AF471 - Nov 16th 2011 - Final Report

[Jack1985]

French BEA have released there final report into the serious incident that occurred on AF471 a Boeing 777-200 flying from Caracas to Paris-Charles de Gaulle on Nov 16th 2011. The Commander was PF, Co-pilot was PNF a relief pilot was seated on the observers seat.

Basically a Summary;


Aircraft was operating in CAT III conditions on ILS approach to runway 08R at LFPG in LVP.

AF471 at 136kts angled at 1 degree nose up, descended through 320ft when the Master Caution indicated that the flight mode had changed from LAND 3 to LAND 2.

Relief pilot calls out ''Alarm'' the PNF calls out ''go-around'', the PF responds by applying full power to initiate the go-around, which disconnected the Auto-Throttle.

Nose-up pitch command on the control yoke is recorded however insufficient enough to disconnect the Autopilot.

The aircraft began to accelerate the attitude changes from +1.15 degrees to -0.5 degrees.

PF orders flaps to 20, pitch decreases further to 2 degrees nose down.

Relief pilot calls out ''Pitch!'' - 10 seconds after the G/A, both crew pull on the yoke and Autopilot disconnects, aircraft pitches sharply resulting in +1.84G's vertical acceleration. Altitude changes from 2 degrees nose down to 7 degrees nose up - Subsequently reducing to 4 degrees nose up.

The speed is now 169kts, the Relief pilot again calls out ''Pitch!''.

The crew apply nose up input on the control yoke, the aircraft reaches its lowest point of 63 feet at 180kts, the nose then rises to 11 degrees nose up in 2 seconds, subsequently 19 degrees nose up and the aircraft climbs to safety.

Climbing through 870 feet the gear is retracted and the crew position the aircraft for a second approach, followed by a safe landing.


The BEA has found the probable cause was - ''inadequate monitoring of flight parameters by the flight crew.''

Whilst there were 3 contributing factors;


1. Partial execution of the go-around procedure.

2. Inadequate management of the automatic systems during execution.

3. The conflict of plans of action between respecting the operators instruction and continuing the landing, which seemed to be safely possible according to the manufacturer.


Report here - http://www.bea.aero/docspa/2011/f-pp111116.en/pdf/f-pp111116.en.pdf

My questions are what the hell is going on in Air France training programs and safety classes?? I've read numerous reports now AF6 (2010), AF447 (2009), AF1620 (2012), AF2184 (2012), AF3093 (2012).

I've never seen such serious breaches of basic safety and so many in the space of 5 years (And there are MANY incidents!) with such a big carrier.

If this was some African carrier they wouldn't be flying in Europe. What's so different about AF? :confused:

[Oakape]

If this was some African carrier they wouldn't be flying in Europe

Couldn't agree more!

My questions are what the hell is going on in Air France

My question as well!

It seems that being a large, western, legacy carrier these days doesn't guarantee a higher level of training & safety. Is excessive cost cutting starting to bite?

[Una Due Tfc]

The 4,400 FPM dive towards the runway in Tunisia was particularly shocking

[bubbers44]

Another automation dependency situation when going to manual flight is considered an emergency for some pilots?

[g109]

Air france should be blacklisted and prevented from flying in Europe.
In my opinion it's not a question are they going to have another crash or not, just when!!!!!

I don't know what's going on there, maybe somebody that works in AF can enlighten us.

My opinion: part of the problem is being French, and the arrogance they display, why did they not learn any lessons from the AF 447 crash, was the chief pilot, head of training, accountable manager been fired?

[awblain]

Lordy.

Down the back, 10 seconds would seem like a long time to be accelerating nose-down close to the ground.

The French version of the report doesn't seem to have any more details, but 63 feet doesn't seem to be many seconds remaining for this world, and even if instantaneous, 1.84g seems to be rather a lot at such speeds, presumably needing a fairly dramatic angle of attack.

Another highlight seems to be 3 seconds of disagreement on the direction to fly from 29:57 to 30:00.

I guess the impact on pitch of full power was handled automatically, or there might have been even more fun and games.

[Jack1985]

Another automation dependency situation when going to manual flight is considered an emergency for some pilots?

Really is upsetting to me personally, because I for one know the feel of the aircraft you have (and I fly just C172's!!) is to me an experience I always look forward to and basically what we have in recent years is people who'd prefer to swtich on the autopilot when airborne and off at the DH on landing - what happened to actually FLYING? It seems many AF pilots lack the understanding to fly effectively, or that they are simply thought to over-rely on automation, the latter seems obvious after AF447.

Another highlight seems to be 3 seconds of disagreement on the direction to fly from 29:57 to 30:00

The mind boggles. :bored:

[Jet Jockey A4]

WOW just WOW!

[CONF iture]

Between 9 h 29 min 57 and 9 h 30 the PFs nose-down input and the PNFs nose-up input on the wheel cancelled each other out: the wheel returned to the neutral position.
As the yokes are mechanically linked, I am not too sure how such information is registered on QAR FDR or is it mainly according to the pilots accounts ... ?

[bubbers44]

How hard is it to just go manual and go 15 degrees up and add go around power? Everybody I know can do it easily. Learn how to fly so you don't need automation. You will never be in a 4400 fpm descent pushing buttons. Learn how to fly, not push buttons.

[vilas]

bubbers
Automation cannot be blamed for this. If you press multiply key instead of divide you don't bame the calculator and throw it away but learn to use it properly.The problem is pilots have stopped monitoring the progress of the flight. They have lost the scan. Flying manually cannot be the only solution in highly automated aircraft. Flying manually 747-400 is no big deal for a classic captain but he should not be cleared till he is at ease with the automation.

[Wizofoz]

Yes, this was Automation mishandling rather than reliance- they WERE doing an Auto-land after all.

All he had to do was push the TOGA switches- that or simply make a one-push down-grade to CATIIIA (Though that seems to be against the operators SOPs).

Why do AF operate with a 20'DH for CatIIIB?

[swh]

All he had to do was push the TOGA switches- that or simply make a one-push down-grade to CATIIIA (Though that seems to be against the operators SOPs).

From that report "The PF stated that he pushed the AT disconnect switch, located on the throttle levers, unintentionally and by mistake, instead of pushing the TOGA engagement switches."

I recall BA coming very close in a 747 as well when a pilot came from another Boeing type where the TOGA buttons were located in a different spot.

