Flight Recorder for Air France 447 found intact [Update]

[Broomstick]

They do make recovery parachutes that are fitted in test aircraft and some production aircraft, which can be used to recover from a deep stall.

True. Unfortunately, there are definite size and weight limitations on such systems such that there's no way to build such a system for an airliner like the Airbus.

At stall, and especially in a deep stall, there is no useful air flowing over the control surfaces so as a result, any pilot control inputs are nearly useless. The usual recovery method is to increase power and pitch down, but if the tail was computer-trimmed all the way up, then it may have kept the nose up no matter how far the pilot jammed the yoke forwards.

There's a huge difference between stall and deep stall. You are in stall when partial or total flow separation at the airfoils occurs, which means that you suddenly lose elevator force and the capability to generate it - the airplane starts falling. You also lose partial or total roll control, as the airelons require attached airflow to function. However, the air flow around the elevator remains normal, so you can control the pitch of the airplane and so pitch down. Pitching down increases the pressure on the up side of the airfoils, where flow separation occurred, and thus the air flow tends to become attached again.

While non-swept-wing aircraft these days are designed so that the air flow separates gradually, both decreasing the abruptness of the stall and permitting some control surface usefulness, at least in the initial part of the stall, this is not the case with swept-wing aircraft. Stalls in an airliner will tend to be more abrupt than in a smaller airplane.

I am inclined to slightly dispute "air flow around the elevator remains normal". While the elevator may remain useful that doesn't mean the airflow is normal. In airplanes with mechanical links for the controls (such as I fly) you can actually feel the buffeting on the control surfaces as a vibration in the cockpit controls, it's the disordered air coming off the wings and hitting the tail surfaces as turbulence. How intense that is various with the airplane design and with how deeply/thoroughly you are stalled. It can diminish the effectiveness of the elevator for a stalled airplane, especially in t-tails. In a fly-by-wire airplane such as the Airbus you can't feel that buffeting (although it exists) because there is no longer a direct pilot-to-tail connection. So they put it back in, in a sense. The stall warning in a big Boeing or Airbus airliner shakes the controls in the pilot's hands, using the pilot's tactile sense as a warning instead of adding to the lights and buzzers going off, hence the term "stick shaker". Stick shaker = stall warning.

In a deep stall, you have total flow separation at the airfoils and at the elevator. The aircraft has no pitch, roll, or elevator control anymore. You're really fucked unless you're really, really lucky. But basically, on a figher aircraft you can eject, on a commercial aircraft you're going down. If this is what happened happened with the Air France flight, there's nothing they could have done.

^ This. An Airbus is not designed for falling leaf stalls. At that point the flight was doomed.

[Chris OFarrell]

I've only briefly read through whats been said right now, but I can't help but wonder if the pilots got into a situation of complete spatial disorientation here. The constant 'nose up! inputs that sound completely against what they should be doing to break the stall and gain control makes me think that we might have had that horrible 'complete darkness and/or complete cloud cover' that can frack with the best of pilots no matter HOW much you look at the artificial horizon...

[Broomstick]

Ah, thank you for correcting me. Your information brings another question though - how do they ensure that the elevator remains unstalled when the wings stall? I assume it's a combination of having an airfoil with better stall characteristics and the angle at which the horizontal stabilizer is attached to the fuselage.

It's mostly the placement of the elevator/horizontal stabilizer/stabilator in relation to the rest of the fuselage and wings. It is possible to stall the elevator - which usually kills all aboard because it's rare, hard to recognize, and requires a different recovery procedure than a wing stall. Usually, there's not enough time before the airplane hits the ground to recognize it's not a wing that's stalled, diagnose the problem, and take corrective action. As I mentioned in a prior post, the risk of this is greatest in T-tails but it's happened with other tail designs.

I found an interesting video of a power-on stall in an Extra 300. The instructor keeps backpressure on the stick and the student is trying to level out the plane with the rudder. http://www.youtube.com/watch?v=lGL_T9Mhtb4

edit: And here's the instructor doing the demonstration of the same thing: http://www.youtube.com/watch?v=ZOPsQn2Mksg

Yep - falling leaves. I said something a LOT more blistering during my first one. The instructor on the clips reminds me of the guy I did mine with. However, judging by the prop speed (or occasional lack of it) I don't think either of those was a power-on stall, I think those were power off. A power-on falling leaf would be much more violent. But I think you can see what I mean when I say the passengers had to notice something was wrong.

