Air France A330 - Missing over Atantic

To add to the confusion, the stall warning stopped when the indicated airspeed got so low, and AoA so high, that it was classified as invalid by the computer. When the crew did try lowering the pitch attitude the result was the stall warning coming back on which would normally lead a pilot to believe that it wasn't helping.

This.

I'm trying to think of in that cockpit, and this happening under all that stress. The stall horn goes off. You make input changes. it starts again. You have 300 people on the plane, no faith in the instruments, no way to see anything outside, and are starting to panic. So you roll back the controls, and the horn goes off. But the controls don't make sense, so you try going to something more normal. The stall horn starts again.

Oh... my... god.

Turning off the stall horn when it's potentially unreliable.... that's horrible design. Instead have it make a different kind of sound, to alert the pilot to the fact that it has been turned off. I am surprised this has not been highlighted more significantly - this issue is all on Airbus, who designed and built the plane with this "killer feature" (literally).
 
This.

I'm trying to think of in that cockpit, and this happening under all that stress. The stall horn goes off. You make input changes. it starts again. You have 300 people on the plane, no faith in the instruments, no way to see anything outside, and are starting to panic. So you roll back the controls, and the horn goes off. But the controls don't make sense, so you try going to something more normal. The stall horn starts again.

Oh... my... god.

Turning off the stall horn when it's potentially unreliable.... that's horrible design. Instead have it make a different kind of sound, to alert the pilot to the fact that it has been turned off. I am surprised this has not been highlighted more significantly - this issue is all on Airbus, who designed and built the plane with this "killer feature" (literally).

Agreed. It's easy to blame the dead pilots. But, seriously, put yourself in their position. All the confusing and even contradictory indications they were receiving from their systems would likely confuse the heck out of the best of us. I also haven't heard anything to the effect that the pilots had a history of incompetence, so I have no reason to believe that they were anything short of the professionals they should have been. Pilot error? Sure. They made the control inputs that brought the plane down. But they made those control inputs on the basis of a bunch of crap information that their "state-of-the-art" airbus was giving them. Sometimes it's the contributing factors that push things over the edge.
 
To play Devil's Advocate, how would you deal with the stall system when the plane is on the ground? The whole idea behind that particular design was to prevent nuisance warnings when the plane was on the ground.

As well as I can remember, the stall system is based on the Angle of Attack vane, and that can be all over the place on the ground.
 
To play Devil's Advocate, how would you deal with the stall system when the plane is on the ground? The whole idea behind that particular design was to prevent nuisance warnings when the plane was on the ground.

As well as I can remember, the stall system is based on the Angle of Attack vane, and that can be all over the place on the ground.

Weight on wheels?
 
To play Devil's Advocate, how would you deal with the stall system when the plane is on the ground? The whole idea behind that particular design was to prevent nuisance warnings when the plane was on the ground.

As well as I can remember, the stall system is based on the Angle of Attack vane, and that can be all over the place on the ground.

Nuisance warnings on the gound are much better than dead people in the sea.

Not being a smartass, just sayin', I hoe that the fear of nuisance isn't used to continue the use of a potentially scary failure model.
 
Weight on wheels?

Yeah, that. It is called an Air/Ground switch on some airplanes.

Nuisance warnings on the gound are much better than dead people in the sea.

True that, but nuisance warnings on takeoff roll are not much better than nothing at all.

Not being a smartass, just sayin', I hoe that the fear of nuisance isn't used to continue the use of a potentially scary failure model.

In Airbus's defense, no one ever thought that a crew would get that far into a stall. It hasn't ever happened before. You can bet there is probably a redesign in the works.
 
Does the Airbus have a shaker or is the stall warning something else?
 
In Airbus's defense, no one ever thought that a crew would get that far into a stall. It hasn't ever happened before. You can bet there is probably a redesign in the works.

Thats because Airbus expected more out of the pilots who would fly their aircraft. Shouldn't have to redesign an aircraft because pilots forget how to fly. Pilots need to remember the basics. Its tough to remember that stuff when flying an aircraft that should fly for you. But in the end all pilots should be able to recognize a stall and recover from it or prevent it, in any airplane.
 
Thats because Airbus expected more out of the pilots who would fly their aircraft. Shouldn't have to redesign an aircraft because pilots forget how to fly. Pilots need to remember the basics. Its tough to remember that stuff when flying an aircraft that should fly for you. But in the end all pilots should be able to recognize a stall and recover from it or prevent it, in any airplane.

