AF447 wreckage found

[SeenItAll]

Bearfoil and xcitation:

I fully agree that the pilots did not have a cakewalk presented to them. And in my previous posts I have always pointed out that things could have been quite chaotic on the flight deck, AB's stall warning was not their friend, and that from a very early point in the three-minute scenario, the plane may have been unrecoverable.

But that said, it does not appear that any concerted attempt was made to try the only thing that might have saved the plane once the sh__ had hit the fan at the top of the zoom climb, and that was sustained nose-down and trade some altitude for airspeed.

The question we face is "why?" Lack of training? Lack of non-AP flying experience? Lack of appropriate CRM? Or did the normal thought processes of the PF, PNF and Captain just simultaneously "freeze?" What I don't see is any way to lay the cause completely at AB's door. Surely the stall warning logic could be improved, but no piece of equipment is always infailable. Something will always go wrong, and that's why we have pilots. We have to figure out why they failed -- exactly at the time when the equipment/software failed -- and figure out what we need to do to fix it.

[stepwilk]

Is is not possible to rescue an unused DC-10 or DC-8 from the boneyard in Arizona, and fly it around Hudson Bay, say, with a retinue of trainees in the back, taking their turns manual flying?

Oh, yeah, that'll do it...sort of like my half-hour of CPR training three years ago. "Where do the hands go, again?"

[Rob21]

The big question

So the question now (for AF & AB) is:
We should give more training to the pilots or "teach" the computers to get out of emergencies?

Computer's way of "thinking" is very simple: "if this, go to that" and so on.
Of course, to "teach" a computer to solve emergencies, like UAS, will take some effort (and money).

Then, when an emergency occurs, the computer shall warn the pilot what is going on and "ask" him if what he is doing is OK.
Computers can "read" the whole Flight Manual in a split second.
The computer opens the page on "Unreliable airspeed", and says: "I am holding 2.5º pitch and applying 85% N1. If you don't agree, please press the DO NOT button".

The problem is, if we "teach" computers to fly these very complex stuff, it's actions will become normal, like today's protection laws are normal. Pilots rely on these laws.
And then emergencies will be "managed" by computers, under the supevision of the pilot. Nice, but if in a dark and stormy night the computer faces an emergency that is not listed?

If flight computing goes in this direction, computer engineers will be thinking that a computer can think. And I am not sure if they are not already thinking that computers can think a "little"...
But they can't (at least to my knowledge).

Wouldn't be easier to allow pilots to hand-fly a little?
Wouldn't be easier to give pilots more SIM time?
Maybe if we change "easier" for "more economical", companies will be "smarter"...

[Rob21]

Victor2, are you telling me ATC does not trust pilots any more?

Wouldn't be a nice deal hand-fly in an airspace where you can not gain or loose one hundred ft?

As I recall, in my commercial pilot flight check ride I was allowed to gain (or loose) only 50 ft on level flight, and turns. And I had to be quick going back to the right altitude.

Wouldn't be nice to show computers that pilots can fly too?

Thanks!

[Lonewolf_50]

Fox 3

No emergency procedures are verbally identified. No drills are done. No cross-checks of Attitude, or anything else, are called.
I'm not flying Air France again.

I have suggested to my brother, the international traveler, that he not use them at all until their process of correction (I guess a year or two) is gone through.

I have some experience lecturing on FBW flight controls, and instructing in a simulator. Aircrew tend to follow A drill, rather than do nothing. The usual reasons for following the wrong drill are rushing into action based on only one or two bits of information (probably the audio warning and the speed readout failure in this case), reverting to type (flying the aircraft/action they are most familiar with, e.g. TOGA loss of IAS)

YES. Seen it as well. Hell, done it myself.

Power, Attitude, Trim.

Unless in direct law, the trim is provided by HAL.

The PF is flying for 15 degrees NU, power is variable as he's not sure of speed. There's no trimming, probably because he's forgotten it.

HAL is trimming, not the pilot. That's a design feature. No coolie hat on the stick to trim with ...

In fact, I think he's going for at least 12 degrees Nose-Up attitude right from the start of the problem.
Question please, doesn't everybody memorise the Power/Attitude combinations for t/o, climbout, cruise, descent and approach any more?

