Announcement

Collapse
No announcement yet.

Boeing's opinion on the Airbus side stick...

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    The Airbus is apparently more challenging to fly in cross wind landings.

    Comment


    • #17
      Excellent analysis of feedback pros and cons

      Gabriel,

      Thanks for your insights. They certainly come into play when understanding the AF A330 crash just north of Brazil. Inadequate training, coupled with faulty Pitot sensors that AF had decided to NOT upgrade in a timely fashion, played a significant role, but the fact remains that the copilot and pilot were unaware of each other sending conflicting (and opposite) instructions.




      Kishore




      Originally posted by Gabriel View Post
      I'm sorry I missed this thread back then, but I'll offer my version now.

      First, a little bit about control forces and displacement. As an example I'll use the elevator but a similar reasoning may be made for the rest.

      The force needed to move the elevator is proportional to the displacement from it's natural equilibrium position and to the airspeed squared. PERIOD.

      No matter what type of control system the airplane has, someone or something will have to make that force, be it that you grab the elevator directly with your bare hands or that it is an hydraulic piston acting according to the orders of a flight computer.

      Now, the forces that the pilot has to do differ depending on the type of control system.

      So, in the beginning, and still in use, the control system is mechanical, with the stick or yoke linked to the elevator by pull cables or push rods, in a way that the movement of the elevator is proportional to the movement of the stick or yoke.

      In this way, the force needed to be applied on the stick or yoke to displace the elevator is proportional to the force applied by the system on the elevator, that is proportional to displacement and speed squared. And as said just before, the displacement of the elevator is proportional to the displacement of the stick or yoke.

      Now, this has a number of good characteristics:

      A given displacement of the elevator (and hence of the stick/yoke) always makes the same change in angle of attack, regardless of the speed and the force needed to make such a displacement. In that way, the distance that the pilot displaces the stick/yoke is a direct feedback for angle of attack.

      Because the force on the elevator (and hence on the stick) is proportional to the the speed squared and the lift is also proportional to the speed squared, the force applied to the elevator (and hence to the stick/yoke) is proportional to the increase in Gs, regardless of the speed and the displacement needed to do such a force. There is a key handling characteristic that is called pound per Gs. It is very important because the pilot has to apply a given force BEFORE the airplane sustain a given G, and in that way the force applied is a direct feedback for Gs. A lot of force means a lot of Gs and a risk to break the plane.

      Finally, the relationship between force and displacement gives a direct feedback of speed. If you need to make a lot of force to displace the yoke a little you are going fast, if you need to make little force to displace the yoke a lot (mushy controls) you are slow.

      That may sound complicated, but if you think a minute it's the same than a steering-wheel in a car.

      Turning the steering wheel a certain angle will make the car turn with the same radius (assuming the car doesn't skid) regardless of the speed or of how much force is needed to move the steering-wheel that much.

      Applying a given turning force on the steering wheel always puts the car (and the occupants) under the same centrifugal force (or lateral Gs) regardless of the speed or of how much the steering wheel is turned to get that force. Again, this is a key characteristic because that force gives a direct feedback to the driver of how close is the car from skidding.

      Finally, when you go slow (but not too slow) you can turn the steering-wheel a lot with little force, and when you go fast you need a good force to turn it just a couple of inches. So the relationship between displacement and force is a feedback for speed.

      Enough for mechanic controls. Then came hydraulics, or better servo-hydraulics.

      In servo-hydraulic systems the yoke moves a servo-valve which proportionally controls the displacement in the hydraulic piston that moves the elevator. The problem here is that moving a servo-valve doesn't require any force and any force applied to the elevator doesn't transmit to the servo-valve and hence to the yoke. Bye-bye feedback.

      This is very dangerous, because the pilot could for example displace the yoke enough to apply a lot of Gs and realize only when the plane (and the pilot itself) is already under those Gs, or when the plane breaks under those excessive Gs.

      This problem is so serious, and the mechanical system gave such a good feedback, that the airplanes manufacturers, all of them (including Airbus in the early models), decided to design something that they called "artificial feel" and that would be the equivalent of the "force feedback" in modern PC joystics and steering-wheels. And they designed the artificial feel so that the forces on the yoke were proportional to the displacement and to the speed squared! Not by chance that's exactly the same than the mechanical controls.

      Then came fly-by-wire.

