Air France 447 - On topic only!

[Black Ram]

Yeah, usually when a plane crashes it's because many things go wrong. Then that leads to the discussion of all the things that added up just right to cause the crash...and then, most logical folks go to the "what ifs"

It's really not that complicated. I said it already - there is no evidence one of those three pilots knew what was going on and thought Bonin was addressing the issue by pushing the sidestick down. Robert did not follow through with his monitoring. They had altitude and AI. They knew what altitude was safe, permissible with less margin, out of reach...now they are saying they needed just one more piece of information - the position of the control inputs - and they would have surely handled this properly. But this is pure speculation. Especially considering how much they overlooked.
What I don't like about Sully's interview was him clearly stating/implying "this would not have happened on a Boeing". Yet, no credit was given to the hated Airbus during his Hudson landing. To me this shows a trend and lacks objectivity. I could care less what Sully likes to fly, but how many people actually go and read either the AF447 or US1549 reports? It's easier to just listen to this guy, especially since he is now a celebrity.

[Deadstick]

Here's the deal. Aside from all the hindsight of "he could have made it back to pavement as shown by a simulator" and what he and Skiles may have forgotten to do (ditch switch) and did he or didn't he command extreme nose up before impact, the facts remain that in the incredible pressure of a twin engine failure at 3000 feet (a 10:1 glide ratio gives him 30k feet forward, about five miles and a couple of minutes) he made the correct decision on his best alternative and put it down in a totally survivable water landing in which all 158 souls got on to dry land. I wouldn't bet on me being in the left seat and having that outcome.

And to my limited little flying knowledge of mostly gliders with some SEL mixed in, it strikes me that Air Busses (Airbi?) are over engineered to compensate for poorly trained aviators flying the craft in second and third world countries, and when things go south in all of that automated protection stuff you're likely to find yourself just north of the southern tip of Chile, and I think Sullie's assessment of a yoke type aircraft being less likely to be found contributing to the problems in the AF 447 situation is indeed correct. If I was sitting in the left seat with a yoke shoved up my lap I think I would ask if that kind of control input was necessary as I was watching the altimeter spin down and listening to repeated "stall, stall, stall" coming from the overhead speakers.

So I'll grant you that everything had to line up perfectly for all to survive Cactus 1549, but some of that alignment (if not most) came from the guy in the left seat on that day.

No?

[Black Ram]

Source? (I agree with it anyway)

It's actually some blog I found by accident, from impartial Australia

http://blogs.crikey.com.au/planetalking/2012/04/29/af447-did-an-airbus-death-grip-kill-228-people-or-was-it-air-frances-fault/

Then there was another "response" somewhere on the internet, titled something like "how crappy cockpit design crashed an Airbus, killing 228 people"...

I agree with you though, especially with finding the root causes. I'm not saying the pilots were solely at fault for their incompetence. Of course it is a system failure. And I support interface improvements, optimization. But do you really think they are going to change the sidestick design?

[3WE]

...What I don't like about Sully's interview was him clearly stating/implying "this would not have happened on a Boeing". Yet, no credit was given to the hated Airbus during his Hudson landing. To me this shows a trend and lacks objectivity. I could care less what Sully likes to fly, but how many people actually go and read either the AF447 or US1549 reports? It's easier to just listen to this guy, especially since he is now a celebrity.

Take a deep breath. Maybe you should be directing your hatred towards the media and not Sully.

The side stick makes for a interesting story. But it's pretty rare that the evening news could cover a 300 page report, and thoroughly review all aspects.

Discussion forums are where we discuss all the crazy swiss cheese alignment nuances. The evening news is where we get the quickie that maybe there's an adantage to Boeings old fashioned control deisgn and a flaw with independent side sticks...and that that might have saved the day.

...And face the facts, Sully's a lot more interesting at delivering that message to the entertainment-needing public than you are.

[Black Ram]

Here's the deal. Aside from all the hindsight of "he could have made it back to pavement as shown by a simulator" and what he and Skiles may have forgotten to do (ditch switch) and did he or didn't he command extreme nose up before impact, the facts remain that in the incredible pressure of a twin engine failure at 3000 feet (a 10:1 glide ratio gives him 30k feet forward, about five miles and a couple of minutes) he made the correct decision on his best alternative and put it down in a totally survivable water landing in which all 158 souls got on to dry land. I wouldn't bet on me being in the left seat and having that outcome.

