Join 3,381 readers in helping fund MetaFilter (Hide)


What really happened aboard Air France 447?
December 7, 2011 2:29 PM   Subscribe

What really happened aboard Air France 447? With the discovery of the cockpit voice recorders comes a tale of Air France 447's final moments. Previously and previously
posted by thewumpusisdead (144 comments total) 38 users marked this as a favorite

 
well that's fucking terrifying.
posted by nathancaswell at 2:39 PM on December 7, 2011 [18 favorites]


I saw this earlier today and it's a sad read. You can vicariously feel their panic and it's frustrating not to be able to reach through time and help them.
posted by neuromodulator at 2:39 PM on December 7, 2011 [19 favorites]


Most hair raising thing I've read in weeks.
posted by nutate at 2:40 PM on December 7, 2011 [2 favorites]


Shit. I didn't know about the Airbus' asynchronous controls. What a horrible joke.
posted by Burhanistan at 2:43 PM on December 7, 2011 [3 favorites]


Some people I know who fly Airbuses commercially nicknamed them the "Scarebus" in the 1990s for their propensity to give pilots accustomed to more control the heebie-jeebies during casual use.
posted by meehawl at 2:48 PM on December 7, 2011 [1 favorite]


Wow. I hadn't realized the pitot tubes de-iced so early in the sequence. Had the plane not been programmed to deactivate the autopilot when it stRted getting anomalous airspeed info, or if it had been programmed to switch back to auto as soon as normal data resumed, everything would have been fine. It's such a harrowing thing to read--partially because anyone who is at all honest can see themselves in it. Talk about human, all too human.
posted by yoink at 2:49 PM on December 7, 2011 [1 favorite]


I went to read TFA and then very quickly decided I probably did not want to read TFA. What an emotionally difficult job it must be to be an accident investigator.
posted by jepler at 2:50 PM on December 7, 2011 [3 favorites]


accustomed to more control

Except that the problem here is precisely that the pilots had full controll. If only the plane had not given them complete control, everyone would have lived.
posted by yoink at 2:51 PM on December 7, 2011 [4 favorites]


Fascinating, and frightening. It's so easy to sit in a warm chair at a computer and think "what a maroon, I'd never do something like that!" At night, trying to recover from an upset, while inside a violent thunderstorm, with warning bells going off and everyone yelling, I'd probably find it hard to find my ass with both hands let alone control a 100-ton airliner.

The writing could be better. Besides the usual "pilot/co-pilot" BS (there is no such thing as a co-pilot; captains and first-officers are all fully qualified pilots) I had to smile at this:

Aside from the loss of airspeed indication, everything is working fine.

Airspeed is probably the most critical thing about flying.
posted by phliar at 2:51 PM on December 7, 2011


What a great, well-written, almost beautiful piece. Enough technical explanation to feel like we're getting some semblance of the full picture without going needlessly egghead, expertly chosen detail to add terrifying atmosphere (e.g. the St. Elmo's fire), minimal speculation about motives and thought processes to add color without going overboard, just enough expert commentary that answers questions without disrupting the story.

Scary as hell (can you imagine the horror at "Damn it, we're going to crash... This can't be happening!"), but also one of the best reads I've run across in a while.
posted by eugenen at 2:51 PM on December 7, 2011 [5 favorites]


Except that the problem here is precisely that the pilots had full controll.

More like the co-pilots were continually battling for control, and the one who didn't know what the hell he was doing won.
posted by mudpuppie at 2:52 PM on December 7, 2011 [13 favorites]


Fantastic article. The asynchronous control thing was new information to me and is utterly baffling to me. Particularly in alternate law when flying by feel seems like it'd be pretty important.

One thing aviation training is big on is hazardous attitudes, a collection of dangerous mindsets that lead to accidents. The attitude described in this article is a little different from the five that are the main curriculum. It sounds like incredulity, like they just couldn't believe and understand what the systems were telling them was happening. And instead of quickly getting creative they froze up. What a shame.
posted by Nelson at 2:53 PM on December 7, 2011 [1 favorite]


Had ...[X happened when Y]... everything would have been fine.

You can say this about any accident. It's always a chain of events, and most of the time you can't really say unambiguously that foo was the cause.
posted by phliar at 3:01 PM on December 7, 2011 [1 favorite]


What a harrowing read. It seems much of the problem was that the right-seat guy, Bonin, had his controls pulled back the entire time, and no one else knew it until it was too late. They knew they were in bad trouble and couldn't understand why.

Those poor people.
posted by Gelatin at 3:02 PM on December 7, 2011


This is shattering. So it really all comes down to human error after all... I cannot begin to imagine what was going on in the rest of the plane.
posted by kinnakeet at 3:05 PM on December 7, 2011 [1 favorite]


As the plane approaches 10,000 feet, Robert tries to take back the controls, and pushes forward on the stick, but the plane is in "dual input" mode, and so the system averages his inputs with those of Bonin, who continues to pull back. The nose remains high.

I'm trying to think of any situation where averaging the inputs from the two sticks would have a positive result and am coming up blank. It seems completely insane. Why on earth would they design that system that way?
posted by enn at 3:06 PM on December 7, 2011 [29 favorites]


I've read a lot of the articles about this crash. The difference this time is learning about the two stick inputs. What I don't get is why the hell Airbus takes both stick inputs and averages the two. Had the sticks not been asynchronous there would have been no way they could have fought over control without knowing they were fighting each other. Can anyone with flying experience explain why anyone would want to average? Because it seems ridiculous in my head...
posted by Mister Fabulous at 3:07 PM on December 7, 2011 [4 favorites]


Unlike the control yokes of a Boeing jetliner, the side sticks on an Airbus are "asynchronous"—that is, they move independently. "If the person in the right seat is pulling back on the joystick, the person in the left seat doesn't feel it," says Dr. David Esser, a professor of aeronautical science at Embry-Riddle Aeronautical University.

Are you kidding me???

That seems, on its face, a situation that just invites trouble.
posted by Cool Papa Bell at 3:11 PM on December 7, 2011 [6 favorites]


JFC, harrowing read is right.

This is one Metafilter link I'm not forwarding to my wife. She's already deathly afraid of flying and this won't help matters at all. We just visited Beijing last week and she was a nervous wreck for months beforehand thinking about the flights.
posted by kmz at 3:11 PM on December 7, 2011 [1 favorite]


Mister Fabulous: "Can anyone with flying experience explain why anyone would want to average? Because it seems ridiculous in my head..."

Crowdsourcing!
posted by Joakim Ziegler at 3:13 PM on December 7, 2011 [21 favorites]


Unable to finish reading the transcript, too much like a snuff film to me.
posted by Old'n'Busted at 3:15 PM on December 7, 2011 [2 favorites]


This was bad enough when it seemed like the pitot tubes icing up were the root cause but now it's just horrible. There were so many opportunities to recover, and none of them were taken.

That said, it's pretty clear that the overall system design of the Airbus seems faulty, as if it was designed by non-pilots with no experience with large planes and how they react outside of normal flying conditions.
posted by tommasz at 3:15 PM on December 7, 2011 [1 favorite]


I'm trying to think of any situation where averaging the inputs from the two sticks would have a positive result and am coming up blank. It seems completely insane.

What I don't get is why the hell Airbus takes both stick inputs and averages the two. Had the sticks not been asynchronous there would have been no way they could have fought over control without knowing they were fighting each other.

Are you kidding me??? That seems, on its face, a situation that just invites trouble.

I agree with each of these statements. What gives?
posted by Gelatin at 3:15 PM on December 7, 2011 [2 favorites]


02:10:55...The problems that occur from this point forward are entirely due to human error.

Jesus. My heart just filled with despair when I read that part.
posted by lord_wolf at 3:16 PM on December 7, 2011 [1 favorite]


More like the co-pilots were continually battling for control, and the one who didn't know what the hell he was doing won.

There's only a fairly brief period when they're directly battling for control--and they're already in a power of trouble at that point. Both take turns getting it wrong before that.
posted by yoink at 3:16 PM on December 7, 2011


Really a combination of incompetence and bad design. But. When you have something of this complexity, it seems almost inevitable that there will be serious flaws and it's just a matter of time and luck when that flaw has a deadly outcome. Put in a pilot with a particular weakness together with a design flaw, and there you have it - disaster. This is why I always felt distaste when I saw the Quantas bragging about their perfect record. A big heaping of luck. That's not to say that objectively huge flaws and massive incompetence don't also exist. Like here.
posted by VikingSword at 3:19 PM on December 7, 2011


BTW Putain does not mean "damnit."

(it means "whore")
posted by zomg at 3:20 PM on December 7, 2011 [4 favorites]


Literally, yes, "putain" means "whore". In the field, however, it's used exactly as "fuck!" is in English.

While we're here, "éventuellement" doesn't mean "eventually" (as the article has it) either. It means "possibly".

So endeth the lesson.
posted by Wolof at 3:24 PM on December 7, 2011 [11 favorites]


zomg: Yes, but it's used as an interjection much the same way we use "damn." A literal translation ("Whore, we're going to crash... This can't be real!") sounds nonsensical.
posted by dhens at 3:24 PM on December 7, 2011 [2 favorites]


02:10:55...The problems that occur from this point forward are entirely due to human error.

Jesus. My heart just filled with despair when I read that part.