      Fly-by-wire means that the control that the pilot manipulates, regardless of what it looks like, sends digital signals to a computer which then send orders to a hydraulic system to move the elevator. The "dictionary" in the flight computer that translates the the pilot's input into commands to move the elevator is called "flight control law", or control law for short. And here is where manufacturers start to diverge.

      The Boeing 777 is fly-by-wire. But the control law is simple: The computer will command the elevator to move proportionally to the displacement of the yoke, just like in the mechanical or hydraulic systems. Also like in mechanical and hydraulic systems, the force needed to displace the yoke that much will be proportional to the displacement and the speed squared, of course again with the help of an "artificial feel" system.

      Airbus followed a different philosophy. They decided to take advantage of the power of computers and the flexibility of software programs to create some advanced features.

      To begin with, they went for a joystick type of control instead of a yoke. In my opinion, that's the most irrelevant difference (unless there are some issues with the range of motion available).

      Then, they got rid of the "artificial feel" system. The joystick is simply a spring-loaded self centering joystick.

      Now, with a small control that has no mechanical linkage or hydraulic servo-valves and doesn't even has a force feedback system, the whole device got small enough to be placed virtually anywhere. So they put it at the side thus creating the side-stick. That location has several advantages: It gives the pilot an easier access too and egress from his seat, it lets the pilot hold a chart, a laptop or a tray of food on his lap, it lets the non-flying pilot to cross his legs, and it doesn't obstruct the view to any portion of the instrument panel.

      Because the spring-loading and lack of artificial feel, now force and displacement of the joystick are always proportional one to the other. That means that force and displacement now feed back the same piece information, instead of three different pieces of information as the other cases (including the fly-by-wire B-777). Now which piece of information would that be will depend on the design of control law.

      Airbus decided not to keep the proportionality between the joystick displacement and the elevator displacement, as had been the case up to then. Instead, they decided that the computer will move the elevator as needed to achieve a G load proportional to the joystick displacement. That means that the same joystick displacement will produce different elevator displacements in function of things such as speed, aircraft weight, and bank angle.

      So now the Gs are proportional to the joystick displacement and to the force applied to the joystick. So now both joystick force and joystick displacement are two different feedback for Gs. The feedback for angle of attack and the feedback for speed are gone.

      So to prevent that the pilot inadvertently stalls or overspeed the plane, they programmed restrictions in the control law. If the pilot pulls up to command a G load that, at the current speed, would make the airplane stall, it will not comply and will command a pull up just before the stall. If the pilot pushed down into a dive or advances the throttles in a way that would cause an overspeed, the computer won't comply and will pull up to reduce speed before an overspeed happens.

      Despite being there two different feedback for Gs (joystick displacement and joystick force), perhaps because the range of displacement is small and the control forces are low they added also a restriction to the Gs. The computer won't command a pull up past the design max Gs despite of how hard the pilot pulls up trying to do so.

      With the pilot pulling up hard to avoid a mountain, the computer will fly into the mountain rather than letting the plane break-up due to overstress or stall.

      That restricts the pilot authority, but in many cases can be good.

      For example, if the pilot finds himself flying into a mountain he can simply apply full power and pull up to the stops without worrying about stalling or breaking the plane. The computer will pull up as much as it can be done without that happening. If that doesn't save the day, nothing will. In any of other systems, the pilot has to "fine tune" his pull-up to achieve that optimum performance. If he just scares and pulls-up hard he can stall and crash in a situation where a max performance escape maneuver might have saved the day. It's like the ABS in your car. It won't let you brake so hard to lock the wheels.

      Another important advantage of this philosophy is standardization of the handling characteristics: Moving the side-stick one inch back always requires 0.5 pounds of force and always produce a 1.2Gs pull-up regardless of whether the plane is going fast or slow, is heavy or light, has the center of gravity on the forward or aft limit, or it is a small A 319 or a huge A 380. (the numbers there are not real, I just made them up for the example). That simplifies training and transition between types a lot, and hence saves money to the airlines.

      Another difference is that the side-stick is out of the field of view, specially the other pilot's sidestick, and the side-stick are not linked one to the other so when one pilot moves his joystick the other joystick just stays still. That removes the visual feedback for the non-flying pilot to know what flight control inputs the flying pilot is commanding.

      Finally, another difference between the Boeing philosophy and the airbus philosophy is how the automation (autopilot and autothrust) controls the systems.

