Hello Machinbird;

Re your post #869

"Is it proven that the Quick Access Recorder data cannot be resurrected from its time in the depths? I don't remember the answer anymore."

The QAR box etc is located on the main radio rack known as "800VU". It was brought to the surface. There is no information directly stating that the QAR has been found or what condition it may be in. It was stated that the two FMGC memory modules would be examined, (to my recollection the first time that was done was the AA965 at Bogota).

Though not crash-protected, the QAR on this aircraft is almost certainly solid state (vice optical etc) as well and if it survived the initial impact and subsequent water pressure it could be readable.

It's my experience that data frames for QARs is generally far more detailed and sometimes has higher sampling frequencies for some parameters. Let's hope all this comes to pass.

HN39;

Re, " I must admit that my understanding of FD commands is very limited, and I'm just assuming it 'mirrors' the commands that would have been given by an active autopilot."

Yes, that's the way the FDs work - they display the orders that the AFS is issuing, either to the AP or to the pilot flying.

Overspeed flying (not recognizing the stall)?
 

Hi,

mm43,

Is it possible (after stall) their thinking was just of being flying at high speed (and high ROD) just requiring to reduce (the speed) with NU? The inputs they received during the descent could "match" this modeling? Or this is too simple to be considered? (below a K.I.S.S. "safe" threshold). As we understand it seems they never acknowledged a stall. So...

:confused: Is it possible?

PS

Like as if the plane "entered" an state they never realized as something plausible. And without enough confidence in the available (considered not reliable, stable and erratic) resources to ascertain "what was happening"

PS2

Again, this is a risky exercise without more factual info. I know.

This wreck started with duff speeds, and continued with duff speeds, and duff speeds killed them.

They HAD "crazy speed". The problem is that too much of it was vertical. No STALL recovery was discussed or attempted. It is as you say, and clear as day, they thought they had an overspeed, and whether they sussed their PITCH accurately or NOT, they all signed on to pull, after much apparent confusion.

Too much has been made of how "simple" the soluton was. It was NOT. What does a heavy sound like at 15k fpm descent with an AoA in the roof?

Spinkleman, hear this: I have NOT accused BEA of manipulation of data. What they have done is release too little, and set the stage for a miscarriage of the TRUTH. "OFFENDER"? How so? I counsel doubt, and scepticism, which to me is far short of purposely slandering dead pilots. Make a list of what is MISSING in sounds, conversation,and discussion. Then drive your truck through it.

Assess the state of UAS, STALL training, and high altitude upset recovery. Read Post #66. Ignore what I say, consider only what is missing from a fair stage.

We'll have to see. Patience, there is no other way.

Lyman, you were claiming at one point that obviously the VS had fallen off, because Europeans don't understand glue (so far as I understood it).

I understand,and appreciate, your desire to get to the root cause of this crash. It happened in flight school :- when the PF wasn't told to go away.

Hi F4 and JCJeant,
I stand corrected, in French no less. (Wish I could understand WTH they're talking about!) Reading all of the threads, I mistakenly got the impression that most here feel that the stall could have been corrected, mostly from earlier posts. I confused that issue with a separate one: the stall should not have happened in the first place.

I hate to add more useless noise to the discussion by posting this, on the other hand, I feel it is important to acknowledge my mistakes. I only like to appear arrogant if I actually know what I'm talking about! (Happens occasionally.)

Carry on gentlemen. :ouch:

Just that Airbus/EASA were concerned enough about that configuration to publish an AD about it. AD No.: 2010-0271
http://ad.easa.europa.eu/blob/easa_a...AD_2010-0271_1

For at least 15th time on AF447 related threads: yes, it is not just possible, it happened before. There was at least one other crew mistaking stall buffet for overmach:

ASN summary of Northwest Orient B727 N274US accident near Stony Point on DEC 01 1974 with link to NTSB report included.