I agree and disagree. This was an incredibly complex situation when you factor in the control laws designed by Airbus. That being said, there should have been and there will be training in the future to address this exact issue due to its complex nature. Airbus design is not synonymous with any other type of airliner. The old saying that it's just another airplane doesn't fit into this scenario. That is one thing that Boeing has done very well in my opinion, but this isn't a Boeing v. Airbus debate, and for the record I think they both build terrific airplanes. It just isn't as simple as say the Colgan crash which was very simply pilot error. Just my $.02.
 
I think my question remains, though. Would a regular stall warning ever happen when you push the nose down? If the stall warning did come on here when pitch was decreased, it's if anything an indication that things aren't working correctly or that they are not working in the way I expect or understand.

In either case, I would ignore the stall warning and get the nose down. I admit it would be incredibly confusing, but if I had somebody else to talk to about what was going on, I wonder what would get me to conclude that the stall warning was right and that the fundamental principles of physics were wrong....

I don't think I understand the system well enough to truly understand what was going on here.
 
I think my question remains, though. Would a regular stall warning ever happen when you push the nose down? If the stall warning did come on here when pitch was decreased, it's if anything an indication that things aren't working correctly or that they are not working in the way I expect or understand.

In either case, I would ignore the stall warning and get the nose down. I admit it would be incredibly confusing, but if I had somebody else to talk to about what was going on, I wonder what would get me to conclude that the stall warning was right and that the fundamental principles of physics were wrong....

I don't think I understand the system well enough to truly understand what was going on here.

They would have been reluctant to just push down because they could overspeed the aircraft. As high as they were,the difference between the stall speed and mach 1 is very small. Non-supersonic planes don't like going transonic. I'm sure the pilots were concerned about that.

Still, I think they should have set a known attitude and known power and let it ride. My crusty old primary flight instructor (who had 10's of thousand's of hours in big jets) drilled me with "Proper pitch and power produces predictable performance".

I wonder what is in the "Unreliable IAS" checklist.
 
Greg,

Here are some other crashes that occurred due to super stall/deep stall condition. IMO it is a situation that designers should be aware of.

http://en.wikipedia.org/wiki/Stall_(flight)#Deep_stall

The A330 doesn't have a t-tail, so I don't think that is at issue. The elevator and trim were commanded full up, which should drive the plane into a stall. I'll bet it would've recovered pretty quick once the AoA was lowered to something reasonable.

[EDIT] OK I read a little more. The wing would not have been eclipsing the stabilizer, but it could be that they drove the AoA so high the the stabilizer had stalled and could not get the nose back down. 40 degrees AoA is almost falling not flying.
 
Last edited:
I understand the overspeed issue, but wouldn't there be another alarm for that? Why wouldn't they have at least tried it?
 
To play Devil's Advocate, how would you deal with the stall system when the plane is on the ground? The whole idea behind that particular design was to prevent nuisance warnings when the plane was on the ground.

As well as I can remember, the stall system is based on the Angle of Attack vane, and that can be all over the place on the ground.
I'd inhibit the low speed disable unless the gear was down or maybe unless there was weight on the wheels.
 
When the air speed indicators fail, aren't the pilots supposed to use a specific throttle setting and a specific pitch attitude?
[/I]

Ok, but if I have my facts straight, the airplane was at 10 degrees nose up, (a reasonable pitch attitude in a climb) but the angle of attack was at 40 degrees, which isn't reasonable. But they have no way of easily knowing what their angle of attack was. So here they are, with what they thought was erroneous airspeed indications. They are at 10 degrees nose up at basically wide open throttle, which SHOULD have given them pretty good climb rate. But they weren't climbing because they were in fact in a stall.

All of that to say that a given pitch with a given power setting will not ALWAYS produce the result you would expect.
 
They would have been reluctant to just push down because they could overspeed the aircraft. As high as they were,the difference between the stall speed and mach 1 is very small. Non-supersonic planes don't like going transonic. I'm sure the pilots were concerned about that.

Still, I think they should have set a known attitude and known power and let it ride. My crusty old primary flight instructor (who had 10's of thousand's of hours in big jets) drilled me with "Proper pitch and power produces predictable performance".

I wonder what is in the "Unreliable IAS" checklist.
My neighbor who's been flying the A330 for quite a while said that as a result of the AF accident their recurrent sim training includes a high altitude reversion to direct law with loss of airspeed indication. IIRC it's something like TOGA power and +7° pitch up. He said that's supposed to put the plane's IAS safely between min and max speed.
 