      In the Boeing philosophy, the autopilot and autothrottles move the yoke and the throttles like the human pilot would, and then the yoke and throttles transmit the commands in the same way as if it was the pilot who had moved them. This gives the pilot the visual feedback of what the autopilot is doing. The human pilot actually sees the yoke move back when the autopilot pulls-up, and the throttles move forward when the autothrottle adds power.

      In the Airbus philosophy the autopilot and autothrust actuate directly on the systems bypassing the yoke and throttle levers. The autopilot tells the flight control computer to pull up, the side-stick doesn't move. The autothrust tells the FADEC to increase power, the throttle levers don't move. So here again, a visual feedback is removed.

      It is important to note that this difference have nothing to do with intrinsic characteristics of the side-stick or the fly-by-wire concepts.

      The Cirrus SR-20 is a general aviation piston single that is flown using a side-stick, but it's not fly-by-wire. That side-stick is mechanically linked to the elevator, so it shares all the feed-back characteristics of the mechanical control systems.

      The Boeing 777 is fly-by-wire and uses a yoke that emulates a mechanical control system. They could have chosen to use a side-stick and still emulate a mechanical control system.

      Most Airbus use a sidestick with no force feedback and that doesn't move when the other pilot or the autopilot make control inputs. They could have kept those concepts and the same control laws with a yoke instead of a side-stick.

      So whoever at Boeing wrote that, he was very careful to word the sentence as to be correct:

      "Existing commercial side sticks offer no visual or tactile cues to the pilot and must have restrictive performance limits."

      "Existing commercial side sticks" basically means "Airbus side-sticks".

      And yes, Airbus sidestick, coupled with Airbus flight laws and the other design features mentioned, offer no visual an little tactile cues to the pilot and have restrictive performance limits.

      Still, a yoke could be like that too, and a side stick doesn't need to be like that.

      Those are not intrinsic characteristics, but design decisions.

      And of course, that doesn't mean that "existing sidestick" are worse than "existing yokes". Neither the contrary.

      Comment


      • #18
        Originally posted by kishoreajoshi View Post
        Gabriel,

        Thanks for your insights. They certainly come into play when understanding the AF A330 crash just north of Brazil. Inadequate training, coupled with faulty Pitot sensors that AF had decided to NOT upgrade in a timely fashion, played a significant role, but the fact remains that the copilot and pilot were unaware of each other sending conflicting (and opposite) instructions.

        They must have been aware of the conflicting instructions, because there is a "dual input" warning, and also the pilots used the side stick priority button a few times. I can't accept this claim that the pilots were some sort of victims to a faulty control input design, and I was massively disappointed when Sully himself made that claim (and other claims that he has made). The fact of the matter is, there have been other stall events in commercial airplanes, with interlinked control columns, where the crew f-d up massively. And crashed. And a few of these have happened from cruising altitude. On a few of these, the captain was telling the PF to keep pulling up and keep the wings level (sound familiar? ) So to me, shitty airmanship and shitty CRM will be shitty no matter what control input system is used. And given that the crew of AF447 had all the necessary information to recognize and deal with the situation (even though they should have followed the standard procedure, and even if they didn't, they should have avoided it), I find Sully's comments not only biased and with a hidden agenda, but also very dangerous. Not to take away from what he did, but he should go through his own Hudson landing FDR transcript.

        Comment


        • #19
          Originally posted by Original post and Black Ram
          ...Sully...
          (No disagreement with Black Ram's comments)

          I am struggling with Sully and Air France- and don't think he really belongs in a comparison.

          Sully basically did what he was supposed to do with the side stick.

          If Sully had been flying a 737, I bet he would have basically done what he was supposed to do with the yoke.

          I think maybe the Airbus deserves some kudos for it's waterskiing ability and structural stoutness...ditching's often turn out not so good.

          The Air France bunch did something that seemingly defied common sense...

          The control question sneaks in when you consider that the Airbus philosophy makes the airplane behave almost game like AND removes a big bunch of feedback.

          To do a 30,000 ft stall in a Boeing will require that both yokes be way up in two pilots chests, pulling hard against someone's hand, AND a stick shaker going off...The Airbus stall let one of the pilots make a much more modest pull on his side stick, and the clues to the other pilot were much more subtle...