This situation should have been met with.
Take control
Set 2.5 degrees nose-up attitude and 90% power.
Cancel warnings, identify loss of IAS readouts.
Get PNF to confirm diagnosis.
Run through drills.
Recall Captain to flight deck.

What it's met with is:
Take control
Fly unreliable IAS drill calling for 12/15 degrees nose-up attitude
Push power up when stall warner sounds (instinctive)
No correct communication with PNF

Succinctly put.

I do think the Captain's failure to formally divide the tasks before leaving the cockpit is important to the last point.

From a formal perspective, yes, I agree, however, given the reaction of the PF in this case, as the various bits of CVR are dribbled out to us, I suspect it may not have mattered.
There is some vidence of "do the wrong drill, and do it now" as the first domino to drop after the Auto Pilot stepped aside.

RWA

they didn't have speeds - and both the Flight Directors kept cutting out at intervals, so very probably they didn't have working artificial horizons either for long periods

Are you supposing that the Artificial Horizon is linked to pitot static system, or do you refer to the FD function rather than the "nose pitch" and "wing roll" attitude indication?

Does it exists a procedure that could have allowed the PNF to take control by strengh after the PF has re-taken control without annoucement yet the PF had admitted having lost the plane ?

Reach across the cockpit and punch him in the head with some force? That's a bit old school, but it has been known to result in a change of controls. (Don't ask me know I know ...)
JJFFC

2:11:37 : PNF takes controll after the PF having said 'I have lost the plane"
2:11:39 : PF re takes control from the PNF witout annoucement
2:11:42Captain enters
2:11:45 : End of CONTINUOUS stall
So the capt could believe the PF has save the plane but from 2:11:39 the PF is no longer a professional pilot.
In an army, had he survived, he proubably would have go to the martial court for insoumission.
Before 2:11:39, the PF was a bad pilot. From 2:11:39, he could well be judged as a criminal...

Ouch

[infrequentflyer789]

Originally Posted by predictorM9

I don't think that the airbus design philosophy of a sidestick without feedback has caused the accident. Also, how do you generate the feedback when the speed data is lost? You can synthetize something based on the AOA vane sensor, but the fact that there was icing in this case precisely meant that all data was suspect. IMHO generating a feedback based on faulty sensors is even worse than no feedback at all.

You are right about control feedback being wrong in UAS (it can only be thus), but Conf wasn't talking about that sort of feedback but rather visual feedback of what the other guy is doing with the controls.

Whilst I agree that is a potential disadvantage of side stick, I actually don't think it is relevant here - PNF seemed to know perfectly well when PF was climbing or overcontrolling roll (in fact he seemed more aware than PF himself).

See also Bearfoils post(s) on the other thread: http://www.pprune.org/6619313-post1529.html

But that doesn't remove the fact that the stall alarm design is completely unnatural and misleading, which is a big design flaw.

It is, but I believe there are reasons behind that design and it would be foolish to just call for changes to it without understanding all the logic fully. Clearly the designers never thought a bus would be <60kts in the air, or maybe they thought if it was, it was already lost (may be right), or that the crew must have had enough warning of stall already by that point.

In this case, by the time the SW goes on vacation, the AOA is up at 30deg or more, airdata is going haywire due to high AOA, and the crew have seconds to recognise stall (where they didn't in previous minute of continuous warning) and push nose down over 30degrees (from full NU trim) to get anywhere close to a recovery. With no outside visual reference. And we're so far outside flight envelope we have no real idea how the airframe would have responded to controls. Was it really, practically (not theoretically) still recoverable at that point ?

[grimmrad]

"How can AF explain that there is no-one clearly in charge in the cockpit after the Capt leaves?"

Wouldn't one assume that that is automatically done through seniority or hours on type? Are there no SOP at AF that clarify that?

[jcjeant]

Hi,

grimmrad

Wouldn't one assume that that is automatically done through seniority or hours on type? Are there no SOP at AF that clarify that?

Read this ... maybe it can help .....

http://www.pprune.org/tech-log/45687...ml#post6619789

[bearfoil]

I am no stranger to simplification, so here goes. It strikes me that what is involved here is less to do with hand skills than Pilot Attitude.