Oddly, I felt the reverse. It seems better somehow that such an obvious (with hindsight) human error caused the crash because now we know the chain of errors and the spread of that info means that it will likely never occur again. In contrast, if a defect in a relatively cheap mechanical part could bring down an aircraft through zero fault of the pilots, then the whole theory that modern aircraft are safe because everything is designed to be double and sometimes triple redundant would be out the window.
posted by modernnomad at 3:25 PM on December 7, 2011 [9 favorites]


Oh of course, guys. But people need to know the literal as well as the colloquial
posted by zomg at 3:26 PM on December 7, 2011


Ugh, that was awful to read. Bonin freaked out and, like people under intense stress, reverted to what he was most practiced and familiar with; gaining speed and altitude at sea level. Unfortunately they weren't at sea level, they were way up in the thinner air so it was exactly the wrong thing to do.

Yeah, it was a strong of errors and problems, but that string looks a lot simpler than they usually are for these major disasters. Three things cover almost all of it;

1) The least experienced pilot was in charge of the controls during a severe crisis.
2) That pilot did exactly the wrong thing because under severe stress he couldn't think through the problem, that he can't successfully generate speed and ascend at the same time at that altitude.
3) The controls were for shit and the guy in the left seat had no idea the guy in the right seat was pulling back full force on his stick while he was pushing forward because there was no feedback.
posted by Justinian at 3:26 PM on December 7, 2011 [5 favorites]


modernnomad, after reading your post, I thought about it some more, and now I agree with you. Thanks for the counterpoint!
posted by lord_wolf at 3:30 PM on December 7, 2011 [2 favorites]


I'm gobsmacked trying to figure out how they didn't know they were in a nose-high attitude for most of the time. The pitot tubes freezing up shouldn't have affected their attitude indicators. Though in the confusion they may not have known which instruments to trust.

I've taken flying lessons, and stall recovery was part of them. And the instructor couldn't have been more clear: You don't pull back on the yoke to recover from a stall. I can't imagine what I'd have done in their place, but I also can't imagine why a professional airline pilot would keep hauling back on the stick like that.

For that matter, the fact that no one seemed to acknowledge the stall warnings at all is baffling.

Justinian, I'd add that the captain didn't seem to be in full control of the situation; he certainly didn't know that Bonin was pulling the nose up hard. What rotten luck that he had gone for a rest before the emergency occurred.
posted by Gelatin at 3:31 PM on December 7, 2011 [2 favorites]


Both take turns getting it wrong before that.

For example:
02:11:37 (Robert) Commandes à gauche!
Left seat taking control!

At last, the more senior of the pilots (and the one who seems to have a somewhat better grasp of the situation) now takes control of the airplane. Unfortunately, he, too, seems unaware of the fact that the plane is now stalled, and pulls back on the stick as well. Although the plane's nose is pitched up, it is descending at a 40-degree angle. The stall warning continues to sound. At any rate, Bonin soon after takes back the controls.
Yeah, Bonin fucked up the most, but Robert was not on top of the situation either.
posted by yoink at 3:31 PM on December 7, 2011 [1 favorite]


The article speculates that the reason the pilots may have been unwilling to believe the plane stalled was that they may have believed it impossible for the plane to stall because of the automatic systems because they didn't understand that those systems shut themselves off when the pitot tubes went out.
posted by Justinian at 3:35 PM on December 7, 2011


The asynchronous controls is the scariest thing I've learned this month. Maybe this year. At the very least there needs to be some sort of visual indicator to the other seat what the controls are doing, and a way to override a panicked pilot.
posted by maxwelton at 3:36 PM on December 7, 2011 [4 favorites]


Yeah, attributing the problem just to "human error," when the controls seem designed to make it impossible to catch your co-pilot's errors if you're not deliberately looking for them, is facile. The design of the plane was absolutely part of the disaster.
posted by Holy Zarquon's Singing Fish at 3:36 PM on December 7, 2011 [12 favorites]


I have a friend who's wife and family was on that flight. I simply don't know what to think he must be feeling knowing this now.
posted by michswiss at 3:36 PM on December 7, 2011


Did any hardware change recommendations come out of this?
posted by maxwelton at 3:36 PM on December 7, 2011


The article speculates that the reason the pilots may have...

That was possibly the worst sentence I have ever written. Wow. Just... rewrite it in your heads.
posted by Justinian at 3:37 PM on December 7, 2011


The account seems a bit reminiscent of another 2009 crash, that of Colgan Air Flight 3407. A bit of icing, the pilots lose situational awareness, end up stalling the plane, even though there is nothing mechanically wrong going on, at which point the right and left-hand seat pilots start working at cross-purposes.
posted by Numenius at 3:39 PM on December 7, 2011 [1 favorite]


God that was chilling. As an engineer, I've been in the same situation - something is going wrong, and I'm either too close, too removed, or too inexperienced to figure it out. Except when things go wrong in my field, we break something or set back our schedule in some way.

The most nightmarish part for me is at the end, when the inexperienced copilot asksBut what's happening?
posted by muddgirl at 3:41 PM on December 7, 2011 [2 favorites]


From the article:
Just then an alarm sounds for 2.2 seconds, indicating that the autopilot is disconnecting. The cause is the fact that the plane's pitot tubes, externally mounted sensors that determine air speed, have iced over, so the human pilots will now have to fly the plane by hand.
It seems like it would be a good idea to try to figure out alternate ways for the plane to measure wind speed, maybe with some kind of radar system or backup anemometers that are heated forcefully enough to avoid icing?
It's quite possible that Bonin had never flown an airplane in alternate law, or understood its lack of restrictions. According to Camilleri, not one of US Airway's 17 Airbus 330s has ever been in alternate law. Therefore, Bonin may have assumed that the stall warning was spurious because he didn't realize that the plane could remove its own restrictions against stalling and, indeed, had done so.
Interesting. With something like this maybe there should be a reminder that the controls had gone unreliable.

What's amazing is that airplane pilots have to almost be engineers, most of the serious work they do is just sitting around waiting for the autopilot to screw up, and when it does figuring out what's wrong and diagnosing the problem -- and if they screw up, everybody dies.

The dual-stick thing... it would be easy to add some force-feedback to the sticks so that each side knows what the other is doing. The pilots thought they had lost control of the plane because their inputs weren't working, they didn't realize they were working at cross purposes.
posted by delmoi at 3:43 PM on December 7, 2011 [1 favorite]


I'd like to read more background on the asynchronous control design. I tried Googling a bit but everything's dominated by this one account.

the captain didn't seem to be in full control of the situation ... What rotten luck that he had gone for a rest before the emergency occurred.

This part of the narrative baffles me, too. Thunderstorms are serious business for all planes, small or large. Even the biggest planes try to go around major weather systems. There's been a fair amount written about why AF 447 went through this storm, a combination of bad weather reportage and having no good alternative. But why would the captain leave the cockpit at that exact time?
posted by Nelson at 3:44 PM on December 7, 2011 [1 favorite]


I Googled "asynchronous controls airbus" and got pretty much just this story (and commentary thereon) and various aeronautic pages which clearly use the term to mean something completely different (having to do with modes of data processing). I wonder if there might be a misunderstanding here. Any pilots who could weigh in?
posted by yoink at 3:45 PM on December 7, 2011 [1 favorite]


Ha! Jinx.
posted by yoink at 3:46 PM on December 7, 2011


I'm trying to think of any situation where averaging the inputs from the two sticks would have a positive result and am coming up blank.

This of course depends on what you think the stick does. Is it a force input? (That is, the airplane should effectively be pushed proportionally to the force applied to the stick.) Or is it a displacement input? (The airplane's orientation is changed in proportion to the distance the stick is moved.) Or is it something else? (Moving the stick changes the camber of the appropriate aerodynamic surface, and the airplane then does whatever the laws of aerodynamics dictate.)

In an old-style airplane (e.g. a jetliner from the 70s) the last is true, because, well, that's all we knew how to build. It's also really complex and weird, and hard to learn and internalise. The Airbus approach is that we have a computer, let's use it: what do we think is the right thing for the stick to do?

I don't think you can argue from first principles that any of those approaches is necessarily superior. I, being a pilot, am obviously prejudiced and think that Option 3 is the only right thing; but that is only the result of the training I've had. I can recognise that flying a 100-ton machine efficiently at 40,000 feet (where the air is a quarter as thin and terribly cold) may be different from my own limited experiences.
posted by phliar at 3:46 PM on December 7, 2011 [2 favorites]


yyyyeeeeow I am not going to sleep tonight. I'm glad I read that (and, yes, there are a few glitches in the translation to English), but holy hell that was stressful.
posted by LMGM at 3:50 PM on December 7, 2011


In any case, I put the disaster squarely on the shoulder of the people who designed the flight control system. Letting the pilots operate the controls in a conflicting manner like this where they have no idea what the other is doing is a disaster waiting to happen. I'm only surprised it took this long.
posted by mullingitover at 3:51 PM on December 7, 2011 [1 favorite]


Year after year I've become less willing to get in an aircraft. This may have been the straw that broke the camel's back.
posted by tomswift at 3:53 PM on December 7, 2011


Averaging the control inputs probably makes a lot of sense if you assume that the two pilots have roughly the same information (from functioning instruments) the same purpose (flying the plane to its destination then landing safely) and the same plan to achieve that purpose (because they're both well-trained pilots).

It probably even makes sense in some of the "cross-control" use cases the designers considered -- a skilled pilot trying to keep the plane in the air working against a skilled pilot trying to crash it has a decent chance of at least keeping things straight and level (since you can average control inputs to zero).

Here, not so much. I presume that this will be used as input (a new use case, if you will) to the future programming of the control system.
posted by Vetinari at 3:54 PM on December 7, 2011 [1 favorite]


In any case, I put the disaster squarely on the shoulder of the people who designed the flight control system. Letting the pilots operate the controls in a conflicting manner like this where they have no idea what the other is doing is a disaster waiting to happen.

