AF447 wreckage found

[Loose rivets]

I also do not mean to come across that the pilots on flight 447 were in anyway incompetent.

Seems you don't have to. But, it's so, so easy to fly those last few minutes from the comfort of one's armchair. It must have been enough to throw a good mind into chaos for a moment or two, but for those with very limited time handling real hardware in extreme conditions, it seems what they had in front of them was more than they could unravel. It takes a lot of willpower to shut out a vista of sophisticated equipment and concentrate on one small and comparatively primitive device.


BBC News - Air France Rio crash: Pilots 'lacked training'

[Mac the Knife]

Sorry, very ignorant person confused here.

Did they not have an attitude indicator/artificial horizon working (I thought that they did)? Would that not have shown that they were markedly pitched up and given them a clue to what was going on?

[Loose rivets]

As far as I know after some weeks torn from the internet, the standby horizon was unaffected by the computer derived errors. Although the main systems were probably okay, the idea of the standby is that its power supply should be maintained and its internal workings unaffected by the plethora of information inputs. If in doubt, check that little unit.

Just a video of a random 330. Start just after a minute in. Worth looking at the scan of first the standby horizon, then captain's screen, then it goes through standby to the right. A bank has been started.

Even the standby is electronic. I would love to see a gyro in there somewhere.

‪A detailed flight deck view of A330‬‏ - YouTube

Probably a translation plus press-speak, but some of the dialog must be from the transcripts. Totally horrifying.


Cockpit terror of jet's 38,000ft death plunge - News, Frontpage - Herald.ie

[Mark in CA]

Just reading the NY Times report on this (http://goo.gl/IKmQy), which ends with a paragraph that says:

Since the accident, both Airbus and Boeing have modified stall-recovery procedures with guidance from safety regulators in the United States and Europe. Safety experts say those procedures now instruct pilots to first lower the nose of the aircraft, regardless of altitude, and, if necessary, reduce thrust to avoid excessive acceleration. Previously, the standard procedure when nearing a low altitude stall was to raise the nose by around 5 degrees and maintain thrust.

First, this just so elemental it seems to be rediculous. But that last sentence really surprised me. In my extremely limited experience as a GA pilot, I was always taught that the first thing you do when approaching a stall is lower the nose. Can someone explain why the "standard procedure" has been to raise the nose?

[ChristiaanJ]

Originally Posted by Loose rivets

As far as I know after some weeks torn from the internet, the standby horizon was unaffected by the computer derived errors. Although the main systems were probably okay....

So far there are no 'hints' either that the main attitude displays were affected at all.

...the idea of the standby is that its power supply should be maintained and its internal workings unaffected by the plethora of information inputs. If in doubt, check that little unit.

Not relevant here, but the "old-fashioned" electro-mechanical standby AH would remain stable and reliable (because of the high gyro rotor speed) for minutes after everything else went "kerplunk". It's still credited with saving a Caravelle when the entire electric systems went belly-up.

----

Just a video of a random 330.

Not quite random, since it has an independent 3-inch standby A/I. The AF A330 that crashed had an "ISIS", a bigger standby 'instrument cluster' on a separate screen (there should be a photo somewhere on this or the other AF447 thread). Hence I assume the video was not of an Air France A330.

Even the standby is electronic. I would love to see a gyro in there somewhere.

I don't know what the type/manufacturer of that particular standby A/I is.
But I think that you'll discover the old and well-known fully electro-mechanical SFENAs have now left the stage, but that behind that electronic display there's still an independent gyro. Switching to an electronic display just made the mechanical bits simpler....

Probably a translation plus press-speak, but some of the dialog must be from the transcripts. Totally horrifying.

The dialogue that the Herald quotes is indeed based on translated excerpts from the transcripts.

[Two's in]

Can someone explain why the "standard procedure" has been to raise the nose?

I believe the (flawed) training analysis is along the lines that if you learn to recognize the onset of the stall early enough, then an increase in thrust (and hence airspeed) will fly you out of the stall threshold WITHOUT any reduction in altitude or requirement for a nose down attitude.