[vilas]

bubbers 44
Your enthusiasm for manual skill this time makes you forget the reality. This was a CAT III approach and people who only love to fly manually or are uncomfortable with automatics should not be in the cockpit. He made a mistake of pressing A/THROTTLE disconnect button instead of TOGA levers that can happen but incidents rarely happen due to one mistake. He desired a G/A but he did not ensure that was happenning. Had he disconnected AP and not monitored the GA the result would have been more serious. In automated aircraft any pilot who keeps disconnecting AP to manage the situation because of poor knowledge needs to be retrained. You don't get cleared for CATIII for manual skills do you? You cannot get airborne without automation the whole world is RVSM.

[Machinbird]

Yes he needed to have pressed the TOGA button/lever.

Once P/F disconnected A/T instead (inadvertently) will pressing TOGA create the desired go around?
Will A/T then re-engage or must power be set and A/T re-selected?

Do you ever practice how to recover from an incorrect procedure in a time critical situation?

He could have disconnected the A/P and continued with the miss manually,
but it seems his brain was slow to shift into Go Around mode. Definitely behind the aircraft at that point.

[Trackdiamond]

Am sure the said Pilot was trained appropriately and previously demonstrated proficiency in consistently maintaining the requisite standard..on this occasion there might have been Human Factors issues at hand...Fatigue?Subtle incapacitation due to...??? Just wondering!
Also..was this Captain ex FBW Airbus pilot? Regressing to old habits under pressure syndrome!

[Wizofoz]

Once P/F disconnected A/T instead (inadvertently) will pressing TOGA create the desired go around?

Yes, it would still put the AFDS mode into TOGA.

[bubbers44]

Yes, cat lll approaches require automation but somewhere in the cockpit at least one pilot has to know what to do if pushing a button doesn't work. If not then we will hear more repeats of this incident. It is your choice.

[mightyauster]

As the yokes are mechanically linked, I am not too sure how such information is registered on QAR FDR or is it mainly according to the pilots accounts ... ?
Yes, the yokes are mechanically linked, but there is a "break out" mechanism on the torque tube between them, so it is possible to force the yokes in opposite directions. There are pitch transducers on the bottom of each yoke, so the DFDR will pick up the position of each yoke.

[Zippy Monster]

Why do AF operate with a 20'DH for CatIIIB?

My operator (European lo-co) has to use 25' for a CAT IIIB in France, despite using "no DH" everywhere else. I can't remember the exact reason our flight ops people gave to us word-for-word, but it is something to do with the French not allowing "no DH" operations.

[A4]

The NO DH was removed about three years ago - why, I have no idea.

It strikes me as particularly stupid because now at 20' or 25' right in the middle of the RADALT 50-40-30- you get a "MINIMUMS" with a requirement to confirm you can see 1 light, respond with "Continue" or "Go-around Flaps". Additionally the FLARE call comes in at 30-40 from the PNF. It's a mess.....and only in France -why? :ugh:

Regarding this incident.....lack of understanding of A/c systems as root cause? If you ONLY have to push the TOGA buttons to initiate (does the AP automatically pitch to 15 degrees?) I don't really see how it could be made any simpler. Airbus - just push the levers till they won't go any further and the autos (if functioning correctly) do the rest - easy. If they're not working and you are unable to pitch to 15 manually/promptly, perhaps you should consider an alternative career. Are the TOGA buttons and A/THR disconnect buttons close together on the 777 levers?

[one post only!]

I was told the reason was that the French authorities wanted the pilot to make the decision and not the aircraft. Hence the DH.

They clearly wanted the crew flying the aircraft and not the aircraft flying the crew.....ironic eh.....

[A4]

I'd heard that as a reason....but dismissed it as too ludicrous to be true. By introducing a decision at 25', a press of the wrong button......followed by "analysis" of the error.....doesn't exactly leave a lot of time. And, again, why is France the only Country in Europe to introduce this? They must know best - right?

[Cows getting bigger]

Fortunately there appears to have been one pilot in the cockpit who was quite happy and capable of telling the two 'in control' where to focus their attention.

[Zippy Monster]

I've never understood the "one light" visual requirement on a CAT IIIB with DH either. Like you say, A4, it doesn't leave a lot of time at all - and what happens if the "one light" you see (and on which you base your decision to continue) turns out to be, for example, a runway edge light, or something on the airfield periphery?

(Admittedly it's unlikely, such is the accuracy and reliability of both ground and airborne equipment these days, but then you come back to the question - why bother with the "one light" requirement at all?)

[CONF iture]

He saw the ground at 250 ft. Having visual references and, knowing that the landing was possible, his plan of action was to land. The NOLAND3 warning, making mandatory a go-around, led to confusion and to a change in his plan of action. He thus progressively moved the throttle levers forwards.
It is not clear why the PF did not call :
CONTACT - CONTINUE

[RAT 5]

I am not a B777 pilot, but I have flown Boeing a/c with Land 2 & Land 3 annuciations. Land 2 did not require a G/A it just changed the DH from Cat 3A to CAT 2 and it was still an autoland. So why the G/A? I would have expected that before executing an LVO ILS there would have been an extensive briefing of what to do normally and what to do if a common mis-hap occurred. Why was there a sudden surprise? Equally, the gotcha of pushing the wrong TOGA button is not a first and I'd have expected the captain as PF to be extra alert to where his fingers were close to DH. Does AF fly LVO as a monitored approach? Can the F/O call G/A above DH? Do not get diverted by a discussion about a 20' DH. It is irrelevant to this. I haven't read the report, but from the details given in the first post the incident was created by some strange behaviour above DH. Without knowing the SOP's in AF we can not make reasoned comments. Let's keep any comments to relevant facts.

[IcePack]

An auto go-around at 14 to 22 ft probably means the aircraft will touch down anyway so why have a DH. QED

[PEI_3721]

This incident begs several questions – with hindsight.
Why GA. What is the procedural rationale for GA if some capability remains; does it assume only the worst case visibility? Similarly, would the procedure apply during a VMC practice approach? How are the procedures evaluated, taught, and reviewed?
Are the crew expected to manage such dilemmas?

If there have been similar inappropriate selections of AT disc opposed to GA, were there any safety reports, did the manufacturer know, was the design evaluated and approved with such actions considered? Assuming safety reports are submitted, who considers and acts on them?

Why didn’t the AP initially disengage with pilot overpower. Normally this is a certification requirement to prevent back-driving the trim, as in this incident. AP approach settings might use higher torques thus requiring more overpower force, but due to the proximity to the ground other monitors are normally provided, if so what, when, why not?
Perhaps the investigation should have asked these questions, also the others above.

The ASAGA study referred to in the report is at http://www.bea.aero/etudes/asaga/asaga.study.pdf
Highly recommended reading of Sections 3 Flight Crew Survey - Problems, 4 Simulation, and 6 Analysis, all of which identify opportunities for improving safety.