It seems unlikely to me that this can possibly be as "obvious" a problem as we seem to want it to be. I mean, it's not as if its a cost-cutting thing for airbus to have done this. In the absence of actual expert commentary on the reasons for this design, I think we would be wiser to suspend judgment.
posted by yoink at 3:54 PM on December 7, 2011 [5 favorites]


This is like nightmare fuel for interface designers: realizing that all the incredible bullshit processes that go into the making of almost all modern products (that designers suffer and fight and tear our hair out over every day) also applies to the critical systems that we entrust our lives to.
posted by danny the boy at 3:57 PM on December 7, 2011


Year after year I've become less willing to get in an aircraft. This may have been the straw that broke the camel's back.

Really? They really had to fuck up pretty heroically to crash this plane. I mean, if you're going to avoid all situations that might result in death if someone does something extraordinarily stupid over and over and over again then planet earth is a bad place to be. You sure as hell won't be driving anywhere ever again.
posted by yoink at 3:59 PM on December 7, 2011 [7 favorites]


... alternate ways for the plane to measure wind speed, maybe with some kind of radar system or backup anemometers that are heated forcefully enough to avoid icing?

Pitot tubes are already heated, and an airliner will have several. (If you can manage to get external access to an airliner or bizjet, you'll notice warnings around the pitots that they are heated and will burn you.) Unfortunately Nature can always come up with icing more severe than anything a puny human can build.

Radar is no good because that will measure your speed over the ground (groundspeed), which is not relevant. To control the aircraft you need to know the airspeed.

Incidentally, to all the people expressing outrage over the averaged "asynchronous controls" -- if both sticks are being pulled/pushed, what should happen? Average them? Take the larger one? Always let the one on the left override the one on the right? Whichever one you choose, I can come up with an emergency scenario where that would be bad.
posted by phliar at 4:01 PM on December 7, 2011 [7 favorites]


phliar: Incidentally, to all the people expressing outrage over the averaged "asynchronous controls" -- if both sticks are being pulled/pushed, what should happen? Average them? Take the larger one? Always let the one on the left override the one on the right? Whichever one you choose, I can come up with an emergency scenario where that would be bad.

I'm actually okay with the averaging... I can see that being very useful in a lot of situations. The disturbing part is the lack of feedback. Even the lowliest flight simulator enthusiast has a joystick with force feedback. I can see the designers not wanting to mess with the physical stick, so force feedback is out... but a little display that shows how your partner has their stick positioned? Or even a light or audible alarm when the sticks inputs are somewhere near diametrically opposed?
posted by gilrain at 4:06 PM on December 7, 2011 [6 favorites]


Incidentally, to all the people expressing outrage over the averaged "asynchronous controls" -- if both sticks are being pulled/pushed, what should happen?

Not have asynchronous controls? Have force-feedback indicating you are fighting each other? A big flashy red light with the words "hey assholes, stop it!" playing on an alarm?

While lack of communication was obviously a contributing factor, there was seemingly no other indication that they were both putting inputs on the stick.
posted by Mister Fabulous at 4:08 PM on December 7, 2011 [3 favorites]


phliar: "Incidentally, to all the people expressing outrage over the averaged "asynchronous controls" -- if both sticks are being pulled/pushed, what should happen? Average them? Take the larger one? Always let the one on the left override the one on the right? Whichever one you choose, I can come up with an emergency scenario where that would be bad."

Well, I lack the domain research to answer this confidently, but a couple of naive possibilities come to mind:

1. Defer to seniority: I'm guessing Captains consistently sit one side (left?) and the F.O. the other side. If you're in the big chair, your commands always supersede junior officers. Just like they do in the cabin.

2. Visual/force feedback: Here we assume that the main issue to resolve isn't arbitration of commands, but delivering relevant information without relying on humans to remember to do so in a stressful situation.

3. Manual selection: Always have one chair, and one chair only in command, set manually by the crew.
posted by danny the boy at 4:13 PM on December 7, 2011 [1 favorite]


I would caution people against a quick "Well, that's clearly a bad design, they should obviously just do X" reaction. I do some work in this sort of field and I have to tell you, the people that design these things aren't dumb. There may well be some problems that need fixing but the chances that Airbus is missing an obvious and intuitive fix are very, very low.
posted by LastOfHisKind at 4:14 PM on December 7, 2011 [12 favorites]


Stories like this are scary, to be sure, and I can't say I blame anyone who feels less inclined to board an aircraft after reading it, but personally I find the streets and highways of Toronto* (filled as they are with idiots with no respect for human life, apparently including their own) far more terrifying.

* feel free to insert the name of your own city of residence here
posted by The Card Cheat at 4:18 PM on December 7, 2011


Year after year I've become less willing to get in an aircraft. This may have been the straw that broke the camel's back.

The whole point of the article was demonstrate the spectacularly rare chance of this event happening. You have a much greater chance of dying by driving cross-country than you do in flying, no matter how great of a driver you think you are. In the year that AF447 crashed, almost 800 million passengers flew without incident in the US alone. Over 30 000 people died in car crashes in the US last year.

Powered flight is a remarkably safe way to travel.
posted by modernnomad at 4:18 PM on December 7, 2011 [5 favorites]


Where's an avionics expert to set us straight?
posted by Mister Fabulous at 4:20 PM on December 7, 2011 [3 favorites]


Incidentally, to all the people expressing outrage over the averaged "asynchronous controls" -- if both sticks are being pulled/pushed, what should happen? Average them? Take the larger one? Always let the one on the left override the one on the right? Whichever one you choose, I can come up with an emergency scenario where that would be bad.

Once again, I agree with the several comments following this statement. It isn't so much the behavior of the controls as the fact that the left-seat pilot had no way of knowing* that the right-seat pilot's actions were killing them all.

*Okay, he could have asked, but still, he shouldn't have to.
posted by Gelatin at 4:22 PM on December 7, 2011 [1 favorite]


I was following this very closely when it first happened. It is good to see a short and coherent summary of the incident.

I found it quite fascinating to read along with the on-the-fly analyses of the accident and the investigation at pprune.org, the various threads found here and here, with their internal links back two years. It's a huge amount of armchair analysis, but almost all the commentors are professional pilots or aero engineers. Indeed, they had deduced most of the major elements of the crash well before the black boxes were found.
posted by Rumple at 4:23 PM on December 7, 2011


I think "incompetence" is a pretty harsh allegation. Even the best of us make bad decisions. Be thankful yours don't cost over 200 lives and try not to be so smug.

Honestly, I can't help but think this is an inevitability that comes with greater automation. In the aggregate it makes things a hell of a lot safer, but those safety controls become so ingrained that operating outside of them only happens in such extreme or untried situations that they become impossible to account for. The good news is that this particular oversight will never be repeated. The bad news is that there will be some other one in the future, but maybe with increasingly less frequency.

(Which is already pretty damn low.)
posted by absalom at 4:27 PM on December 7, 2011 [1 favorite]


OK, this is a supremely naive question from somebody whose only piloting experience was playing some ancient version of MS Flight Sim eons ago, but why are there multiple inputs anyway? I can understand having two sticks so one pilot can take over for another one without having to shuffle seats, but then you could have a selector so that one seat or the other has control. Why have simultaneous non-identical inputs?
posted by kmz at 4:29 PM on December 7, 2011


LastOfHisKind: "There may well be some problems that need fixing but the chances that Airbus is missing an obvious and intuitive fix are very, very low."

Well of course everything is complicated, but that doesn't mean it's not possible that this was indeed a gross oversight that has an obvious fix, and has so far been ignored because in 99.99% of situations it wouldn't matter.

I believe people can be smart but still make mistakes. I believe organizations can be filled with smart people, but have fundamentally flawed processes. In fact, that sounds a lot more like the real world, to me.
posted by danny the boy at 4:29 PM on December 7, 2011 [3 favorites]


I think "incompetence" is a pretty harsh allegation. Even the best of us make bad decisions. Be thankful yours don't cost over 200 lives and try not to be so smug.

I hear you, and your point about automation is well taken -- the article indicates that the pilots at the controls couldn't quite believe the automated systems would allow a stall to occurr, and so they disregarded the information -- in the form of multiple and ongoing warnings -- that indicated that's exactly what was taking place.

It's just amazing that the pilots apparently had been so trained to depend on said automatic systems that the right-seat man, at least, seems to have completely disregarded the basic aeronautical skills he must have learned when first getting his license (or, at least, that I was taught in the couple of dozen flight lessons I've taken).
posted by Gelatin at 4:32 PM on December 7, 2011 [1 favorite]


1. Defer to seniority

This was the Way Things Worked, until Tenerife^ -- when the co-pilot saw the captain make a mistake, but assumed the captain knew what he was doing (he wasn't just the pilot, he was the airline's top trainer). The captain then flew the plane directly into another plane already on the runway.

Since then the industry has moved toward CRM -- crew resource management. The concept here is that all persons in the cockpit should be working together to solve the problem. I think the biggest failure here is not the Airbus control system, but CRM. Bonin, in particular, never communicated his stall-inducing pull-back until very late, and the first officer never asked him. If they had worked the problem as a team, they would have vastly increased their chances of survival.
posted by dhartung at 4:36 PM on December 7, 2011 [10 favorites]


.
posted by Xoebe at 4:37 PM on December 7, 2011


Pitot tubes (and my own lack of flying knowledge) aside, this looks like a UI clusterfuck:

1) No bright/loud indicator that the control override had shut down and the pilots now had complete and undamped manual control.

2) No replacement for the lost visual and tactile feedback from the change to async controls.

3) Limited feedback from the instruments on the physics acting on the aircraft - a combination of attitude, altitude loss and speed loss should have been part of some sort of combined indicator. Pilots: is the stall warning enough indication of this for an aircraft of this type?

