Concours77

" lorsque le pilote automatique se déconnecte, un gradient de vent latéral concomitant (20 kt en 4 s) engendre le départ de l’avion en roulis à droite ;" so we do know the source of the uncommanded roll. Fig 64 of the French version shows a simulation of what the lateral response of the aircraft to this turbulence would have been with and without the actual pilot input. Note that this data does not exist in the English version.

What did Robert know, and when did he "know" it?

Do you have an opinion on the Trimmable Horizontal Stabilizer cycling "Full Nose Up" but only reaching -13.2 degrees? Is there any chance that the jackscrew got stuck short? Any chance of an "Alaskan" type fatal Pitch?

KayPam

If this was true you should be dead by now.
Except if you were flying some sort of highly maneuverable aerobatic or jetfighter aircraft.

Owain Glyndwr

When did They know and what's missing from the CVR?

All this was described in second by second detail in the FIRST expert's report, but that has never, to my knowledge, been translated into English.
You have to be careful to differentiate between when they could have known of the law change and when they actually spoke about it. The expert's report shows an ECAM page timed at 2:10:08 that says "Alternate Law (Protection lost)" but then at 2:10:22 it says:"02h 10.22 Le PNF continue la lecture de I'ECAM : « Alternate Law, protection slo.." so they could have known about it 3 seconds after the actual law change.
The intervening time seems to have been taken up by other activities, because as the experts concluded " Face â l'incompréhension de la situation, le PNF cherche dans la lecture partielle et désordonnée de I'ECAM une justification â ce qu'il perçoit, et revient à son projet d'action initial de prévention des conditions givrantes" - loose translation "not understanding the situation, the PNF resorted to a partial and confused reading of the ECAM to find a justification of what he was seeing and went back to his initial actions for prevention of icing conditions"
From other remarks one can infer that the experts attributed the low level of cockpit conversation to the increase in stress of the moment.

With respect to your other posting, I would have thought that the last thing a pilot, flying in turbulence and faced with a sudden aircraft movement would want to know would be how big was the gust that hit him. He "knew" there was a roll diversion immediately..

I don't know where you got 13.2 degrees from; the published DFDR graphs are really to small to get that much precision unless digitised; but in any case the THS did not "cycle". Its position is determined by a time integral of elevator deflection and without knowing the integration constant and doing a careful integration with the actual elevator input one cannot say what the final THS deflection should have been.
But no, no chance of an Alaskan type pitch.

Machinbird

At the time of the event, I was seriously concerned about dying as you suggest.
The aircraft was a North American T-2A which was a straight wing jet trainer. I was number 6 to roll with 10 seconds interval between takeoffs as briefed.
With gear and flaps up, accelerating through 200 knots at about 50 feet up, the aircraft began to roll right against full left aileron and then full left rudder too and reached about 85 to 90 degrees of bank.
I remember thinking, what do I have left???
The thought came-Aft stick! The aircraft then popped out the side of what must have been a multi-aircraft vortex and began to respond to control. With a wingspan of 38 feet, there wasn't much time or altitude to play with. The ejection seat was useless in that attitude.

Concours77

"You have to be careful to differentiate between when they could have known of the law change and when they actually spoke about it"

With respect, that is my point, precisely. You also state that the second ADR failed at 2:10:05...and that AL was instantaneous. That leaves two seconds before the Auto Pilot quit, and perhaps another two seconds for the crew to observe the alarm, and institute input. The "gust" likely complicated a crisp decision, perhaps adding another two seconds of distracted focus, and stick input. Six seconds to start manual flight, and stabilization of flight path.

You leave unaddressed my question about when they were cognizant of loss of Protections and a major change in handling (Roll DIRECT). Did they not lose both PFD's?

Do you fly this aircraft? I have an acquaintance who flew the A320. He flew as Captain for ten years. I asked him at one point "how many times have you flown this plane in Alternate Law exactly?" He answered, "Zero".

I asked him if he knew what AL2b is? He answered: "I have no idea."

I repeat, the first time we know absolutely when the pilots were aware of the Law degrade is "...We've lost the speeds......Alternate Law...." Time: 02:10:22.

Seventeen seconds perhaps of Mode Confusion. It is my belief that it is in this initial time period the aircraft was lost....do you disagree? It is certainly within this time frame that we see the "commanded climb" ?

That the THS deployed to Full Nose Up, in the Stall and the Stall Warning was inhibited whilst airborne are but two extraneous conditions that are inexcusable even though not material.

Finally, I need to repeat my thought that though we all can take some ersatz authority from the data post mortem the flight crew had no such luxury. This is the truth, for me, and I make note of this "data" approach by folks who are without the empathy to place themselves wordlessly in this flight deck "on the day"...

Many thanks, Owain Glyndwr.

Are you connected in any way with Airbus?