[CONF iture]

Why didn’t the AP initially disengage with pilot overpower. Normally this is a certification requirement to prevent back-driving the trim, as in this incident.
the efforts to apply a nose-up input by the PF on the control column were insufficient to disconnect the autopilot.
There was no trim back-driving, the AP kept the aircraft on the ILS centrelines and applied a nose-down input on the THS as speed was increasing. Most probably FD were centered.

To me, a more informative call from the relief pilot after the amber warning could have been "NO LAND3" to which the PNF could have confirmed "LAND2" FMA indication and the PF who was visual "CONTACT - CONTINUE"
Not being visual at that time, and following AF procedures, the PF would have called "GO AROUND" and execute it.

Yes, the yokes are mechanically linked, but there is a "break out" mechanism on the torque tube between them, so it is possible to force the yokes in opposite directions. There are pitch transducers on the bottom of each yoke, so the DFDR will pick up the position of each yoke.
Interesting, it would be nice to see a architecture diagram and also some FDR listing. I have now read you can achieve a split of 8 deg (?) but then what next ?

[PEI_3721]

CONF, I can appreciate that there would not be any trim-follow-up (as in manual flight) whist the AP was engaged. However from page 1 of the report, with the AP engaged (9 29 50) back stick appeared to result in nose down THS, possibly due to an autopilot generated trim demand to balance the opposing stick input.
The AP was still attempting to follow the ILS, but the mechanism of control had to balance the computed demand for the flight path with that of the disturbing demand of nose up stick input. Thus THS movement opposed the stick input and the AP managed a semblance of GS control.
Following the AP disengagement the aircraft was out-of-trim nose down, thus a tendency for nose down pitch until both pilots applied back stick. Presumably thereafter in manual flight there was some manual trim-follow-up.

There is still no explanation as to why the an opposing stick input (fighting the AP) at such a low altitude did not result in AP disengagement or provide a further alert in order to avoid a hazardous out-of-trim condition.
In systems where the AP design requires a higher force before disengaging then a combination of force-duration may be used, but again in this situation with a low force, the duration appeared unacceptably long for the low altitude.

[Xulu]

We read many accident reports where the observer or PM have said nothing.

Well done that man for spotting and shouting it out repeatedly. I'm sure it saved a lot more than 63ft.

Air France :yuk:

[Lonewolf_50]

Thus THS movement opposed the stick input and the AP managed a semblance of GS control.
I am trying to sort out whether or not that is a good thing. :confused:

[Chris Scott]

Stretching memory back to my pre-Airbus-FBW days on several types, including the A310, I'm sure that it was normal for the AP to use autotrim to combat any yoke displacement by the pilot.

[Lonewolf_50]

Thank you, Chris. Point being, I presume, to maintain last ordered/selected state or performance. If pilot is to fly it by hand AP disconnect would be a usual first step. Do I have that right?

Since there are various models and schemes, maybe this question isn't precise enough.

[CONF iture]

CONF, I can appreciate that there would not be any trim-follow-up (as in manual flight) whist the AP was engaged. However from page 1 of the report, with the AP engaged (9 29 50) back stick appeared to result in nose down THS, possibly due to an autopilot generated trim demand to balance the opposing stick input.
So we don't have a similar appreciation of the situation, as thrust increase and its pitch up effect + speed increase alone justify the THS nose down movement to properly trim for the negative pitch necessary for the aircraft to follow the GS.

[Chris Scott]

Quote from Lonewolf_50:
"Point being, I presume, to maintain last ordered/selected state or performance. If pilot is to fly it by hand AP disconnect would be a usual first step. Do I have that right?"

Definitely. Trying to override the AP is a no-no. If it's not doing what you want, you must disengage it. **

I infer that this PF must have thought that (a) he had operated the TOGA switches and/or (b) he had already disengaged the AP (he had, of course, only disconnected the A/THR). Meanwhile, the autotrim was countering his gentle pull-up by down-trimming the HS by 1.65 deg in 6 seconds.

Quote from CONF_iture:
"So we don't have a similar appreciation of the situation, as thrust increase and its pitch up effect + speed increase alone justify the THS nose down movement to properly trim for the negative pitch necessary for the aircraft to follow the GS."

Quite. Perhaps a B777 pilot will also tell us what the HS would normally have been doing if the go-around had been conducted properly - with or without AP? Experience on other big twins with under-slung engines suggests that down-trimming is normally required to counteract the thrust increase AND the partial retraction of the flaps. So maybe the failure of the a/c to rotate to a go-around pitch when the PNF pulled hard enough to disconnect the AP was caused purely by the PF's unexplained pushing on the yoke.

In terms of hours, this was a very experienced crew - both in terms of total hours, and hours on type... :sad:


** Detractors of Airbus-FBW should note that, with no AP connection to/from the sidesticks, the action of pulling a sidestick back would have no effect on the AP or autotrim until it was pulled hard enough to disengage the AP. At AP disconnect, the THS would still be trimmed for the vertical profile the AP was last conducting.

[CONF iture]

Stretching memory back to my pre-Airbus-FBW days on several types, including the A310, I'm sure that it was normal for the AP to use autotrim to combat any yoke displacement by the pilot.
It was "normal" but unwise.
In 94, after Nogoya, Airbus made mandatory the Flight Control Computer modification to allow AP disengagement through manual controls input.
So maybe the failure of the a/c to rotate to a go-around pitch when the PNF pulled hard enough to disconnect the AP was caused purely by the PF's unexplained pushing on the yoke.
This is not what happened. AP disconnected when both were pulling on their yoke.
Only later on the CPT has been unexplainably pushing on his controls.
Detractors of Airbus-FBW should note that, with no AP connection to/from the sidesticks, the action of pulling a sidestick back would have no effect on the AP or autotrim until it was pulled hard enough to disengage the AP. At AP disconnect, the THS would still be trimmed for the vertical profile the AP was last conducting.
Again, nothing here for that 777 indicates that the THS was not trimmed for the vertical profile the AP was last conducting.
Actually, it would be very interesting to know if the too weak action of the CPT on the controls to disconnect AP, had any effect on the THS ... ?
It is a question that Boeing could answer.

[flyhardmo]

Quite. Perhaps a B777 pilot will also tell us what the HS would normally have been doing if the go-around had been conducted properly - with or without AP? Experience on other big twins with under-slung engines suggests that down-trimming is normally required to counteract the thrust increase AND the partial retraction of the flaps. So maybe the failure of the a/c to rotate to a go-around pitch when the PNF pulled hard enough to disconnect the AP was caused purely by the PF's unexplained pushing on the yoke.