(In a training context, or in situation where you set out to do these, where you have minimized the variables, are in an airplane suited to the maneuver, and know exactly what's going on these can become fun because your chances of getting killed are minimal although it's the edge of the sort of flying where a parachute is a good idea. In an airliner, to have it happen unexpectedly? I think I just turned three shades paler merely typing that.)

Remember, the report mentioned that the guys on Flight 447 reduced power? This sort of state is one of the few times where more power is not an advantage. Gotta get the airplane under control, then you can pitch down, then you apply power. You have to change the angle of attack first. The instructor on the video said nothing about increasing power, he talked about angle of attack - in a stall AoA is more important than power for recovery. All applying power does is decrease the time of a full recovery to normal flight (usually straight and level) after you resume flying. You can completely recover from a stall without using any power at all, you'll just lose more altitude while doing it. Of course, losing altitude can be bad, hence the standard technique is to apply power, especially in something like an Airbus that comes out of the sky extremely rapidly. But more power before the AoA problem is corrected will actually cause more harm than good.

[Colonel Olrik]

I am inclined to slightly dispute "air flow around the elevator remains normal". While the elevator may remain useful that doesn't mean the airflow is normal. In airplanes with mechanical links for the controls (such as I fly) you can actually feel the buffeting on the control surfaces as a vibration in the cockpit controls, it's the disordered air coming off the wings and hitting the tail surfaces as turbulence. How intense that is various with the airplane design and with how deeply/thoroughly you are stalled. It can diminish the effectiveness of the elevator for a stalled airplane, especially in t-tails.

True, I agree with you. It's the fact that there's (usually) a gradient between normal flight near stall (no flow separation), beginning of stall and total stall. In the beginning of stall the flow separation at the airfoils is minimal, and naturally the impact of the turbulent flow is minimal at the tail. As the airfoil flow separation increases, the impact will be greater at the tail until at worst possible case the airplane is at total stall.

To note that as a logical consequence of how pitch control works, when you have total flow separation at the elevators, the elevators can produce no lift, which means that the aircraft will tend to pitch up (remember, the tail falls), not down as would be desired. Like Broomstick just said, losing elevator control has horrible consequences.

[Broomstick]

First of all, a correction to my post: by "The aircraft has no pitch, roll, or elevator control anymore" I meant to "write pitch, roll or yaw control". Obviously, pitch and elevator control are the same thing:
pitch - elevator controlled
roll: airelon controlled
yaw: rudder controlled

Actually, you lose the yaw control last in a stall. In a stall your rudder becomes pretty much your only working flight control, and you use it not only for yaw but for what roll control you have. Yeah, it's kinda weird. And you need a LOT of movement to get any effect at all. By that I mean instead of the normal, slight pressure with your toes you're shoving the rudder pedals to the stops. No flight control has anywhere near normal effectiveness in a stall, even if it is still working.

I'd also like to point out that an airliner has, in proportion to the rest of the airplane, a much smaller rudder than a smaller airplane does. If I understand the engineering correctly (and remember, I'm an empirical pilot more than an engineer) that's because normally enough air flows by it in normal flight that only a slight amount of deflection is required to gain the desired effect. In normal flight an airliner doesn't need a huge rudder. In a stall, though, that smaller rudder because a disadvantage. It makes it harder to recover an airliner than a smaller airplane.

Canard designed airplanes have the elevator before the wings. Then, the elevator can stall as a positive wing pitch is caused by a positive elevator pitch. Incidentally, this is an advantage of the Canard design, as the loss of lift at the elevator causes a pitch down so the airplane starts recovering from the stall instantly without pilot input.

However, there have been a few instances where people have managed to stall a canard - it's very rare, and one instance I recall involved wake turbulence which is air moving in quite unorthodox ways compared to normal flight. When that happens it's almost impossible, if not truly impossible, to get the canard unstalled. They don't spin, but they do head groundward at a very high rate of speed which is usually fatal. (There have been a couple survivals, usually in very small, very lightweight canards that weren't high enough to build up maximum speed, but the occupants were severely injured)

Even so - nothing is as stall resistant as a canard configuration. They do have other issues, however. All aircraft involve compromises.

