AF447 wreckage found

[Lyman]

Coagie. Coagie. The THS was effectively dormant for the climb. The PF accomplished the climb in ordinary and unassisted ways? He also attempted to stop it. The THS enters, stage left, basically at STALL...

[Coagie]

The PF pulled back on the stick for a long while, causing the trim to increase so much. It wasn't just the extra thrust from the underslung engines causing the sudden climb.

[deSitter]

Recovery?

Can we start a thread on how these guys could have recovered? What immediately occurred to me was to generate asymmetrical thrust to perhaps roll over, or at least enough to get into a steep bank that would lead to a nose-dive. Once the nose was really down the problem would be solved. You would need to physically understand throttle response to time things.

[Clandestino]

Can we start a thread on how these guys could have recovered?

Eeeerm.... by pushing the stick forward?

Ladies and gentlemen, did it ever occur to you that the lift equation, Cl vs Cd curve, power required vs. power available et al. are not just some scribbling on paper or PDF that are to be learned by rote to pass the ATPL exams but extremely accurate descriptions of principles that keep us flying?

I am going on my silly sailing vacation, since I don't fly DC-4 but Q400, it will be four days instead of three weeks. I wonder what you'll make out of this thread by then.

[Lonewolf_50]

I am going on my silly sailing vacation, since I don't fly DC-4 but Q400, it will be four days instead of three weeks. I wonder what you'll make out of this thread by then.

Hash?

Fair winds and following seas, Clandestino, hope the time at sea is a joy.

(Also, suggest you avoid Eastern Seaboard of the US, seems that Irene has arrived and is blowing hard).

[iceman50]

TheShadow

Not quite sure what you are trying to say in your post but there are so many inaccuracies in it that it was not worth the time reading it!

The discussion around CLB thrust and 5 degrees is more to do with the climb phase up to CRZ level. Below 10000' it is 10 degrees. Cruise level the attitude it is around 2.5 degrees, the NORMAL cruise attitude so there is NO need to go to 5 degrees, any thinking competent pilot should be able to work that out, otherwise they should not be in the flight deck.

When the A/THR disconnects it goes to THR LOCK and keeps the thrust set at the time until the pilot moves the thrust levers, so no great dramatic thrust pitch up is induced. The autotrim will only work to compensate for PILOT input and it went up due to pilot demand!

deSitter

There is No need to start a thread all they had to do was get the nose down. If they had held the input the nose would have gone down and the autotrim would have then assisted with a sustained nose down input. They could even have helped by using manual nose trim themselves[/COLOR], as taught in UA recoveries. As for trying to roll / yaw the nose down.

Lonewolf

If you can explain to me, in plain language, why a stall warning system goes dormant while the aircraft is stalled, in flight, and tell me why this is allegedlyl a good design, I'd sure like to hear it.

It only went dormant <60kts IAS, what manufacturer / regulator would think that a properly trained Airline Pilot would get a Transport Category A/C into a deep stall at less than 60kts!!! Having ignored the warning for >50 seconds prior to that.

[Lonewolf_50]

iceman, you didn't answer the question. I understand how it happened. Given that a pitot static system can fail and give erroneous inputs (they can, this has been known for a long time) I was asking for a defense of the design decision, considering that a number of aircraft features are routed through a weight on wheels switch ... there may be a solid answer, but I've yet to see one.

[AlphaZuluRomeo]

Hi

Originally Posted by Clandestino

Procedure that required setting 5°/CLB by heart was in force at the time of the accident. Page 59 of interim 3 report in English refers.

That procedure seems to have been badly phrased, as there are so many questions about it.
Anyway, I think the way its writer intended it, it was :
- if there is a danger for the flight safety => apply memory items (which is as you quoted is, 5°/CLB), then refer to QRH for the "fine tuned" attitude/thrust to apply, then manage the failure.
- if there is no danger => refer to QRH, no immediate action (i.e. implicitly, fly ahead until you know which pitch & thrust to apply), then apply it, then manage the failure.

In AF447's case, there was IMO no danger. This would explain why, as soon as june 2009, AF published a note "reminding" to its crews that they were not to apply the memory items in high altitude cruise.

For the rest (i.e. is or isn't 5°/CLB "dangerous" and/or will it or not trigger the stall warning in HA cruise), I have no clue and won't comment.

[HazelNuts39]

Originally Posted by Lonewolf 50 #3244

Doesn't it bother you that you got a stall warning for no good reason by keeping the nose up in a pitch climb for no good reason?