Could be, but so far I see no proof to it as: a) three preliminary reports are quite elaborate on the technical matters while being pretty thin on HF side b) being competent pilot does not come granted after the training obstacle course has been negotiated. It is matter of keeping oneself fit to fly; there was many a pilot with five-digits total times who perished after making a beginner's mistake c) there's the additional issue of CM1 not recognizing that pulling 15° ANU at typical cruise level is very bad idea or what is the significance of computer shouting "STALL STALL STALL". It's all very well discussing cockpit authority gradient but we are looking at failure of CM1 to realize the gravity of the situation and that CM2 actions will turn out to be lethal pretty soon. Idea of having two pilots is the same one behind having doubled or tripled ADC, IRUs and NAVs - CRM and MCC does not mean we use each other as a crutch but rather that as long as both pilots independently arrive to same conclusion where the aeroplane is, where it is going and where it should be going there's good chance they're correct.

Second notion is totally correct, first notion is correct up to the certain point in time which cannot be precisely determined as no other A330 has ever ventured into AoA achieved by AF447. Given the pitch-down of the A330 when the sticks were released, chances are aeroplane could be recovered successfully until there was no sufficient altitude available to recover from dive. While it is certain that A330 can be recovered from stall if recovery is initiated as soon as practicable after G-break, we don't know what was the latest time/level at which recovery from extreme AoA had to be initiated. It could be FL300 or FL200 or even FL100, perhaps there will be enough data for aerodynamists to say their verdict, perhaps we'll never know.

Thiells 727 (misleading) similarities
 

Hi,

Clandestino:

Quote:

---

There was at least one other crew mistaking stall buffet for overmach:

---

Indeed there are similarities on this ferry flight with AF447 case. I commented this in an earlier post, IIRC on UAS and Pitot generating false (misleading) inputs to the crew.

Their (heater off) Pitot's "failure mechanism" generated false (increasing) speed misleading the crew during the climb.

I will locate my post and will be back with a comment on what i see common between both cases.

Anyway in both cases the stall buffet probably played an important role in their thinking, i agree with you. A dangerous one (the role).

In AF447 what is impressive is how (and why) they failed to even understand (timely) they simply (after stall) were not flying. Just performing a "free fall" trajectory (at near terminal speed).

The role of the man machine interface "performed similarly" to the (much simpler) one of the 727: Not helpful.

Will think and compare HF aspects in both before continuing.

Hiya RR NDB

I do not know if this is pertinent, but on 330, Stall buffet is encountered after the STALL WARNING, so buffet should not be confused with high Mach.

(In ALT and DIRECT LAW, obviously).

ciao paisan

Once the airplane was deeply into the stall, I think a distinction has to be be made between whether the airplane was physically able to recover, and what psychologically the pilots could reasonably be expected to accomplish.

As I gather with my scanty understanding of the French language, AB Chief Test Pilot J. Rosay states in the video (post #863) that the airplane's physical capabilities at extreme AoA have not been determined in flight test and thus will remain somewhat speculative.

Apart from the question whether the simulator reflects the real airplane, the psychological aspect is amply illustrated in "post #66" of the other thread:
For example, around 02:13:20 the airplane descended through FL120 with pitch attitude +10° and FPA -40°, i.e. AoA= +50° (assuming still air). To unstall and stop the stall warning, the airplane would need to be pitched down until the AoA was less than approximately 10°. Even if the airplane would respond to forward sidestick by pitching down, the lift and drag forces and hence FPA and airspeed would not change significantly until the airplane was unstalled, i.o.w. the airplane would have to pitch down to an attitude of -40 +10 = -30° to stop the stall warning. That attitude would then have to be maintained for some time until the airspeed had increased enough to permit a pull-up from the dive.

Hi,

Nothing to do with AF447 A330 but this recovery of the Tarom A310 show how the pilot go in a steep dive to recovery ....

Phugoid Curves
 

The diagram is from NASA TM X- 73,229 which is about recovering hang gliders from dives, but has broader implications as well as a fair bit of aviation history behind it.
http://home.comcast.net/%7Eshademake...OID-CURVES.jpg
Do you see a connection with the TAROM Airbus over Paris, the Interflug A310 on approach to Moscow, and for that matter what the AF447 crew would have faced in recovering their aircraft to level flight if the trim did not run down quickly enough?
Just trading potential and kinetic energy.

Trading potential and kinetic energy
 

Hi,

Machinbird:

Tarom Airbus A310-325 24 SEP 1994

Interflug Airbus A310-304 11 FEB 1991

Feedback systems in action. And for Ho you need 3Ho

Question: What about air density change effect during the descent? Seems to me makes easier the recovery. Dampens, provides more lift, ground effect :}
(it comes to my mind Ekranoplan. You mentioned Interflug :E :)

Machinbird,

Thanks for an interesting reference.

