TJHarwood / airtren
 

TJHarwood

No comment!

airtren

Best to leave a stall recovery thread until after the final report. Besides, in the meantime, best to stick to and focus in on what went wrong after the temporary UAS leading to A/P and A/T disengagement and before things really went pear shaped and "real airmanship" was required at FLT -100 per min.

DozyWannabe
 

I think the Wingman was referring to "revolutionary" in the narrow context of the control columns, as the SSs would have been revolutionary to any pilot (even an A300 pilot) entering the flight deck of a new A320! As Boeing are still using control columns in the 21st century, even on FBW aircraft such as the B777, might it be to reflect pilot mindsets......?

With the benefit of hindsight, would it have been better for all manufacturers to use the same control column system whilst jointly addressing the Stony Point "stick shaker"/stall issues? We now have nearly half the industry using a different system, with flaws appearing.

AF447 stands out. The loss of Capt Warner was a test flight, with all that entails, and no more indicates an A330 fault than the RR engines FOHE issues indicates a B777 fault. The Air Afrique A330 loss is the least of Libya's priorities, so no satisfactory investigation is assured.

AF447 is critical so, unless pilots are better trained at high altitude manual flight, and with all the additional cost that entails, the pitot tubes and ADIRUs had better be damn reliable!

DozyWannabe
 

Thanks for your comments. A few points:

(1) Definitely one for the human factors experts, but unlikely PNF grasped just how far back/NU the PF's inputs were. Not relevant if CRM is perfect, but there already seems on these threads to be a certain amount of consensus on CRM shortcomings to a differing degree....

(2) As TJHarwood has alluded to, and as someone who "lived" during my early piloting years by my AoA indicator, best not to get me started on the design philosophy that resulted in a stall alarm being repeatedly switched off when the aircraft remained ......... in a stall (however far outside of the designed flight envelope).

(3) I wasn't thinking in terms of the A300, although there were revolutionary aspects, but more in the context of what Lockheed, Boeing and McDonnell Douglas were producing. A different "philosophy".

(4) In this world of aviation deregulation, more than ever, there needs to be a global (i.e. especially USA/EU) regulatory response. No airline will take an expensive lead. AF, for example, had 3 hull losses (2 with heavy fatalities) in a decade, but retains a loyal customer base. It has got to be extreme for an (insured) airline before corporate reputation dictates huge costs on a root and branch training and manual flight expertise revamp.....

Additional costs will have to come down as a level playing field, or it is just about hopeless.

I'm sure I read on here earlier that the original B777 design used sidesticks, but the launch customer's (UA) pilot's union objected - certainly not representative of pilots as a whole!

Sums it up for me.

Colour me cynical, but I'd be surprised if most airlines couldn't purchase a small airfield hangar and a bunch of trainers (jet and prop) for a fraction of the CEO's annual bonus!

Anyways, I'm spent for now - later.

DozyWannabe
 

Touch wood, but CFIT is not the killer it once was (if still too frequent).

LOC has been the real concern in recent years.......

And one very good reason is the 320/family CFIT escape solution. I am a big fan!

This quote from Joe Sutter popped up on a blog yesterday:

@ Jazz Hands:

1/ Your link appears broken. Too many stars, it seems ;)

2/ I'll ask his messenger, as Mr Joe Sutter seems out of reach:
- Where exactly did the Airbus A330 of AF447 tell the pilot what to do?
- Where exactly the Airbus A330 of AF447 did not do what its pilot asked for?

@Jazz Hands, AZR,

Joe Sutter is enough of a legend that he's perfectly entitled to his opinion, even if it betrays his misunderstanding of how a competitor's product works. Hell, if I made it to 90 I'd be inclined to be as opinionated as I liked whenever I liked, even if I knew I was talking horlicks!

The stuff of urban legend.....

One goal of urban legends is to entertain, and the other is to distract. Comedy has to be generic, or not enough people 'get' it.

