HarryMann

More like... PF heard a 'different' airflow sound, due to grossly high incidence, never heard it before , incorrectly associated it with high airspeed sound ?

HarryMann

What is 'auto dropout' ?

Mike X

Sorry for any assumption.

'When the automatics leave the party'.

jcjeant

Hi,

It was not about this message but about the message with discrypancies about CV's in report 1 and 3
BTW yes my position remain the same about Air France .. suffice to read Vienna and Montreal conventions rules for know who bear the responsibility in such event

HarryMann

Mike,

Lots of peeps have mentioned the initial response to A/P disconnect, and PJ2 having direct experience of the type, went to some trouble a while back explaining his approach, and the whys and wherefores of leaving alone as far as possible as there may be better things to be doing, although establishing wings level may indeed feature

takata

Sure, a brutal right bank increasing to 40° while full left stick was maintained continuously for 50 seconds, and a sharp pitch change of -28° nose down while full aft stick was maintained continuously for 40 seconds, is certainly related to aircraft stability when stalling with no speed worth to be mentioned.
But what changed the previous equilibrum (16° NU, TOGA, continuous oscillations in roll) seems to be the sharp reduction of thrust.

No no. PF stick did not move from full left and full aft from ~0211:40 to 0212:15, while aircraft started to pitch down from 16° nose up after 0211:45 and reached 12° nose down in about 10 seconds (0211:55).
PF declared feeling a "very high speed" at 0212:02, and in fact, the aircraft accelerated by ~50 kt but from an airspeed close to zero.

It happened right after the captain came back on the FD. He should have been quite puzzled to see the horizon moving down and up again to zero, while V/S increased from 150%, with airspeeds/Stall Warnings disappearing (NCD) and reappearing (valid)...
TOGA put the aircraft back in stable attitude, close to zero pitch for a while and wings could be put barely level again.

bearfoil

As flattering as it is to be confused with PJ2, My remark had to do with PF's first input after a/p loss.

With NOSE UP, (and LEFT ROLL), if PF was impatient for NU response after initial pull, he may have accentuated it with more pull. The a/c is heavy, and having taken manual control in these conditions w/o alot of experience (did anyone?), his exaggerated pull may have caught the airframe responding, and been responsible for the 1.6 gee rotation?

The traces tell this?

Indarra

Long time lurker, retired intelligence analyst. Not a pilot but I am trained to think out of the box and approach issues laterally.

I would like to take up more explicitly the suggestions that a few others have made about the visibility of PF actions with the sidestick. I do think care needs to be taken in considering this factor. Unfortunately, it perhaps has become a taboo or “jump over” issue because of unproductive AB vs Boeing arguments, and it has tended to be quickly dismissed in this forum. Similarly I do not see it picked up in explicitly in the BEA reports.

But it seems to have been a significant factor in this accident at two points, both the incident start and subsequently in the cockpit in finding a resolution

My understanding of the operation of a two person crew in aviation is that the PNF should monitor the actions of the PF and express views and act when appropriate. In this case we had a relatively inexperienced crewmember as FP and the most experienced person on type as PNF. However the PNF does not seem to have been privy to the sidestick actions of the PF, so that his checking role was stymied. Similarly the Captain when he arrived at 2:11:43 could not see the sidestick actions.

2:13:40
PF: Mais je suis à fond à cabrer depuis tout à l’heure
Capt: Non non non ne remonte pas

The BEA is undertaking a study with psychologists and psychiatrists but I hope it gives consideration to the reactions of the PNF and Captain given the incompleteness of the information that they had regarding sidestick operation. Did they automatically tend to assume the sidestick was being handled in the way that THEY would have handled it, or did they tend to omit a focus on sidestick operation because it was not visible?

What I am getting to a need for a simple panel display so that both pilots can easily view the sidestick operation. For example, a simple color display showing one or two of the following colors as appropriate: blue (up), orange (down), red (left), green (right), white (neutral). This should inform and hopefully empower the PNF (and anyone in the 3rd seat) to the necessary degree.

A number of incidents come to mind in this connection. Of course, the availability of a yoke didn’t ultimately save EgyptAir 990 or Silkair 185. There was an earlier suicide attempt with JAL 350 (DC8) in 1982 on landing approach. I understand that in that case PNF grabbed the yoke and forced it back, reducing the crash impact and allowing the majority on board to survive. I also wonder with AirBlue 202 whether a sidestick display could have given PNF a few more seconds to realize what was happening. Visibility of sidestick actions will surely be an issue from time to time in the decades ahead.

takata

It seems that autothrust reduced N1 from ~100% down to 85% just before disconnection (reduction caused by turbulences or Mach selected, diconnection by UAS) and it stayed there until it was unlocked by the pilots (~0210:25 -> CLB).
This caused at first a pitch down of about 3°, down to zero or slightly below (~0210:00 - 0210:05)... which could have delayed the aircraft response to PF NU orders.
Thrust increased to CLB between 0210:25 and 0210:35, from 85% to 100+, it would have added some pitch up before the PF tried to reduce aircraft V/S.
N1 was reduced again to 85% at 0210:45, quickly followed by TOGA after 0210:51 (at second stall warnings sequence).