Originally Posted by Machinbird Most likely not, but you may not focus on it as much, depending on a lot of things. It also depends on what scan you are using. While I cannot fathom a scan that does not depend upon primary reference to the Attitude Indicator (Artificial horizon) that takes a primary reference pitch and roll, and then crosschecks from there), even that scan basis, if rusty from disuse, can break down under stress.
While I would hope not, it seems that IF he was paying attention to his pitch, THEN he had a target pitch in mind. 12 def? 15 deg? 10 deg? Hard to say. As was discussed some threads ago, an element of the scan breakdown may have been him used to "flying the bird" (FPV). If the bird wasn't behaving as usual, or had dropped out, etc, then getting his scan into operation using the pitch/roll, cross check all else, might have taken him a few cycles to realize needed doing. If he ever did.
What he was seeing didn't match what he expected, that much seems to be true.
Heh, that made me laugh. :ok:
Retired F4 Thanks for this:
Flying in a flight regime they weren't trained for. That has killed more than a few ... :(
Maybe, but I think it's fair to say that the training and qualification system at AF seems to have some play in this accident.

1. How well equipped was this crew for the UAS drill? CVR seems to show that the response didn't quite fit what was in the book at the time. Training issue?

2. How well grooved was the instrument scan of either pilot?

3. CRM question: how does the LH seat pilot's scan, or lack thereof, and his attempts to get RH seat pilot back into the game reflect on the currency and training at AF?

4. As asked with some frequency before, how much time hand flying at 37K did any of the AF pilots have?

Hi,

I think we could very quickly have a simple answer to this question if some Air France pilots (who certainly read this forum) would dare answer ...

All,

I have read all and I mean all the pages and grow dizzy with the content but, with no great confidence, do I recall correctly reference at some point by someone no doubt better informed than me, to AF SOPs (et al) specifically forbidding hand flying up there without a d*** good reason?

Hi,

Why this interdiction ?? .. it's so dangerous .. ? .. so .. this can explain why AF447 had some ... troubles .....

You just have to be smooooooth to fly in cruise at altitude. Actually no big deal once you relax. Unfortunately, AF447's PF was anything but relaxed or smooth. The unexpected cavalry charge set his heart racing and it appears he never did begin to fly cooly and deliberately. He continued to beat snakes with the stick all the way down.:rolleyes:

Unless you are in RVSM airspace, pilots should be grabbing a few minutes of stick time just to keep the feel of the aircraft. What kind of back up system can you be if you don't actually fly?

Sim time twice a year? Give me a break. Flying is one of those skills that needs more frequent refreshing. Sims are close, but not quite the real deal. On an 8 hour+ cruise, the extra fuel used because you are hand flying for 5 minutes can probably be measured easily with a cup, that is, if you are smooth:).

Machinbird

Yes, and no. My neighbor flew the baby bus for ---Lines for many years. He was out of NORMAL LAW, exactly: not once. I am not sure what manual stick (NORMAL) accomplishes up high. The sim cannot duplicate what these three inherited. I could argue that it is a negative, perhaps imparting a false security when LAW degrades for real (which is hardly ever).

BEA say the a/c rolled right to 8 degrees, they do not mention trend, and therefore people assume it stopped at 8 degrees, right? The a/c started a rapid and cyclic ROLL that you have eloquently described, and hand flying in NORMAL LAW is nothing like it. It is NORMAL Law itself that wants mayonnaise, and AL that wants smoothe; so practicing mayo up high is exactly wrong, and could be the source of the entire problem? DIRECT ROLL with mayonnaise is poison.

There is one salient point that is my take away, that BEA mentions very directly.

Pilots of this aircraft MUST have an AoA reference, PERIOD.

cheers sailor.

Speaking for myself, I don't believe it stopped there by itself, but as a result of PF's control (over reaction) input. There really is nothing to stop a roll rate in Alt 2 except a bit of opposite turbulence, a bit of counter control input from aileron or rudder, or the aircraft's natural stability once it begins to slip a bit.

If PF had been accustomed to flying smoothly at altitude in normal law, he would have just had to fly even a little smoother than usual in Alt 2. As it was, he was completely clueless in that environment.