So from an internet comet comes further misunderstanding. "Tell the PILOT what to do?" Well, Yeah. By doing it without him/her noticed, warned; by stealth/surprise (ready for flash bang, here)

Bottom line is this, probably. As the a/c "gives it back" it "hangs on". It is not a "secret", for it is published and trained. It is counter intuitive, however, and ultimately leads to disaster, along with other "eccentrics".

POV? Yep. Maybe PJ2 was right, Maybe "Human Factors" is the landscape for the discussion. Although the problem ultimately, is a design that requires a degradation of human skills, and a deeper reliance on Technology that to some extent doesn't allow for an everpresent design consideration. BTW, This "Technology" was new when I was using an APPLE e2.

ROTE learning ending up requiring ROTE handling. There was a fourth "pilot" aboard, and we neglect that influence at our great risk.

Some "one" decided Autotrim was an excellent fallback in recovery from Unusual Attitude(s), and that STALL warning could (should) be defeated at low (!) speed. PILOT ERROR I'm saying, ennabled by a remote command from another time, and another place. What a deadly Intrusion!

So, Yeah. Next time, can we have APPLE?

AZR

"Where exactly did the Airbus A330 of AF447 tell the pilot what to do?"

As BOAC, Lyman and Wingman have alluded to, and wp no doubt to your primary contention that the pilots weren't listening to the A330 telling them that it was stalled(!), the A330 repeatedly told the pilots it was no longer stalled (and at a critical point for stall recovery at FLT 350 as the CDB re-entered the cockpit).........

Lyman

"Although the problem ultimately, is a design that requires a degradation of human skills"

I think that is going a little far, even if there is considerable concern expressed as to what has happened in practice over the past 30 years. There is nothing in the underlying design philosophy itself to preclude pilots from honing their flying skills on light aircraft, line flying by-hand within the flight protection envelope or spending additional time in the simulator. I liked the Wingman's description of Airbus (and the B777 and B787, hence not Boeing v Airbus?) as "airmanship plus", of which the late great David Davies (his fellow countryman) let alone Gordon Corps would approve, which conveys the need for standard flying skills plus additional skills. The problem is far far wider than Toulouse......

PJ2 - any thoughts?

I'll bite:

a) I don't think they turned FD off did they ? Maybe it told them to climb ?

b) [easier] The pilot definitely, forcefully and persistently told the a/c to climb, meanwhile the stupid a/c ignored him and continued to fall out of the sky. Presumably if you tell a Boeing to climb out of a stall, it does just that.

Did I win ? :E

IF789/AZR/Lyman
 

Three A320 hull losses between 1988 and 1992, essentially because the flight crew and the aircraft were on a different "wavelength" i.e. unfamiliarity issues.

Notice any similarities between the Air Inter flight crew's discomfort with having to suddenly make a non-precision landing at Strasbourg and the AF447 F/Os discomfort at having to take manual control at FLT350 (no visual aid, not even moonlight)?

AF447 is so depressing because it is 17 years later, and we have another total disconnect between a flight crew and their aircraft.

Simulator (AB speak) - learn to fly (typerating) direct law, progress to alternate law, round-off with normal law? Gate to gate, every time. All sides of the industry implicated, not just manufacturers or a specific manufacturer (SS/feedback issues notwithstanding vis-a-vis AB).

Commercial pressures? Non fici facio, vera prae ceteris - as Davies would say. Get it sorted. Thanks to automation, properly used, there has never been so little excuse for an air crash.....

Reminds me of planning a driving vacation in the UK where I would have to adapt to driving with the gear shift on my left and round-abouts that go the wrong way for my skill base..

Do I adapt ?

Take my chances at a higher accident rate while learning to adapt

call for a change of driving regs and car design in the UK?

or simply stick to driving only familar cars?

And Airbus changed their training and the FMC interface accordingly.