[Colonel Olrik]

In addition to my previous post, of course in the design of any airplane the shape and location of the wings and of the elevator/tail take into account flow separation. That's in the end the reason why it's very hard to stall an elevator. In a Canard aircraft such as I mentioned before we have the inverse problem, in that the airflow around the elevator will induce disturbances in the flow around the wings, and this disturbances will be greater as the aircraft approaches stall at the elevators. This is a primary reason why Canard designs are relatively rare.

[Broomstick]

I've only briefly read through whats been said right now, but I can't help but wonder if the pilots got into a situation of complete spatial disorientation here. The constant 'nose up! inputs that sound completely against what they should be doing to break the stall and gain control makes me think that we might have had that horrible 'complete darkness and/or complete cloud cover' that can frack with the best of pilots no matter HOW much you look at the artificial horizon...

Entirely likely. Maybe not complete disorientation, but less than full awareness of their situation.

As I've mentioned, I've spoken with pilots who have flown commercial airliners in bad weather. Paraphrased, this is what they have told me:

When flying through a storm like that, you are in turbulence. The pilots are strapped in to their seats, the passengers have been told to buckle up, and quite likely the rest of the crew as well. Anything loose will be bouncing around. At times the airplane will shake so hard it can be difficult to read the instruments due to all the vibration. Most of the time the windows of the cockpit will be completely and utterly black, as if painted opaque, but from time to time lighting will flash across the entire sky, dazzling the eyes of the pilots and perhaps temporarily blinding them for a few seconds. There may be times where so much lighting flashes that it becomes near-continuous for brief periods. Meanwhile rain - if you're lucky it's just rain, but it also might be sleet, hail, snow, ice, or other stuff - is pounding on the aircraft at near deafening volume. A lightning strike on the airplane won't hurt the airplane, but it will be a BANG that makes your ears ring.

Add to that the fact that human senses are inadequate to fly in such conditions. You have to have flight instruments. You can't rely on your own senses, they never evolved for this, and they will give you wrong information. You have to ignore what your sense of balance is telling you, and your perceptions of acceleration and orientation. They are lying, lying, lying. You have to rely entirely on what your instruments are telling you, to the point you will ignore whatever your own senses are telling you. Your life (and the lives of everyone else aboard) completely depend on you putting the machine before your own perceptions.

That's the set up. Got it? Can you imagine it?

OK....

Now your instruments start to fail....

It's not a failure on the ability of the pilots that if their instruments fail they lose spatial and situational awareness in these circumstances. You can't maintain full awareness without that information. It's impossible.

Now, because those instruments are so crucial to everyone's survival they are made redundant. Three pitot tubes instead of one, for example. Pilots are also cross-trained, so if they lose their primary airspeed control they use other instruments, their training, and their own intellect to gather than information in another manner. For example, by memorizing flight control and power settings you can achieve a reasonable approximation of any particular airspeed you desire by setting controls and engine outputs. So in addition to instrument backups there are, shall we say, intellectual backups - but they're harder to utilize, and still dependent on reliable information from at least some of the instruments.

So, it's at night, in a storm, they're probably getting the crap shaken out of them, their instruments are failing, the damn airplane is dialing in nose up trim, they're having to extrapolate information from secondary instruments (and are they wondering if those instruments are reliable or not?), are they nose up because 1) they're nose up or 2) it's monetary buffeting or 3) it's a sensory illusion? Decide, guys, and oh yeah - you have to be right the FIRST time and everyone's life is depending on you getting it right. If you're wrong you die - and not instantly, no, you get to "enjoy" your impending doom for 2-3 minutes of falling and gyrations.

If all the instruments crash, at night, in a storm, you die. I'm sorry, but that's reality. If it's a calm, clear night with moonlight or enough starlight pilots can switch to the Eyeball Mark 1 as a primary instrument and that will get you home. Can't do that in a storm, however. A partial instrument failure, such as they had, isn't quite as certain a doom but it takes little to kill the airplane. The safety margin becomes hair thin. This, by the way, is why airliners usually try to avoid the worst storms, either by going around or over them (if possible - some storms extend to higher altitudes than airliners can reach. The storm Flight 447 went through extend to twice the maximum altitude the airplane was capable of reaching), or by simply delaying or cancelling the entire flight. They're dangerous.

That doesn't mean every rainstorm is a hazard to an airliner. Smaller storms aren't that bad, the airplanes can handle them, but this was a major thunderstorm, near the top of the storm rating scale. Our technology is good, but it's not perfect. Yes, the pilots made at least one major mistake here, but weather was an enormous factor in this crash. If the storm hadn't been bad enough to ice over the pitots and start crashing the instruments the crash never would have happened.