Sounds like it's very psychological for you. You look up to someone, who does something you are convinced you can never do. Admiration. But ask yourself what will happen if everyone in the NTSB followed that path and didn't stay professional and focus only on the facts. The facts are, as I stated, that the plane had a very improtant role in the mechanical aspect of flight, reducing workload, aiding with smoothness, lateral control - very important for a water landing - and culminating with preventing what would have almost certainly ended as a catastrophic loss of stability, control, lift....You can believe whatever you want, but I think the NTSB agrees the plane played a role. I'm not taking away from the crew's role, their crucial decision making. But as much as pilots' union people hate to admit it, computers will always be better at the mechanical aspect of flying.

it strikes me that Air Busses (Airbi?) are over engineered to compensate for poorly trained aviators

You mean sophisticated. And no, it's not just for poorly-traind pilots from third world countries. It's for humans, who have human limitations and sometimes fail to perform under stress, acting instinctively. Humans, who also have biological resource limits and who can multitask only a limited set of things. Just imagine Cactus 1549's case - you have to figure out what's going on with each individual engine, you have to take out manuals, read procedures, carry out those procedures, think what your options are, come up with a main plan, alternative plans, while steering. Isn't it nice to have the plane retrim itself? And while Sully isn't poorly-trained by any means, he activated the stall protection. You are descending into water, you have no airspeed, your instinct is to want to pull up. It's such instincts, which are human, that make those protections useful. Because dropping into the water from 150ft, while dipping one wing would not have been fun.

So I'll grant you that everything had to line up perfectly for all to survive Cactus 1549, but some of that alignment (if not most) came from the guy in the left seat on that day.

No?

Actually, I would agree with the NTSB - this was a great example of how man and technology should interact and work together, bringing out the best of each. Sadly, in this craze of finding the next hero/celebrity, many fail to recognize it as such. Poor HAL doesn't have very strong communication skills and can't go on a bunch of nightly shows to collect his 15 mins of fame

[Black Ram]

Take a deep breath. Maybe you should be directing your hatred towards the media and not Sully.

The side stick makes for a interesting story. But it's pretty rare that the evening news could cover a 300 page report, and thoroughly review all aspects.

Discussion forums are where we discuss all the crazy swiss cheese alignment nuances. The evening news is where we get the quickie that maybe there's an adantage to Boeings old fashioned control deisgn and a flaw with independent side sticks...and that that might have saved the day.

...And face the facts, Sully's a lot more interesting at delivering that message to the entertainment-needing public than you are.

I agree. But I don't have hatred for Sully. I just expected a little more overall objectivity from him. Maybe he wants to be part of the good story on tv. Or maybe, as a pilot, he feels he should express solidarity with the crew and it feels wrong bashing the french pilots. I don't know.
Another question is whether anyone has any plans to modify the sidestick design.

[Black Ram]

I think Sullie's assessment of a yoke type aircraft being less likely to be found contributing to the problems in the AF 447 situation is indeed correct. If I was sitting in the left seat with a yoke shoved up my lap I think I would ask if that kind of control input was necessary

One pilot had said that if they were in the place of the captain of AF447, they would have removed one of the F/Os from their seat to take control. There were many signs for captain Dubois that Bonin was feeling uncomfortable and anxious, prior to the captain leaving for his break, when sitting next to him. But he let him fly both before and after his break. Not to mention the captain left right after realizing he had misjudged the storm (and as later noted by Robert, the captain hadn't set the radar properly to get a good picture of the storm)

[Gabriel]

And I support interface improvements, optimization. But do you really think they are going to change the sidestick design?

No, and not only the sidestick design. Airbus is deeply committed to it's philosophy, especially the parts of it that I don't like, of which the spring-loaded, self-centering, non-force-feedback, independent side sticks is just one of them. (and Airbus approach to autotrim is another).

But I have to admit that this philosophy has shown a level of safety at least comparable with that of Boeing, which doesn't rate it as a good innovation but at the same time makes it perfectly acceptable to say the least.

I would fly on an Airbus any day. Or a Boeing. Or Embraer...

Just imagine Cactus 1549's case - you have to figure out what's going on with each individual engine, you have to take out manuals, read procedures, carry out those procedures, think what your options are, come up with a main plan, alternative plans, while steering. Isn't it nice to have the plane retrim itself? And while Sully isn't poorly-trained by any means, he activated the stall protection.

Actually, now that I think of it, don't you think that the autotrim and the normal law (both with Airbus philosophy) might have contributed to him running out of speed too soon and too high?