          ...and then add to that, the confusion that they had very limited recent experience on how an AIRPLANE acts (they only had the extreme vanilla version)...so when presented with an AIRPLANE, they may have been a bit off their game. [Acknowledgement- I am making some exaggerations to point out some subtleties.]
          Les règles de l'aviation de base découragent de longues périodes de dur tirer vers le haut.

          Comment


          • #20
            Originally posted by 3WE View Post
            To do a 30,000 ft stall in a Boeing will require that both yokes be way up in two pilots chests, pulling hard against someone's hand, AND a stick shaker going off...The Airbus stall let one of the pilots make a much more modest pull on his side stick, and the clues to the other pilot were much more subtle...
            Let me add to that the mechanical interlink between both yokes in a Boeing, compared with 2 independent sticks in the Airbus. The Dual Input would have been of a very different nature in the Boeing. In the Airbus, they fought for who is in charge by pressing the priority button, and when neither was pressing it and both were making inputs, one may have been pushing full down believing that he was commanding full down elevator, while the other was pulling full up believing that he was commanding full up elevator, when both, as a team, were unknowingly commanding neutral elevator.

            In a Boeing, the end result may be the same: One my pull up and the other push down with equal force, resulting in no motion of the yoke and elevator. However, it would have been very clear what the yoke and the elevator was doing, what was the other pilot's intention, and who was winning the fight.

            --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
            --- Defend what you say with arguments, not by imposing your credentials ---

            Comment


            • #21
              Originally posted by 3WE View Post
              Sully basically did what he was supposed to do with the side stick.

              If Sully had been flying a 737, I bet he would have basically done what he was supposed to do with the yoke.
              Sully did a more than good enough ditching, since everybody walked away from it, but, technically, it was far from perfect.

              He didn't touch down with the deck angle, horizontal speed, vertical speed and landing configuration recommended by Airbus. In particular, he did the final stage of the approach too slow, flared to high, run out of airspeed too high, and when the plane started to sink (in the air, due to insufficient airspeed), he pulled up to the stops, and it took Airbus normal law to take over and keep the airplane at the maximum (but below the critical) angle of attack to prevent a stall.

              Whether Sully intentionally pulled fully up when he found himself in that situation, to take advantage of this Airbus feature, or if he pulled up to arrest a descent that was (at that point) physically impossible to arrest and would have stalled the plane if it wasn't for the Airbus feature that kicked in (unexpectedly to Sully), I don't know.

              So would he have stalled the plane had it been a 737? I don't know. I tend to think that he would not have, even if running out of speed and pulling full up taking advantage of the Airbus logic was not his plan and decision. In a 737, first he would have had to pull further and further back (and with more and more force) to keep increasing the angle of attack, and then, at a point, he would have had a stick shaker. I prefer to believe that, even if he reached to that point unintentionally, he would have not kept pulling further up past the stick shaker. As an example of a very dare situation where any pilot would be very tempted to keep pulling up but a good pilot would not, we have the BA 777 that crash-landed short of the runway at Heathrow with both engines failing during the final approach. They didn't know if they were going to make it to the runway or not, or even to the airport property (and there was a highway just out of the fence). They didn't know if they would be able to clear the localizer antenna or crash through it. The FO, who was the pilot flying, was trying to keep the plane gliding as far as possible and then, stickshaker!!! What did the pilot do? He relieved back pressure on the yoke and reduced the Angle of Attack, hence recovering from an approach to stall, and imminent stall if he didn't take corrective action. He desperately wanted to go up (or at least not as down), but he understood that trying to "impose" an unachievable glide path would only end up going further down, and possibly lose control.

              I tend to think that Sully, who was also an experienced glider pilot, is in this teams of pilots that know their stuff, rather than in the "other team" that react like the Air France pilots did, and hence would not have stalled a 737 either.

              In any event, the ditching itself is not the complicated part of what Sully did. Ditching in a long, wide, smooth river is just a matter of keeping enough speed to keep the airplane flying and gliding, and flaring before touch down (What are you going to do? Push down and fly nose first into the water?).

              The brilliant part of Sully's act was not the stick-and-rudder skill, but the managing of the situation as a whole, from the resource management with ATC, his FO and the cabin crew, to the decision making: it is not easy to have a few airports nearby that you are not sure if you can reach or not, and decide for place that is NOT an airport "only" because you KNOW that you can reach that one, especially when this alternative place is wet!!! To accept that the plane will not complete its mission, and it will not land in any airport, and it will end up in the water, and that likely it will be a total loss of the airplane, and that it is possible that people will die (including yourself), and that this is still the best course of action, when you have 2 airports that perhaps are within reach, is a very difficult decision for a Captain, who is the ultimate responsible for the piece of equipment and lives put under his command. And he did it, brilliantly. Many others died (and killed others along the way) trying to "force" the plane to do something it would not do.

