Those two pilots in the commuter plane wreck awhile back did the same thing. They say it’s funny how often they will examine a wreck and find the stick pulled back deep into the lap of the pilot(s). Gotta unload the sucker.
There is a suggestion that the copilot’s instruments may have been giving different readings than the pilot’s instruments.
Ok, there was a pilot and two copilots. The pilot was on a break, and presumably one of the copilots was in the right seat, and the other one took over the pilot’s seat when he left.
At that point, which of those two is in charge?
Also, there’s no suggestion they ever got incorrect altitude data, so if they were watching, they would have seen, at first a climb of about 2,500 feet (which should tell them they’re not in a stall, right?), followed by a leveling off and then altitude loss until the crash.
This part of the report is interesting:
So when they did drop the nose, they got a stall warning. We know that was because they picked up speed and began to get some valid data. But is it possible they thought they were getting flow separation because they were going too fast? I’m way out of my league here - but this quote from the wiki article on coffin corners might be relevant:
So is it possible that, in reality the plane was stalling because it was going too slow, but the pilots thought it was stalling because it was going too fast? This could explain why they kept pulling the nose up.
That is confusing to someone as ignorant of aircraft as I.
Broomstick, where are you? 
I have heard of “high speed stalls” but I can’t say I understand them. I assumed it meant that your wings are so off-axis to the airflow, the airspeed was virtually irrelevant.
I’ll buy that, thanks!
It’s difficult to overcome the natural tendency to pull up if you are going down in an unplanned dive. But that’s what a professional is trained to do.
Yeah, well, it’s a guess, and I hope a real pilot comes along to better explain it.
Anything’s possible when talking about human perception, however it is unlikely because flow separation due to high speed doesn’t trigger the stall warning system because the stall warning monitors angle of attack and high speed separation is not associated with excessive angle of attack.
Unfortunately the FDR and CVR can only tell us what happened, it can’t tell us why it happened. Why did the pilot apply almost continuous nose up inputs?
Could be poor training. It is common for stall recovery training to emphasise minimising height loss, sometimes that emphasis can lead to the pilot failing to recover the stall in the first place.
Could be a poor understanding of the aircraft systems. In alternate law the aircraft no longer automatically protects against the stall, but in normal law the aircraft will prevent the pilot from doing anything to stall it. If he didn’t understand that, he may have thought he could apply full back stick and the aircraft would essentially fix the problem for him.
Could be that they simply didn’t understand what the problem was. If you don’t identify the problem, you won’t be able to fix it.
The Wiki passage does not say too high a speed may create a stall, it says “lose lift”. Lose too much lift and the plane stalls, but there is a middle ground.
In a way airspeed is always irrelevant to stalling. Stalling is caused by the angle between the wing and the airflow (angle of attack) being too great. It means that the air has too sharp a corner to turn as it flows over the top of the wing and so it can’t stick to the wing surface. If the air separates from the top of the wing, the wing doesn’t work any more.
When people talk about the “stalling speed” they are referring to the speed at which the wing’s angle of attack equals the stalling angle when in straight and level flight. You can stall a wing at any speed though, the only important thing is the angle of attack. As an example, when in a turn the you need to increase lift to maintain altitude and you do that by increasing the angle of attack. The aircraft will stall at a higher speed in a level turn than it will when flying straight.
Clarification: A high speed stall is not the same as airflow separation due to speed of sound issues. When approaching the speed of sound the airflow gets disrupted by the shockwave that is formed, it is not related to a high angle of attack. A high speed stall on the other hand is just the same as a low speed stall except it occurs when the wings are dynamically loaded.
Airspeed can’t be completely irrelevant to stalling. A plane in perfectly level flight will eventually stall is the engine stops and airflow around the wings isn’t sufficient to support the weight of the plane.
I have some more questions about the “fly by wire” nature of this plane.
Does the plane give input from one pilot priority over the other? If the two pilots were looking at different instrument readings, and responded differently than each other on the throttles or joysticks, what happens?
Actually, do they have separate throttle controls, or is there one set on a center panel? But joysticks are no longer directly physically linked to each other, correct? In the manual systems the pilots would have actually felt each other’s input to the stick and presumably either one would let go or they would both start pulling/turning in the same direction. With these new sticks that are pretty damn similar to game controllers, how do they know if they are fighting against each other?
Thanks so much for the added information on stalls. It’s very helpful, even if it doesn’t reveal as much as I’d hope about what might have happened in this case.
Out working on a construction site today - I’ll digest the information linked and get back to ya’ll. But, still, I’m pretty darn tired so it might not be until tomorrow morning as I want to make a coherent post.
Thanks! What are you constructing?
I’m building a deck in place of a collapsed concrete porch. Well, helping to build it, I’m part of a crew and not “flying solo”, or even the captain on that one.
Anyhow, the usual disclaimer: I am not a professional accident investigator, nor do I fly airliners. My aviation knowledge is limited and it is possilbe that someone who knows more than I do may come along an correct me. Speculation may change based upon new information. That said…
… this one strikes me as pretty horrible from the standpoint of those who went through it.
The relevant series of events seems to be as follows:
When the autopilot and auto-thrust disengaged this is exactly what is supposed to happen – the pilot flying the plane (PF=”pilot flying”) takes over and announces “I have the controls”. This alerts everyone else in the cockpit that something is up and someone actually is flying the airplane.
I’m assuming the left input was to control the roll to the right. The nose-up seems odd to me, but the fact this was at night, in turbulence, might mean the PF wasn’t fully cognizant of the total situation.
The rapid fall off in airspeed is not a good sign, and certainly not to that degree. There is the possibility that the pilots were getting conflicting airspeed reports from the instruments, which is horrible because then they don’t know which airspeed is correct and which ones aren’t.