Perhaps 2/3 could be combined into a real time digital diagram displaying control input and simplified force moments.
posted by vanar sena at 4:39 PM on December 7, 2011 [3 favorites]


if both sticks are being pulled/pushed, what should happen?

Well, what does Boeing do? The article specified "Unlike the control yokes of a Boeing jetliner..."

Presumably, it's some other behavior or warning system, or it wouldn't be brought up.
posted by Cool Papa Bell at 4:40 PM on December 7, 2011


There's been a fair amount written about why AF 447 went through this storm, a combination of bad weather reportage and having no good alternative. But why would the captain leave the cockpit at that exact time?

Hubris. Certainty that the machines could handle the situation.
posted by ricochet biscuit at 4:46 PM on December 7, 2011


Airspeed is probably the most critical thing about flying.
posted by phliar at 2:51 PM on December 7 [+] [!]


Landing is the most critical thing about flying. But I get your point.

The account seems a bit reminiscent of another 2009 crash, that of Colgan Air Flight 3407. A bit of icing, the pilots lose situational awareness, end up stalling the plane, even though there is nothing mechanically wrong going on, at which point the right and left-hand seat pilots start working at cross-purposes.
posted by Numenius at 3:39 PM on December 7 [+] [!]


Negative. Captain Marvin Renslow was the PF (pilot flying) and in command of the aircraft. The stick shaker was warning him of an impending stall and he overrode it. One of the theories which has been floated was that Renslow might have believed he had a tailplane stall rather than a wing stall, and the recovery procedure for that is opposite of a regular wing stall. But we'll never really know what he was thinking.

I can dig more up on it if you want.
posted by Thistledown at 4:47 PM on December 7, 2011


"In any case, I put the disaster squarely on the shoulder of the people who designed the flight control system."

No, it's pilot error.

"It isn't so much the behavior of the controls as the fact that the left-seat pilot had no way of knowing* that the right-seat pilot's actions were killing them all."

The reason why this is pilot error, and not just Bonin's, is that the single most important factor in any such situation is "crew resource management".

I've a book on my bookshelf called Why Airplanes Crash: Aviation Safety in a Changing World, by Oster, Zorn & Strong. It's more than a bit dated, published in 1992, but I bought it because after doing some research on this I found that it's among the most comprehensive scholarly works on this topic.

And here's the thing: plane crashes are the result of pilot error more than any other single cause, and the largest portion of those cases involve a crucial lack of communication and coordination in the cockpit. This was the case throughout AF447, beginning with Robert not explaining why he was advising a course change—because the radar had been set incorrectly and what seemed like a relatively safe course through the storm was revealed, upon Robert's correction of the radar setting, to be taking them directly into the worst of it.

Then, of course, there's the fundamental failure of Bonin to explain to either Robert, or the Captain, what he was doing, and why. Far more than anyone else, Bonin is to blame for this crash.

"...alternate ways for the plane to measure wind speed, maybe with some kind of radar system or backup anemometers that are heated forcefully enough to avoid icing?"

For those who have followed this crash and what's related to it, the subject of the icing of the pitot tubes has been very important. You need to understand that it's difficult to measure airspeed, and that's the only measurement that really counts. Yes, given a failure of being able to measure it directly, the groundspeed can act as some sort of vague reference—if Bonin or Robert had looked at the groundspeed, they'd have realized that their airspeed was far, far below what it should have been and that they were actually in a stall. Because, by the way, I'm inclined to think that the explanation offered in the article is true: that they disregarded the stall indicator because they didn't believe they could be in a stall. Also, they were generally uncertain about the instrumentation.

But they shouldn't have been. Adequate CRM, which crucially includes what the missing Captain would have provided, involves not just communication but directed problem-solving so that the crew isn't working at cross-purposes against each other and they focus on what are actually the most important problem at hand.

If I recall correctly, a lot of the impetus for CRM was influenced by the book I referenced and by one very important crash, the flight that crashed into the Everglades because the crew was focused on determining if and why the landing gear light was burned out and didn't notice that they were losing altitude. They simply slowly lost altitude until they flew into the swamp.

Anyway, IIRC, it was the case that there was only two or three times that these pitot tubes had all iced up and airspeed indication was lost, despite the redundancy of them and their heating mechanisms. It had happened, but pilots were not, at that time, generally briefed on this and they didn't expect it.

Reading that book, and listening to the recordings and/or the reading the transcripts of the cockpits of airline crashes, it's really quite striking how often crashes result from what are relatively avoidable pilot errors and a lack of communication between the flight crew. That's not to say that crashes result from maintenance errors (second most common) or design flaws; but that human error is most common. And it's generally not the sort of error that was inevitable, it's error that could have been avoided with better procedure and training.

By the way, I need to include the disclaimer that I'm not a pilot. This is just something I'm interested in and have researched to a limited extent.
posted by Ivan Fyodorovich at 4:53 PM on December 7, 2011 [20 favorites]


Yeah, attributing the problem just to "human error," when the controls seem designed to make it impossible to catch your co-pilot's errors if you're not deliberately looking for them, is facile. The design of the plane was absolutely part of the disaster.

Good luck getting a French investigation to admit that a French plane might have been part of the problem, though.
posted by kersplunk at 4:54 PM on December 7, 2011


Far more than anyone else, Bonin is to blame for this crash.

Is my impression that if Bonin had just said something like, "I'm pulling up!" at almost any point in the prior 3 minutes that the crash could well have been avoided? The first time he says something of the sort is, what, 45 seconds before impact and the other pilot's immediate response is, "No no no no!".
posted by Justinian at 4:58 PM on December 7, 2011 [4 favorites]


I think the word he was looking for was "unsynchronized" rather than "asynchronous".

And to +1 everyone else, some kind of feedback seems essential when the control system is receiving two inputs instead of one.
posted by joshwa at 5:00 PM on December 7, 2011


... and some big 'ol flashing light and siren screaming "ALTERNATE LAW! ALTERNATE LAW!"
posted by joshwa at 5:01 PM on December 7, 2011 [4 favorites]


I'm amazed to read that the pilots didn't know about alternate law. When the crash occurred, there was a lot of info on the web about the recovered ACARS messages (automatic telemetry from the plane) which included the string 'alternate law'. At the time, I was curious, and googled it:

http://www.airbusdriver.net/airbus_fltlaws.htm [first hit]

and there it is: The airplane CAN be stalled in Alternate Law.
Apparently the plane even says:"ALTN LAW: PROT LOST". You'd think pilots would be hugely aware of this.
posted by scolbath at 5:06 PM on December 7, 2011


joshwa: "And to +1 everyone else, some kind of feedback seems essential when the control system is receiving two inputs instead of one."

Any input at all, in fact. How on earth could one guy be messing with the elevator for four minutes without the other two people in the cockpit being aware? And who the hell designs a system where he has to say "I'm pulling up" for the co-pilot to become aware that this has been happening? They have dozens of square feet of instrumentation in there, surely they could spare a few inches to show the pilots what the control surfaces are doing.
posted by vanar sena at 5:08 PM on December 7, 2011 [1 favorite]


Yeah, ok. We'll do this again.


A fundamental principle of multi-pilot operations is Only one pilot flys the plane at any time. The sidestick, or yoke for that matter, is never, ever meant to be used simultaneously by both pilots. This is ingrained in the procedures used by pilots every day, in emergency and non-emergency situations alike. One person has control.

There is no good solution to combining inputs from the two sidesticks. If both sidesticks are being used at the same time, there's no way for the aircraft to decide which pilot is entering the "right" commands to the plane. The flight controls do need to ensure that if one pilot does something like passes out and slumps on the sidestick, the other pilot can always override that and bring the plane back to straight and level. Averaging the inputs accomplishes this. If something like this happens, there is a priority button on the sidestick that will disable the other pilots controls. This is designed as an emergency feature though, not as a normal method of resolving which pilot has control of the aircraft. The correct behavior is always for only one pilot to be flying the plane.

When both sidesticks are displaced from the neutral position, an audible alarm "Dual Input" is triggered. A voice in the cockpit literally says that, "Dual Input". This happens several times in the flight, and the warning is audible on the cockpit voice recorder. Both pilots clearly know when the other is making inputs. They don't necessarily know what the other pilot is inputting, but, as I discuss below, there is nothing to suggest that the pilots had conflicting intentions for any significant length of time.

Now one could argue that if the sidesticks provided force feedback about the position of the other stick, then it might have been more likely for one of the pilots to recover the plane. This assumes that the pilot (in the left seat) knew what the correct action for recovery was, which I don't believe to be the case. The first time that the left seat takes control, the command input is a left roll, not nose down. The second time the left seat actually enters a nose-up command, continuing the stall. There's a few exchanges back and forth, but the left seat is almost always commanding nose-up. There's one last nose-down input from the left sidestick, but it's far, far too late to recover.


To quote from the article:

At last, Bonin tells the others the crucial fact whose import he has so grievously failed to understand himself.

BS. The left seat pilot (Robert) told Bonin to climb just one second earlier, of course Robert knows that Bonin is pulling nose-up.

Descend, then... Give me the controls... Give me the controls!

Bonin yields the controls, and Robert finally puts the nose down. ... At any rate, without warning his colleagues, Bonin once again takes back the controls and pulls his side stick all the way back.


According to the BEA report, Robert (left seat) puts the noes down for ~15 seconds, then brings it back to nose up. When Bonin takes control, Robert never attempts to go nose up. Bringing up the differing inputs from the two sidesticks in this case is a moot point, because neither of them were the correct input. To quote the report, "neither of the pilots formally identified the stall situation."