So the facts are:

1. Three experienced pilots on board.
2. Airspeed indications become unreliable due to pitot icing.
3. As a result of the erroneous airspeed and confusing automated warning systems, the pilots induce a stall at high altitude.
4. Throughout the stall and descent the flight crew did not recognize the condition of the aircraft and did not take correct measures necessary to return the aircraft to level flight.
5. As a result all aboard the Northwest Airlines Flight 6231 Boeing 737 are killed.

Thirty-five years later only point 5 needs editing. (I copied point 4 verbatim from the NTSB report, linked below.)

http://www.airdisaster.com/reports/ntsb/AAR75-13.pdf
 
Jim: I'm really torn over this. I'm certainly not disagreeing in any manner: things like this have happened before. Three ATP rated pilots with good levels of experience, the attitude indicator worked from what I can tell, and they had back up analog instruments. Yes, there were a lot of things that distracted them, but it sure is confounding. Just shaking my head.

Best,

Dave
 
I have a small nit-pick... how are you determining that the attitude indicator was working? I've been suspicious of that the whole time...
 
Jim: I'm really torn over this. I'm certainly not disagreeing in any manner: things like this have happened before. Three ATP rated pilots with good levels of experience, the attitude indicator worked from what I can tell, and they had back up analog instruments. Yes, there were a lot of things that distracted them, but it sure is confounding. Just shaking my head.

Best,

Dave

There was one thing distracting them really, every other indication was situationaly correct for being in a fully developed stall. There was really only the stall horn causing confusion, and for that the designer who decided that the system should cancel the stall warning on a single point of data with no indication that in fact a stall does not exist, should be sent to the Guillotine, and I'm serious. That was a design fail of Epic proportions and all those responsible for that system should die.

None the less, the captain was a total failure as well as were the seat warmers under his command. The plane wanted to fly out, all they had to do was let the nose down and leave the power in. Once they had made it into the FL 2xx's that the Mmo/Vs spread opens up enough to be able to safely recover by hand on a straight up- Pitch/Power settings and that is what the captain should have ordered, but he didn't.
 
This recommendation is a sad comment on the state of affairs in airline training. Have we trained a generation of ab initio button pushers?

...make mandatory the creation of regular specific exercises aimed at manual airplane handling.
 
The plane wanted to fly out, all they had to do was let the nose down and leave the power in.
Has that been determined? If the airplane was that deep into the stall would this have worked as a recovery technique? I think that some deep stall situations in transport category jets can become unrecoverable. I don't know if this was the case here.
 
Has that been determined? If the airplane was that deep into the stall would this have worked as a recovery technique? I think that some deep stall situations in transport category jets can become unrecoverable. I don't know if this was the case here.


As soon as he let off the stick and dropped the nose, the stall horn came back on showing it was responding to pitch change, then they pulled the power levers back. Everything about that showed that the plane was still accurately responding to control inputs.
 
Has that been determined? If the airplane was that deep into the stall would this have worked as a recovery technique? I think that some deep stall situations in transport category jets can become unrecoverable. I don't know if this was the case here.

It's my understanding that the A330 transfers fuel to the rear in cruise to make the plane more efficient and that CG shift might indeed preclude a recovery from a deep stall using the elevator to pitch down. That said, the combination of nose down elevator and trim (assuming the "fixed" portion of the stabilizer moves to adjust trim) ought to be extremely powerful and would likely be sufficient to drop the nose.

There's one recovery method that's supposed to work on the most stubborn swept wing designs but it requires a lot of altitude to complete successfully. Basically you let the plane roll over to approximately 90° bank and rudder the nose down. At 10,000MSL there probably wasn't enough room left for that but at 20-25k it ought to work.
 
Has that been determined? If the airplane was that deep into the stall would this have worked as a recovery technique? I think that some deep stall situations in transport category jets can become unrecoverable. I don't know if this was the case here.

This airplane also has a feature that pumps fuel into the tail to move CG aft during cruise. Apparently that makes cruise flight more fuel efficient.

The downside, it would seem to me, is that being heavy in the back would make it more difficult to correct the excessive AOA situation. Using down elevator might reduce the AOA enough to reactivate the stall horn but its not clear to me that there was enough elevator influence to break the stall.

I dunno. I'm just a 180 hr. PPL trying to understand this tragedy.
 
Why do some airliners lack the ability to recover from a deep stall? Is it because of a lack of elevator authority?
 