At AP disconnection (0210:05), she started wings level and she rolled to the right at 4° per second, before first PF imputs two seconds later.

Lazerdog

I wonder if there is a software routine that could have detected the stall that the pilots were unable to? High AOA, sufficient thrust, yet altitude is decreasing steadily for a long period. (Have the computer do the instrument scan and suggest action.) I still think training and new procedures are required to get crew back to being active pilots. If I am on a passenger flight, I sure hope the folks up front have had a crusty instructor pilot throwing everything at them so that the right scans, observations, and responses are learned and quickly applied when things go awry.

STICK N RUDDER

"My students in a J3 cub with three hours knew you couldn't get out of a stall by holding the stick back''

If you ever did fly with this kids some of training programs only wants you to teach them how to recover on the first sign of stall then you got to teach them how to fly jets in pistons...

grity

mm43 I read the ATHR disconect 2 sec later 2:10:07

in the acceleration we can seperate two different parts IMO, the average value and the short peaks

the average value will corespond to the main movements of the heavy airbus, the shorter peaks shows the turbulent influence

between 2:10:07 and 2:10:17 the average acceleration get up from 1.0 g to 1.3 g (and fall back to 1.0 g in the following 10 sec) I think this is caused to the elevator move to -1.0 deg (the pull by the PF) also the g-protection can play a rule here

but the acceleration peaks in this time ( 0.8 ...1.6.....0.85 .......1.5 g) indices strong up and down drafts

this are followed by changes in AoA ( 4....2.....5....2....6) and the stall-approach-warning worked perfekt ("stall" if AoA>5.8 deg)

but I think that the air flow over a wing even with this higher AoA will not break down so fast, (flow-seperation needs his time)

so the wing with AoA=6 is still not in a stall, but with this speed he create a very strong lift and can create strong momentes

my read is that the turbulences can caused a lot of this first pitch up......

mm43

Yes, no doubt about that, and I'll fix the graphic. Thanks.

PJ2

Hello, HN39;
Yes, I agree with your assessments concerning conditions just prior to AP-A/THR disengagement.

On your question regarding disengagement of the AFS, I suspect that the AP - A/THR disengagement occurred two seconds before the loss of airspeed because the Mach parameter dropped to M0.30 for one sample, two seconds prior to the CAS dropping and the engagement conditions for the AP-A/THR-FD systems were temporarily not met.

On the flight conditions, aircraft response etc, in my view the variations in roll, pitch and engine thrust would be considered minor and common in light to more than light turbulence. Even in moderate turbulence at altitude, manual flight in Normal Law is not a problem. I haven't done it in anger in Alternate Law but in the simulator it is a non-issue but as I've mentioned, one has to be gentle. I think the control inputs by the PF were exceedingly brisk and resulted in the 1.5g initial vertical accelerations. But eight-degrees of roll in two seconds, (as noted in the BEA Report's graphs near the beginning of the Report), is not in itself a "control" issue, nor is the slight pitch down then recovery by the autoflight system just prior to disconnection.

What I wish to convey here is, with a loss of auto-flight capability, one simply flies the airplane, maintaining stable flight while sorting out why the system disconnected, then reconnecting when re-engagement conditions are met.

On why the PF's seatbelt may not have been fastened at impact, I believe Mr. Optimistic commented earlier that the PF may have been preparing to get out of his seat to let the captain in. I think this is a reasonable conclusion to draw at the moment - I can see something like that occurring, but of course there is no evidence thus far to support this suggestion.

jcjeant

Hi,

I don't think this is a reasonable conclusion ....
The reasonable (and logic conclusion) was that the captain take place in the left seat (due to his qualifications and common usage)
It's certainly the first place to go who come to his mind if he want to take command
In the CVR report .. nothing (not a word) about captain to take command.
If this was the case .. sure the BEA had reporting the wordings .. as it is a important event ...

AlphaZuluRomeo

Thanks for enlighting / correlating the facts, HN39, much easier to read (for me)
About the AP discon : AP2 was engaged, it's based on the right side "chain" (pitot 2, ADR 2, IR 2, EFCS 2...) unless I'm mistaken. My guess is that the pitot #2 was the first to freeze, the first to give an value dropping, unmatched with the 2 other speeds (ADR1, ADR3).