In a Boeing 777 (also FBW), the autotrim philosophy is to keep the trim speed (correcting for other factors that affect the trim like thrust, flaps, spoilers, changes in CG...). So you need to actively pull up (or re-trim, or change the selected speed) to slow down.

On the other hand, the Airbus logic is keep the load factor (fast speed) transitioning to keep the pitch (slow speed). This means that left alone, the plane will try to hold the vertical speed. If there is no enough thrust to sustain the climb / descent, the plane will first trade airspeed to hold the vertical speed. As it goes through a certain airspeed threshold, it will change to hold pitch (letting the plane reduce the climb rate / increase the sink rate somehow), but it will still not let the nose go down to prevent the speed from further decaying. Only when reaching the alpha prot the flight control law will change to hold AoA. At this point the airplane will lower the nose as necessary to hold the threshold AoA (as long as the pilot doesn't pull up) and hence keep the airspeed. If the pilot pulls up, then the alpha max will be reached and the plane won't go past this AoA no matter how hard the pilot pulls up.

So maybe the Airbus prevented Sully from stalling at the last second, but at the same time maybe it was the same Airbus who induced the plane to get into such a low energy mode to begin with: Unlike with a Boeing (FBW or not, or a C-172 for the matter), with the Airbus it required active steering (pushing down) to hold the speed.

[Black Ram]

Actually, now that I think of it, don't you think that the autotrim and the normal law (both with Airbus philosophy) might have contributed to him running out of speed too soon and too high?

I can see how it can also be a responsibility, but any professional airman would be watching their speed. I think I would like it better if I can point the stick where I want to go and not worry about retrimming. Just fly pitch and heading, watching the airspeed. Especially since you can just let go of the stick once the plane is going where you want it to for the time being. It seems you get one task out and you can use that freed up resource for something else in the emergency.
From what I read their approach was more or less fine until the last few hundred feet, which is where they should have adjusted for aoa in a manual plane. I think the Airbus simulations, where they showed they could have landed much more smoothly, touching down in ground effects is very theoretical and it would have been hard to make it work on the Hudson, where you have so many bridges, ferries. It would have taken a run that is just too long.
Also, wouldn't have autotrim helped for a smooth transition to flaps 1 and 2? And flaps 3, if they decided to use them - I'm not sure why they didn't.
And now here I'm not sure how things are, but if there was no autotrim, wouldn't the nose have started to go down almost immediately? In that case they would have had to been engaged in retrimming, because it would have been too early to start losing their precious altitude.
Another aspect is that sometimes I like hi-tech stuff. But I think there are objective pros, not just cons like no feedback and low observability.

[Gabriel]

I can see how it can also be a responsibility, but any professional airman would be watching their speed.

Awh, come on. Isn't this a circular argument? All this stuff is there precisely to protect professional airmen against themselves. If one would assume that professional airmen will always do what they should then everything would be direct law and there would be no envelope protection.

I think I would like it better if I can point the stick where I want to go and not worry about retrimming. Just fly pitch and heading, watching the airspeed. Especially since you can just let go of the stick once the plane is going where you want it to for the time being.

Well, here I don't agree. I mean, yes this works great in some situations, but not always. When everything goes right and normal and you are well within the envelope, this is good. But when you have a lot of workload and speed and AoA are of concern (and a glide is such a situation), I would prefer to have speed stability over path stability any day.

It seems you get one task out and you can use that freed up resource for something else in the emergency.

And here is the point. In a glide, speed control is of uttermost importance. You don't have too much control over path. Pitch up to extend the glide: the speed goes below glide speed. Push down to aim closer: your speed goes up and you'll have either a high energy touchdown or a too long flare.
The way to control path in a powerless glide is controlling drag. So I'd prefer any day that the plane takes care of the speed so I have one task out and I can use that freed up resource for something else in the emergency, like monitoring the path and see if I need to add more drag to steepen the glide.

Imagine this situation: The plane is aiming too low and the speed is too high, so I grab the sidestick, pull up to establish a new path, and release the sidestick. Then I get distracted by other tasks. The Airbus would have no problem to pull up for me as the speed goes down to keep the vertical speed, to the point of alpha prot (in direct law) or into the stall (in abnormal alternate law). Compare that with the 777 philosophy. I trim for the glide speed that I want and I know that the plane will hold that speed. I prefer that. In a powerless glide, just "aiming past that bridge" just doesn't work.

I think the Airbus simulations, where they showed they could have landed much more smoothly, touching down in ground effects is very theoretical and it would have been hard to make it work on the Hudson, where you have so many bridges, ferries. It would have taken a run that is just too long.