To accept that manual control failed here is to accept too large a wedge of the format. In NORMAL LAW, for goodness' sake, all pilot inputs are Manual.

No, the failure appears, at this early time, to be re: PREPAREDNESS. I will wager that PF had all the drills down, the changeover down the flying down.

He was not ready, and his skills were not ACCESSIBLE. No more "When I learned.....", please. He knew, and he knew he knew.

The wheels came off, and catch up is impossible when one's talents are not available. I am weary of the Mechanicals. There is likely some interesting metallurgy and design issues here, but I think they will be found secondarily. I love to discuss design and fabrication, and I leave open the window for bent or unstowed Spoiler, or compromised Jacks and Fuel issues.

Not too much the matter there. The ship seemed fine at 2:10:05. What happened next (and is still happening, one fears), has to do with some basics, and that, as a pilot, is embarrassing.

I am afraid I have expressed my feelings to all my family. Until further notice........check your options; Air travel is a luxury, address it as such.

[xcitation]

Bearfoil
So have you also told you family to stop driving, drinking and other much riskier activity than flying? Don't make me pull out the stats .

[jcjeant]

Hi,

Bearfoil
So have you also told you family to stop driving, drinking and other much riskier activity than flying? Don't make me pull out the stats

Human life (personnal life) is not driven by numbers or statistics ( we are not already computers).. but by personnal feelings ...
So I can maybe feel in better security in my car on a crowded highway .. than be seated in a AF plane

[bearfoil]

For many reasons, not the least of which I find to be BEA propaganda, Nonsense PR from All Three principals, and a lack of sincere consideration, Long haul flight on AF has become (unwittingly) problematic

OPINION: The fact that remote mathematical odds are thrown about, I consider it whistling past the graveyard. The loss of an autopilot for any reason (and I have NOT accepted UAS or ICE had effall to do with it) means that one of three potential Pilots may be responsible for an aircraft that has ventured into mathematical probabilities of a demonstrable ONE IN THIRTY THREE HUNDRED OF crashing to the surface of the EARTH.

The odds of losing an auto pilot (disconnection as well as through malfunction) might be, one in one hundred? The chance of encountering bunk MET in ITCZ is likely what, ONE TO ONE?

So, 1/33 UAS
1/100 AP LOSS
1/1 BUNK MET

I see a chance of ONE in 3300 of coming to serious grief on AF.

Tell me again how utterly safe AVIATION IS?

[BRE]

As this thread keeps growing faster than I can keep track, bear with me if these questions have been asked before:

1. Had the PNF pressed the takeover button, would the PF have been able to re-take control?

2. Is it normal for a crew at AF:
- not to be using standard calls (maybe even in English), but informal French throughout?
- to be addressing each other with the familiar "tu" in spite of the age and authority gradient?

3. Just for comparison, what is the official cockpit language in say KLM or LH?

[RWA]

Confess to still being ‘bugged’ by the issue of the THS going to ‘full up.’ Unfortunately the BEA doesn’t say much about it in the latest report – and what it does say is rather contradictory.

Page 10 - "At 2 h 10 min 51, the stall warning triggered again. The thrust levers were positioned in the TO/GA detent and the PF maintained nose-up inputs. The recorded angle of attack, of around 6 degrees at the triggering of the stall warning, continued to increase. The trimmable horizontal stabilizer (THS) began a movement and passed from 3 to 13 degrees pitch-up in about 1 minute and remained in the latter position until the end of the flight."

That’s just a repeat of the section in the earlier report.

Page 75 - "Despite some nose-down inputs, the PF maintained nose-up inputs overall. Pitch attitude fluctuated between 11° and about 18° and the angle of attack between 11° and 23°. The THS began a movement that was consistent with the PF’s inputs and reached 13° nose-up about a minute later. It should be noted that in alternate law, the auto trim is still active. On the other hand, it is difficult for the crew to know the trim position and there is no warning to the crew that it is moving."

The important thing there, to me, is that autotrim was still in operation.