The 777 FBW will auto trim the HS during any power or config changes so you don't feel the trim changes. Basically point to an attitude and hold it without feeling any increase in stick force. It's a SPD trim and only needs trimming in manual flight for any increase or decrease in SPD.

From the FCOM about autopilot disconnect force.
When an override force sufficient to move the control column, control wheel, or rudder pedals (LAND 2 of LAND 3 annunciated) is applied, the fly-by-wire flight control system will detect a difference between the cockpit controls and its own estimate of the expected positions, and will force a disconnect of the autopilot. After the autopilot is disconnected, the control system will transition the control surface commands from those of the autopilot to those of the flight crew

From what I understood the PF disconnected the A/T, applied full power whilst still going down the GS because the TOGA SWITCHES weren't pushed which would have put the aircraft in a 1500fpm climb. I'm not sure exactly how much force is necessary to disconnect the A/P but it could have been exaggerated by the increasing SPD.
The PF could have still hit the TOGA switches which would have forced the aircraft to climb as he had already applied full power although at a higher pitch angle. Alternatively he could have just used the A/P disconnect button on the control Yolk but like most people I suspect he thought the A/P was disconnected.

[Chris Scott]

Quotes from CONF_iture:

"It was "normal" but unwise."
Normal for the AP, not for the PF, which is why I described it later as "a no-no"!

"This is not what happened. AP disconnected when both were pulling on their yoke. Only later on the CPT has been unexplainably pushing on his controls."
Quite. You might consider reading what I write more carefully? After the AP disconnected, the PNF's attempts to pitch the a/c up seem to have been hindered by the PF's pushing forward. One wonders if the PF might have been anticipating an over-rotation that never happened.

"Actually, it would be very interesting to know if the too weak action of the CPT on the controls to disconnect AP, had any effect on the THS ... ?"
That was what I suggested, but you argued quite correctly that to maintain the descent the HS would need to compensate for TOGA even without pilot interference.

Quote from flyhardmo:
"The 777 FBW will auto trim the HS during any power or config changes so you don't feel the trim changes."

That sounds straight out of a sales brochure! My question was: does it have to trim forward during a normal go-around, and - if so - roughly how much? In other words, did the fact that it had trimmed 1.65 deg forward before the AP disconnected subsequently make the pilots' manual rotation to a go-around pitch more difficult, or would they have needed to trim forward anyway?

[PEI_3721]

flyhardmo, thanks, some interesting points.
The THS operation in manual flight is understood, but how does the AP involve the THS; AP command of autotrim would be logical to relieve the control system effort as discussed above.

It is assumed that the THS movement was due to the AP’s response to the change in pitching moment – opposing the overriding control input and/or the thrust increase, with possible further input due to configuration change. However, with FBW why would a stick input be used to control the aircraft (to oppose the AP) when the AP is engaged, cf Airbus.
If the override stick input has no control with AP engaged, then the THS change was probably due to thrust / configuration change, which leaves the potential for a residual out-of-trim condition when the AP disengages (disparity between stick position and control surface position) until “the control system transitions the control surface commands from those of the autopilot to those of the flight crew”, i.e. a period of readjustment with the pilot not totally in control, which is probably not significant … … unless there is a large change in pitching moment – woops.

Re the disconnect force; the description suggests that there is a ‘hole’ in the logic. If the crew override the AP following a down grade from a high integrity approach mode (cat 3), in circumstances where the AP does not disengage, then the overpower-disconnect logic changes / is non-existent.
Alternatively, (via Chris) do opposing stick inputs invalidate the overpower-disconnect logic; or does the disconnect description only relate to a jammed (blocked) stick, which may not apply with AP engaged.
What control forces (and movement) are fed back to the stick with AP engaged?

[flyhardmo]

That sounds straight out of a sales brochure! My question was: does it have to trim forward during a normal go-around, and - if so - roughly how much? In other words, did the fact that it had trimmed 1.65 deg forward before the AP disconnected subsequently make the pilots' manual rotation to a go-around pitch more difficult, or would they have needed to trim forward anyway?

There is no trim wheel in the 777 so unless you are looking at the stab trim band you don't have an instantaneous idea how far or fast the trim is working. Like I said, in manual flight you just rotate to set a pitch attitude for GA and the only time you feel that you have to trim is when you increase the SPD. In the drivers seat I have no idea what the ST is doing but I assume it's counteracting the pitching moments with FWD trim. No sales pitch, just the way it works which is most likely why (just having an educated guess) it trimmed 1.65deg forward because the SPD was increasing. ( going down the GP with full thrust).

What control forces (and movement) are fed back to the stick with AP engaged?
The control yolk moves with the A/P engaged as it would in manual flight as do the thrust levers. With your hands on the throttles and yolk you know exactly what the A/P is doing.

I'm still not sure why everyone is making a big deal about the position of the stab trim. I don't think it's really a factor. Disconnecting the A/P by overriding the stick force is not a normal practice on the 777 operation and seemed like it was done in desperation. The PF was fighting an aircraft trying to maintain the GP on autopilot rather than press TOGA or use the A/P disconnect switch.

[Trackdiamond]

There seems to be a spectacle of noise o this thread on "who-done-what -it is".
How about examining the Human Factors that led to the cockup in Automation mismanagement in the firstplace?
This was a perfectly functional aircraft.
The crew had plenty of total experience and hours on type.

What were the crew rosters like till the material incident date?
Were there any phisiogical causal factors that might have undermined the Captain who was PF?
Were there any stressful events enroute (factoring in times of operation and circadian dysrythmia)..?The nominated PM did not appear particularly alert..hence the vocal takeover by the relief FO!At least some one was awake!
Why,however, didn't the relief co-pilot annunciate more succinctly as is the required standard for LVO operations?Why didn't the PM take over as a matter of SOP after observing such a dangerous excursion..and in any case the PMs would have been more heads down and on the instruments and in theory better awareness of flight tragectory whilst the PF(Capt) would have been busy checking external visual cues to make a decision..I.e. the monitored approach model.
Both PF and Pm apparently were not paying atention to the FMA!
I believe this incident centres overwhelmingly on Human Factors.Will check the actual report to see how focal that issue was treated in the final analysis.