Having the stall warning go off is bad enough… having it do so twice is getting really serious, especially combined with rapidly falling airspeed.
This time stamp is what makes investigators think the pilots had conflicting airspeed displays – they didn’t know which one was correct so essentially they didn’t know how fast the airplane was flying. The stall warning is connected to an angle-of-attack detector, so it was going off despite the lack of accurate airspeed.
OK, the airplane is nose up, so it’s trading airspeed for altitude, which only works for a very limited time. The PF correctly made nose-down inputs but I haven’t a clue why he was doing left and right roll inputs from what’s given here. I am guessing this was done to counteract the tendency of airplanes to “wobble” at extremely low airspeeds/near stall conditions, or perhaps in reaction to turbulence, or both. The airspeed went back up, which is good because that’s further from a stall, and the angle-of-attack diminished, which is good for the same reason.
This is where it gets puzzling to me. The stall warning goes off again, but nose-up inputs are maintained? That’s the opposite of what one normally does in a stall. Now, there are some rare exceptions to that rule, but they’re rare. Or perhaps the PF panicked. Pilots don’t like to admit to panic, but it does happen. Again, though, it’s rare – people prone to panic tend to be eliminated from flying early.
OK, the instruments themselves apparently recorded dissimilar airspeeds. That makes flying a lot harder. However, the stall warnings going off and the changes in altitude should have clued the pilots in to the idea you guys are getting very close to a stall.
OK, this sounds like the instrument readings were so off from normal the warning system shut down. On the other hand, there would have been other indications that there were Serious Problems, like what the altimeter was doing. Some airplanes have a particular “feel” prior to a stall, a sort of wallowing, not-responding-well-to-steering sort of feel. I don’t know how much that would apply to this model of Airbus, but if that applied that could be another warning that Something Is Seriously Wrong.
This is where they lost it (in my opinion). A forty degree angle-of-attack? That’s insane for the circumstances. They were descending at 10,000 feet per minute. That can only be described as falling, it’s not flying. Roll oscillations of 40 degrees? The passengers had to be screaming, there was no way in hell they didn’t notice that. Why such rolling motion? They were stalled. The ailerons work poorly, if at all, in a stall and the airplane tends to want to roll. The only thing to counteract that is the rudder, the control surfaces on the wings will do little if anything.
What is can not understand is the sidestick left to the stops – that’s using the aileron, not the rudder – and the nose-up to the stops. WTF? You are in a stall. Nose up can fucking kill you, and everyone aboard. What the hell?
OK, it’s bad when the instrumentation stops giving you information, I believe that’s what the two pilots were saying. They had reduced power. Normally when near or in a stall you increase power, but there are valid reasons to decrease power, specifically when you think increased power is aggravating the stalled condition. In my admittedly limited experience that’s usually when you think you’re either in a spin, or about to enter one, because engine power in a spin tends to make the spinning faster/more violent. I am guessing that’s what these guys where afraid of.
“Level one hundred” is 10,000 feet. They had fallen over 3/4 of the way to the ground in just a couple minutes. Both pilots had to be afraid by this point, which is why I think both started using the controls. The PF yielded control to the other, but at this point I don’t think it matters, or made a damn bit of difference. They don’t have time or altitude to get out of this situation.
The qualification of “when it was valid” on this is scary.
What this means is that they were still in the stall, still nose up although the wings were close to level. They weren’t in a spin although at one point they might have been damn close to one, and they were falling at 124 mph (200 kph) towards the ground.
What this means is that they were alive and concious all the way down, crew and passengers alike. Given some of the gyrations, it was horrible, frightening, and at times violent on the way down. Frankly, reading this gave me a real feeling of horror and fright. I am so glad I was not on that flight, it must have been horrific.
Thanks for the summary. Fascinating, but tragic.
No, it’s damn frightening. Like when I heard some EMT guys from Tennessee on CNN who drove to NY after September 11. Hearing my native tongue made it suddenly, horribly real.
The stalling angle of attack is the same regardless of what the aeroplane is doing, but the stalling airspeed changes depending on the g loading on the aeroplane. You can be at five knots and not be stalled and you can be at 300 knots and be stalled. There is a corresponding airspeed that equates to the stalling angle of attack at a particular g loading and the one published in the flight manual is for 1 g and is the speed the aircraft would stall at in your example. Airspeed isn’t irrelevant, but it is secondary, it does not, by itself, give any indication of how close to the stall you are.
The inputs from the sidesticks are averaged and a light on the glareshield warns that both sticks are being used. The pilot can regain full authority by pressing and holding a priority button on the sidestick.
There is one set of throttles on the centre pedestal.
For further reading on the technical aspects, check out this site: http://www.smartcockpit.com/plane/airbus/A330/
I think they were going a bit faster than that. It’s 107 knots ground speed which should mean forward speed across the ground. They were also going down at close to 11000 fpm which equates to about 107 knots as well. That would put their actual flight path at about 45º down and speed of about 150 knots. It’s a translation though, so “ground speed” might not mean what I’m thinking it does.
I noticed that the data recorder only record the instrument readings from the left side of the cockpit, so I guess they were inferring a possible discrepancy between the pilot’s and copilot’s sides from what was said on the voice recorder.
Does the data recorder record input from both flight sticks? Do they (the crash investigation team) know if the two pilots were entering conflicting input? And if they don’t record both, are they recording the averaged input, or only one side?
I gotta wonder, what if one of the pilots was panicking and hauling back on the stick? It seems possible that the warning light for the control conflict might not even be noticed among all the other flashing and wailing alarms going on.