I do not think that anyone should be misled by thoughts of the gliding qualifications of Bonin (FO, PF). All I have seen is that at some time before the accident he obtained a French glider pilot’s licence. He was qualified (at a point in time) but not necessarily “highly” qualified, nor necessarily current.

More to the point, however, there have been enough stall/spin accidents to glider pilots of all levels of experience and currency to show that no class of person is immune to failure to recognise and recover from such departures when they happen unexpectedly, invariably as a result of pilot error/mishandling. I had access to all the UK gliding accident reports for many years, and one included a fatal accident to a 19000-hour ATPL who converted to gliding and spun in when flying solo. Almost every year in the UK there was a fatal accident from stall/spin, not corrected properly by the glider pilot. Sometimes more than one.

A rare survivor of an inadvertent spin at low level told me exactly how it had happened; he was fixed on a wrong diagnosis of what had gone wrong (as I suspect Bonin did) and never reverted to his training. (In this case, he had a cable break, during a winch launch, and with incorrect pilot reaction, the nose dropped as it would. He wrongly thought that the tail had fallen off, forgot all his stall/spin awareness training, and remained convinced that nothing he could do would restore normal flight. He was lucky to survive.)

By the way, ignoring warning sounds and other indications when one has a wrong idea of what is happening is far from unknown. There are sufficient numbers of recorded events demonstrating this. I am not an expert, but I have read that the auditory channel to the brain is the first to get blocked out under overload. It is well known in instructing circles (or was when I was an instructor) that the only way to get through to a student pilot in brain-freeze mode is to physically get their attention. Speaking/shouting becomes ineffective. Another classic example is the video clip of three people in an aircraft, at least two being qualified pilots, who ignore the persistent undercarriage warning noise and land wheels-up.

tdracer

Interesting - it raises the thought that when considering the range of system reliability the airframe manufacturer must use the limits defined in the equipment manufacturer's Declaration of Design Performance (DDP) not values picked out of a hat. If one probe manufacturer says OK above 30 kts and another above 60 kts then those numbers will be what the airframer will use. Which leads to the question - does anyone know what make/model AoA probes are used on the Boeing fleet and Airbus fleet?

rudderrudderrat

When one puts V/S and ground speed together, the aircraft never got below about 150 kts throughout the whole process. That means the AoA probes as probes were working properly and supplying valid signals (as long as the AoA did not go out of the probe range). As has been explained many times, the problem was that the extremely high AoA took the aircraft right out of any sensible AoA range for which the PEC algorithms might have been defined. Consequently the measured airspeed was wildly different from the actual airspeed and fell to below 60 kts IAS. That, taken with the probe manufacturer's DDP was the reason for the AoA signals to be declared non valid.

I agree that a continued stall warning might have made a difference. I have said it before, but I will repeat it, the aircraft was designed to meet JAR 25 Change 13, which says:

This differs from the latest rules (CS 25)

Clearly, it was not intended (and is not intended) that the stall demonstration should be taken too far into the stall, so there was no requirement to maintain stall warning through a prolonged stall. If present rules are interpreted as requiring a valid signal that angle of attack has been reduced to below stall warning then this would IMO cover the AF447 situation.

As for "Is it really that difficult to have stall warning Valid >60kts AND on ground" I can only say that it would be a futile exercise as it is virtually impossible to stall an aircraft with the wheels on the ground since it would be attitude limited well below any stall AoA.

Dozy

I covered your first point above. The 'flipping' of the AoA signal on the DFDR is I think caused by the AoA going above the range of the instrumenation (50 deg plus!!!)
I agree with jcjeant - the experts panel was convened by the French judiciary, and in fact this is said explicitly in the preamble to the French original.

Hi Owain Glyndwr,

Thanks for your explanation of the present rules CS 25.
The reason for 30 or 60 kts logic is to prevent nuisance stall warnings whist on the ground due to the vanes being blown around randomly.
However it is better to know that you have a false stall warning around 60 kts during the take off than only receiving it once you get airborne with only WOW logic. (see ASN Aircraft accident Lockheed L-1011 TriStar 1 N11002 New York-John F. Kennedy International Airport, NY (JFK)
"...the aircraft lifted off the runway. At that moment the stick shaker activated and the first officer, who was making the takeoff, sensed a loss of performance. The captain than took over control and decided to abort the takeoff. The TriStar touched down again (at a vertical descent rate of 14 feet/sec - the structural design limit being 6 feet/sec -) after being airborne for about 6 seconds."