[porterhouse]

Boeing AP disconnect alarms are softer in comparison My God it is quite loud and very characteristic in tone. If someone misses this sound (or its accompanying flashing red light) it means their mind must be in a la-la land.:uhoh: :rolleyes:

[jcjeant]

BEA Report page 1
Le CVR, préservé par l’équipage, a été effacé par erreur avant qu’il ait pu être exploité par le BEA.CVR, preserved by the crew, was deleted by mistake before he could be analysed by the BEAThis is not the first time this has happened with Air France
So if you want to make something disappear ... give it to Air France .. they are like magicians

[Chris Scott]

Quote from flyhardmo:
"In the drivers seat I have no idea what the ST is doing but I assume it's counteracting the pitching moments with FWD trim."

Are you saying that the B777 has no EICAS or other indication of current stab-trim (horizontal stabiliser) position?

[ATC Watcher]

jcjeant : It is true thance since Habsheim the relations between AF and CVRs are quite complicated :E

On the other hand when one sees that every complete CVR in France is ending in " le Parisien" or "Le Figaro" suggesting totally innapropriate crew behaviour ( Quiberon/Proteus, Pau and more recently AF447, to name only 3 ) the temptation to make a " by mistake " might be high..

[flyhardmo]

Are you saying that the B777 has no EICAS or other indication of current stab-trim (horizontal stabiliser) position?

You can pull it up on the FCTL page but it's not something you normally look at in normal ops. The other indication is on either side of the centre pedestal.

From the FCOM
When pilot trim inputs are made, the PFCs automatically move the elevators to achieve the trim change, then move the stabilizer to streamline the elevator. Stabilizer motion may also automatically occur to streamline the stabilizer and elevator for thrust and configuration changes.

All this is very instantaneous and as a driver you don't feel the A/C trimming through the control wheel.

Porterhouse
My God it is quite loud and very characteristic in tone. If someone misses this sound (or its accompanying flashing red light) it means their mind must be in a la-la land.

Exactly it makes a hell of a sound and even the A/T disconnect gives an audible warning with a caution light and EICAS msg.
This was a classic case of mode (FMA) awareness. There is a big 'A/P' just above the PFD ( primary flight display). It's not very subtle

[Trackdiamond]

Airplanes that have no manual trim wheels or lack tactile feel to trim activity usually have trim indications in the normal field of vision of pilots or incorporated in the normal display page on EICAS/ECAM or centre miscellaneous advisory panel.Additionally some planes flash a pitch mistrim message on the FMA additional to a trim aural whooler if operated more than 1 second manually..and activation of normal pitch trim causes AP disconnect.
That said..Pilots are supposed to have well trained and observant eyes and display sensitivity and competency in instrument scan,interpretation,and flight control.The FMA is there to make life easier..not lazier for us..let's use it like our eyes and ears! And..when the FMA screams at us an alarm and we miss it through panic stricken motions...and we end up causing an upset both in the cabin and in the office..or after the final analysis give reasons for the crash detectives to promounce "pilot error"..and tabloids and hollywood to portray us for less than what we are supposed to be...

[wiggy]

You can pull it up on the FCTL page but it's not something you normally look at in normal ops. The other indication is on either side of the centre pedestal.

Agreed. The analogue (tape) trim indication sits on the centre console out of any normal scan. You'll also get an EICAS warning for any uncommanded trimming.

The 777 can certainly do some interesting things if you foul up a "high level" go around :\, but regardless of FMA scans, warnings and HST trimming all that matters at low alt is that if you hit TOGA the thing powers up and pitches up towards 15 degrees..... At least the "heavy" pilot was on the ball.

[totempole]

What I write is gonna get a whole load of barking from the PC crappies!

The777 is designed to be flown by 2 pilots. It should never need 3 pilots ( except during training with a trainee who is yet to be qualified as fleet/type functional, or of course a line check, audit check, etc ).
Introducing a third pilot ( no matter what the PC crappies aver, recommend or insist ) introduce an " unknown " in the form of " psychological pressure ", " snooping with later bar room trashing ", " nosey peeking " etc. Sometimes the third pilot input has a "STUN " effect or " STUMP " effect, or even an" IRRITANT " effect NO MATTER WHAT THE ANAL RETENTIVE CRM advocates aver!

This all affect the " psychological " state in a modern 2 man cockpit. You can all justify all you want, but the non-requisite 3rd man is a silent, stealth liability...the world of good it did in the OZ214, the best it could have done is to lull everybody into a false sense of security.

Had the AF471 skipper ignore the " stun " effect of the 3rd pilot's call and continued, it would have been a non event with a LAND 2 fail passive autoland. With runway sighted as noted, he could have even disconnected the AP and landed manually. The " STUN/STUMP " effect was largely responsible for the brain fart that cause him to activate the AT disconnect button instead of the TOGA switches. During a deeply mentally focussed Cat 3 approach, the PF's faculties are HARDWIRED to get input and callouts from the RIGHT SEAT PILOT, not a voice from the back! The Cat 3 training is such! Any voice from the back can and will cause a STUN/STUMP effect leading to all sorts of brain farts.

Now I am sure there will be dissenters and anal CRM crappies jumping onto this; believe me, YOU ARE ALL STUMPED! And you will remain brain stumped as long as you believe the crap of the so called " safety benefits of a third pilot " in a COCKPIT/TRAINING DESIGNED FOR 2 PILOTS!!!:ugh:

[olasek]

I can't speak for Boeing...so let's hear it for those that have flown bothIf you can't speak for Boeing why you were putting forth a proposition that Boeing's audible signal is perhaps too "soft".:ugh:
Also it is not a competition who's alarm is louder :confused:. It must be sufficient, which it is. Plus you completely disregarded presence of the visual warning as well.
To date there was no accident report which would suggest that Boeing's AP disconnect alarm is insufficient.

[Kilda Ste Hilda]

Totempole has a point there...I always never get to my A game whenever we have someone nosing around in the jumpseat no matter how quiet, smart or competent she/he is. They are a distraction however well intended the policy of having a third guy in the cockpit. Somehow they don't fit into a 2 person cockpit concept.

I have briefly flown in a 3 person cockpit in a A300 B4; the 3 person crew concept in that aircraft was superb because the training and procedures were developed for a very effective 3 crew management. However in a cockpit designed for 2 pilot operation, the third person is a potential stealthy liability however well intentioned her/his presence. Check flights are a different kettle of fish...any argument bringing that in is just being argumentative.

[Hogg]

Yes very valid and exact point Totempole. Even on 767 "jumpers" can confuse things with a 2 man/woman! crew.
When I flew 3 man (B727) it was different but my FEs had great eyesight! twas fab having a 3rd set of eyes being a crewmember not one thats out to score points as it happens nowadays. Always the small ones! I keep em out like my colleagues do now.

[exeng]

I agree with Kilda here.

The three crew concept with a fully integrated Flight Engineer/Third Pilot was the best operation I was ever involved with - it just worked well in my opinion..