Not that long. He used that extra energy to bleed of speed at say 100ft instead of skimming the water. Okay, doing it low would have taken a bit more of room because of the reduced drag in ground effect, but not that much. And the Hudson in long, there were no bridges in the vicinity of the touchdown zone, and the ferries, if there was any, you can manage by lateral control (the Hudson is quite wide too).

Also, wouldn't have autotrim helped for a smooth transition to flaps 1 and 2? And flaps 3, if they decided to use them

Yes. And in the 777 it would have helped too. Only that with a different philosophy.

And now here I'm not sure how things are, but if there was no autotrim, wouldn't the nose have started to go down almost immediately? In that case they would have had to been engaged in retrimming, because it would have been too early to start losing their precious altitude.

Yes, and with an autotrim with the 777 style too. In the 777 when you move the trim wheel in normal law what you are doing is changing the selected speed, the one the autotrim will hold after that. So what you would have done in a 777 is trim for glide speed and let the autotrim take care after that. If your speed at the time was higher than the trim speed, the plane would have climbed to trade that extra speed for altitude. If it was below, it would have lowered the nose to gain speed. Optimimum energy management.

Another aspect is that sometimes I like hi-tech stuff.

That's okay. But note that the 777 (or 787) is also FBW and also has normal, alternate and direct law, envelope protection and what not. They have no less hi-tech stuff than the Airbus. Maybe even more, since:
- the systems will move the controls as a pilot would have to to achieve a given performance, thus providing feedback to the pilot about what the plane is doing.
- The yoke is not just spring-loadad, but has an artificial feel system that gives the pilot important feedback and cues about what he is doing with the plane.

So it's not a matter of technology. Just a different philosophy.

What I like about the Boeing approach to FBW over Airbus one is:
- That the airplane response in normal, direct and alternate law is almost the same between them and compared to a non-FBW plane, thus giving a continuity of applicable stick.and-rudder skills since the first hour in a Cessna 172.
- That The AP and AT move the controls like like the pilot would have to do to achieve the same performance.
- The linked yoke.

But I think there are objective pros, not just cons like no feedback and low observability.

Of course, and one of the pros is that, at least in normal law, the airplane will simply not let the pilot do things like stall, exceed the structural limit load factor, or roll inverted. All these things would be possible in a 777, while hard because the artificial feel system will exaggerate the feel as you approach the edges of the envelope, being needed a disproportionate force to stall or roll inverted.

But one nice feature in the Airbus philosophy is that, in a maximum performance maneuver (low level stall, windshear, CFIT escape), you can simply firewall the throttles and pull up to the stops, and the plane will give you max performance. You don't have to worry abut being pulling less that what you could or more than stall.

It's also a nice feature in a UAS event if correctly managed by the pilots: You apply the memory items (5° nose up, climb thrust) and then you can release the sidestick and the plane will keep it for itself. Even further, even without applying the UAS procedure, since the plane was flying straight and level before the AP disconnected, the only think you have to do is not touch anything and the plane will keep flying straight and level, at least initially. You would then have to monitor if the plane is adding pitch to keep the VS at zero, a clear sign that you are slowing down. But since the UAS events typically take less than a minute (and AF was no exception), chances are that it would have cleared beofre you've bled off too much speed. It's interesting that one things that the pilots of AF could have done to survive the UAS event was... nothing at all. The plane would have behaved better with no pilot at all in the cockpit. (of course, applying the UAS procedure would have been the right thing).

Another aspect is cost: no wonder that two small, independent, spring loaded, non-feedback, non servo sideticks are much cheaper than the linked yokes with artificial feel and AP-induced motion.

No wonder that, all in all, I prefer Boeing approach. But all that is a matter of preference and philosophy. And, as said, I would climb on any Boeing or Airbus plane any day with no concerns for my safety (the operator is a much more major factor than the plane)

The only one thing that I find unacceptable in the Airbus philosophy is that in the abnormal alternate law that AF was flying, there is no speed or AoA stability (as required by the FAR) and there is no low speed or stall protections. While it was not a factor i the AF case, it could have been, since the plane left alone would have no problem to pull up into a stall and then keep pulling up while trying to keep a vertical speed or pitch attitude. Yes, it sound contradictory to what I've said tow paragraphs above, but it's not. it's the pros and conns.

[3WE]

...I think I would like it better if I can point the stick where I want to go and not worry about retrimming...

On a Boeing, that would be called the autopilot.