Not really - for a start the AF447 PF doesn't make any reference to discomfort until the roll oscillations start - there's no evidence that he's aware he's doing the wrong thing.

Secondly, while the main technical revision to come from the Strasbourg crash was making the Honeywell FMC display the difference between V/S and FPA select in a more obvious way (by the simple addition of illuminating two zeros to the right of the indicator in V/S mode), the background of the two incidents couldn't be any more different, particularly with regard to human factors. Air Inter dealt exclusively in very short flights with very quick turnarounds - very different to Air France's long-haul operations where the possibility of making up lost time is much easier. This is important because Air Inter in the '80s was placed in direct competition with SNCF's new TGV rail service and therefore the penalties for not making schedule were potentially severe.

In the case of the Strasbourg accident, both the Captain and F/O were experienced pilots, but both had less than 100 hours on the A320 - something which would not be allowed today. When Approach control told the Air Inter Captain that he could not use his preferred approach and would have to land on the reciprocal runway, the report says that his tone of voice on the CVR becomes increasingly agitated, possibly due to a combination of frustration at not making schedule and uneasiness with the fact he'd be making an NPA at night for the first time on this type.

Then we have the fact that ATC gave the crew an incorrect vector, putting them off-course laterally, the aforementioned bad interface design making it possible for them to set a 3300 ft/min descent rate as opposed to a FPA of 3.3 degrees and for the worst bit of luck, a random pocket of turbulence causing the A320 to descend even more rapidly.

What this adds up to was that the situation in the flight deck was tense, but they weren't aware of how severe their situation was until they were a few feet away from the mountain. Compare this to the AF447 case where they were well aware they had problems almost as soon as the PF began overcontrolling, but they did not or could not formulate or action a plan to correctly resolve the situation.

The other big difference is the fact that the Strasbourg A320 was in autoflight almost all the way down to the ground, whereas all evidence indicates that the AF447 A330 was manually controlled into trouble.

I'd argue that the more pressing worry in the case of AF447 is the flight crew's disconnect from *each other*. They could see what the aircraft was doing and so unlike the Strasbourg crew they were not "behind the aircraft" as such except for one critical detail, and that was that the aircraft was stalled.

To be fair to the aircraft, it did tell them it was approaching stall, and then that it was stalled, for nearly a minute before the readings caused the stall warning to trip out. Why did they not respond accordingly, whether the Captain was there or not? Why did neither of the F/O's mention to the Captain that they'd had a minute of stall warning prior to his arrival?

I'm not saying that the stall warning logic doesn't need an overhaul, because it clearly does, but ultimately the warning was there and continued long enough for someone to take notice and do something about it (to say nothing of the oscillating bank angle, rollercoaster pitch changes and the rapidly unwinding altimeter)?

Emphasis mine, and therein lies the rub!

[ Addition : Air Inter's usual practice, because of the unique way they used their aircraft, was to frequently bat their Caravelles around at 300KIAS+ below 10,000ft (sometimes considerably below!), which tended to play havoc with the GPWS hardware available at the time, and for this reason they received a special dispensation to not have GPWS fitted to their A320s, despite incoming regulations requiring it. Shortly after the accident, GPWS became mandatory and was retrofitted to their fleet. ]

lomapaseo / TJHarwood
 

Re: incremental change.

Weight dictated FBW on any SST, with the Concorde's famous rams's horn control columns and "feedback". Built by the predecessors of Airbus.

CONCORDE SST : FLIGHT SYSTEMS

Love the analogy (don't you realise we mainly build roundabouts to stress American tourists, particularly if they are not from Utah....?!) Do Ford/Chrysler/General Motors, when they sell cars in the UK, either:

(1) move the steering wheel to the right of the gear shift; or

(2) abolish the steering wheel in favour of side sticks without (artificial) feedback?

Incremental, even when it involves a bit of a leap....