When another Pilot is on the jumpseat/observing, or whatever, they do not seem to 'add' to the operation as they are not 'part' of the operation.

The 'third crewmember (F/E or Pilot) were part of the operation and therefore took some responsibility for what happened. An observer as such adds nothing and may be quite destractive.

[wiggy]

Working for an outfit who are very keen that the relief pilot is "involved" I'd have to say that I am in some agreement with the above, however looking back at the OP's bullet point 3

Relief pilot calls out ''Alarm'' the PNF calls out ''go-around'', the PF responds by applying full power to initiate the go-around, which disconnected the Auto-Throttle.

I'd suggest the die was cast once the PNF had made their call...

[glofish]

I very much agree with totempole.

Flying for an outfit who displays its infinite wisdom by not providing any seat possibility for augmenting crew but the two jump seats, i could give ample testimony of "stunned" or "stumped" operations induced by moronic utterings from an overcrowded cockpit.

I am all for having more hands on deck if the skipper so feels the need. My former employer handled it correctly by stating that it is his call to have the supernumerary in or out of his cockpit, but this implies a proper seat for the supernumerary (which by the way is the only civilised and respectful way for airlines).

As has been mentioned, the training in a two man cockpit layout is exactly that: Two men/women. OZ heroic handling or not (I actually don't think it increased safety, only flight time .....)
with the ever increasing fatigue factor on such cockpits, we should stick to what we learned and trained. Every deviation is a disturbance factor when tired and stressed.

No excuse though for this AF crew, only an explanation for the hick-up.

[falconeasydriver]

The NO DH was removed about three years ago - why, I have no idea.


As was explained to me, in simple terms by someone involved in the decision…"to pass liability to the operator/PIC"

[pax2908]

glofish, please clarify, are you suggesting that the presence of the relief pilot in the observer seat, during the final phase of this particular flight, might have been a distraction rather than an increase in safety?

[misd-agin]

Which manufacturer's A/P disconnect alert is louder? Both are loud. So one is louder and the other is louder. No one can miss either alert.

[VNAV PATH]

Quoting "zippy monster" on page 1:


My operator (European lo-co) has to use 25' for a CAT IIIB in France, despite using "no DH" everywhere else. I can't remember the exact reason our flight ops people gave to us word-for-word, but it is something to do with the French not allowing "no DH" operations.


Quoting "A4" on page 2:


The NO DH was removed about three years ago - why, I have no idea.


No , one french operator has been using no DH concept for many years on 747-400 and AF A 380 are using it too .


Seems to be a case by case regulation ; more exactly a company/aircraft type case by case.

[LYKA]

Think it's also airport specific too. LFPG AIP doesn't reference DH with CAT 3, LFLL does.

[Winnerhofer]

Airbus A320 CAT 3 autoland approach and landing at Paris Charles de Gaulle Airport - YouTube

[PEI_3721]

Has it been determined what caused the ‘No Land 3’ alert? This appears to be the initiating factor in the event.
Might the crew have been looking for a cause or contributing (distracting) factor?

[RAT 5]

Has it been determined what caused the ‘No Land 3’ alert? This appears to be the initiating factor in the event.

But it still begs the question what is AF SOP's. In a previous life & a different a/c LAND 3 = Fail operational autoland. LAND 2 = Fail Passive autoland. In both cases it was continue and autoland. Why the G/A? What was briefed before the approach. Some airlines had the policy Land 3 = CAT 3 & Land 2 = CAT 2. That would require a change of DH. Why for heavens sake did they have this idea? HE only knows, but they though it was simpler for the crews to remember. Either way it was still a continue and autoland.

I'd like to know the human factors involved in this event and what the SOIP's are. If the G/A was not necessitated by SOP's then why did they do it? Was the call "Alarm" an SOP by any pilot, and following such a call is the SOP to G/A?

I would have expected a CAT 3 approach to have been briefed fully and well understood by all. This brief would have included the "alarm" call if it was an SOP.
Was the G/A initiated by any confusion caused by such a call? When in doubt etc. etc. but it still begs the question……...

[wiggy]

In a previous life & a different a/c LAND 3 = Fail operational autoland. LAND 2 = Fail Passive autoland. In both cases it was continue and autoland.

Regardless of the RVR's passed at/before 1000 feet?

[A4]

@VNAV PATH - we always used to be able to CAT 3B NO DH ops in France until approx 3 years ago. Now, without exception, all airfields in France have 25' DH for 3B ops. Nothing changed in our outfit (major European LoCo), but suddenly we weren't allowed. So I don't see why AF A380 ops are exempt.....but thinking about it, may be I do :hmm:

[RAT 5]

Wiggy: If the company declared that Land 2 = CAT 2 then this would include the CAT 2 DH & RVR. If Land 3 changed to Land 2 <1000', and the last RVR given was <CAT 2 there are options. Ask for the latest RVR and decide what to do; say nothing and continue to CAT 2 DH; or what ever the company SOP said. There is no a/c technical reason to make a G/A. Remembering I'm not a B777, but ex-B767 pilot and I suspect they have a similar family of systems.

[wiggy]

If Land 3 changed to Land 2 <1000', and the last RVR given was <CAT 2 there are options......

Interesting.

Just for clarity in our case (EU-Land, 777 ops, but not AF) for a "No Land 3"/"Land 2" below 1000' we need to have been passed CAT 3A compliant RVRs at/before 1000' to allow continuation on down to the 3A DA (usually 50'), otherwise it is a go-around - there aren't any options (except for force majeur :bored:).......

[PEI_3721]

RAT, wiggy, et al, the points are all very interesting, but; “what caused the ‘No Land 3’ alert”.
Perhaps type rated specialists could speculate or add examples; is this type of alert a frequent occurrence?
Was a fault identified, was the aircraft returned to service NFF, has the alert happened since?

[wiggy]

wiggy, et al, the points are all very interesting, but; “what caused the ‘No Land 3’ alert”.
Perhaps type rated specialists.......

Err, well I may not be a specialist but as I hinted at I'm type rated on the 777 and have done a few uneventful autolands. In my time of both the 777 and the 744 which had (superficially at least ) similar architecture "No land 3's" were very very uncommon. If pushed to quantify it I reckon I've seen the message on approach no more than two or three times in over twenty years and never seen it during an auto land.

I don't see the point in me speculating in what can caused this particular "No land 3", there are a handful of failures I know of which will produce a "No land 3" but you'd need an engineer to come up with a full list...if there even is one. The AF engineers will have known the real story once they interrogated the Central Maintenance Computer post flight....maybe one of them will post here....