PJ2 describes the proper effort as little more than a light squeeze on the stick if I remember correctly, although he also indicates that the aircraft is a bit goosey in roll in Alt 2 (my words, not his). I expect he will jump in if I have misinterpreted his intent.:}

Anyone who thinks an airliner needs an AOA gage is not going to get an argument from me. It is an excellent performance indicator. Just don't hide it away in some corner of the panel. It belongs right in front. You want it where you can incorporate it into your scan easily should a problem arise.

Machinbird;

re, "PJ2 describes the proper effort as little more than a light squeeze on the stick if I remember correctly, although he also indicates that the aircraft is a bit goosey in roll in Alt 2 (my words, not his)."

That's just how I described it: light squeeze, if anything at all, and that's how the sim performed. Taking one's hand off the stick would have been by far the best decision and then gently correct the right roll with one small motion. Tough to do? You betcha. This is just hindsight.

Really, this initial roll and the few back-and-forths were entirely a non-event...in the sim one rolled back and forth a bit just like we see in the data, primarily because it is sensitive, but as Owain Glyndwr has stated and I completely agree based upon experience with hand-flying the airplane at cruise altitudes (in Normal Law), the PF learned quickly and got this bit of roll under control quite nicely.

The rolling response of the airplane is all but inconsequential to any of this including the initial, instant pitch-up. This pitch-up was instant, steady (uncorrected, unchallenged) and held for a relatively long time. It went from what may have been perceived as the "correct" response (to the UAS) to an urgent pull-up to arrest either the descent, or what may have been later assessed as an overspeed, (I think that notion is difficult to support). The lateral stick forces which produce the minor rolling we see in the data (+/- 8deg) are tiny but the half-pull (10deg NU IIRC) to produce such pitch-up requires a much stronger force - I don't think it is a result of 'being distracted'.

The eight degrees is nothing to be concerned over, nor are the back-and-forths; such minor gyrations don't lead to a complete loss of control, and he got it under control. Now, over thirty-five degrees bank at that point would be something to comment upon.

I've never understood the fascination with this initial roll and subsequent minor PIO. There's just nothing in it - it's what the airplane would do with just a bit of lateral stick and he very quickly got it right. But it doesn't stall the airplane.

While it has generally been agreed that the sim cannot - does not reproduce aircraft behaviour at and post stall, nor does it behave wildly and unexpectedly in the stall and subsequent descent if one keeps the stick back. The stall is however quite recoverable, taking about 20,000ft or so. The positive lift from the THS even at 13deg NU and the elevators 30deg UP has been acknowledged as contributing to a ND tendency which, if full ND stick is applied AND HELD, with pitch at 15-18deg ND until the airplane is clearly flying again, (rapid increase in CAS, disappearance of buffet, increasing wind noise), the stall is exited before FL200. That's the sim, (Level D). I am not an engineer but my pilot instincts tell me that below FL200, opportunity for a successful recovery rapidly reduces, mitigated only by the thicker air. It would take very aggressive action on the controls to do so and we're just not trained to do that. That means the guy at the controls at that very late point in the event would be of a mindset that recognizes only one way out with seconds to decide..., "15deg nose down, 15,000ft away from earth doing 18,000fpm down is my only chance" and that capacity to think and do that is, I expect, rare.

After all situational awareness was lost and reassessing what had occurred just prior to the apogee, (meaning, why the airplane couldn't be in a high-speed condition given its initial pitch up after stable flight), the loss of energy & post-stall entry became extremely difficult to assess and respond to. there just isn't the instrumentation to tell you that and I firmly believe that an AoA indicator would not have saved this flight at that point. It may have up to the apogee but the psychology of perception/assessment/agreement rapidly changed after the descent began.

The recurring stall warning was a sad happening in an extremely complex and impossible-to-assess set of circumstances but that occurred very late into the descent, (about FL200 to 180) where recovery was rapidly becoming impossible even with extremely aggressive actions.

Something learned?
 