The serious point, when Airbus embraced automation ahead of Boeing (not instead of Boeing - see B777 onwards), is that there were various options.

Some fast jet pilots were familiar with sticks (side or otherwise). Most civil airline pilots were not.

SSs and the absence of "feedback" were not critical to the automation drive, and don't let anyone tell you otherwise - see Concorde/A300. It was a decision on one aspect of automation.

Captain Harwood,

The degradation of skills I see is not limited to direct neglect, nor would it be. A "Thick through Disuse" type thing, to where the skills are vestigial, if any remain at all. A great deal of hype re: the platform, at the expense of Laissez faire, le Pilotage?

Now that would be acceptable, should the numbers firm up a sound reason for allowing skills to atrophy. There are none, as you aptly show, 447 is a sad example of a precious penalty for allowing a "blind spot"?

One small addition, the market will not support levels of auto didactic skill, the investment will be ignored, and the entry level, de minimus, will prevail. Thus it actually is a direct degradation, though in this case, market driven.

There is one additional disincentive to "hiring" the self taught (self invested) man/woman: "That one Clive, looks like a troublemaker, to me!"

:ok:

@Welsh Wingman - I think a more appropriate analogy would be :

"When we went from horse-drawn carts to motor vehicles, did we attach ropes to the front axles and steer with that (because after all, that's what people are used to), or did we design something more appropriate for the technology?"

and later

"When we went from gigantic steering wheels required for the leverage to move worm-gears, via rack and pinion, to power steering, did we keep the wheels artificially large (because, after all, that's what people were used to), or did we design something more appropriate for the technology?"

If we're to believe the earlier poster, Boeing only retained the yoke under pressure from a single pilot's union (the one that belonged to their launch customer). There has been no unequivocal evidence that the SS/feedback issue has ever played a role in any recent LOC incident (not least because several of them involved yoke-equipped airliners). This argument feels like a cul-de-sac to me, and I think it's a shame we're having it again.

[PS. WW - I'm sure you know which type the sidestick concept was originally tested on, right? ]

DozyWannabe / TJHarwood
 

I don't wish to be accused of putting words into TJHarwood's mouth, but I interpreted his comparison between the crashes as limited to the similarities between the instant discomfort of the PF upon losing his automation. I don't think that was a happy cockpit from the A/P disconnect, and the roll oscillations weren't long in coming....

DozyWannabe - but you raise a point that TJHarwood appears to have overlooked i.e. PNF "nagging". Wasn't the PNF in 1992 concerned about the PF's positional shortcomings (sadly horizontal, not vertical) in much the same way as the AF447 PNF was concerned about the PF's inputs and the climb? Good point re: disconnect between pilots on both occasions.

And whatever one think about an AoA indicator or BUSS as an option, I think we can all agree that a GPWS should never be optional (whatever SNCF's TGV competition).

Human factors, and that must focus on the interface and training. Automation, properly used i.e. "airmanship plus".

The exact answer to that quiz question is of course "F-WTSB".
Without asking Google, I'm not sure whether the original 'joystick' idea dates back to the Viper (F-16) or to the early Atari video game consoles...... i.e., what inspired what?

The wright flyer used a side stick;)

Thanks, nice one!

The Air Inter captain never entirely lost his automation, just his pre-programmed approach. He never got as far as having to manually control the aircraft, instead following the radar vectors and initiating descent via FMC input.

Thanks! So - if I recall correctly - yes, the Captain started to say something to the effect of the altitude or V/S not looking right, and in an unfortunate coincidence, at the exact same time the F/O interjected to point out that the display had them to the left of the airfield (which was due to ATC turning them onto final prematurely), which then preoccupied both pilots (in much the same way that an unfortunately-timed radio transmission distracted the Kegworth Captain as he began to re-evaluate the symptoms).

Of course, with 20/20 hindsight!

No argument from me there!