Owain Glyndwr

@Concours77

Look, I'm not going to get into a fruitless discussion on things which have already been covered time and time again in the twelve AF447 threads and the relevant official documents. If you were to read the accident reports and expert analyses carefully you would find the information you seek.
I will quote just one point to illustrate what I mean. You have just hypothesised that there was a delay of six seconds to start of manual flight and stabilisation of flight path.
But if you had read the BEA Final report you would have seen on Fig 26 the following:
2:10:05 A/P disengaged
2:10:06 First Officer (on the right) "I have controls"
2:10:07 read from chart 1st application of pitch control: 1st application of roll control
Bear in mind that there may be a delay between action and recording of that action because of DFDR characteristics

If you had read the official report you would know that your remarks are incorrect.

No, I do not fly the A330

"It is my belief that it is in this initial time period the aircraft was lost ....do you disagree"

Yes; it is my view that the aircraft was lost at 02:10:50 when, from an attitude of 6 deg in more or less level flight and an airspeed of 216kt the FO initiated a pull up to 17.9 deg.

Am I connected with Airbus? Check my age - at 81 I don't work for anyone!

My view is that "what" happened is adequately covered and not really in dispute. "Why" it happened that way is best left to the PROFESSIONAL airline pilots that contribute herein

Concours77

From OG

"Bear in mind that there may be a delay between action and recording of that action because of DFDR characteristics"

Thank you.....I have allowed for that. Unless crew were expecting a Law Degrade with loss of autopilot, the recorded data suggests this exact thing (delay in recording).

I repeat that the only (first) evidence we have of the crew's recognition of these important aspects is seventeen seconds AFTER the DFDR (sic) reporting of them.

I believe I can make a reasonable statement relative to the (recorded) difficulty in managing Roll... Turbulence? Possible. Patent aircraft roll response as sensitive in Alternate Law? Certainly, Airbus is explicit about that. Uncommanded Roll as a result of approach to Stall? Again, possible, the StallWarning demands we consider that.

"...We've lost the speeds....Alternate Law...." At 02:10:22, spoken by David Robert, PNF (pilot not flying......)

Seventeen seconds after the first reported manually effective control inputs by Pierre Cedric Bonin, and the defining moment of onset of "confusion....." We know that they are at odds with the annunciation PFD data, and each other, concerning ascent...

That is all completely patent, I believe. The last vocalization I believe was from Captain du Bord duBois:".....Pitch ten degrees...." The purpose of this dying man's final words? Remarkable in that he does not mention his mother, utter terrified profanity, or other recorded last words.

He is trying to tell us something. With two seconds to live, he knows the DFDR has recorded Pitch, and Angle of Attack is not available, the aircraft doesn't tell the pilots this, only the DFDR gets the AoA data. He wants us to know they know attitude, that they are without a solution, and he wants posterity to know what the problem is.

He knows the aircraft is nose high, he knows the rate of descent is fatal, and he may as well have said, although discretion may have prevented this: "She is not responding in Pitch...."

You don't care about "WHY". Nor do I, nor do most professional pilots, I would say...

"WHY" is for Priests, the word we concentrate on is "HOW".

My initial conclusion remains that the pilots did not know, nor were they privy to how the aircraft was behaving, and confusion ensued. Once confused, reason becomes difficult, and solution development becomes perhaps impossible.

Owain Glyndwr

The delay I was referring to is the internal DFDR delay - maybe quarter to half a second, no more than that. It was simply to explain why it might be read that there was a whole second between his announcement of taking control and evidence that he did so.

Please don't put words in my mouth. I did NOT say I don't care about "WHY". On the contrary, I care very much and I am amazed to hear your opinion that most professional pilots don't care either! I just happen to think that professional opinion outweighs that of enthusiastic amateurs every time.

Most people would agree that the pilots did not understand what was happening to them, but not perhaps with your argument on the reason.

KayPam

DFDR recording is VERY complicated.

For instance, some parameters will be refreshed 5 times per second in the DFDR, and be recorded only 4 times per second, so you skip one in every 5 possible recording.

Parameters are not at all refreshed as much as you would expect them. The best ones being refreshed 8 times per second (they're very rare)

For instance if you push your button "event marker" in the cockpit, in order to mark an event in the DFDR, it will only work if the parameter is refreshed (written) while you're pressing it. Since it is (say) at 1 point per second, you could be unlucky, press it for 0.9s and miss it !

To finish with, some parameters will be refreshed very regularly but the parameter itself can be late due to some reason (gear squat parameter can be as long as 2s late)

roulishollandais

Why is it so bad?

Owain Glyndwr

roulis
It's difficult to get a litre of wine into a standard 75cl bottle!

Ian W

It is a pity that these days when electronics are far more capable, smaller and cheaper that the DFDR/CVR and communications links to them are of such antediluvian design. The only assumption that can be made is that these recordings are not considered important enough to spend time and money on.

Are there any other push button selections in the cockpit that have such a hit and miss design? Of course not - "but this is only DFDR so it doesn't matter"?

Concours77

"Are there any other push button selections in the cockpit that have such a hit and miss design? Of course not - "but this is only DFDR so it doesn't"

Likely assumed to be the cynic here, there is this:

DFDR is basically "default quality control". CVR is sufficient, and generally is more efficient in post accident conclusions, since the interface is a driving metric?