The video highlights a new addition to the airline's flight-analysis protocol, which Air France implemented in 1974 and today analyzes operational data from the carrier's 1,500 daily flights.
Each flight records some 100 different data points, including altitude, speed and trajectory. The new feature is expected to reduce the time necessary to identify and address areas of operational safety concern.
Air France employees, called gatekeepers, will provide pilots with faster access to flight-analysis data.
"With the gatekeepers," explains Gilles Laurent, vice president of Air France's flight safety division, in the same video, "we can question the crew about contextual elements just a few days later and not a few weeks later."
Gatekeepers' duties range from simple discussions with pilots about safety concerns to recommendations of extra flight training where warranted.
While pilots normally cringe at the idea of flight-data gathering, fearing possible disciplinary actions, this new Air France safety protocol was established with the full agreement of the SNPL, the French arm of the Air Line Pilots Association.
Data gathered in the program will also be reviewed on a bi-monthly basis by an Air France committee with a mandate to suggest improvements to operational procedures, equipment, infrastructure, crew training and aircraft maintenance programs.
While Air France's overall safety record is quite good, the airline has suffered two high-profile fatal crashes in the past 12 years: the Concorde crash in Paris in 2000 and the 2009 loss of an Airbus A330 in the South Atlantic.

Source: AINSAFETY

Rich and dense post (HF as the key)
 

Hi,

PJ2:

Quote:

---

...psychology of perception/assessment/agreement rapidly changed after the descent began.

---

HF, as the "key" to/of/from, this accident.

PS

Excellent post :ok:

PS2

On how a inadequate behavior may led to "non manageable" situations:

Quote:

---

and it appears he never did begin to fly cooly and deliberately. He continued to beat snakes with the stick all the way down.

---

Fast feedback to crew
 

Hi,

A33Zab,

Excellent approach:

"Air France is the first airline to have implemented a prevention system based on the analysis of recorded flight data." :=

This was published in carrier corporate site. And it sounds like a phrase generated by marketing personnel perhaps without enough info. on what was done also by other carriers many decades ago. And applied by competitors with similar objectives.

Quote:

---

Something learned?

---

A33Zab, :) Good question. :)

PS

Thanks PJ2 for "limiting the time" duration of the incorrect info. in my post.

RR_NDB;

Re, "Air France is the first airline to have implemented a prevention system based on the analysis of recorded flight data."

No, it isn't. Their statement on the page to which the link is provided is incorrect and misleading and is spin. BOAC/BA was doing flight data analysis in the late fifites and pioneered the entire field and remains a leader today. BA today has the most enlightened and advanced FDM Program in the industry. I know some of the people involved in their work and we took many lessons from them and from QANTAS when developing our own program in 2000.

Many airlines are doing FDM/FDA/FOQA especially in the United States where FOQA is regulated and privacy of data supported in law.

It seems they are implementing one of the most important learning processes which is using pilot gatekeepers to contact crews to learn more about the circumstances of the events being seen. As may be imagined, such processes require a mature, trusting relationship which guarantees that recorded flight data will not be used against individual pilots for disciplinary or legal proceedings. Except for Asia, these enlightened approaches seem to be the norm in FDM Programs.

Management never does this work and do not have access to actual data due to conflicting priorites and trust, again except at BA where, I understand, FDM data is even shared with the regulator, (de-identified).

Perhaps BOAC or another PPRuNer can tell us more, but I wouldn't consider an implementation date of 1974 "late".

PJ2

Hi,

PJ2
Indeed ... and is to remember that in all the cases of a Airbus stall identified (Toulouse-Perpignan and AF447) the result was a crash ...
Coincidences ?

RE: "Gatekeepers".

Don't go looking for applause for doing what you should have been doing all along....

It makes one look foolish.

When they (AF) started the program?
 

Hi,

Bear:

Let's hope the results of the program to be much better than the way they are doing the marketing of this "basic feedback tool" as PJ2 mentioned being used since long time ago.

Or much worse.

Lyman and RR,
I don't give a fig if they're "looking for applause." If it serves to improve flying skills and safety for the public (as I have no doubt it will) then it is a Good Thing. Let's not condemn, out of cynicism, what looks like a good effort.
:=

Back to flying my armchair.

Howdy.

Safety is like humility. If you splash it about, you don't have it. Real time monitoring of a/c is done extensively, and doesn't need a Jeanne come lately to validate it.

Clueless.

Hierarchy of this Feedback mechanism
 

Hi,

PJ2:

And the issue (Quality) is Rich! Addressing what is most important to "accomplish the mission". In "search of Excellence" as i understand it.