@CJ : I think the first modern sidestick came with the Mercury project at NASA. :)

DozyWannabe
 

I haven't been following this thread from the beginning, so not sure from where the "story" emerged that B777 yokes was purely down to United line pilots. It most decidedly did not, and there was widespread consultation with Boeing's 747 Classic customers (the target market for the B777, in competition with the A330/A340). My 747 Classic F/Os, that are still flying, are predominantly on B777s.


777-Yoke Vs. Sidestick


With all due respect, and I really value your contribution to this thread, and which is always insightful, I think a better analogy would be the migration from horse and cart to the motorcar - whether reins or wheels, the retention in front of you over which you clearly control (whether horse or machine).


We all know that "feedback" to yokes is artificial, and has been for decades, but a yoke emphasizes control (pleases Joe Sutter!).

The argument, for those that are not "fans", is that the SS regime inadvertently disconnects the PF from the aircraft and (at least without interconnectivity to the other SS and with artificial feedback through the SS to the PNF) from the other pilot - it is a system operator tool, rather than a pilot control tool, if I may put the argument in those terms.

I am undecided because overall Airbus are excellent aircraft with an excellent safety record (particularly if you remove the "human interface" crashes from their list of casualties, and Boeing have had more than a few LOCs). I also take your point that the SS came out of the USA through the F-16.


None of this would of course matter to this thread, if the pilots were at one with the A330 comprising AF447. But this is currently very much a point of contention...

WW,
Thanks for the link.

The stick v yoke argument is irrelevant to this isn't it ? Ditto analogue v digital displays. The logic that increased automation has reduced the necessary total human capacity on the flight deck would seem more fundamental (the arithmetic supporting the argument that any given level of safety can be achieved with fewer people eg no FE owing to the automatics). Fine when the automatics are present and correct, not so solid in an accident sequence.

Careful, the early cars were actually steered with tillers - just as people were used to in boats. Steering wheels came later, but again the pre-existing nautical usage is not hard to spot...:)

Nice :D
But, as you noted, that's an aircraft. I do not know of any manufacturer which have already made a better one (i.e. not prone to human errors, nor physics laws)

CJ
 

A complete tangent, for which I apologise in advance to all others, but what did Andre Turcat make of a SS on the test SST?

We ended-up with ram's horn yoke, FBW and elevons with artificial feel.

Was there any significance to the final choice, given that aircraft type went under rigorous testing unheard of today (sadly to post-oil crisis commercial disaster, compare with hurried into service DC10)?

I'm not even sure Turcat himself flew F-WTSB with the sidestick.
But the test flight reports were positive (IIRC, will have to look through the report again), which says a lot, considering the general praise of other pilots about the manual handling of the aircraft with the conventional 'ramshorn' controls.

The side-stick on F-WTSB was never really intended for use on Concorde. After the certification flight tests, F-WTSB was used for several other 'non-type-related' tests, and the side-stick tests were part of that, and indeed led to the A320 story.

Mr Optimistic
 

It has only become "SS v yoke", which is somewhat of a misnomer, because the AB implementation of the former does not mirror the traditional characteristics of the latter and the flight crew never diagnosed the stall.

The human factors/CRM/cockpit discipline issues that arise out of the AF447 tragedy, as regards the SS, are the absence of artificial feedback (to the PF and through interconnectivity to the PNF) and apply equally to the throttle feedback.

Enough to fill in the missing pieces of the jigsaw for the PNF, if only enough to brief the returning CDB to get him over the line....?

Diagnose the stall and get to work on ND and the THS "problem", as per the related thread on AF447, although it feels somewhat esoteric and academic in the context of a flight crew that never diagnosed the stall. But an interesting post by Owain Glyndwr on that thread, for those of you aerodynamically inclined......

WW, thanks, yes appreciate that many things didn't help and that such help shouldn't really have been needed. A bit surprised (given the significance of the Capt's absence) that the more general question over whether two is enough isn't on the table. Noted that another company operates dual Capt/FO type arrangements - this sounds comforting to me as Pax.