The flaw in this training philosophy is that it assumes you will always recognize and react at early stall onset, so the full stall never develops. If you never couple this training with full stall recovery techniques, eventually pilots end up not knowing what to do if a full stall develops.

This originates in flawed training analysis where the fear of losing any height at all during stall recovery on an approach has the effect of removing the basic skills required to recover from a fully developed stall.

[Loose rivets]

Yes, I think it is based however, on a full compliment of engines running at say, cruise power, and an attitude that wouldn't be that much of a change when hurriedly pulling to +5. The power increment could then be metered out by erm, skilled judgment.

The crew were perhaps dealing with not only vigorous turbulence, but also more than a few man-made excursions from the horizontal. Also trying to recognize a stall with massive lifting and down-droughts would certainly be very much more difficult. Stalled one moment, and flying with reduced wing loading the next.


What I find hard to understand is the 'Deep Stall' nature of this decent. There does not seem to have been a succession of conventional stalls, yet the aircraft is not one of the T tailed types that could lock one into such a decent. I can only conclude a lot of time was spent with the systems blurring the issue, indeed applying power with vectored thrust 'shaping' the angle of stabilization to some deceptive angle.

Is there a clear recording of the flight-deck ambient noise? Once height was reduced enough to give a good gap between stall speed and overspeed, the ambient noise would have been quite different. The silence should have been deafening.

[Capt Turbo]

The previous recovery procedure was based on the fact that all widebody aircraft with hi bypass engines is capable of stall recovery with minimal altitude loss due to quick engine response and massive thrust. If slats is extended at the same time the aircraft can be flown out of the stall without altitude loss if at low level.

An A330 in degraded law at high altitude stalls at 7 degrees AoA (vs 15 degrees at low altitude), max thrust = climb thrust, and the recovery pitch is somewhere between 0 to -5 degrees and requires 4000īif done correctly. If you do not initiate recovery promptly, but hold the stick back, lack of elevator response will cause the THS to trim nose up, just as the report suggest.

In this situation, if the AoA becomes extreme, you may have to use rudder to get the nose down, and it takes a very low nose down attitude to unload and regain airflow.

Now, with this attitude (more than 20 degrees n.d.), once the airflow is back, the acceleration is huge - especially in manual TOGA - and you [B]must[B] start a smooth recovery immediately while avoiding over-stress ( fortunately the Airbusses have g-meters), secondary stalls and Vmo excursion. Vd excursion with structural failure is a distinct risk.

Recently, we have done the AF profile a number of times in a very good CAE SIM: recovery is possible at 37000īif done properly. After holding the stick back and letting the THS trim aft, the ensuing deep stall is IMHO not recoverable for the average line pilot, and that brings up a new question:

To which extreme situations shall we select and train commercial line pilots?

Oh, we all want to keep our cosy straight-and-level job, and the operator just want to fill a seat, so the mental break point SIM assessments has gone out of fashion, and no one knows if the co-pilot is really up to it, if the poop hits the fan.

What is left is training, and while some operators are doing a great job, some are not, partly because proper high altitude recovery training takes time, and the allocated training time is already full of NPA, V1 cuts, ECAM work, Evacuations etc. etc.....

Hopefully this tragic event will trigger some thoughts in the training departments and among the regulators, so we can swop some of the endless checklist reading with some hands-on training. Getting extra time....? In your dreams.....

And BTW, 25 years ago I initiated a hi-level training program for the old captains; many of them didnīt have a clue of what was going on up there unless they had survived a tour on one of the "interesting" fighter types of the day, and only a few of them had actually recovered from a stall in the Starfighter, the Phantom, the Lightning or the Mirage.

[Greek God]

Correct me if I am wrong (I'm sure someone will!) but I thought the revised stall recovery was as a consequence of the 737 stall on approach into Bournemouth and similar incidents which application of TOGA could cause an uncontrollable pitch couple especially if out of trim?