With the benefit of hindsight of AF447, I still believe it would be better to have had the stall warning on continuously whilst they were stalled, rather than be turned off simply because IAS<60 kts. It must have sounded like a spurious "computer" glitch - and hence be initially disregarded.

rudderrudderrat

I agree with you there.


I'm not sure I follow the logic here. It would surely be better to have no false stall warnings at all? I stand to be corrected, but I thought that the present AI system has no WOW logic for stall warning, just the 60 kt discriminant on valid AoA signals. That way if there is a faulty AoA signal it becomes apparent immediately one passes 60 kts, but random nuisance warnings are suppressed, so doesn't that satisfy your requirements?

Hi Owain Glyndwr,
I agree. Unfortunately no one has yet designed a system which never fails.

No.
As you so eloquently explain in your earlier post despite "the aircraft never got below about 150 kts throughout the whole process."

Therefore the 60 kts IAS logic inhibited the valid AoA probe information.
The stall warning should be independent of IAS when airborne.

Think you are being a bit harsh there Dozy calling it 'rubbish', unless it's about the possibility of a female on the FD in which case fair enough. the Cpt didn't return with a 'what do you want', or 'what's the problem', or 'can't I leave you for even 10 minutes without calling me back'. His opening gambit is entirely consistent with him challenging the crew about a situation he has discerned himself and not with being summoned back.

Relevance as to where he had been ? Well if it was a mandatory rest break and he wasn't resting: whether other factors bore on his decision as when to have his break: whether 'social' tensions affected CRM (including his choice of who was to command): the mental alertness/fatigue arising from how they spent the stopover, etc.

On the AoA, a continuing stall warning may have helped the Cpt diagnose the situation but there were sufficient other factors surely, even excluding the period when it did warn. Wouldn't be excoriating the design on that one.

rudderrudderrat

And never will, but in this case (take-off) we were addressing a situation where there were no system failures, just vanes moving as a result of atmospheric turbulence I think.

Strictly speaking the stall warning cannot be independent of IAS because one needs some sort of speed signal to cater for warning threshold variations with Mach Number.

I agree that the logic inhibited valid AoA data, but as I said earlier keeping the stall warning sounding irrespective of measured IAS until there is a valid return to AoA below the threshold would fix that. I assume here that excursions in measured AoA outside the S/W threshold below 60 kts and on the ground would be transients so that the conditions for switching off the warning would be quickly satisfied.
Perhaps I should also say that so far as I can see, it would not be necessary to change the existing AoA validity logic to make this work. All that is required is that the stall warning logic be changed to remove the warning only when a VALID AoA signal shows that safe conditions are restored.

Not necessarily. e.g. On a parallel taxiway to the runway, whilst heading towards the holding point in a stiff breeze will cause the vanes to align 180 degs away from zero. Hence the need for the 60 kt logic when on the ground.

Unfortunately no one ever imagined anyone could pull an aircraft into such a stalled AoA without realising it. However this crew succeeded. But why were they so confused? Intermittent warnings didn't help the Captain to diagnose the problem. One tends to think that particular problem has been solved when the warning stops, and so move onto the next problem. e.g. Why are the altimeters winding down to electrical zero?

Edit. Hi Owain. Changed " didn't help the crew to diagnose" to "..Captain .." since the two FOs hadn't a clue between them.

rudderrudderrat

OK, I accept the case, but I think the vanes would go to one or other of their mechanical stops not rotate through 180 deg.

That said, refer to my edited version of my previous post where I suggest that the 60 kt inhibition could in fact be maintained as is.

I agree intermittent warnings don't help diagnose the problem, but this only applied to the latter stages after the captain's return to the cockpit. 54 seconds continuous operation is hardly intermittent, although this is not affected by the 60 kts affair.

. Just for accuracy - that may well be true for the A330 but not all aircraft are attitude limited.

DozyWannabe
My conclusion that they had no idea of stall recovery is based on the action they took to solve the problem. Had they been trained properly for stall recovery they would have been instinctively pitching well below the horizon for the altitude they were flying which never happened instead they applied TOGA and zoomed into the sky. Their action was perfect for GPWS warning and not a stall warning. In any case subsequent monitoring of pitch, altitude,ROC/ROD, or asking for UAS procedure nothing was done.The airline's training programme deserves more blame than the pilot. Pilot doesn't decide what he should be trained for before being cleared as relief crew. Problems during cruise high altitude stall recovery and especially unreliable speed should have top priority in training of cruise captain. He doesn't do approach and landing.