CJ
 

Interesting. I can see a transition from yoke to SS as evolutionary, with the right training, but the loss of artificial feedback (including interconnectivity to the PNF, and also the throttles) makes it revolutionary and makes me, as British, about as enthusiastic as Margaret Thatcher having to go to the bicentennial of the storming of the Bastille!

Remove the flight engineer (the one flight deck officer devoted to aircraft systems), insufficient training for high altitude (including emergency), limit by-hand flying, lack of urgency about replacing a part critical to the aircraft system (pitot tube) and then give the pilots no blatant mechanical feedback through the SS and throttle at the moment of maximum sensory overload in the middle of the night during a reinforced oceanic cruise (sadly without a visual horizon, and sufficient to ignore a stall alarm albeit complicated by the 60knt/AoA design floor). I am not saying the flight crew should not have done better, because my thoughts can be deduced from my previous posts, but to keep chipping away at the margins...?

Would you deliberately design this from the outset, irrespective of the CRM issues on this flightdeck? Lyman (or whatever alias!) would not, and quite a few of you might find yourself uncomfortably in agreement with him over this.......

Again, a valid opinion - but certainly not an incontrovertible fact based on the evidence available. We've had an A330 successfully deadsticked into the Azores, an A320 very famously deadsticked into a near perfect landing on water, and conversely we've also had several yoke-equipped aircraft (B727, HS Trident, B757) go into a full stall and crash despite all manner of stick-shakers, stick-pushers and so on giving tactile feedback to the crew (which at least one crew mistook for Mach buffet). It was discussed at length in both the Tech Log thread and this one if you want to check it out.

The upshot of this admittedly anecdotal evidence is that the tactile feedback channel may not be the panacea for communicating aircraft behaviour to pilots that the pro-feedback brigade seem to think it is.

@WW (above, regarding "would you design...") - As I said in greater detail in an earlier post, Gordon Corps certainly didn't mind - ultimately he was the senior engineering pilot who signed the design off, and he was nothing if not a first-rate, safety-conscious pilot of the old school. The design itself was specified with input from pilots of all levels, from the line, through engineering pilots and test pilots.

@CJ (below) - when would it activate? I could see it being annoying if it was on all the time - perhaps when the trim passes a certain threshold which varies with altitude and AoA?

Yes, no.....
Personally, I'm still wondering about the absence of the "bicycle bell" on the THS trim wheel.
Not tactile feedback, of course, but still feedback.

Mr Optimistic

Now that is optimistic!

A fully reinforced (i.e. duplicated) flight crew would be preferable, ceteris paribus, but there are the same cost issues that ended the flight engineer (who would also have to be reinforced). Even a relief CDB and F/O is not the end of it, because you have the Van Zanten/Meurs CRM issue at the other end of the spectrum to AF447.

Takata set out the AF procedures at the time of AF447, and how the junior F/O in the RHS came to automatically become PIC. Interesting, and thankfully no more....

DozyWannabe

There will be nothing "incontrovertible" about AF447! Look at the number of pages on this secondary thread....

I believe, and have strong views on the point, that the need after Stony Point was to ram into every pilot's head that, before discounting a stick shaker (yoke or SS), that an uncorrected stall = death. End of. Be absolutely certain, as you watch your VSI plummet.

A SS with the same function is not a philosophical problem for me, even if it is to others, but a deadstick makes me nervous. As you state, you can land an A330 "glider" with a deadstick and airmanship. It is the additional scanning at the moment of maximum stress in an emergency that is unwelcome, when feedback and interconnectivity aid cockpit discipline in a cockpit that needs assistance.

If Boeing and Airbus, in unison (even if they preferred feedback yokes and feedback SSs respectively), had addressed stall issues, I would like to think that 9L/CJC 3407 could have been avoided (let alone AF447).