Page 77 - throughout the flight, the movements of the elevators and the THS were consistent with the pilot’s inputs,

That’s the contradictory bit. As far as I can see, from the chart on page 111, the THS movements simply weren’t ‘consistent’ with the pilots' inputs at all. Indeed it appears only to have made the one movement – to ‘full up’- during the whole episode.

On Page 111, the pale blue line under ‘STABILIZER POSITION’ appears to show that it didn’t move at all at first. Then, at about ‘2.10.50,’ it began to pull up. This movement seems to have started during a period when the PF (shown as ‘CAPTAIN’) was putting in quite a lot of small movements both ways – none of which seem, on the face of it, to have been large enough to trigger such a huge movement by the THS? The THS movement finished (i.e. reached ‘full up’) at about ‘2.11.50. From then on the THS seems just to have remained at ‘full up,’ and didn’t respond to any of the further stick inputs made by both the pilots?

Can anyone with more knowledge of the aircraft and its systems – bearing in mind that the BEA says that autotrim was still in operation - explain why the THS would have behaved in this (on the face of it, very strange) way?

Is it possible, for instance, that the autotrim was responding not to the pilots’ inputs but to the very low Indicated Air Speed, or to other irregularities ‘reported’ by the sensors?

[bubbers44]

Sounds like Airbus safety could be enhanced by low time pilots getting high altitude hands on flying so they don't zoom 3,000 ft into a deep stall when the autopilot fails. Also explain to them that holding back on the SS once in a deep stall is not a good idea. Just a thought.

[zekeigo]

Authority gradient??? In Europe the FO's think and act as they have the same status as the captain, and the EU captains accept that. Had many attitude behavior problems flying with European FO's, but when the S hits the fan they say "You're the captain, it's your problem"…
Until they understand the "Leaders and Followers" rules, conflicts in the cockpit will exist and flight safety will be jeopardized.

[mm43]

Originally posted by RWA ...

Can anyone with more knowledge of the aircraft and its systems – bearing in mind that the BEA says that auto-trim was still in operation - explain why the THS would have behaved in this (on the face of it, very strange) way?

I believe the outcome is due to the THS being 'g' driven. In other words, the SS command was given that allowed the actuator to take the elevator to maximum NU, but the expected 'g' didn't occur and the THS moved to (try) correct that situation.

In explanation - the THS will always attempt to neutralize the position of the elevator, and if necessary readjust according to elevator demand.

If I am wrong, you will soon find out.

[HazelNuts39]

My understanding from the Perpignan report is that, with autotrim in operation, the THS only moves when the elevator goes past the neutral position: nose down elevator will command the THS to move more nose down. The elevator follows the (g-driven) SS orders. So a small nose-down command from the SS that causes the elevator to move from nose-up to less nose-up without going to nose-down will not cause the THS to move.

[rollnloop]

1. Had the PNF pressed the takeover button, would the PF have been able to re-take control?

Yes, pressing the take over button himself. Last one pressing has commands.

2. Is it normal for a crew at AF:
- not to be using standard calls (maybe even in English), but informal French throughout?

No. This shows much stress.

- to be addressing each other with the familiar "tu" in spite of the age and authority gradient?

Standard use. At AF the last in F/O will use "tu" in cockpit with the most senior pilot in airline if they're to fly together. It spares a few syllabs. After all, "you" can mean both "tu" and "vous", it's practical to decide for only one word and "tu" helps forming a team spirit. All crew uses "tu", except in cabin where "vous" is used. Sometimes F/A use "vous" talking to captain, mostly when they're white/grey haired.

[eSpoiler]

Horrified

As a very part time passenger and model plane pilot, I have followed this threads with interest. The CVR and FDR data is horrifying - I would have thought that professional pilots would have behaved very differently in such a situation. I could imagine myself as thrown into that situation and doing no worse than the PF did - perhaps better when it came to stick inputs. That is scary.

On a recent flight I noticed the issue mentioned in a couple of comments - that is the loss of sense of aircraft orientation. I could have sworn the plane was level but when I looked out saw thet the wings were banked and the plane was turnning left. I can only imagine how hard it must have been for these pilots to sense what the plane was actually doing withouit any horizon on a pitch black night, with what appeared to be unreliable instrumentation.

I guess that is when training and professionalism is meant to take over...