Why do some airliners lack the ability to recover from a deep stall? Is it because of a lack of elevator authority?

http://goliath.ecnext.com/coms2/gi_0199-11989369/Deep-stall-aerodynamics-you-didn.html

When flying at a near the airplane's minimum drag speed ([V.sub.MD]), an increase in angle of attack causes drag to increase faster than lift. The airplane begins to sink. In turn and when maintaining a constant pitch attitude, this tendency results in a rapid increase in . angle of attack as the flightpath trends more downward and less forward.

Once the stall has developed and a large amount of lift has been lost, the airplane will begin to sink rapidly, accompanied by a corresponding rapid increase in angle of attack. This is the beginning of what is termed a deep stall.

As an airplane enters a deep stall, increasing drag reduces forward speed to well below normal stall speed. The sink rate may increase to many thousands of feet per minute. The airplane eventually stabilizes in a vertical descent. The angle of attack may approach 90 degrees while the indicated airspeed may drop to zero. At a 90-degree angle of attack, none of the airplane's control surfaces are effective.

That condition can occur without an excessively nose-high pitch attitude. On some airplanes, it can occur at an apparently normal pitch attitude, something that can mislead the pilot because it presents itself similar to the beginning of a normal stall recovery.
 
Squat switch on the gear works great.

That would be of little use since a stall detection and warning systems are all based on an alpha detection system. A squat switch would not be good enough to inhibit stall detection on the ground since something like a bounced landing will not inhibit the stick pusher since the weight would be off the gear at that time. You don't want a stick pusher pushing you're nose down when you bounce a landing :).

You can think of this system as being similar to a gear retraction inhibit system on most light twins. You can not have just one input to account for all the variable. You'll want to inhibit the stall detection system by a combination of inputs from a few various systems. Squat switch would only be one part of the equation :wink2:.

There is a reason why systems like these are not overly simple systems :)
 
Why do some airliners lack the ability to recover from a deep stall? Is it because of a lack of elevator authority?

I think the easiest way to explain it without getting into the aerodynamics behind it is to just say that swept wing stall characteristics do not yield to easily reattaching the airflow over the wing. The reason is that on swept wing aircraft the center of lift moves forward as the wing stalls and causes an upward pitch for the plane.

Bob
 
There's one recovery method that's supposed to work on the most stubborn swept wing designs but it requires a lot of altitude to complete successfully. Basically you let the plane roll over to approximately 90° bank and rudder the nose down. At 10,000MSL there probably wasn't enough room left for that but at 20-25k it ought to work.

Will the Airbus even let you roll over to 90º bank? :dunno:
 
As soon as he let off the stick and dropped the nose, the stall horn came back on showing it was responding to pitch change, then they pulled the power levers back. Everything about that showed that the plane was still accurately responding to control inputs.

The stall horn came back on at 60kts, still a looong way from recovery. The THS, trimable horizontal stabilizer, had gone to full pitch up. The elevator may not have been able to overcome this, and the only way to have corrected this was to use the trim hand wheel, something that just isn't normally done in an AB.

I have read a lot about the technical situation these poor souls faced. In our armchairs this upset is recoverable. However I do not think it's that simple. Might some pilots been able to recover? Probably. Could most? I doubt it.

There is lot of blame to be thrown around. To put it all on the pilots is a political copout.
 
The stall horn came back on at 60kts, still a looong way from recovery. The THS, trimable horizontal stabilizer, had gone to full pitch up. The elevator may not have been able to overcome this, and the only way to have corrected this was to use the trim hand wheel, something that just isn't normally done in an AB.

If holding aft stick caused the trim to go full nose up, why wouldn't holding forward stick cause it to go full nose down?

I have read a lot about the technical situation these poor souls faced. In our armchairs this upset is recoverable. However I do not think it's that simple. Might some pilots been able to recover? Probably. Could most? I doubt it.
With the proper training I expect most AB pilots could have easily avoided this trap.
There is lot of blame to be thrown around. To put it all on the pilots is a political copout.
But that's the French way.
 
If holding aft stick caused the trim to go full nose up, why wouldn't holding forward stick cause it to go full nose down?

Been a long time, but I am pretty sure the system doesn't work that way. Why would it trim into a stall?
 
With the proper training I expect most AB pilots could have easily avoided this trap.
I think that with the proper training most pilots could and should have avoided getting very far into a stalled situation. They should have at least recognized that they were stalling. However I'm not convinced that many pilots could have recovered once the airplane was falling and not flying. Maybe it was possible, maybe not.

I don't think it has anything necessarily to do with Airbus either. None of these types of airplanes are certified to be recoverable from all situations.
 
Last edited:
Back
Top