I personally had a great opportunity in my career to understand the importance of this (type of) feedback. May represent the "differential" of a good Company.

As relevant and "capable to correct" as any well designed feedback. In this case with hierarchy to almost everything in the carrier ops.

This type of feedback mechanism is much more than a basic tool: IS A VALUE
 

Hi,

Bear,

Safety is like humility.

:ok:

An intrinsic value.. That acts silently.

Organfreak, my position was clear:

Let's hope the results of the program to be much better than the way they are doing the marketing of this "basic feedback tool" ...

One can but hope. Fecklessness and chauvinism can be dropped as a weight at any time, is there the will?

The WILL
 

Hi,

Bear,

This is the most important question.

Intrinsically connected to A33Zab one:

The title of his post

@Lyman:

Who is clueless? Was that aimed at me? We disagree; does that make me clueless?

Safety is not at all like humility. So what if real-time monitoring of a/c is done extensively? What really matters is what is done with that info.

Organfreak. Forgive me, of course I did not mean that for you. For Air France.

"We are missing an a/c....", "Fortes Turbulences.." "Wait, no, Lightning, yep lightning..." "No, it was the Radars, the poor pilots were not lucky..." "They flew into a Storm, right, it was the Storm..." "Whatever, Non?......" "Pitots?" "Certainement, the Pitots..." "Oh, we were to have replaced the Pitots?" "Then certainly not the pitots, it was Airbus..."

et plus la meme chose.

Air France is so used to hiding, they have forgotten how to "get in front".

Hence, clueless.

I value your postings, and would never insult. You are well informed, and articulate. Again, please excuse.

Lyman,
Ah, I took it too personally. Thanks for clearing that up. I don't disagree.

Zut alores!

Premature to comment. Sounds very good.
 

Hi,

Looking to what A33Zab post (#904) i think this is a potentially powerful "mechanism" being introduced by the carrier.

If well applied with absolutely essential aspects mentioned by PJ2 like:

It could represent a "very good answer" to the question put in the title of A33Zab post.

In my opinion this came in time to show AF probably learned important lessons and seems committed to a "process of change".

PS

The concept using "pilot gatekeepers" is new?

PS2

PJ2:

Sure, is not.

Hi RR_NDB;

None of this is "new". Flight data analysis and monitoring has been around for fifty to sixty years at BOAC/BA. The rest of the industry began getting on board in the '70's. Some just don't get it and don't implement it for some strange reason but in this day and age of cheap equipment, massive processing and storage capacity and the ability to "know" vice "not know" and the notion of the accountable executive, there simply is no excuse for a major carrier to resist or ignore such programs. But believe me, it happens and all kinds of excuses are given. It is complex, takes a very long time to design, create and implement and for those who like to think that their operation is "pretty good" it is an unwelcome and difficult wake-up call to accept. No carrier is immune and no management has the right to think that they don't have such events occurring on a regular basis.

Air France was a very early entrant and user of FDM and they run a good program from the information and presentations I have seen them create.

The notion of the "gatekeeper" is equally not new. The data process tells us "what", but not "why" something out of the ordinary occurred on a particular flight.

FDM data is trended, graphed and presented to Flight Ops managements, (de-identified) and other stakeholders including the pilots' association.

Gatekeeper contact, once established and trusted, is an extremely valuable process which puts pilot association member, the pilot(s) and the data together in a confidential private process of discussion. Such discussions may be straightforward or may be quite detailed depending upon the event. In almost all initial contacts the captain is contacted first. Very often the response is relief and a willingness to be a partner in finding out what occurred and trying to improve those areas which were felt to be contributing factors.

It may be appreciated that this is no ordinary conversation. It requires great comprehension on the part of the pilot association FDM representative and also requires great confidence in such a system by those so contacted, and it further challenges the Operations people to trust that such a process is being done and is providing through appropriate channels the necessary and timely feedback to address problems where they may have emerged in the data. This involves the training/checking process as well as the SOP and policy documents.

FDM people also keep an eye on the industry. Sometimes an occurrence is not monitored such as "split thrust levers at landing", (a result of the TAM accident), and a new FDM "event" is created; all of a sudden an airline learns something new about its operation where it was "blind" before.

For those who want to know more about this safety tool, try CAP 739, FAA FOQA, and the Flight Safety Foundation for starters.