The stupid comments that the author makes at the end are completely unsupported by any of the evidence. Really, it's just such a "these dawgone new-fangled planes these days!" sort of paragraph that I can't even put together a response. The pilots of AF447 never recognized that they were in a high altitude stall, or at least never made any serious effort to avert a stall. To therefore suggest something vaguely scary and menacing about aircraft with modern flight control systems is downright irresponsible.
posted by kiltedtaco at 5:09 PM on December 7, 2011 [31 favorites]


"This was the Way Things Worked, until Tenerife^ -- when the co-pilot saw the captain make a mistake, but assumed the captain knew what he was doing (he wasn't just the pilot, he was the airline's top trainer). The captain then flew the plane directly into another plane already on the runway."

Yeah. The importance of CRM was to create training and procedures to avoid the situation where a senior pilot makes a mistake that's obvious to a junior pilot, but the junior pilot is afraid to correct him.

A better example of this, although I can't remember the exact incident and it's possible I'm misremembering it, was also on an island and the senior pilot taxiied into another plane, or attempted a take-off on an occupied runway (like Tenerife) while the other pilot saw the problem and said something about it at least twice, but didn't intervene because of seniority.

"...at almost any point in the prior 3 minutes that the crash could well have been avoided?"

Yes, not just because they didn't know he was pulling back on the stick (and could then do otherwise), but also because almost everything that was happening becomes instantly comprehensible when they learn this. Robert also pulled back on the stick, but he didn't know that Bonin had put the plane in a stall and in the attitude it had.


"... and some big 'ol flashing light and siren screaming 'ALTERNATE LAW! ALTERNATE LAW!'"

Okay, I'm no pilot. And I had the same sort of reaction to the competing flight stick problem and just the (apparent) inability for the other pilots to be able to tell what Bonin was doing with his at a glance. But that reaction is very much a function of the fact that we're not big airliner pilots.

Because, take a look at those cockpits. In fact, one of the problems that pilots have, and one of the contributors to crashes, is that pilots are overloaded with information, anyway. And you can't say, well, but this is critical information and most of that other stuff isn't. Because the vast majority of it could be critical information, too, under the right circumstances.

Basically, designing a good functional user interface for pilots of these horrendously complex machines is itself a horrendously complex problem. This is one reason why there's a move to integrated flight displays that can display the supposedly most important information without all the clutter. But that poses it's own problems, too, of course, not unlike the whole automation thing.

The thing is, and no one wants to hear this, but at this point with these planes, fewer crashes would happen if they were fully or almost fully automated. Some would still happen, of course, when the automation can't cope with something its designers didn't anticipate. But that would be outweighed by the crashes that would be avoided by human error. This is one of them. Even though the loss of the pitot tube data is a big problem for the automation, it could still have moved to a maximally safe flight profile given all the other available data. But we're still pretty fearful of letting transport machines run themselves.

"The pilots of AF447 never recognized that they were in a high altitude stall, or at least never made any serious effort to avert a stall. To therefore suggest something vaguely scary and menacing about aircraft with modern flight control systems is downright irresponsible."

Right. I do disagree a bit with you, though, about the import of Bonin's information. Yeah, Robert had just told Bonin to pull up. But Bonin said something to the effect that he'd been pulling up the whole time...that's the significant fact that Robert didn't know. Like you said, they didn't realize they were in a stall and they would have if they had communicated.
posted by Ivan Fyodorovich at 5:13 PM on December 7, 2011 [2 favorites]


Yeah, on re-reading it doesn't look like anyone recognized they were in a stall until it was far too late. The pilots in the pprune.org threads differ on whether the crash was avoidable even if they had recognized it earlier; Some of the commenteriat believe it may have been recoverable down to almost 10k while others think you'd have to nose down closer to 30k.

So, yeah, that one of the pilots apparently recognized a stall down around 4000 feet is almost immaterial.
posted by Justinian at 5:13 PM on December 7, 2011


Oh, I see what you're saying Ivan F. That one of the major reasons Robert didn't recognize a stall was that he didn't know Bonin had been pulling up the whole time.
posted by Justinian at 5:14 PM on December 7, 2011 [2 favorites]


I need to clarify my first statement: When I say that only one pilot controls the plane at one time, I'm saying that not because I think the pilots of AF447 violated that principle, but because I don't think that the (brief periods of) conflicting inputs were a substantial contributing factor to this accident. The pilots knew how to switch who is piloting the aircraft. They didn't always do it perfectly, but it's completely impossible to say "If only the pilots knew what the other was doing, the crash would have been avoided".
posted by kiltedtaco at 5:15 PM on December 7, 2011 [1 favorite]


kiltedtaco: "The pilots of AF447 never recognized that they were in a high altitude stall, or at least never made any serious effort to avert a stall."

(That's useful info in your comment btw, thanks)

So is there any kind of consensus on why they failed to realize they were in a stall, despite the stall warning? Just blind panic?
posted by vanar sena at 5:15 PM on December 7, 2011


Er, I wrote before:

"Yeah. The importance of CRM was to create training and procedures to avoid the situation where a senior pilot makes a mistake that's obvious to a junior pilot, but the junior pilot is afraid to correct him."

What I meant was that the lesson of that particular crash in its influence on CRM was...what I wrote. CRM solves more problems than just that one.
posted by Ivan Fyodorovich at 5:15 PM on December 7, 2011


yoink: "Except that the problem here is precisely that the pilots had full controll. If only the plane had not given them complete control, everyone would have lived."

A explained to me, the problem was precisely the on-off nature of the fly-by-wire controls, that they were modal without much warning of switch between modes, and that (at least on the early systems) you would be used to having full control, then find you were flying by wire, then when you had become accustomed to that, the system would occasionally flicker off and the system suddenly expected full manual control whereas psych wise you were still on oversee mode. Also, the digital systems (at that time) lacked the fine, analog feedback and visceral physicality that the pilots were used to and appreciated on older planes.
posted by meehawl at 5:17 PM on December 7, 2011


meehawl: "A explained to me, the problem was precisely the on-off nature of the fly-by-wire controls, that they were modal without much warning of switch between modes"

What sort of indication do the pilots have that the mode has switched?
posted by vanar sena at 5:27 PM on December 7, 2011


My retroactive, and futile, wish for the passengers on that flight would be this.

"Wow, this is a bumpy flight!"
"This is a really bumpy flight!"
"Shit!. . .this is my worst flight ever!"

Instantaneous and irreversable loss of consciousness upon impact with the ocean surface.
posted by Danf at 5:30 PM on December 7, 2011


They have dozens of square feet of instrumentation in there, surely they could spare a few inches to show the pilots what the control surfaces are doing.

Since they were ignoring the constant, audible stall warning, I'm not sure that they would have noticed a few inches on the (assuming) already crowded display.

on preview, it seems the only mode warning was a 2.2 second audible.
posted by MikeKD at 5:30 PM on December 7, 2011


A harrowing read. My heart was pounding by the end, knowing their fate.
posted by hockeyfan at 5:40 PM on December 7, 2011


kiltedtaco: "When both sidesticks are displaced from the neutral position, an audible alarm "Dual Input" is triggered. "

I didn't see this in the transcript in the article. Does anyone have a link to the full raw transcript? Some rudimentary googling has not been fruitful.
posted by joshwa at 6:11 PM on December 7, 2011


I just learned that not only is some dual input allowed without sounding the alarm, it's also involved in some really pilot petty turf wars.
posted by vanar sena at 6:20 PM on December 7, 2011 [1 favorite]


That should read "petty pilot turf wars"
posted by vanar sena at 6:21 PM on December 7, 2011


As someone mentioned upthread, many accidents like this read as a series of negative events that might have been interupted at any point, avoiding the incident entirely. In the fire service, I read Line of Duty Death reports that are similarly heartbreaking; you want to reach back through time and tweak any one of the small failures that led to someone losing their life. Crew Resource Management is now taught in the fire department as well, as a means of interrupting the chain of events. It crowdsources the safety of team members, so that even though authority for actions comes from rank, responsibility for safety comes from all sides.

Unfortunately, like much of what we do, it's easily forgotten when human beings are stressed and it extremely harrowing situations. In those cases, we default to simple and well-practiced strategies. I'm interested in knowing how flight simulator training in the airlines builds in CRM concepts to maximize the possibility that crews will use it under stress.
posted by itstheclamsname at 6:28 PM on December 7, 2011 [2 favorites]


Is my impression that if Bonin had just said something like, "I'm pulling up!" at almost any point in the prior 3 minutes that the crash could well have been avoided?

Justinian nails it. This is an inevitable result of an extremely poor training program. Not just technical flight physics and procedures-poor, though it certainly is that. Pulling back on the stick with a stall warning blaring should seem so wrong that it's almost impossible to do even when you want to. But that's the obvious thing.

The main failure is in the leadership and cooperation, or the complete lack of both, in the flight station. Nobody at any point said, out loud, time out. Here is what I'm seeing, this is what I think is why, this is my intended action. I think Robert knows the whole time what to do. He wants to descend. He gets Bonin to agree, but the plane doesn't do it. But then Robert does not ask out loud why will the plane not go nose down? Are you going down elevator?

When they think things are improving, they get airspeed indication back, again nobody says ok, time out. We had a loss of airspeed indication and an altitude excursion. Why are the stall warnings going off? How is that even possible in an Airbus?

Bonin's brain froze and he did the wrong thing reflexively. It happens. But the two men could have figured it out together if they had had the... I don't even know what you call it. Team problem solving skills? to assess the situation and think about a solution. They had plenty of time.