[JPJP]

Whilst the nuances of minima and LAND 2 are fascinating - the fact that two experienced crew nearly drove a 777 at Go Around thrust, straight down the glide slope and into the ground, is disturbing.

If I'd been on that Jumpseat during this multi second display of confusion, I'd have been a little tempted to reach forward and hit TOGA.

:sad:

[wiggy]

f I'd been on that Jumpseat during this multi second display of confusion, I'd have been a little tempted to reach forward and hit TOGA.



Funnily enough during a quiet moment in the cruise we looked at this the other day. We reckoned with the thrust levers forward ( which it seems they were in the AF case )the jumpseater would probably have to unstrap to get anywhere near the TOGA switches.........

[JPJP]

My length of my reach is directly proportional to the strength of my fear ;)

[JammedStab]

@VNAV PATH - we always used to be able to CAT 3B NO DH ops in France until approx 3 years ago. Now, without exception, all airfields in France have 25' DH for 3B ops.

I thought it was 14 feet. Does it vary by operator?

[acbus1]

Lots of the usual PPRuNe waffle. Only 200+ pages to go before the Mods close the thread down.

Maybe Air France should simply introduce a new flight deck announcement before departure:

"Bienvenue à l'émission de Muppet".

If I recall the exercise numbers correctly:
AF447 = Private Pilot Licence exercise 10
AF471 = Private Pilot Licence exercise 13

[PEI_3721]

Thanks for the contributions of the ‘informed’ knowledge.
The technical inquiry is if there would be any other indication as to why the system degraded, e.g. rad alt fail; and if so would there be any other flight deck alert, independent of the autopilot.
As for ‘no land 3’, I presume that this is actually a change of status (system integrity / capability) for which there is no alert other than the change of annunciation. Thus an alert ‘warning’ call might be unwarranted; whereas the change of state – ‘LAND 2’ call could aid system awareness and cue remaining capability.

The lack of GA – mis selection, contributed to the incident, and appears to be associated with a weak design / switch location. Just because everyone else manages to select GA does not mean that the design is adequate for all situations, especially those with surprise.

The crew procedures should be considered in the context of the situation. If the crew choose to autoland in clear conditions – a practice approach without LVP, then an unwarranted change of state might be anticipated. A pre landing briefing should cover the alternative courses of action – land 2 or manual, etc.
Alternatively, as in this instance, the autoland was with LVP but > cat3 , thus a change of state might be less lightly, but still significant; the pre landing briefing again should specify the choice of action, e.g. if land 2 then continue, if AP disc then GA.
For visibilities requiring full Cat 3 capability then the alternatives would be further limited, but in all cases the annunciation call / change of state could be the same – consistency.
The GA flight procedure might vary - auto / manual, but the monitoring should be similar. The PNF should be able to select a change of configuration, quickly scan pitch, power, speed, then check back to the configuration.
In these instances it is not that the tactical situation management was weak, but more of the strategic plan. What was the pre landing briefing, what monitoring was expected - relative importance of modes vs aircraft flight path, what parameters should be scanned, why, when (cf BEA AGASA study).
Are crews prepared to be surprised?

[wiggy]

As for ‘no land 3’, I presume that this is actually a change of status (system integrity / capability) for which there is no alert other than the change of annunciation. Thus an alert ‘warning’ call might be unwarranted; whereas the change of state – ‘LAND 2’ call could aid system awareness and cue remaining capability.


Yep, if "No land 3" appears (on the EICAS) the system is telling something has "broken" and you've lost a level of redundancy. If that happens to occur whilst on an ILS approach, "LAND 2" will also immediately appear (on the PFD) because of you've had a change of capability.

[PPL Hobbyist]

This is a question and maybe a little bet of speculation I hope you don't mind me posting this here, and please feel free to correct me if you think I am wrong.

In 1972 Eastern Air Lines, Flight 401 crashed. The probable cause was: "The failure of the fight crew to monitor the flight instruments during the final 4 minutes of flight, and to detect an unexpected descent soon enough to prevent impact with the ground. Preoccupation with a malfunction of the nose landing gear position indicating system distracted the crew's attention from the instruments and allowed the descent to go unnoticed."

However in the final report, http://libraryonline.erau.edu/online-full-text/ntsb/aircraft-accident-reports/AAR73-14.pdf the NTSB alluded to the possibility that the captain may have caused the autopilot to disconnect when the captain bumped his control column while turning around and ordering his co pilot down to look at the visual nose gear indicator, down a hatch in the cockpit. This only required 15 lbs, or 7.5 kg of force to deactivate the autopilot's pitch control. This might have been a spit second bump on the control column.

I am wondering if Boeing saw that report and decided that more assertive inputs were needed on the control column to kick the auto pilot off to prevent a similar accident from happening again? The NTSB made loads of recommendations, but nothing like that.

[RAT 5]

Guys, we're still arguing in circles around the crux of the matter. Why did Land 2 necessitate a G/A? The G/A procedure was messed up, but why was it performed in the first place? Perhaps someone inside AF can enlighten us as to their SOP's. If the G/A was not necessary then that decision is the root cause of the subsequent problems. Why, how it was messed up; autopilot disconnect forces, the position of switches, etc.etc. are all consequent of the initial action and worthy of discussion under another topic.
In direct discussion of this incident I'm curious if the G/A decision was appropriate yes/no.

[flyhardmo]

Guys, we're still arguing in circles around the crux of the matter. Why did Land 2 necessitate a G/A? The G/A procedure was messed up, but why was it performed in the first place? Perhaps someone inside AF can enlighten us as to their SOP's.

Section 2.3 of the report states the operator instruction for a CAT 3 final App.
A GA is mandatory for any warning or caution light.
An automatic landing was still possible as Land 2 indicates the autoland system does not have redundancy for a triple channel autoland.

The lack of GA – mis selection, contributed to the incident, and appears to be associated with a weak design / switch location. Just because everyone else manages to select GA does not mean that the design is adequate for all situations, especially those with surprise
Boeing spends a lot of money on ergonomics and design to make sure it feels and works right based on scientific research and not personal opinion.

[Capn Bloggs]

Boeing spends a lot of money on ergonomics and design to make sure it feels and works right based on scientific research and not personal opinion.
Arr, a good boffin. Pilots, not scientists, fly aeroplanes. To think that Boeing gets it right every time is a bit naive, methinks...

[porterhouse]

, contributed to the incident, and appears to be associated with a weak design / switch location
Nothing of the sort is in this report.