I have some difficulty with any argument from anybody on this forum that, if the LHS SS had been moving in tandem with the RHS SS as a result of the PF's inputs, it would have made no difference to what the PNF would have said and done (particularly in the initial critical pre-stall phase). PNF was "nagging" PF, without such feedback assistance, instead of calling the UAS drill.....

P.s. re: "design" - I mean the totality of that paragraph. What Gordon "signed off" is not necessarily what he would have preferred, and he could not possibly have envisaged the industry changes over several decades on the human interface issues that have arisen.
P.p.s. I would almost go so far as to secondary alarm the THS trim wheel whenever you depart from normal law in an AB, to keep the flight crew cognisant of the manual trim issues.

CJ
 

Please don't get me started on the THS trim wheel tonight, or I will never get to bed..........feedback, yes........but leave that one for another day.

But didn't the PNF take control only to have it snatched back, and this with the Captn present ?

The direct cause of 447's crash will not be "direct". To get to a single vector is impossible.

A starting point would be the untoward (and shamefully anticipated) dump of the a/c control into the hands of the PF. It had happened, though likely not ever in such piss poor conditions. This makes the foot dragging inexcusable.

There are no explanations, only excuses. Because all the precautions, all the preventions, were inadequate. How about that, then? One cannot focus on a single iteration; they were without exception, re-iterative.

In a perfect World? OK. PARK IT. FIX IT. Won't happen. I cannot say it cannot, but it won't.

That is why TRAINING is such a whine. And since it is a whine, it is dishonest, and arguably criminal. It is also self incriminating, the Pilot group does not self train.

Airframe?

The two or three glaring problems won't be fixed, too expensive.

What was known, and when did it become so? Ask that question of Socrates, or some modern day Solomon. It will be ignored.

PITOT. STALLWARN. THS. UAS Procedure. (This last, alone, food for an action at LAW that would cripple the players, and indict the Authority.)

When "corrections" can be applied before the engines have cooled, does one not think that the problem was a priori?

Yes he did, but it was approximately 5 seconds before the Captain arrived, the PF took it back after less than 2 seconds, the PNF presumably turns to talk to the captain as he comes through the door and in the intervening time (approx 2 seconds after the Captain enters the flight deck - so the Captain should have heard it, albeit briefly) the stall warning stops.

I've not put this together before, so it raises some very interesting points and questions.

• At the start of the sequence, as autoflight trips out, the PF makes the callout "J’ai les commandes" ("I have the controls").
• The PNF was preoccupied with the return of the Captain (and unless I'm reading it wrong, increasingly unhappy with how the aircraft seems to be being handled) before finally deciding to take the controls himself, just before the Captain arrives. He says "Commande à gauche" ("Controls on the left"), takes priority and makes a small nose-down and left correction.
• Now I thought that statement was ambiguous to start with, and was roundly told that it meant he was making a left-stick input. Now I've seen he takes control, makes the callout *and* makes that input at the same time, I'm unsure again.
• Was he aware that the PF had taken back the controls without announcing it? (interconnection aside, that's a flagrant breach of procedure on the part of the PF *unless* he interpreted the PNF's "Commande à gauche" callout as an order to apply left stick and as far as he's concerned still has control. Can anyone clarify what the correct callouts should have been?)
• The stall warning stops just after the captain arrives, and just after the PNF took the controls and made a couple of small corrections (which were correct, but nowhere near enough). Does he think his corrections have solved the problem?
• Either way, that's an unfortunate set of time-critical coincidences.
• Five seconds after that, the PF makes his "crazy speed" remarks and starts to throw the speedbrakes out, prompting an emphatic warning from the PNF not to do so. How much more pressure and cognitive load can one guy take?

Hmmm...

@WW - I think the lack of properly defined command gradient is a much bigger problem in the case of this incident than the feedback issue. I say again - same thing has happened plenty of times with interconnected yokes and it made no difference to the outcome.