[TioPablo]

B4-Landing said:

Until bean counters quit ruling the earth which I do not think is going to happen anytime soon we have to stick together and do the best we can with the crap we are given.

Thanks B4

[Clandestino]

Originally Posted by Loose rivets

The crew were perhaps dealing with not only vigorous turbulence

Vertical acceleration trace is not consistent with flying through turbulence. There is +1.6 G spike as the aeroplane starts its climb, after that, it remains between +1.25 and +0.7G for the remainder of the flight.

Originally Posted by Loose rivets

What I find hard to understand is the 'Deep Stall' nature of this decent.

Trimmable horizontal stabilizer went to maximum nose up deflection. For most of the upset, both elevators were at their maximum nose-up deflection, too. Twice they start moving towards nose-down, first prompted by command from RH stick, second time by LH, but they only get halfway to neutral before new pull on the sidestick sends them back to their NU stops. Seemingly FBW was in Altn law so elevators were trying to satisfy G demand from pilot controls. It goes to show that even when there's no direct stick displacement to control deflection things work out in quite conventional way.

[misd-agin]

Have to wonder if the FB(in the left seat) and CA(jumpseat after returning to cockpit?) could tell that the FO had the stick in the full nose up position?

With a sidestick, especially at night, I wonder if anyone could see the FO's sidestick inputs?

[larssnowpharter]

I work for a major oil/gas and petrochemical company and a fair amount of my job involves investigating what we call catastrophic events. For this reason I follow investigations into civil aviation accidents as they are probably the best investigated accidents available and, generally, set a standard for other industries.

I have a background in military and sport aviation although not commercial flying.

This interim report from BEA has resulted in headlines in the press that basically lays the fault/blame on the crew. Indeed, it would seem that errors were made and I can see at least 6 of DuPont's Dirty Dozen raising their combined heads above the parapet. No doubt the cockpit transcript will provide excellent training in Human Factors in the future. Sad but true.

I have no idea of the BEA's remit in terms of how far it can go in a proper RCA of this event. So far we seem to have a good idea as to the WHAT and the WHO. However, when we get to the WHY we have the all too common finger pointing at the training system. Sure, a failure in training can be rectified relatively easily if the will and money is there. However, my view is that this is an easy cop out and that the true root cause can be disguised all to easily by following this route.

So, let's ask ourselves a further WHY: Why had the training system NOT identified the need to train pilots to deal

with inaccurate IAS indications at high level

especially in the light of previous events as well as known reliability issues with the pitots.

I would like to see this next step taken. There was clearly a decision making (or, more likely, a lack of decision making) process involved here that may be the real root cause of this tragic event. Was complacency also a factor here?

Lastly, I find AF's comments re the lack of AoA indications (is not the stall warning an AoA indicator?) specious and more likely made in what will no doubt be a blame game between Airbus and AF decided in the courts.

[TioPablo]

Sure, a failure in training can be rectified relatively easily if the will and money is there.

Do that! Donīt dare to doubt please! It will pay out in time... It is just a question of time... Dare to invest in the future, but... Donīt be afraid... You have a lot of good ppl going around.... You are one of them

[before landing check list]

Thanks B4

De nada dude.

[MountainBear]

I deleted my prior post because frankly it was hot-tempered. Let me try to be more civil and to the point.

At this stage in the aviation industry's development it is clear that airline accidents have become nothing more than blame game with software designer/engineers on one side, hardware manufactures on the other, and the pilots stuck right in the middle. Rather than either of the primary culprits taking responsibility the industry's response to every accident is to trot out PILOT TRAINING as the grand panacea that will cure all the ills.

This accident is the poster child for pilots taking the heat when the real culprits are (a) poorly thought out software design decisions and (b) flaky hardware. That doesn't mean that I think the pilots performance was mistake free. But it's obvious that they were set up to fail and MOAR TRAINING is not the answer to the problem.