[Black Ram]

No wonder that, all in all, I prefer Boeing approach. But all that is a matter of preference and philosophy. And, as said, I would climb on any Boeing or Airbus plane any day with no concerns for my safety (the operator is a much more major factor than the plane).

Thanks for explaining all this.

I'm really not debating which type of airplane is safer. I
m debating whether the plane did its part in the Hudson landing and whether Sully was fair in his remarks.
Keep in mind US Airways doesn't have 777s. The only other FBW type they have is Embraers. They have 757, and what would have likely substituted the A320 from the Boeing camp, older 737s. For me, the difference between the 737 and 777 is tremendous, they are not standardized like the airbuses (though the NG makes some difference).

What I don't understand about your glide analysis - do you really disregard the airspeed and let the plane take care of it in manual trim airplanes? Is monitoring the airspeed really a resource you should be freeing up? And haven't most glides in history been more erratic than just trimming the plane to keep the speed stable? You said it yourself - you can't just "aim past the bridge", so wouldn't you be watching your speed anyway? And can't you adjust the airbus sidestick for a glide, though with more frequent (yet very subtle) corrections, which however eliminate two separate inputs and combine them into one? Yes, I understand that if you forget about flying the plane, it will keep your selected attitude until it bleeds the speed off, up to alpha prot. I was just hoping that in a dual-engine failure with two available pilots, one would be flying the plane. And lastly, isn't there a way to manually trim an airbus (I don't mean on the ground).
Gabriel, you are really making me wanna save up for an a320 sim and try all this out.

[Black Ram]

On a Boeing, that would be called the autopilot.

Mea culpa. I really forgot only Boeings are blessed with this innovative feature called an autopilot.

[3WE]

What I don't understand about your glide analysis - do you really disregard the airspeed and let the plane take care of it in manual trim airplanes? Is monitoring the airspeed really a resource you should be freeing up? And haven't most glides in history been more erratic than just trimming the plane to keep the speed stable? You said it yourself - you can't just "aim past the bridge", so wouldn't you be watching your speed anyway? And can't you adjust the airbus sidestick for a glide, though with more frequent (yet very subtle) corrections, which however eliminate two separate inputs and combine them into one? Yes, I understand that if you forget about flying the plane, it will keep your selected attitude until it bleeds the speed off, up to alpha prot. I was just hoping that in a dual-engine failure with two available pilots, one would be flying the plane. And lastly, isn't there a way to manually trim an airbus (I don't mean on the ground).

There is only one maximum glide speed- go faster OR slower, and your glide steepens. A traditional airplane tends to stay on speed, and will let you know if you are fighting against that trimed speed by feedback/resistance on the controls. An Airbus will dutifully stay on attitude- and then the airspeed can move well away from the best glide speed (when the engines are on- autothrottles maintain speed on either aircraft, but when descending towards the hudson....) So, when your two engines are out and you are trying to remember to switch on the ditch switch, and realizng that you may or may not have less than 5 minutes to live, you can rest assured that the plane is going to tend to stay at the same speed.

[Gabriel]

Black Ram,

The only reason for me to compare with the 777 is that it's a FBW plane. many times people say "Airbus xxxxx because it's FBW". And that's wrong. It's Airbus' approach to FBW what is different from Boeing's approach to FBW, and not the FBW itself.

Now, to your questions:

Yes, speed control in a glide critical in any plane. It's just that in a speed-stable plane this task is much easier and demands much less workload.

Yes, in a two persons crew, in an emergency one pilot is suppused to be fully dedicated to fly the plane while the other one should run the procedures, checklists, etc.

Yes, of course in an Airbus you can and will control the glide path and speed with the sidestick. The difference is that in a non-FBW plane or a Boeing-like FBW, other than the airspeed indicator, you have another strong cue regarding whether you are slowing down or speeding up from your trim speed: The force on the yoke (no force = trim speed). In fact, this cue is constantly used for example in every approach (in manual flight, of course).

AFAIK, yes, you can trim an Airbus manually (in fact, I think that if you dare to touch the manual trim, the auto trim is lost for the remining of the flight), but what is "trim" in a plane that works on a load factor response to control inputs? The comment by 3WE ("Boeing calls it autopilot") might be more relevant that what it seems. Sometimes I have the impression that the Airbus philosophy is like an AP always on, and that the sidestick is just a data input device for the AP like any of the knobs and switches in the AP: Pull/push the sidestick to dial-in the desired climb/descent rate, and then let it go.