For no good reason? Two apparently very brief occurences of stall warning were observed at 02:10:10 and 02:10:13 (page 29 of BEA#3). The reason for those is clearly shown in the traces for normal acceleration on page 42 (Note that normal acceleration is proportional to AoA at a given airspeed). The purple line is the DFDR recording, the blue line is what Airbus has calculated for the pilot's control input without turbulence. At 02:10:10 the pilot pulled slightly over 1.3 g, and a gust increased that to slightly over 1.6 g. Similarly at 02:10:13 the pilot was pulling 1.4 g (increasing towards 1.5 g at 02:10:15), and again a gust increased the pilot commanded 'gee', this time to less than 1.6 g. I would expect the pilot to recognize the 'gusty' origin of the brief warnings, and he should have done nothing except maintain pitch attitude. Whether that is 2.5 or 5 degrees is really of secondary importance.

[Welsh Wingman]

At least until we have received the final and more comprehensive BEA report in 2012, could we please not get sidetracked on Boeing v Airbus ideological debates?

The main aviation safety issue in recent years has been loss of control, with the main focus in relation to automation and the "human interface" (specifically concerns over training and the absence of line manual flying), and across all aircraft manufacturers. Even to a Boeing veteran like myself, there are only two "Airbus specific side issues" that this thread has thrown-up and neither should have resulted in themselves in the downing of AF447.

(1) The historic role of Airbus in creating the impression to line management that "planes fly themselves", even if partly inadvertently, and the knock-on effect upon the training culture.
(2) Several additional "complications" in an Airbus cockpit when things go awry, if cockpit discipline is not tight i.e. feedback issues (e.g. SS v control column, and the throttle and trimming) and the PNF is visually less aware of how the PF is by-hand flying. I suppose you could add under this heading the <60knt stall warning design "issue".

This is a training/human behaviour issue otherwise the temporary UAS on AF447 would have been temporary by-hand P+P flight and another logbook entry on the Thales pitot tubes being phased out......

[stepwilk]

aircraft manufacturers have been claiming the ability of their aircraft to "fly themselves" to some degree ever since the introduction of INS-enabled autopilots

Not quite the same thing, but remember Cessna's "Drive it up, drive it back down" advertising campaign of the 1950/'60s?

[DozyWannabe]

Originally Posted by Welsh Wingman

(1) The historic role of Airbus in creating the impression to line management that "planes fly themselves", even if partly inadvertently, and the knock-on effect upon the training culture.

Well, let's be fair here - if they'd asked the engineers rather than the sales guys they would have got a more honest answer going all the way back to 1988. As a slightly tangential aside, aircraft manufacturers have been claiming the ability of their aircraft to "fly themselves" to some degree ever since the introduction of INS-enabled autopilots - in fact the first time I ever heard the term used (albeit in a fictional setting) was as an annoying kid watching the annoying kid in "Airport 1975" saying the very same thing about the 747. I'm sure the Boeing sales team would not have gone out of their way to qualify that remark. And the movie, though utterly terrible in any kind of objective sense, certainly demonstrated via some rather nifty stunt flying that the 747 was very stable in environments where it would never normally be used.

Another quasi-fictional allusion to said ability crops up in John G. Fuller's "Flight 401" book, this time in reference to the L-1011 - and no matter what you think about the supernatural bobbins included therein, he did his research on aircraft and line operations pretty well.

(2) Several additional "complications" in an Airbus cockpit when things go awry, if cockpit discipline is not tight i.e. feedback issues (e.g. SS v control column, and the throttle and trimming) and the PNF is visually less aware of how the PF is by-hand flying.

Not completely unaware, mind... PJ2 confirmed a few posts back that most of the FBW Airbus pilots he knew didn't really see the feedback thing as an issue.

My interpretation of the CVR, for what that's worth, is that the PNF saw what the PF was doing by watching the attitude of the aircraft on the ADI, but for reasons that are likely to be endlessly debated in human factors forums for years, he either failed to accept that the guy to his right really was mashing the controls about that cack-handledly* (would you, or any pilot on here for that matter not have a moment of "this can't be happening"?), or indeed came to that conclusion but again, for reasons that will be debated for years, felt he didn't have the authority to tell his colleague to get his hands off the d*mn stick right now. We've seen instances of this before, with KLM4805 at Tenerife and Birgenair - but previous instances have tended to involve an F/O who felt he couldn't overrule his Captain. With a poorly-defined command gradient, is it possible that an F/O can also feel he or she does not have the right to give orders to a colleague of the same rank?