An onboard "fire" is required to have a probability of "once every one billion flight hours"

Since there are demonstrably more hull losses than chunks of fleet flight hours in the billions, a fire is less common than losing a hull full of people....

The first bit of wreckage (447) pulled up, and spirited immediately away to Toulouse was the avionics equipment bay contents. "What's that smell?" We are to believe it was St Elmo's fire, a common source of O3, the aromatic that can prompt "What's that Smell?"

Obviously no post flight fire, due to its position in a sea of extinguishing fluid, the wreck had important digital data to give up...

"Command and control" is not limited to the military.

There are no, repeat NO unbiased people present at the scene of an airplane crash....

Only police, military, and "invited parties". Assuming an accident might lead to losses, including economic, political, and prestigial, it is only reasonable to see this as common....

At the inception of each hull loss, a representative of the flying public, (the Raison d'être of the industry) must be present, call this person(s) an ombudsman if you will.

If a United States President can concoct a story to allow an immense invasion of a foreign country (pick one of many, Iraq, Syria, Vietnam?) does one believe it is outrageous for a multi trillion dollar industry to merely "collect and protect all evidence, then issue a report"?

And be honest about it? What are we smoking?

tdracer

Ian, it's not just the flight data recorders - most of the avionics on an aircraft are several generations behind the electronics in your phone. The reasons are simple - it's insanely expensive to certify new aircraft avionics, and more critical the system, the more it costs (the certification costs to change something in FADEC software is north of a $1 million USD - and that's just the certification costs - it doesn't include actually making the s/w change). Further, if you 'upgrade' something you need to make sure it's a drop in replacement - otherwise all the systems it talks to have to change as well. That's why most 747-400s are flying around with the same flight deck display system that was designed 30 years ago.
Further, while there are occasional 'block' changes of avionics, unless the regulators mandate it, they are seldom retrofit. As a result, there are still lots of commercial aircraft out there that still have the old analog foil tape recorders. I was involved in investigating a 747-200 freighter crash about 5 years ago that was the result of multiple engine failures after takeoff - foil tape FDR, the only engine parameter was EPR
I've looked at dozens of DFDR files over the years - usually incident investigations but occasionally accident investigations. Trust me, having something like the DFDR in an A330 or 777 is wonderful - they are head and shoulders better than the stuff from even 10 years earlier. Yes, they have their limitations but they are not that difficult to work with once you understand those limitations.

tdracer

Concours, until you've actually been involved in investigating the cause of death of hundreds of people, I suggest you shut up.

Concours77

I have, Except for the hundreds part. On one MVA, our defendant was prosecuted by a District attorney who withheld evidence, and by investigating police who sat on an eyewitness's testimony. The testimony was explicitly exculpatory, and when another investigator discovered it, the case was dropped.

This case in France is one of manslaughter. I have seen principals in manslaughter cases violate law. The defendant in the case I describe was looking at fifteen years in prison....yet the "authorities" wanted a conviction, at any cost. Don't lecture me....

tdracer

Concours, you're talking about suppressing the cause of an accident that caused hundreds of deaths, and it doing so putting additional thousands at mortal risk.
I've worked with perhaps a hundred other aviation professionals who have been involved in crash investigations at one time or another. They all had one overriding concern during the investigation - finding the root cause so corrective action could be taken and no one else need die.
Stop slandering those people...

Concours77

As an investigator, I am familiar with the profession. My comment you cite has to do with on scene; an emotional, raw, and highly challenging format.

My problem is not the dedicated professionals to which you refer. Examination and interpretation of evidence is a fine art. My issue is with the "after the fact" partisans, the agencies, the politicians, the investors, etc. and the lawyers, not to mention the spinners.

I cast too broad a net, and I apologize....

IcePack

Tdracer wasn't it a Swiss company that proved the FDR was switched during an early Airbus crash investigation?
Methinks in this day & age their is too much money involved to have absolute reliance on ALL investigators ethics. Concours I am sure is not slandering, but just being sceptical.

tdracer

Fair enough - apology accepted. It's just something I've very sensitive to - the accusations than invariably occur whenever the cause points to something other than the aircraft - that Boeing/Airbus/etc. 'rigged' the data. While the manufactures have dedicated air safety investigators, they regularly pull in subject matter experts from the various engineering areas. I was only intimately involved in one really bad accident - the Lauda 767 that crashed when a thrust reverser deployed during climb. It was perhaps the hardest thing I ever did in my 40 year career - and the reverser wasn't even my system. At first the investigation wasn't that bad because everyone thought it was bomb, but that changed. The DFDR was destroyed and didn't yield any useful information, but we were able to get a printout from the FADEC non-volatile memory. That memory dump was the smoking gun that the T/R had deployed at 24k/Mach 0.78, and I was one of the first people to see the evidence . A deployment it flight wasn't supposed to be able to happen, but somehow it had and 223 people had died because we'd missed something.
I went home that night and put away the better part of a bottle of Scotch . I honestly don't know how my friends who do crash investigations for a living do it...