The problem with human cognition is that it is often not 'logical'. A continuous aural warming can be disregarded as the aural cognitive channel overloads. The cessation of the warning however can have an immediate - what's that just happened' effect. So if the aircraft had sat in deep stall effectively the wrong side of the drag curve with the stall warming sounding the PF and PNF in that situation could well not actually realize the sound was there (as was shown on the infamous wheels up video). The captain however on return to the cockpit would have not been so cognitively overloaded and would have noticed the stall warning.

The advantage of stick shakers is that they are a haptic input (feeling and touch) this cognitive channel does not shut down so fast and as instructors well know a sharp clip around the back of the student's head is often the only way to remove the tunnel vision (attentional tunneling) of an overloaded student.

The effect of cognitive overload seems to be overlooked despite a considerable mass of research demonstrating how it is possible for humans even in relatively low stress exercises to ignore the blindingly obvious. It may even be that the more noise and cavalry charges, flags and flashing lights there are, the more the stressed individual actively disregards them and focuses on the one aspect that they are trying to control. This is when PNF is meant to break in 'haptically' perhaps, but in this case his 'focus' seemed to be recalling the captain.

He didn't. Nor did he link it to the other symptoms, i.e. buffet, low airspeed, high rate of descent with a NU attitude, abnormal motions in roll and in pitch, apparently difficult to control.

Cognition
 

Dubois was overloaded with personal problems.
Torn between his ladyfriend and strife on the home front, his spirit was elsewhere.

Winnerhofer;
Such statements cannot be taken seriously as they come from unidentified sources, which typically signals that there is an agenda behind the statements.

Describing and interpreting psychological aspects of crew members behaviour is where angel's fear and as such, is where a very conservative approach is warranted and even demanded.

For me anyway, unsubstantiated rumour without sources, particularly regarding a flight crew member's state of mind prior to an accident is emminently worth ignoring and has no place in a Tech discussion.

Owain & rudderrudderrat, I am particularly enjoying your exchanges.

Ground speed has nothing to do with air speed, never has, never will. :=

Speeds
 

Bonin: "On a pas une bonne annonce de vitesse" when he should have at said "...vitesses douteuses..." (UAS).
"...annonce..." is never used
Further, he said: "...on a une vitesse de fou.." when he should have said "...on est en survitesse..." (IAS Overspeed).

Personal Problems
 

Crash de l'AF 447*: la fatigue des pilotes mise en cause

Mais pourquoi une telle fatigue?
Le matin même, l'équipage s'est offert une virée en hélicoptère dans la baie de Rio. Selon le pilote de l'hélicoptère, que j'ai interrogé sur place, l'équipe montre déjà des signes d'épuisement. En réalité l'un des pilotes s'est rendu à Rio en compagnie de sa femme. Le commandant de bord, lui, un homme d'environ 55 ans, en instance de divorce, est également accompagné de sa maîtresse. Les hommes de l'équipage et leurs compagnes descendent au Sofitel de Copacabana. On peut penser qu'ils se rendent à Rio dans un esprit plus festif que professionnel, et que ce jour-là, après une nuit trop courte, le commandant de bord n'était pas en état de réaliser ce vol.

But why such a strain?
The same morning, the crew went on a ride in a helicopter in the bay of Rio. According to the pilot of the helicopter, I asked on the spot, the crew was already showing signs of exhaustion. In fact one of the pilots went to Rio with his wife. The captain himself, a man of about 55 years, divorcing, was also accompanied by his mistress. The men of the crew and their companions were staying at the Sofitel Copacabana. Presumably they went to Rio in a festive rather than professional spirit, and that day, after a short night, the captain was not in shape for the return flight.

Ozlander
I beg to differ - ground speed correlates directly with airspeed in level flight and zero wind conditions.

If you have been following earlier exchanges on the various threads of this voluminous tale you might remember that at one time it was being suggested that ground speed, or at least some sort of GPS derived speed, could have been used to give the crew some idea of their actual speed in the absence of genuine airspeed data. What I was trying to convey was that you cannot just take V/S when assessing what the airspeed might have been, but you can combine V/S with an inertial speed to get an approximation of the actual airspeed ignoring any wind effects. No more than that. I am well aware that doing that does not yield airspeed, but it serves in the absence of valid pitot static data, and certainly is a better approximation to the truth than the airspeed deduced from the standard system at 50 deg AoA or thereabouts. Read what I wrote:

The sole object of mentioning the real airspeed was to point out that the AoA probes were operating in an acceptable speed range. In that respect any "errors" from neglecting wind effects are not important.

In fact the use of the IRS groundspeed as displayed on the FMS MCDU is formally recognized and as such is recommended in several A330 flight crew training manuals. The earliest I can find such recommendation is from early 2005.