Also, I don't think flight deck discipline and airmanship in general can be completely separated - if you don't maintain proper CRM and organise your flight crew effectively then you're probably going to end up in the cacky eventually no matter what aircraft you fly.

I suppose you could add under this heading the <60knt stall warning design "issue".

Fair comment, but as of now we don't know how many other types also have the same issue or similar. I'd be prepared to bet money that it's more than a few... One of the perennial issues that has dogged stall warning technology in jetliners since the days of the Comet, Caravelle, 707 and DC-8 is the number of hull-losses attributed to the crew incorrectly diagnosing a stall warning as false when it turned out not to be. I find it hard to fault any manufacturer making a design assumption like that because this is the first time in decades that an airliner has ended up so far outside of it's design parameters.

This is a training/human behaviour issue otherwise the temporary UAS on AF447 would have been temporary by-hand P+P flight and another logbook entry on the Thales pitot tubes being phased out......

Absolutely agree. Though having said that it offers intriguing technical questions about all modern fourth-generation airliners - and I wonder if Airbus and Boeing will be willing to examine their designs to see what potential gotchas lurk when the aircraft is taken that far outside of it's envelope.

* - I happen to be cack-handed/southpaw/lefty, so I can use that phrase.

[BOAC]

Welsh - a good summary, I would, however, make the stall warning item the third type-related issue, and really what other types/manufacturers 'do' or 'did' is not relevant to this occurrence.

As others have said, 2.5 deg/5 deg.emergency/non-emergency - no matter, none would have caused this crash, but 11deg did initiate it. There are enormous holes in the 'French' cheese in this accident, and Dozy sums up the HF neatly in his middle para. I am still greatly puzzled by the lack of assertiveness of PNF.

Lots of new cars for the lawyers.

[PJ2]

HN39;

Whether that is 2.5 or 5 degrees is really of secondary importance.

BOAC;

As others have said, 2.5 deg/5 deg.emergency/non-emergency - no matter, none would have caused this crash, but 11deg did initiate it.

I agree with both these statements.

[Welsh Wingman]

DozyWannabe/BOAC

(1) To be fair, I have added "inadvertently" to point the "blame" in the sales direction. I am acutely aware that Airbus's head of training had a battle just to keep training at pre-existing levels (discussed previously by PJ2, on the main AF447 threads), when heavy automation is basically "existing airmanship plus".

(2) I was intrigued by PJ2's comment on this aspect, as he is an AB veteran. Takata has set out the AF command structure in place at the time of this incident and facing the CDB, which was not ideal for any "unhappy" PNF in the LHS without the "unsatisfactory" handover on AF447. Would the PNF have been more assertive if he could have seen what the PF was specifically doing through a clearly visible RHS control column (or at least if the LHS SS was moving in tandem with the RHS SS movements commanded by the PF?)? The throttle moving also? Would the PNF at least have been better placed to properly brief the returning CDB? Might this have overcome CRM shortcomings and saved the day?

(3) "Design assumptions" are always dangerous. The Titanic was "unsinkable" because how could White Star possibly flood the first five watertight compartments.....? What floats, can sink. What flies, can stall. A stall warning really should stay on until an aircraft is no longer stalled. The stall warning ceasing just after the CDB returned to the cockpit was, at best, "particularly unhelpful".

(4) We Brits naturally prefer "evolutionary", which is why I value PJ2's viewpoints on the more "revolutionary" aspects of the AB design philosophy. From Stony Point through Aeroperu and Birgenair to Colganair, there are stall warning issues and pilots forgetting their training and grimly pulling back on their control columns. Post-Stony Point, was the industry ready for another system? AF447, and the ignored stall warnings, tend to suggest not I would submit. Hopefully, with control column pilots now acutely aware of the "stick shaker" issues after numerous hull losses, there won't be a SS repetition.

As I said, I don't want to get dragged into any AB particular issues.

This is an across the board training issue. The less routine manual flying that pilots do on any aircraft (within the flight envelope), the harder it is for them to suddenly "ride to the rescue" in a degraded flight envelope emergency. That must be self-evident........

[Lyman]

apropos not much,; had the crew recovered, the landing back at CDG would have been interesting, especially in a crosswind.

[DozyWannabe]

Originally Posted by Welsh Wingman

(1) To be fair, I have added "inadvertently" to point the "blame" in the sales direction.

I know, just thought I'd reinforce it for the benefit of non-native English speakers who may not pick up the inference as easily as I did.