@ bear (post 1115) : I'm ready to accept other causes than UAS/icing... providen an alternative theory shows itself either probable/logical or demonstrated (facts, recordings...)

@ grity (post 1120) : Thanks for the pic with the stick moves, very valuable.

xcitation

Having lost the IAS due to pitot they then did a zoom climb. This traded their velocity for altitude gain and resulted in the slower speed and stall problem. We can presume that the initial zoom climb was accidental as no additional thrust was applied, no announcement given.
For me this is the most puzzling part of the incident. Perhaps an accidental climb whilst trying to level wings in the chop. But +2500 feet? When instrument scanning you do your utmost to maintain them at the desired position. Corrections are done frequently and in very minor stick adjustments.
When they lost the IAS why did one of them not glance at the ground speed for a sanity check? You have it right there so why not.
They would have seen it change from >400kts to <50kts. C'est une grand problem, non? Perhaps the PAX seat back display would show ground speed change and they could even deduce a stall.
I say this not in cricticism of the pilots, but more as a question of what would I do if I lost my speed. Maintain attitude and power and treat her gentle as egg shells to buy time so my buddy can figure it out and call it in. Keep my scan going because that's all I have. No doubt it is more difficult because of the chop and warnings.
My gut say there is more to the puzzle. However I have a feeling we will never know. How do the more knowledgeable ones here feel about the PFD switch by the PNF and BEA want to record the instrument panel?
I was hoping to get an answer why the PF nosed up so radically. Reading the report 3 it looks like not yet, perhaps not ever.

Owain Glyndwr

I agree you have to pitch down enough that the aircraft accelerates. The curves of course only relate to AoA and aircraft trim, but in practice, until the dynamics have sorted themselves out and a new flight path angle stabilised, AoA and pitch move together. I think that when you have pitch down from some steady (stalled) state, the first thing that happens is that the flight path angle becomes more negative at more or less constant AoA, and the aircraft starts to accelerate. This would be a steadily escalating effect I think, so it would not be necessary to pitch all the way down to 45 deg nose down in the first instance. But then I am not a pilot, so perhaps someone else could comment?
The AIAA curves were just tunnel data, so no thrust effects were included. There certainly would be an effect such as you describe, which would depend on altitude as well as airspeed. (Thrust increases with decreasing altitude and decreasing airspeed; the overall aerodynamic moments other than thrust decrease also with airspeed, so the pitch effect from the engines is biggest at SL and low airspeed.
I don't think I agree with the "barely no airspeed" bit. To me it seems that the credibility of the measured CAS is going steadily south as AoA increases and above about 35 deg AoA you cannot believe a number of it. I say this because it seems to me that the position errors are wholly unknown for any AoA above the flight tested stalling value and that when you are looking at AoAs of 30 deg or more coupled with roll and sideslip excursions the 'measurements' are completely unreliable. Where the measured CAS can be relied upon the recorded ground speed converted from TAS to CAS is in pretty good agreement with the measured value (and yes, I am aware of wind effects, but the AI analysis needed only a 15 kt wind correction and this, I assume, would have been 15 kt TAS, so much less as a CAS). If you carry that process forward to the regions where the recorded CAS is crazy you find that the aircraft was fairly close to 110 kts CAS all the way down.
A back-of-an-envelope sum for 110 kts and FL350 suggests that going from Flight Idle (essentially zero thrust) to TO power would add 8 degrees to the AoA. In the only sequence where power was reduced to idle the pitch changed from +15 to -10 but unfortunately one cannot say what effect that had on AoA because the AoA recording has hit the stops, as has the flight path angle recording. If you were VERY imaginative and believed the AoA out of the IRS, you might persuade yourself that the AoA was reduced from about 46 deg down to about 39 deg

jcjeant

Hi,

Can gurus explain to a graphics dumb reader ......



What about the movements of elevator and movements of stick ..
Do you see the elevator mimic the stick commands in the 20 - 30 first seconds after AP tell goodbye ?

paull

Since in Normal law it goes where you point it, I guess in the vast majority of cases pitch up equates to going up so I can understand the lack of precision most of the time, but surely if the PNF or Captain did know what was happening (stall)they would have said "push the nose down (or forward)" rather than "Go Down".

1- Can we read anything into PF and Capt's use of Up/Down rather than nose Up/Down that would let us know what they understood the situation to be and when?

2- In a similar situation would explicit references to pitch be more helpful to the PF than saying Up or Down? After all, if I were frightened and could see from the rate of descent that I was going down fast, someone telling me to go down would not be something I would want to hear. A statement of "Get the nose down" might get through to me. (There again "The wing is stalled, get the nose down" would have probably been better still!)

I think of stall recovery as being like turning across the slope on skis, you know the theory, but the first time you turn back towards the fall line and it gets worse before it gets better, most of us fall on our backsides.