              And this part has nothing to do with yokes, sticks, fly-by-wires, control laws or design philosophies.

              --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
              --- Defend what you say with arguments, not by imposing your credentials ---

              Comment


              • #22
                Originally posted by 3WE View Post
                ...Sully basically did what he was supposed to do with the side stick...
                Originally posted by Gabriel
                [A typical, lengthy statement of disagreement]
                I read your reply...twice...I am aware of your normal reaction to say that Sully's ditching was not perfect.

                I would ask that you re-read my comment, and pause on the word 'basically'.

                I restate my opinion that Sully's ditching is not a great example to support or dispute side sticks and control philosophy...at least not in the same discussion forum post with the Air-France eye-rolling-relentless-pull-up.

                I think your last sentence says this...that the control philosophy, with Sully showing decent (no, not International Super Genius) airmanship, is not that important.
                Les règles de l'aviation de base découragent de longues périodes de dur tirer vers le haut.

                Comment


                • #23
                  Originally posted by 3WE View Post
                  I read your reply...twice...I am aware of your normal reaction to say that Sully's ditching was not perfect.

                  I would ask that you re-read my comment, and pause on the word 'basically'.

                  I restate my opinion that Sully's ditching is not a great example to support or dispute side sticks and control philosophy...at least not in the same discussion forum post with the Air-France eye-rolling-relentless-pull-up.

                  I think your last sentence says this...that the control philosophy, with Sully showing decent (no, not International Super Genius) airmanship, is not that important.
                  That's because my typical, lengthy statement of disagreement was typical and lengthy but was not one of disagreement. Rather, I was basically agreeing with you, just adding a little bit (read tons!) of details to support my opinion which, again, basically matches yours.

                  In short (yes, short!):

                  1- Clearly Sully and the AF crew don't belong to the same club, and even if Sully's landing was not perfect and was assisted by Airbus philosophers, Sully basically did know what he was doing with that stick (unlike the AF crew).

                  2- I also believe that Sully would have known what to do with the yoke of a 737, if that was what he had in his hands.

                  3- Sully's ditching does nothing to favor A vs B or B vs A, in terms of flight control philosophy. That said, Sully's comments and criticism on the A's flight control philosophy (as opposed to B's) was in the context where we was opining about the AF accident, not his own ditching.

                  --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
                  --- Defend what you say with arguments, not by imposing your credentials ---

                  Comment


                  • #24
                    Holy thread revival!

                    Comment


                    • #25
                      Originally posted by ATLcrew View Post
                      Holy thread revival!
                      ...and rehashing and some dead-horse beating, for good measure.
                      Les règles de l'aviation de base découragent de longues périodes de dur tirer vers le haut.

                      Comment


                      • #26
                        I continue to be impressed with the amount of control inputs pilots make when landing planes in 'average' weather.

                        Regardless of stick, yoke, act-like-a-plane-FBW or depugoid-and-neutralize-everything-and-go-where-I-point-while-limiting-me-from-bad-and-extreme-stuff philosophy.

                        In the cockpit of the Air Arabia A320 for a flight from Peshawar Pakistan to Sharjah UAE. Find the full film at http://www.worldairroutes.com/arabia.htmlFOLL...


                        TWA B727-200 LANDING R/W 30L STL, VISUAL, (Added Dec. 2) When this video was made, the intro music of "St. Louis Blues" seemed appropriate. The source of t...


                        Then, we have this: https://www.youtube.com/watch?v=erttyI7WOKo (gonna get some double mileage out of this one)
                        Les règles de l'aviation de base découragent de longues périodes de dur tirer vers le haut.