(2)...Would the PNF have been more assertive if he could have seen what the PF was specifically doing through a clearly visible RHS control column (or at least if the LHS SS was moving in tandem with the RHS SS movements commanded by the PF?)? The throttle moving also? Would the PNF at least have been better placed to properly brief the returning CDB? Might this have overcome CRM shortcomings and saved the day?

Well, that's a difficult question - and as I said, one that the human factors bods will be debating back-and-forth till kingdom come. Common sense suggests that it might have made a difference, but on the other hand there's the CVR that suggests that the PNF might have been aware that his colleague was overcontrolling even without feedback, as well as the historical cases (Birgenair, NWA) where the PNF in both cases had the yoke in front of them reflecting the PF's inputs, and still failed to put two and two together.

(3) "Design assumptions" are always dangerous. The Titanic was "unsinkable" because how could White Star possibly flood the first five watertight compartments.....? What floats, can sink. What flies, can stall. A stall warning really should stay on until an aircraft is no longer stalled. The stall warning ceasing just after the CDB returned to the cockpit was, at best, "particularly unhelpful".

Of course they are, but as I'm sure you're aware, design and engineering are and have always been underpinned by the art of compromise. In this case (as I said above) there was a history of incidents where the number of false stick-shaker events led to an assumption on the pilots' grapevine that most stick-shaker events were false, with fatal consequences when a real stick-shaker event happened. Manufacturers were then faced with the task of weeding out the circumstances in which the stick-shaker was a false alarm, and Airbus's logic was based on a set of parameters that were so outside the flight envelope that they couldn't see it happening with the aircraft aloft - it's possible other manufacturers have done the same and at the risk of repeating myself I hope they all band together to find out.

The Titanic (another minor obsession of mine as it happens) and the Comet 1 were both examples of designers working at the limits of contemporary knowledge (regarding worst-case maritime collision scenarios in the former and metal fatigue in the latter). Now we know that it is possible to stall a modern airliner to such a degree that it intersects the stall inhibition parameters, it's likely that designs are going to have to change.

(4) We Brits naturally prefer "evolutionary", which is why I value PJ2's viewpoints on the more "revolutionary" aspects of the AB design philosophy.

Now my old Prof considered the A320 more "evolutionary" than "revolutionary" once he got to poke around inside, the only difference being that the aircraft it evolved from (the A300/A310) was already probably the most technologically advanced airliner flying up until that point (though again, as I said before, the L-1011 came close). It depends on how you look at it - most of the circumstances that made two large, connected central control columns necessary in the first place (largely to do with leverage when the flight surfaces were directly connected by cable) were no longer applicable by 1972, let alone 1982 when the A320 was being specified.

From Stony Point through Aeroperu and Birgenair to Colganair, there are stall warning issues and pilots forgetting their training and grimly pulling back on their control columns.

Aeroperu was slightly different - loss of pitot data affects airspeed indications, and as long as you click the A/P off, it's relatively straightforward to manage. Aeroperu had the static ports taped over, which was a whole other ballgame in terms of what it did to the instruments.

Hopefully, with control column pilots now acutely aware of the "stick shaker" issues after numerous hull losses, there won't be a SS repetition.

Fingers crossed!

This is an across the board training issue. The less routine manual flying that pilots do on any aircraft (within the flight envelope), the harder it is for them to suddenly "ride to the rescue" in a degraded flight envelope emergency. That must be self-evident........

Absolutely - but the pressure must be brought to bear on the airlines, specifically management, that the status quo is not on, and again - as I've said before - who will be the first to stick their head above the parapet?

[TJHarwood]

BOAC

Given that he was CATOBAR trained on the BBS1 and the BBS2 for the old Ark and Eagle, I suspect the Wingman has very decided views on the AoA going to 0 and ceasing the stall alarm at less than 60 knts and is showing remarkable restraint in keeping his views to himself......

[airtren]

Iceman50,

Trying to get the ND, is only the beginning of the first part of a Stall Recovery.

There is more to Stall Recovery than that, and I read "deSitter's" post, as a suggestion on discussing all parts or aspects, and as far as I am concerned that seems to be an interesting idea.

Originally Posted by iceman50

....
deSitter

There is No need to start a thread all they had to do was get the nose down. If they had held the input the nose would have gone down and the autotrim would have then assisted with a sustained nose down input. They could even have helped by using manual nose trim themselves[/COLOR], as taught in UA recoveries. As for trying to roll / yaw the nose down

[Welsh Wingman]

TJHarwood / airtren

TJHarwood

No comment!

airtren

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