Then the captain comes in and apparently does not make them describe what's going on? "We've lost control of the plane" is non-helpful. How about, "I can't get any airspeed and the plane keeps descending even though I'm giving it all the up elevator I can?" Also the captain seems to recognize what to do but does not make an authoritative command like "PUT THE NOSE DOWN AND INCREASE AIRSPEED NOW." Maybe if he had, Bonin would have snapped out of it, or at least followed instructions. Too late anyway, though, probably.

You don't need to feel what the other person is doing if he will just fucking tell you.
posted by ctmf at 6:42 PM on December 7, 2011 [3 favorites]


joshwa: "kiltedtaco: "When both sidesticks are displaced from the neutral position, an audible alarm "Dual Input" is triggered. "

I didn't see this in the transcript in the article. Does anyone have a link to the full raw transcript? Some rudimentary googling has not been fruitful
"

Found the full CVR transcript (french, PDF). At 2 h 13 min 23 the first of several "dual input" alarm sounds (page 102 of report).

So the pilots, in theory, had some indication that they were both attempting to fly the airplane, though not what those inputs were.
posted by joshwa at 6:52 PM on December 7, 2011 [1 favorite]


All aircraft warnings should be voiced by Samuel L Jackson.

"STALL, motherfuckers! Do you speak it?"

"I'm sorry, did I break your concentration? I didn't mean to do that. Please, continue, you were pulling up on the stick while that motherfucker was nosing down."

"Normally, both your asses would be dead as fucking fried chicken, but you happen to pull this shit while I'm in a transitional period so I don't wanna kill you, I wanna help you."
posted by obiwanwasabi at 7:03 PM on December 7, 2011 [12 favorites]


Interesting bits from vanar sena's link:

* There's a "sidestick priority" button that a pilot can use to completely override the other pilot's sidestick in the "pilot passed out" situation

* Pressing this button also causes an indicator light to go green (not sure if it's on the PFD or something phyiscally near the sidestick).

From the Airbus manuals:
When the Pilot Flying (PF) makes an input on the sidestick, an order (an electrical signal) is sent to the fly-by-wire computer. If the Pilot Not Flying (PNF) also acts on the stick, then both signals/orders are added.

Therefore, as on any other aircraft type, PF and PNF must not act on their sidesticks at the same time.

If the PNF (or Instructor) needs to take over, the PNF must press the sidestick takeover pushbutton, and announce: "I have control".

Note that despite the "dual input" warning on AF447, neither pilot exercised control by pressing the sidestick priority button. Which further underscores the "nobody really took command" CRM argument.
posted by joshwa at 7:03 PM on December 7, 2011


Here's the English translation of the report. The CVR transcript starts on page 87.
posted by cowtown at 7:04 PM on December 7, 2011 [2 favorites]



Regarding the switch to alternate law, there's an indication on the EIACS, the center screen with the engine indications and the caution/warning messages, and the pilot flying also called out the change when it occurred, so they were certainly aware of it.

On preview, cowtown beat me to the link
posted by kiltedtaco at 7:07 PM on December 7, 2011


This blog has some decent comments which hypothesize better than I can. One such comment:

... Unfortunately, airline pilots rarely practice hand-flying at high altitude, and almost never do so without autothrust engaged. As a result, we forget that the aircraft is very sensitive to control inputs at high altitude, and overcontrol is the usual result. Because the Airbus sidestick provides no feedback "feel" to the pilot, this problem is dramatically compounded in this aircraft.

I believe the Air France pilot grabbed the sidestick, made an immediate input (because as pilots, that's what we tend to do), and quickly became quite confused as to what the aircraft was truly doing. This confusion likely was exacerbated by fixating on airspeed indications that made no sense while trying to find a power setting with no airspeed guidance.

So why did the Air France pilot find himself at the limits of sidestick travel, and then just stay there, maintaining a control input that simply could not logically be correct? When things go really bad and we are under intense pressure, it is human nature to revert to what we know from previous experience. Remember, the Airbus flies like no other aircraft in that the sidestick provides no feedback to the pilot. It is a video game, not an airplane.

I believe the Air France pilot unintentionally fell back on all of his previous flying experience, in which aircraft controls "talked" to him when he moved them. Distracted by many confusing inputs, he instinctively expected to be able to control the aircraft by "feel" while dividing his attention to address other matters. I've seen it happen in the simulator, and in an Airbus this is a sure way to lose control of the aircraft and is possibly the most dangerous aspect of Airbus design philosophy.

One last note: Airbus pilots often claim that the aircraft "can not be stalled." When the flight controls are in "normal law" this is a reasonably true statement. However, in "alternate law," as was the case here, stall protection can be lost. If we ever practiced this in the simulator, I don't remember it.

posted by RobotVoodooPower at 7:28 PM on December 7, 2011 [4 favorites]


Terrifying but comforting to know it pretty much came down to pilot error. The plane and its instruments and controls gave them a lot of chances to save themselves.
posted by bardic at 7:58 PM on December 7, 2011 [1 favorite]


From the article:
"But he is not at sea level; he is in the far thinner air of 37,500 feet. . . .

. . . Though the pitot tubes are now fully functional, the forward airspeed is so low—below 60 knots—that the angle-of-attack inputs are no longer accepted as valid, and the stall-warning horn temporarily stops."

I know nothing about flying. But am I reading that right? That the plane essentially descended from 37K+ feet at around 70MPH nearly the whole way down? Terrifying.
posted by quadog at 8:30 PM on December 7, 2011


What a read. I'm another one who's been fascinated by this story from the first time I heard them on the morning news announce that an Air France flight had "gone missing". I keep thinking about what it must have been like.

I've always been taught that "there's no such thing as user error" and that in the medical design field this is a priority. If you give the user a chance to fuck something up, then you're the one who fucked up, and the interface just isn't good enough. If there's lives on the line, nothing is too much.

But in this case.. like everyone had said, there were so many mechanisms in place. Even special psychological training to break up groupthink. But one failure caused a cascade effect, causing every single one of those mechanisms to fail. By the time you account for everything, then yeah you have a totally self-flying plane. I don't know what the answer is. Maybe make disaster training the priority? I don't know.

But this is one of those things that doesn't really have a good answer so we want to spend a lot of time on it because we want there to be a good answer.
posted by bleep at 8:36 PM on December 7, 2011


Hit post too soon. Anyway the point is that flying an airplane is a complex intersection between human psychology, human technology, and good old physics. We can do our best to nail every single permutation of where those three things come together but we're never going to get them all and things like this have to happen sometimes. If we were able to make a perfect machine as complex as this that never failed then we wouldn't be human anymore, we'd be something else.
posted by bleep at 8:40 PM on December 7, 2011


(Instructor Pilot in multi-piloted aircraft and aviation safety guy here.)

I'll preface this with saying that we here on the blue don't have all the information, so we're speculating, not really analyzing, and speculation can take you wildly off course, especially as stuff like confirmation bias takes hold of your cognitive processes. That being said...

In aviation safety there's something called the swiss cheese model of mishaps. Think of multiple slices of swiss cheese stacked on one another. You need the holes in each slice to line up in order for the mishap to occur. You've got organizational influences, supervisory failures, preconditions, and unsafe acts, with the unsafe act being the causal factor for the mishap.

Here we've got a multitude of things gone wrong before the 2:10 mark that represents the real start of the emergency:

- Preflight planning: other flights saw that the intertropical convergence would occur between SALPU and TASIL and they requested flight paths around it. (organizational influence: what was Air France policy on translatlantic flights and weather?)
- Non-existent CRM between the co- pilots. (Preconditions: What was the CRM like for two co-pilots? Had they flown a lot with each other? Did they have personality conflicts? Who was actually in charge? Was anyone left in charge?)
- Alternate law never called out and acknowledged by the pilots. (Preconditions: crappy CRM?)
- Both pilots fixate on the loss of the airspeed indicators, ignoring the 7000 feet per minute, 18 degree nose-up climb. (Preconditions: Bonin pulls back on the stick. Was it unwittingly? Did he ever know he was in alternate law?)
- Neither pilot ever called for stall procedures to be executed. (Organizational influence: when you hear the stall warning, you push the stick forward and command TOGO power, without fail.)
- When transferring controls, Bonin never physically lets go of the sidestick. (Unsafe act: the copilot did not adhere to established procedures. )

So on and so forth. Safety reports devote pages and pages to analysis of this sort, but in the end it looks like a combination of overreliance on the normal law to keep the plane in a safe regime of flight, fixation on the loss of airspeed to the detriment of any sort of awareness of what else might be going on.

Human factors continues to be the number one cause of mishaps, and for good reason: when you design a better system, the users adapt to it, finding new and interesting ways to goon things up. One hopes that pilots follow procedures to the letter, but as we see here, that's not always the case.
posted by squorch at 9:08 PM on December 7, 2011 [17 favorites]


Also, flying with a failed pitot-static system - the system that tells you altitude and airspeed - is not hard. You maintain a combination of power and pitch angle that will keep you level for your gross weight, which the computers know and display, or you can manually calculate in a few seconds. What kills me is the indications of stall (stall warning, slow groundspeed & airspeed, extremely rapid descent) were not recognized by anyone, even the captain!
posted by squorch at 9:16 PM on December 7, 2011


How about an alarm that sounds when the two pilots' stick input is divergent to a significant degree for, say, three sustained seconds. The computer would then say, "Divergent stick inputs." This would tell the two pilots that they are dealing with the situation differently (and so might have two different ideas as to what is happening).
posted by jabah at 9:16 PM on December 7, 2011


Only one pilot should be on the stick at a time. This is actually something that is drilled into you from the get go in flight school.
posted by squorch at 9:26 PM on December 7, 2011


I know this is a fairly stupid posit, but given the plane "knows" it's in stall, seemingly it wouldn't be hard to look at the two different stick inputs and see which one is going to fix the problem?