[before landing check list]

Bear, maybe the (or A problem) is software is needed for basic functions. I do think that if there were some real instruments requiring hardware such as a tube that was relaying ram air pressure from the outside DIRECTLY to the inside to an instrument which would translate that to a circular scale and call this an IAS dial, and somewhere close to it have a real gyro to give attitude information with a battery back up and another one with another tube DIRECTLY reading the pressure outside the aircraft to another circular dial and call this an altimeter. Then SOMEWHERE on the panel mount these where the computer operators can see them and before they climb into the cockpit give them training on how/when to use them. While you are at it have a freaking big red button between the seats that when pushed there is no right stick did this, left stick did that and if the moon is at a certain light level and it is Tuesday the software still may or may not give you what you ask for and in doing so the computer operators now turn into pilots.

Face it, the freakin system as a whole was designed to eliminate pilot mistakes. It was not expressly designed to lighten workload, was already had all we needed for that. Not only was it to eliminate our possible mistakes but to also cover up our lack of basic flying skills with designers and software engineers who thought they knew more then we did on how an aircraft is supposed to be operated. We were sold a wagon full of crap now we must do our best to operated within this absurd criteria.

Dudes you need to rise up with your unions and take your positions back, this is not working.
Now let us hear the bitching from the My god is my software and I shall covet no other people.

[Zorin_75]

While you are at it have a freaking big red button between the seats that when pushed there is no right stick did this, left stick did that and if the moon is at a certain light level and it is Tuesday the software still may or may not give you what you ask for

I had already asked you this earlier, but now that we have a much more detailed record of 447s final minutes maybe you're ready to answer - at which point didn't "the software" give them what they asked for?

[before landing check list]

Zorin, hold on a sec. I did miss your prior question to me. So here goes an answer flown an AB aircraft;
Flight 296 in Paris. I do not think the pilots flew into those trees on purpose.

Quoted from misd-again

Have to wonder if the FB(in the left seat) and CA(jumpseat after returning to cockpit?) could tell that the FO had the stick in the full nose up position?
With a sidestick, especially at night, I wonder if anyone could see the FO's sidestick inputs?

So if you were to jump into the right seat and pushed forward on the stick whom would the computer attempt to obey? Either way somebody would not be getting what they asked for and worse that person would never know because there is no feedback. They would just be assuming for whatever reason the inputs they were trying to give were not working now it is time for plan B, C, D etc. All of this happens in seconds adding to the confusion. So you end up with the left hand not talking with the right hand and the computer through lack of feedback not talking with either of them.

Do you think the pilots wanted the THS to go full ANU or do you think the computer assumed one (not both) of the pilots wanted it? I say both because who really has command when the side sticks are displaced in different locations? I really do not know.

In fairness to the PF (whichever one it was) the only 'nose-up input' he (The PF) applied at the onset of the accident was immediately after the sign-off. The BEA states that the 'zoom climb' started at least 11 seconds after that - and there is no mention of the PF moving the stick either way until he applies 'nose-down' to counteract the climb. Does this mean the PF did not cause the climb but it takes several seconds for the AB330 to respond to control movements?

[Hand Solo]

The software gave them what they asked for at all times, but what they asked for was dumb. Regrettably to me this boils down to the fact that if you lose your airspeed instruments at altitude, close to your performance ceiling, initiating a 7000fpm zoom climb is about as wrong a response as you can have. Having done that if you then find yourself descending rapidly with the nose high you should recognise the stall for what it is. Its no good blaming the software or the system. If you did the same things in a 737 you'd get the same result.

Edit: if you have dual inputs on the sticks it will tell you, and you can override the opposite stick with the priority button. This accident was not caused by confusing dual inputs but by inappropriate and sustained solo inputs. The computers don't make assumptions, they follow commands. What the THS is doing is largely immaterial so long as the aircraft retains the capacity to deliver the commanded load factor, which it did throughout. Pulling and holding full aft stick in an Airbus is akin to pulling full back stick and holding full aft trim on a Boeing. There is no mystery or confusion about it, it's Airbus FBW 101.