                        Comment


                        • #27
                          Originally posted by 3WE View Post
                          If Sully had been flying a 737, I bet he would have basically done what he was supposed to do with the yoke.
                          Maybe, we don't know. But here's what we know:


                          Originally posted by 3WE View Post
                          Sully basically did what he was supposed to do with the side stick.
                          And the FBW flight controls did what they were supposed to. The report shows the plane avoided a stall and possibly a crash by attenuating the pull-up commands. On the ACI episode (dumbed down, I know), a NTSB person talks how the plane helped by "stabilizing itself" and the plane and crew were working well together. And yet, after the incident, Sully goes on to blame the bus FBW flight controls for landing too hard. One example here:



                          Basically blaming the plane for avoiding a stall. Not to be cynical, but that kind of sounds like Cpt Asseline, and to me it's very disappointing.


                          To me it's not so much about what Sully did, but about what he said - about his incident, as well as AF447 (and that's why I relate him to AF447- because he made the decision himself to comment and make such a strong statement). He seems to either be confused, or have an agenda. Either way, it's biased and dangerous....

                          Originally posted by 3WE View Post
                          To do a 30,000 ft stall in a Boeing will require that both yokes be way up in two pilots chests, pulling hard against someone's hand, AND a stick shaker going off.

                          And yet, it's happened. Not just necessarily Boeings, but airliners with traditional control columns. It's happened on a 727 at near cruising altitude (Northwest 6231), on an MD-80 (West Caribbean 70, it's happened at twice on a TU-154 - just the examples I recall.


                          Of these, Pulkovo 612 is especially interesting. I looked into the accident again recently and I was shocked. I always thought the stall and flat spin were unrecoverable because of the airplane characteristics, and that the crew did everything they could once their initial mistake of entering deep stall. NOT! Their initial mistake was not a momentary lapse of airmanship, but a sustained series of dangerous pull-ups. With "critical AoA" (stall) alarms in the background. Five crew in the cockpit (a captain, a trainee with another F/O behind him, who was supposedly a former aerobatics champion, a navigator and a flight engineer). Two very traditional control columns, a functional stall alarm, and......an actual AoA indicator. Like the one that actually shows you what the exact AoA is, with an arrow, and with a very clearly marked red zone. And yet, they pushed the plane to climb, repeatedly hitting and entering that AoA red zone, they went into pitch-up several times (and when I say "they", I mean there was no argument or opposition - in fact, the navigator was a big proponent of the plan to climb over the storm). Stall alarms were going off, and the captain was basically saying "all is good, keep doing what you were doing". Then they entered deep stall and a flat spin - they knew they were descending like crazy, and they knew they were spinning in a circle (but apparently did not know why). And yet, not a single command or discussion to bring the nose down that I could find. "Pull up, keep the wings level" were the words of the captain. And that until they hit the ground from FL410.


                          Maybe a peculiar incident and not very interesting to everyone, but clearly showing that not only traditional control columns, a present captain from the very beginning of the events, extra pilots in the cockpit, but even an AoA indicator (discussed after AF447) is sometimes not enough.


                          Therefore, I can not accept Sully's remarks, and I find them dangerous. There is obviously something else going on here. It has to do with human factors and CRM, and that's where attention should go.



                          I had a Microsoft Force Feedback joystick once (I think when they were relatively new on the market). Honestly, I can't see how linking two sidesticks - at least artificially - would be that hard or expensive. Or even adding artificial feedback. The question is, is it necessary? I think more resources should go into researching human factors and improving CRM, then depending on feedback.



                          There is a 2003 incident mentioned in the AF 447 report, previously unreported, involving a TAM A330. It reads just like AF447 - UAS, confusion, an instinctive pull-up, and then a STALL alarm. But thankfully, all resemblance ends there. Once they heard the STALL alarm, the crew lowered the nose and flew the plane, with no more pull-ups. That flight could have been AF447 - it had all prerequisites if we look on the surface - but it ended very differentyl. Why?

                          Comment


                          • #28
                            Black Ram, what you are posting is all very interesting.

                            I had not heard Sully talk about the role of Airbus' philosophy to FBW in his accident. I had heard him talk about the AF case.

                            Let me repost your link starting some seconds earlier than yours:


                            Apparently, he is "blaming" the Airbus FBW for limiting his ability to keep raising the nose, but the words that he uses are more or less these (paraphrasing):
                            - The FBW protections never kicked in, because we never got to the limits of the envelopes where the protections trigger.
                            - I was pulling back commanding more lift, there was more lift to give because we had not reached the alpha max yet, and the system prevented me from getting more lift. Because of that, the impact was with more vertical speed, there was more damage to the underside of the plane, more water ingressed sooner, and the flight atendant in the back was hurt.
                            - The reason the plane limited the nose-up command was not the protections (again, the limit had not been reached) but a phugoid mode that was known only to a few Airbus software engineers and no pilot or airline.