[Hawkwings]

To note that as a logical consequence of how pitch control works, when you have total flow separation at the elevators, the elevators can produce no lift, which means that the aircraft will tend to pitch up (remember, the tail falls), not down as would be desired. Like Broomstick just said, losing elevator control has horrible consequences.

Don't horizontal stabilizers apply a pitch-up moment to the aircraft during normal flight? The aerodynamic center of the wings is usually behind the center of gravity of the airplane, so to counteract that, they stick the horizontal stabilizers on a much longer moment arm, flip the airfoil upside down, and have the horiz stab producing lift in the "down" direction.

If the tail is stalled, then it's not producing as much negative lift anymore, and the plane would tend to pitch down. Of course, if the tail is stalled then the wings probably would be as well, so that's a moot point. Ahh, possibly what happens is that since the wing is stalled more than the tail, the pitch-up moment of the tail overpowers the pitch-down moment of the wings, resulting in a net pitch-up motion of the aircraft.

Broomstick: I think the power is on during those videos, but the camera is filming at a speed close to a fraction of the rotation speed of the prop, so you see it "moving" slowly, reversing direction, etc.

And your description of flying in a storm alone is terrifying, to say nothing of the emergency situation.

[Broomstick]

Broomstick: I think the power is on during those videos, but the camera is filming at a speed close to a fraction of the rotation speed of the prop, so you see it "moving" slowly, reversing direction, etc.

The other reason I think it's a "power off" stall is because I've never heard of anyone practicing falling leaves with power on - it's not impossible that might happen, but it's certainly not common. As those videos were clearly a demonstration and an initial practice of the maneuver it is most likely it's power off.

I will also add that "power off" doesn't mean the engine is stopped, or the prop. What it means is the engine is at idle. At idle, a prop in that sort of airplane is still going around at around 800-1000 times a minute. For that matter, even if the engine isn't working the prop will likely keep spinning several hundred times a minute due to airflow. I know this from experience, although I don't care to go into the details at the moment.

And your description of flying in a storm alone is terrifying, to say nothing of the emergency situation.

Yes. It's a situation that you should be afraid of. As I said, in most cases if you're on a commercial airplane flying through rain it's not nearly that bad, but a major storm - yeah, it's bad.

[AniThyng]

My understanding of the events around the crash was that there were quite a few other airliners in the same flight path at the time - why did they not encounter the same dangerous conditions, or did they also do so but were luckier then AF447?

[Starglider]

This, by the way, is why airliners usually try to avoid the worst storms, either by going around or over them (if possible - some storms extend to higher altitudes than airliners can reach. The storm Flight 447 went through extend to twice the maximum altitude the airplane was capable of reaching)

Another reason to mourn the failure of SST development. Concorde cruised at about FL600, the Boeing 2707 would have been around FL650, the B-70 cruised at FL750. That doesn't completely eliminate weather issues vs FL370 but it certainly greatly reduces them.

[Broomstick]

My understanding of the events around the crash was that there were quite a few other airliners in the same flight path at the time - why did they not encounter the same dangerous conditions, or did they also do so but were luckier then AF447?

First of all, the storm varies in intensity over time and by location. By jogging right instead of left, or choosing another altitude less prone to ice formation, those other airliners might well have avoided the worst of the storm and had somewhat less severe conditions to fly through.

Also, keep in mind the pitot tubes on Flight 447 were considered obsolete - there literally were replacements waiting in France, if the airliner had made it across the Atlantic they probably would have been replaced just after it landed. If those other airliners had better pitot tubes they would have experienced less ice build up and thus never lost their airspeed information.

[Chardok]

This, by the way, is why airliners usually try to avoid the worst storms, either by going around or over them (if possible - some storms extend to higher altitudes than airliners can reach. The storm Flight 447 went through extend to twice the maximum altitude the airplane was capable of reaching)

Another reason to mourn the failure of SST development. Concorde cruised at about FL600, the Boeing 2707 would have been around FL650, the B-70 cruised at FL750. That doesn't completely eliminate weather issues vs FL370 but it certainly greatly reduces them.

Also - I would think that the sheer speed that SST's fly at would prevent/eliminate icing in the first place (doesn't friction cause the skin of the aircraft to heat up significantly?)