Also, given the real-time telemetry dumped by the plane, what about the idea of a "voice of god" call from a control center whose job it is to monitor all flights and be automatically alerted to a huge string of warnings/failures like this? Another level of CRM, essentially...even if only to inject a rational voice into the conversation. "Hi, we see you're in a stall and off auto-pilot, with conflicting control inputs, please report." Or whatever. Impossible?
posted by maxwelton at 9:45 PM on December 7, 2011


Unfortunately I don't have time to read the article or this discussion right now but I wonder: was this a normal accident?
posted by neuron at 10:17 PM on December 7, 2011


02:14:25 (Bonin) Mais qu'est-ce que se passe?

Seconds from crashing, the person who has had the most physical control over the airplane asks this? Terrifying.

All of the fancy engineering doesn't matter when an emergency happens and people have to act. Training matters, personality matters and, what seems to me to be the major problem here: communication matters. If they had talked to each other and talked through the situation it very well might have been very different. I was shocked at how little dialogue there was there. We talk more through a minor software release than that.
posted by marylynn at 11:03 PM on December 7, 2011 [1 favorite]


The real critical period for recovering from the stall until crash seems to have been 3-4 minutes, with increasing pressure and likely panic, that's not a lot of time.
posted by thefool at 11:08 PM on December 7, 2011 [3 favorites]


"... and some big 'ol flashing light and siren screaming 'ALTERNATE LAW! ALTERNATE LAW!'"


How about an alarm that sounds when the two pilots' stick input is divergent to a significant degree for, say, three sustained seconds. The computer would then say, "Divergent stick inputs." This would tell the two pilots that they are dealing with the situation differently (and so might have two different ideas as to what is happening).

I'm sure these things would have helped when the pilots on the flight deck were ignoring a blaring alarm and a voice repeating they were in a stall.

Also, given the continual stream of OMG AIRBUS PLANES ARE WEIRD AND UNSAFE AND FLY-BY-WIRE IS AN INHERINTLY WEIRD AND DODGY TECHNOLOGY ONLY AIRBUS USE that seems to be the mainstay of discussion on US forms and no-where else in the world, I don't know what Boeing pay their PR people, but it's worth every penny. Hell, there were even people insisting that there was no way the safe landing on the New York river a year or so back could never have happened with an Airbus plane because they're so impossible to fly.
posted by rodgerd at 1:32 AM on December 8, 2011 [5 favorites]


From the linked article:

The outside temperature is much warmer than forecast, preventing the still fuel-heavy aircraft from flying higher to avoid the effects of the weather


Could someone explain why the external temperature would be an issue, cause I'm not getting why it would be, particularly given that flying higher would mean a reduction in temperature anyway. Something to do with the intake of air into the engines?
posted by Hello, I'm David McGahan at 2:47 AM on December 8, 2011


Hello, I'm David McGahan:
Presumably it is due to the lower air density when it is warm.
posted by rycee at 3:01 AM on December 8, 2011


RE: Airbus side-stick. Am surprised it doesn't provide feedback. Boeing is FBW too for 777 and above, right? Even though it isn't a side stick but a more traditional layout, it is also a "video game" but with different interfaces, correct? Do boeing planes have feedback on their controls? How about those new Embraer E models?

What is scary, and surprising to me, is that it seems the entire flight crew were, well phoning it in, until it was too late. Other flights deviated around the worst of the weather, the radar "wasn't set correctly" until, I think Robert noticed, then they deviated, zero CRM, from the transcript, it seems like the captain took his time getting to the cockpit and then didn't necessarily take over.

Is piloting jets now so routine that a cavalier attitude is inevitable? I can see that happening. Where is Patrick Smith when you need him?
posted by xetere at 3:52 AM on December 8, 2011 [1 favorite]


quadog: the airspeed, or the speed of air going over the wings, was 70mph. Their land speed may have been much greater.

maxwelton: when you're that far out over the ocean, you are essentially out of radio contact with the land (scary as it sounds!)
posted by scolbath at 4:43 AM on December 8, 2011


I am so used to driving that sometimes when I sit in the passenger seat of a car that is approaching a corner too fast, I instinctively press my for on where the brake would be. I imagine that pulling back on the stick is similar, but that then implies a training problem to me.
posted by knapah at 5:58 AM on December 8, 2011 [1 favorite]


Got this far, and had to stop:
Despite the gap in seniority and experience, the captain leaves Bonin in charge of the controls.
It reminded me too much of this:
In spite of the fact that the weather was very poor (wet with low visibility, causing the cancellation of many flights that day at La Guardia), Captain Martin permitted First Officer Constantine Kleissas to make the takeoff on a short, wet runway, even though he would be conducting his first non-supervised line takeoff in a 737, and had not conducted any takeoffs during the last 39 days.
Only two people died that time. RIP aunt Ayles and cousin Betsy.
posted by MrMoonPie at 6:55 AM on December 8, 2011


02:14:25 (Bonin) Mais qu'est-ce qui se passe?

Seconds from crashing, the person who has had the most physical control over the airplane asks this? Terrifying.


It might help to understand, as this has been brought up a couple times now – this is an expression more equivalent to our English "what the heck???" than its more literal translation "but what's happening?/what's going on?" People here say it all the time when perplexed, it's a rhetorical question and practically automatic when you feel helpless (I say it often myself now, after more than a decade living in France). Just as we don't actually need to know, literally, "what the heck". "Mais", "but", is used as an interjection to add emphasis much in the same way as we use "what", even all by its lonesome. (You should hear all the single-syllable "Mais..." muttered by my French colleagues during the day. In polite situations, no one really says "quoi" by itself, because used alone it's considered rude, you're supposed to ask "Comment?" instead. "Mais" is unintrusive.)

this has been your French language tangent for the day
posted by fraula at 7:26 AM on December 8, 2011 [3 favorites]


- Alternate law never called out and acknowledged by the pilots. (Preconditions: crappy CRM?)

Actually it was called out by Robert (at 2:10:22) but not acknowledged by Bonin.
posted by hat_eater at 7:57 AM on December 8, 2011


But then Robert does not ask out loud why will the plane not go nose down? Are you going down elevator?


I really do think this is a design issue (as well as human error, obviously). The fact that Robert cannot feel that Bonin is pulling the stick back just seems remarkably unwise to me. Maybe he was just overloaded with the alarms going off, maybe he thought it was impossible that his colleague could be making such a fundamental mistake that it didn't even cross his mind to ask Bonin what he was doing with the stick, either way it just seems like fundamentally poor design to not have some non-visual, non-auditory indication of what's being done to the elevators that's evident on both sticks.

So fucking sad. Those poor people.
posted by longdaysjourney at 10:27 AM on December 8, 2011


I'm trying to think of any situation where averaging the inputs from the two sticks would have a positive result and am coming up blank. It seems completely insane. Why on earth would they design that system that way?

enn, it's called "overconstraint", and it's a fundamental engineering flaw. Why this sort of "engineering basics" isn't taught in engineering universities is beyond me - or maybe it is to Mech E's (my B.S. is E.E.). Stunningly stupid design, and it has nothing to do with what phliar suggested is "what you think the stick does."

The stick is an input device with a single output (U-V angle); to have multiple independent input sources, it must display feedback (which it doesn't). Anything else leads to an unstable system... which means gravestones, of course.

In my opinion, the engineers who designed this system should be held liable for criminal negligence. Seriously.
posted by IAmBroom at 12:21 PM on December 8, 2011


The stick is an input device with a single output (U-V angle); to have multiple independent input sources, it must display feedback (which it doesn't). Anything else leads to an unstable system... which means gravestones, of course.

This plane has flown millions of miles and has an excellent safety record. That seems to empirically disprove the claim that this design is inherently dangerous.

"Multiple independent input-sources" did not cause this crash. There is only a very brief period at the very end of this clusterfuck when one stick is forward and one back and it's unlikely that the disaster was avoidable at that time. When Robert did take the controls from Bonin he pulled back on the stick as well:
02:11:37 (Robert) Commandes à gauche!
Left seat taking control!

At last, the more senior of the pilots (and the one who seems to have a somewhat better grasp of the situation) now takes control of the airplane. Unfortunately, he, too, seems unaware of the fact that the plane is now stalled, and pulls back on the stick as well. Although the plane's nose is pitched up, it is descending at a 40-degree angle. The stall warning continues to sound. At any rate, Bonin soon after takes back the controls.
Neither pilot understood the situation or what to do to correct it.
posted by yoink at 1:52 PM on December 8, 2011 [2 favorites]


Also, given the continual stream of OMG AIRBUS PLANES ARE WEIRD

They have a unique design. The confusion scenario described here is a literal "can't happen" with the Boeing controls. The newer Boeing aircraft are fly-by-wire, but they still have the flight yokes that are physically linked together.

It would be obvious to the other pilot that the stick was pulled full back and when he tried to push it forward that the other guy was fighting him. There is a visual and tactile correspondence between the controls and the control surfaces that is lost in the Airbus.

Here are some comments from a guy at Boeing on the design difference.

Between the no force feedback controls and the control laws, it's easy for the pilot's mental model of what the aircraft is doing to become separated from reality in a stressful situation.
posted by bitmage at 1:59 PM on December 8, 2011 [2 favorites]


I talked with a head engineer at boeing about this flight before the flight recorders showed up, but it was already clear to him that this was a high altitude stall. He said he just didn't understand how a pilot could not push forward on the stick in a clear stall situation. Seeing the flight recorders, it makes a little more sense- bonin's brain shorted out and the other pilots didn't know it.