                            Now, as I said, I had never heard Sully say all this, and I have never heard of this "phugoid mode" in the Airbus flight control software (I do know what is the phugoid mode of longitudinal movement, and oscillatory motion where airspeed and vertical speed are exchanged one for the other in successive turns over and over, but that's something that is still there in a balsa plane or paper plane with fixed control surfaces (and hence no controls, let alone a control law or philosophy).

                            If he is right, then you are not right about Sully blaming airbus for keeping him from stalling, and we have something interesting and new to discuss.

                            I say this is a mission for Evan.

                            Regarding the A vs B approach to FBW and sticks/yokes, I still think that B is better for all the reasons discussed here (and mentioned by Sully regarding the AF case). When you have human pilots, feedback is very important for the humans to be integrated in "what is the automation doing now". and A's philosophy removes a very good bunch of feedback (especially intuitive one, as opposed to SMS messages sent to a TV screen) leaving the pilot more out of the loop.

                            You are right that stalls happened in Boeings with convetional controls, and also with FBW controls (the Aisana 777 landing in San Francisco). On the A side, AF was not the only case either. But the difference is that you need a good level of degradation to be able to even start a stall in an Airbus. So before talking about recovery, we should see how many stalls in an A didn't happen in the first place because the plane refused to follow the pilot commands into one. I do thing that A makes a better job on that side (for example, the Turkish 737 accident landing in Amsterdam would likely not have happened in an Airbus), as well as I do think that B makes a better job at keeping the pilots in the loop of what is the automation or the other pilot doing (again, in the Turkish accident, the automation actually pulled the thrust levers back and keep adding nose-up trim turning the very visible and noisy trim wheel; the fact that this was not enough for these pilots these days doesn't mean that it is not better than doing the same actions "stealthily").

                            --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
                            --- Defend what you say with arguments, not by imposing your credentials ---

                            Comment


                            • #29
                              Originally posted by Black Ram View Post
                              Of these, Pulkovo 612 is especially interesting. I looked into the accident again recently and I was shocked. I always thought the stall and flat spin were unrecoverable because of the airplane characteristics, and that the crew did everything they could once their initial mistake of entering deep stall. NOT! Their initial mistake was not a momentary lapse of airmanship, but a sustained series of dangerous pull-ups. With "critical AoA" (stall) alarms in the background. Five crew in the cockpit (a captain, a trainee with another F/O behind him, who was supposedly a former aerobatics champion, a navigator and a flight engineer). Two very traditional control columns, a functional stall alarm, and......an actual AoA indicator. Like the one that actually shows you what the exact AoA is, with an arrow, and with a very clearly marked red zone. And yet, they pushed the plane to climb, repeatedly hitting and entering that AoA red zone, they went into pitch-up several times (and when I say "they", I mean there was no argument or opposition - in fact, the navigator was a big proponent of the plan to climb over the storm). Stall alarms were going off, and the captain was basically saying "all is good, keep doing what you were doing". Then they entered deep stall and a flat spin - they knew they were descending like crazy, and they knew they were spinning in a circle (but apparently did not know why). And yet, not a single command or discussion to bring the nose down that I could find. "Pull up, keep the wings level" were the words of the captain. And that until they hit the ground from FL410.


                              Maybe a peculiar incident and not very interesting to everyone, but clearly showing that not only traditional control columns, a present captain from the very beginning of the events, extra pilots in the cockpit, but even an AoA indicator (discussed after AF447) is sometimes not enough.
                              That incident was (and is) very interesting to me, not least because the MAK report on it is surprisingly brief and shallow.

                              Comment


                              • #30
                                Originally posted by Gabriel View Post
                                Black Ram, what you are posting is all very interesting.

                                I had not heard Sully talk about the role of Airbus' philosophy to FBW in his accident. I had heard him talk about the AF case.

