of time before the THS will respond in anything other than increments which are barely perceptible at the resolution in which the graphs are presented.

In all honesty I don't know, but if you look on page 111 of the French report, you can see the sidestick traces, directly below which is the elevator response, and directly below that is the THS, which gives a nice representation of how the autotrim works. Notably, the big THS movement happens during a time in which some quick "blips" of nose-down are made, but the overriding trend from the sidestick and elevators is for nose-up, and so the THS follows that.

Well, this is an *interim* report after all, and in the FDR traces, below the indicated airspeed (both "conventionelle" and "I.S.I.S"), they have included the ground speed trace, which gives a better picture of what the actual speed would have been.

@JD-EE - I'm pretty sure that "Assiette" can refer to pitch or trim, but in this case I think it means pitch. If it is a valid trace then yes, the nose fell down, but I suspect not in a controlled manner. In any case, the THS position, along with that of the elevators meant that as soon as the nose was down it would immediately come back up due to aerodynamic forces.

mm43

JD-EE;

ASSIETTE (>D- A cabrer)[DA] = Attitude trace, and you are right regarding the excursions into ND. Also check the Stall Warning trace below!

bearfoil

What is bear getting at? Merely this. The SW was alive for ~54 seconds, seemingly without effect re: crew actions. The loss of altitude was grave, the VSI reads were alarming, and yet there is a definite disregard for hard data to effect a recovery FROM STALL. The focus may not have been on STALL. Because we know what happened ( yet continue to gripe and argue about the hard data!), does NOT mean that is what the crew were addressing as their immediate challenge. The CAPTAIN is unsure about whether to PUSH or PULL!



Now all this is from a redacted and selective release of conversation absent tone, inflection, and strict sequence. The only explanation relative to 447's demise based on BEA has to do with how moronic and clumsy these gents acted. I do not for a second believe that to be the case. Instead, I choose the other option, one that makes more sense based on BEA and hopefully their honest representations of the Data, and not an attempt to slander the crew. A DIVE. TOO MUCH SPEED. A reliance on the g protections available in ALTERNATE LAW 2. The only thing they could not get is why she refused to climb when commanded, screw the STALL HORN.

rant, fading into........

@Lonewolf. Finally, It matters less than a gnat's whisker whether you 'get' or I 'get'. Our disagreement is nada. What matters is what THEY thought. I am not seeking drama, I am asking why would the crew speak twice of Altitude, than finally, "hoping a Pull Up, we are at 4000". ? I'm trying to wear their moccasins, in an effort to understand: Why the "Nonchalance"?

What matters is what they think of what THEY spoke! "@9000......@8000"

one second after the other. Does one of them not 'get' the VSI, and someone is trying to REMIND? At 8000, than 4000 feet?

To me, and it is a GUESS, it means they are worried about Overspeed, which means at any split second they can start a successful recovery, even at 4000 feet. No NEED to reload the wings, spend 10000 feet and then have to worry about tearing the wings OFF.

That is why, In my opinion, they were not speaking STALL!

As in: Captain : "N'est-ce pas possible!!" "****in A, we're STALLED"

JD-EE

Lone - "I empathize."

To say the least. Those guys are under incredible pressure. And based on what I see here with people apparently missing important clues they must be working VERY hard.

GarageYears

PEI_3721:
I am quite familiar with both stick-shaker and stick-pusher, however I suspect you are not really accepting that the system designers have generally NOT considered the region deep into a stall. Certification tests are carried out up to and approaching a stall, but never stalled. Trying that in a large transport category aircraft is not recommended or tested (except perhaps using computer modeling).

The stall warning systems currently on the aircraft are therefore designed as preventative devices - i.e. if you are hearing this, then things are starting to go wrong, so perform the appropriate actions.

Consider whether, in the normal course of aircraft operation, it seems likely that an aircraft would be *flying* with an IAS of 60knots and a stall-value AoA? Such a realm is easy for the armchair quarter-back to criticize... "why shut off the stall warning when the IAS drops below 60 knots" ---> because such a condition was/is inconceivable for 'normal aircraft operations'.

But my suggestion a page or so posts back now, was to introduce a new warning, not just an aural, but a bloody great block of text flashed across all glass displays in the cockpit every 5 seconds or so, suppressing all other warnings that may or may not be going off. "AIRCRAFT STALLED, AIRCRAFT STALLED, REDUCE AoA!"

As some other contributor noted, it's not very helpful hearing a "pull-up, pull-up" warning if you're fighting a stalled aircraft - if you're stalled with an AoA of 60+ degrees and vertical speed of -10kft/min then it would certainly not help much now would it?

In the Buffalo crash the pilot pulled through the stick-shaker, but in that case there were issues that were arguably deeply rooted at a skill level.

A simple question:

Would something per my suggestion - "AIRCRAFT STALLED, REDUCE AoA!" flashed in red text on the glass displays have alerted the AF447 crew that they were stalled? - triggered from AoA and V/S?

There are two possible answers - "NO" and "MAYBE" - if it is the latter then proceed with a human factors study, etc. I am not proposing a rushed introduction.

My personal opinion is we, as humans, are bombarded with technology today, but look at the trending.... we are moving away from email to Facebook, and T w i tter (with a limit of 140 characters per message), we TXT msg each other in short-hand code, our attention span is shrinking. When things go wrong we tunnel vision, shutting down our senses - therefore at moments of critical decision making we NEED the available technology to prioritize our input stream for us.

henra

Agreed !
Was imprecise on my part.

Although the measured Airspeed was very likely a consequence of the AoA exceeding certain values leading to collapse of the flow around the pitot and not so much representing the real Airspeed.