I guess the design lesson is that engineers should not allow one panicked pilot to secretly crash a flight. Having 2 pilots flying at the same time could work very well provided that conflicts are called out to everyone and a switch controlled who was operating the plane. The averaging solution is worse than mystifying though- e.g. imagine if they were flying straight at an obstacle and each pilot pulled in opposite directions to avoid it.
posted by efbrazil at 2:07 PM on December 8, 2011


Here are some comments from a guy at Boeing on the design difference.

From your link:
Pilots who fly sidesticks think they're the greatest thing since sliced bread, and pilots who fly yokes say there's no other way to go. In fact, both systems are excellent methods of connecting a pilot to an airplane.
The notion that the cause of this crash was that Robert did not realize that Bonin had the stick back is simply unsupported by the transcript.
posted by yoink at 2:16 PM on December 8, 2011


Having 2 pilots flying at the same time could work very well provided that conflicts are called out to everyone and a switch controlled who was operating the plane. The averaging solution is worse than mystifying though- e.g. imagine if they were flying straight at an obstacle and each pilot pulled in opposite directions to avoid it.

They weren't both flying the plane at the same time--not until the flight was already doomed in the final seconds. If you read the transcript you will see that they are quite clear about handing control back and forth, and you will also see that Robert screws up when he has control.
posted by yoink at 2:17 PM on December 8, 2011


The notion that the cause of this crash was that Robert did not realize that Bonin had the stick back is simply unsupported by the transcript.

I'd dispute that, or at least qualify it. When Robert takes control and pulls back, the plane has already begun to drop precipitously. Given that he doesn't seem to realize that the source of their problems is a stall (my personal guess is that they both assumed the “STALL” warning was further mechanical error, possibly caused by the computer not being able to determine airspeed), pulling up would be the natural thing to try there. If he'd known Bonin had been holding the stick back the whole time, then depending on when he knew that he might have (a) told him to level off or push forward before he stalled the plane, or, less likely given the pilots' apparent panic, (b) realized during the fall that they were in fact stalled and taken action to correct that.

Neither of those is certain. But Robert made different mistakes from Bonin, and I don't see anything in the transcript that suggests he would have thought pulling up was a reasonable thing to do until they were already falling.
posted by Holy Zarquon's Singing Fish at 2:36 PM on December 8, 2011


I know this is a fairly stupid posit, but given the plane "knows" it's in stall, seemingly it wouldn't be hard to look at the two different stick inputs and see which one is going to fix the problem?

This is more or less how Airbus aircraft behave in Normal Law. When the aircraft is approaching a stall, the control systems engage a variety of protection mechanisms that should prevent the aircraft from becoming uncontrollable. However, because of the initial instrument failures that preceded the crash, the aircraft switched to Alternate Law. Essentially the aircraft is acknowledging that it may not have all the information necessary to compute the proper flight envelope protections, and so it allows the pilot greater control to exceed those boundaries. The theory behind this is that you don't want the aircraft to cause a crash by thinking it is in a drastically different situation than it really is, and then erroneously limiting the pilots ability to maintain safe flight.

imagine if they were flying straight at an obstacle and each pilot pulled in opposite directions to avoid it.

They pilot flying that particular flight would maneuver the aircraft, and the other pilot would literally have his hands off the sidestick. No excuses. Same exact thing in a Boeing.
posted by kiltedtaco at 3:15 PM on December 8, 2011 [1 favorite]


Aviation Week has a discussion of factors relating to the crash.
posted by bitmage at 4:11 PM on December 8, 2011


In any case, I put the disaster squarely on the shoulder of the people who designed the flight control system. Letting the pilots operate the controls in a conflicting manner like this where they have no idea what the other is doing is a disaster waiting to happen. I'm only surprised it took this long.
The plane is designed with the expectation the pilots should know how to use it. It's not like a car where we don't really expect users to have much understanding or awareness. They should have both realized the correct thing to do was nose down rather then continuing to pull back in hope of gaining speed.

Also, technology moves really slowly in aircraft terms, The A330 was first flown in 1992, and I imagine work would have started in the 1980s. So we're talking about Nintendo/Apple II levels of tech in terms of what can be done with user interfaces. It's not like a cell phone where the UI gets revamped and refined every six months and you never use tech more then 20 years old. 20 years from now, Airplanes will probably have even more safety features.
How about an alarm that sounds when the two pilots' stick input is divergent to a significant degree for, say, three sustained seconds. The computer would then say, "Divergent stick inputs." This would tell the two pilots that they are dealing with the situation differently (and so might have two different ideas as to what is happening).
As it's been established, they had that. A "Dual Input" warning.
posted by delmoi at 4:54 PM on December 8, 2011


But Robert made different mistakes from Bonin, and I don't see anything in the transcript that suggests he would have thought pulling up was a reasonable thing to do until they were already falling.

Pulling the stick back wasn't a reasonable thing to do at any stage of the proceedings. It wasn't the reasonable thing to try when Robert did, in fact, get control of the plane and did, in fact, pull back on the stick. At that point all the instrumentation in the cockpit was sound and was all giving correct information. I see no indication whatsoever that Robert's problem was a lack of information--or, at least, that what would magically have made the difference would have been the information conveyed by dual yoke controls. Both pilots seem to have gotten their minds stuck in pretty solidly information-blind modes of reaction to their situation.

The notion that sidestick controls are wild experimental notions that invite death every time the plane leaves the ground is disproven by the simple fact of the Airbus's safety record. Unless, of course, the theory is that Airbus pilots are just extraordinarily better trained than Boeing pilots, and that somehow evens out the otherwise deleterious effects of their deathtrap design.
posted by yoink at 5:57 PM on December 8, 2011


The notion that sidestick controls are wild experimental notions that invite death every time the plane leaves the ground is disproven by the simple fact of the Airbus's safety record.

Stated like that, it certainly is. But won't you agree that if Robert had used proper stall recovery procedure, this design would become contributing factor to the accident?
posted by hat_eater at 3:11 AM on December 9, 2011


But won't you agree that if Robert had used proper stall recovery procedure, this design would become contributing factor to the accident?

Worthless hypothetical.

At some point, the three pilots in the cockpit have to decide what they want the aircraft to do. Nose-up, or nose-down. If the pilots in the left and right seats disagree to the extent that they would enter opposite commands to override each other, then there's absolutely no way to fly the aircraft, regardless of whether the commands are entered via sidesticks or yokes. The decision about what to do has to be made by the crew. The aircraft has no role in deciding which pilot to "believe".
posted by kiltedtaco at 6:30 AM on December 9, 2011


> Stated like that, it certainly is. But won't you agree that if Robert had used proper stall recovery procedure, this design would become contributing factor to the accident?

You mean apart from the fact that, if Robert had used the proper stall recovery procedure, there wouldn't have been an accident? :-)

Seriously, though, I don't think so. As stated above, the aircraft does provide feedback when there are multiple stick inputs - just not in a tactile form. If Robert had realised what the situation was and tried to put in the correct inputs, but had been hampered by Bonin's inputs, the aircraft would have (and did, on occasion) give the Dual Inputs warning. Either a verbal confirmation of who had the plane or, in the worst case, using the override button on his stick, Robert could have gained control of the aircraft regardless of Bonin's inputs.


On preview: What kiltedtaco said.
posted by Nice Guy Mike at 6:37 AM on December 9, 2011


At some point, the three pilots in the cockpit have to decide what they want the aircraft to do. Nose-up, or nose-down. If the pilots in the left and right seats disagree to the extent that they would enter opposite commands to override each other, then there's absolutely no way to fly the aircraft, regardless of whether the commands are entered via sidesticks or yokes.

This is true, and I realize I am totally out of my element here, but if a system, whether sidestick or yoke is built so that there is tactile feedback so that pilot A has a tactile reference to what pilot B is doing, it might make it easier to cut through the panic, confusion, and fog. I think that is the crux of what people are saying.

I mean if I am Robert, and I grab my sidestick and I see and feel it is all the way back, well I could still panic but I could also try to move stick forward, now granted I don't know what would happen in that scenario, I mean they could've been in the same situation with left and right overriding each other. That Egyptair plane when the co-pilot put it in a suicide dive, I think the captain and he were using opposite inputs, but again, that was deliberate.

There must be a reason why every manufacturer of commercial planes, not just boeing but Embraer, Canadair etc, use a yoke, and I have a feeling that all the newer ones are FBW as well.
Granted Embraer and Canadair build small regional jets but those E190s are sure sexy, and they "could've" been built with sidesticks. But they were not.
posted by xetere at 6:37 PM on December 9, 2011


That Egyptair plane when the co-pilot put it in a suicide dive, I think the captain and he were using opposite inputs, but again, that was deliberate.

As an aside, I have always found it extraordinarily odd that the NTSB refers to this as the EgyptAir Flight 990 "accident" when their own report determined the cause of the crash to be the relief officer deliberately flying the plane into the ocean. How is that an accident?
posted by Justinian at 9:00 PM on December 9, 2011


Apologies if this has already been asked, but it seems like most of the tactile feedback questions regarded the dual-control feature:

Does the sidestick employ a stick shaker, or some equivalent, when the plane enters a stall? My tiny tiny lay understanding is that, beyond serving as an alarm that's hard to ignore on a conscious level, yoke stick shakers can activate a more-or-less reflexive physical response on the part of the pilot (e.g., stick shakes violently, push forward). Is a similar system in place with sidesticks? Or not, because they lack the physicality of a yoke?
posted by evidenceofabsence at 12:11 AM on December 10, 2011


« Older We call hot wings 'sassy' here,” he explains. “'Ca...  |  This B.A.S.E. jump went very, ... Newer »


This thread has been archived and is closed to new comments