[PEI_3721]

Porterhouse – similarly there was no reference to the nature of a system ‘failure’ in the report which triggered event (thanks wiggy), thus the absence of evidence of a weak system design is not necessarily evidence of good design … etc, it depends on the situation, crew awareness, and choice of action.

Bloggs :ok:

Incident reports are not unchallengeable, the authors are subject to a range of human views like us (hopefully less so due to experience), but none of this should prevent us considering what if – to speculate, to learn.

During the next quiet moment in the sim, set the thrust levers in normal approach range. (Do all pilots know what that is by look / feel?) Then with the seat in ‘normal’ approach position, reach out and check the GA and AT Disc switches – differentiate between them; repeat the check when looking out of the window.
Fix those experiences in memory; they just might help in a future situation.

[porterhouse]

thus the absence of evidence of a weak system design is not necessarily evidence of good designInteresting logic!
Guilty until proven innocent. :ugh:
similarly there was no reference
Because most likely it was found to be non-issue, it is you who have to present evidence (not guesswork) to the contrary. :ouch:

[PPL Hobbyist]

Quoting Rat 5:
Guys, we're still arguing in circles around the crux of the matter. Why did Land 2 necessitate a G/A? The G/A procedure was messed up, but why was it performed in the first place? Perhaps someone inside AF can enlighten us as to their SOP's. If the G/A was not necessary then that decision is the root cause of the subsequent problems. Why, how it was messed up; autopilot disconnect forces, the position of switches, etc.etc. are all consequent of the initial action and worthy of discussion under another topic.
In direct discussion of this incident I'm curious if the G/A decision was appropriate yes/no. The report states the following:

2.3 Operator instructions for CAT 3 final approach
According to the currently applicable operational documentation (GENOPS,
TU777, QRH), “Regardless of the meteorological conditions, an automatic approach
and landing are always possible. However, in both cases the following conditions
must be met:
ˆˆ The aircraft equipment must have LAND2 or LAND3 status;
ˆˆ The QFU used must have an ILS CAT2/CAT3;
ˆˆ The automatic landing limitations must be respected;
ˆˆ The task-sharing must be that of CAT2/3 approaches.”
…
“The following events result in a single “WARNING” callout making a go-around
mandatory:
ˆˆ Any audio warning;
ˆˆ Any Master Warning or Master Caution (display lights on);
ˆˆ Any degradation in capabilities;
ˆˆ Any non-compliant flight mode annunciator (FMA) display or excessive deviation
below 500 feet.”
….
“The go-around is performed by the captain, the co-pilot monitors the flight path, performs the standard callouts and the announcement to ATC.”The following elements are based on recorded data (QAR, FDR) and accounts from the flight crew. The CVR, which was safeguarded by the crew, was erased by mistake before it could be read out by the BEA.How convenient! Since when do crew remove or handle the CVR or FDR from the scene of a crash??? I thought that was the job of whichever authority was investigating the case/incident/crash.

The information in the report from the pilots testimony leads me to understand that the captain could see the runway, and he wanted to continue the approach visually. The first officer had decided to follow the above SOP to the letter and go around.

What isn’t in their testimony is:

1: The captain never said he had control.
2. He never called that he had the runway in sight.
3. He never called “Continue” at minimum.

In my evaluation based on the official report, a contributing factor would be lack of communication in the cockpit. No CRM.

[tubby linton]

Yet another AF problem with manual flight.
Incident: Air France A321 at Paris on Jul 20th 2012, speed drops to alpha prot on approach (http://avherald.com/h?article=455c3eae/0000&opt=0)

How many incidents have AF had now with basic airmanship and flying the aircraft?? They really need to look at this before they kill anybody else.

[RAT 5]

2.3 Operator instructions for CAT 3 final approach
According to the currently applicable operational documentation (GENOPS,
TU777, QRH), “Regardless of the meteorological conditions, an automatic approach
and landing are always possible. However, in both cases the following conditions
must be met:
ˆˆ The aircraft equipment must have LAND2 or LAND3 status;
ˆˆ The QFU used must have an ILS CAT2/CAT3;
ˆˆ The automatic landing limitations must be respected;
ˆˆ The task-sharing must be that of CAT2/3 approaches.”

I ask again: therefore why was a G/A required? I read the 2nd paragraph of the post and it talks about warnings etc = G/A. But a simple down-grade from Land 3 to Land 2 does not require a G/A. On B767 there is no Master Caution for such a change; is there on B777?

[Samba Anaconda]

A lot of posts and " pontifications " in this thread reveal the emperor without clothes. Sigh, stfu if you have never had Cat 2/3 training or expertise!

Unless the skipper had called visual or continue, they are duty- bound to go around once the automation degraded to LAND 2. Their lowest minimums then changed to Cat 3A minimums and

A) they are below a gate ( usually 1000' AGL ) to change their " minimums "
settings.

B) the actual reported weather was below Cat 3A.

C) if they had not briefed on downgrading to Cat 3A.

Any of the above three provisos would have predicated the PM to call for Go Around.

I agree with totempole that the 777 is designed, certified for 2 man operations and the CAT 2/3 training is based on that. Introducing a 3rd person making a call out from a location the PF not expected to hear from in a " tunnel vision " scenario on a deeply focussed Cat 2/3 approach can have a " stun " effect leading to brain farts. I am sure a lot of pilots feel this way BUT WITH THE MODERN CRM CRAP, it is not politically correct to say so. You will deemed to be " not a team player, a maverick " going against new conventional " wisdom ".

[silverstrata]

What were the crew rosters like till the material incident date?
Were there any phisiogical causal factors that might have undermined the Captain who was PF?
Were there any stressful events enroute.


More to the point, when did they last practice a normal all-engine GA in the sim? In my experience, a normal GA is the least practiced and therefore the worst executed manoeuvre in aviation.

In training...
The single engine GA occurs at decision; is sedate due to low power availability; and is practiced endlessly. The all-engine GA is always unexpected; the aircraft can climb like a f...t in a bath; the MA altitude may be very low; and the manoeuvre is never practiced in the sim. The result being surprise combined with not entirely organised chaos.

Disregard the fact that this should have been either a (manual?) landing or a simple TOGA-push GA - it became more complex the second the wrong button was pushed. The reaction was then to go into a manual GA. Thats not a bad decision, as the PIC may not have realised why the a/c did not GA automatically when the (wrong) button was pushed.

The problem therefore comes down to an inability to do a manual GA under pressure. The primary reason for that may well be that the last time this pilot did an all-engine GA was ten years ago. The secondary reason appears to be a complete inability of Air France pilots to hand fly their aircraft. Unless all Air France pilots are given a 2-week gliding course, I for one will not be flying with them.