                                Let me repost your link starting some seconds earlier than yours:


                                Apparently, he is "blaming" the Airbus FBW for limiting his ability to keep raising the nose, but the words that he uses are more or less these (paraphrasing):
                                - The FBW protections never kicked in, because we never got to the limits of the envelopes where the protections trigger.
                                - I was pulling back commanding more lift, there was more lift to give because we had not reached the alpha max yet, and the system prevented me from getting more lift. Because of that, the impact was with more vertical speed, there was more damage to the underside of the plane, more water ingressed sooner, and the flight atendant in the back was hurt.
                                - The reason the plane limited the nose-up command was not the protections (again, the limit had not been reached) but a phugoid mode that was known only to a few Airbus software engineers and no pilot or airline.

                                Now, as I said, I had never heard Sully say all this, and I have never heard of this "phugoid mode" in the Airbus flight control software (I do know what is the phugoid mode of longitudinal movement, and oscillatory motion where airspeed and vertical speed are exchanged one for the other in successive turns over and over, but that's something that is still there in a balsa plane or paper plane with fixed control surfaces (and hence no controls, let alone a control law or philosophy).

                                If he is right, then you are not right about Sully blaming airbus for keeping him from stalling, and we have something interesting and new to discuss.

                                I say this is a mission for Evan.


                                Right, and we did discuss this with Evan back in I believe 2011, I think in the AF447 thread. I thought it was established that the plane was in alpha prot mode and that it didn't have sufficient speed to flare the way Sully wanted it to. Now, alpha prot is before alpha max.

                                I skimmed through the report (https://www.ntsb.gov/investigations/...ts/AAR1003.pdf) by searching for "prot", and I found out that alpha prot activates not just at a certain AoA, but also at a certain speed, and they were at that speed. Here's an interesting bit:


                                "2.7.2 High-AOA Envelope Limitations

                                The airplane’s airspeed in the last 150 feet of the descent was low enough to activate thealpha-protection mode of the airplane’s fly-by-wire envelope protection features. The captainprogressively pulled aft on the sidestick as the airplane descended below 100 feet, and he pulledthe sidestick to its aft stop in the last 50 feet, indicating that he was attempting to raise theairplane nose to flare and soften the touchdown on the water. The A320 alpha-protection modeincorporates features that can attenuate pilot sidestick pitch inputs. Because of these features, the airplane could not reach the maximum AOA attainable in pitch normal law for the airplaneweight and configuration; however, the airplane did provide maximum performance for theweight and configuration at that time.

                                The Airbus simulation indicated that the captain’s aft sidestick inputs in the last 50 feet ofthe flight were attenuated, limiting the ANU response of the airplane even though about 3.5° ofmargin existed between the airplane’s AOA at touchdown (between 13° and 14°) and themaximum AOA for this airplane weight and configuration (17.5°). Airbus’ training curriculadoes not contain information on the effects of alpha-protection mode features that might affectthe airplane’s response to pilot sidestick pitch inputs. The flight envelope protections allowed thecaptain to pull full aft on the sidestick without the risk of stalling the airplane."


                                "Deliberately or inadvertently slowing the airplane into the alpha-protection mode mayresult in an attenuation of pilot nose-up stick inputs, making it more difficult to flare theairplane, even if AOA margin to alpha maximum exists"



                                So Sully may have a point, but it doesn't seem to be some hidden ant-phugoid feature. The pane was in alpha prot. I'm not sure what "maximum performance" means, so it remains an open question whether the plane could have landed more nicely without the protections. I'm not the one who can say what would have been the difference if the plane landed with 17.5° AoA instead of 14°. But it seems those protections did work as designed, with no hidden surprises.



                                Originally posted by Gabriel View Post
                                Regarding the A vs B approach to FBW and sticks/yokes, I still think that B is better for all the reasons discussed here (and mentioned by Sully regarding the AF case). When you have human pilots, feedback is very important for the humans to be integrated in "what is the automation doing now". and A's philosophy removes a very good bunch of feedback (especially intuitive one, as opposed to SMS messages sent to a TV screen) leaving the pilot more out of the loop.

                                You are right that stalls happened in Boeings with convetional controls, and also with FBW controls (the Aisana 777 landing in San Francisco). On the A side, AF was not the only case either. But the difference is that you need a good level of degradation to be able to even start a stall in an Airbus. So before talking about recovery, we should see how many stalls in an A didn't happen in the first place because the plane refused to follow the pilot commands into one. I do thing that A makes a better job on that side (for example, the Turkish 737 accident landing in Amsterdam would likely not have happened in an Airbus), as well as I do think that B makes a better job at keeping the pilots in the loop of what is the automation or the other pilot doing (again, in the Turkish accident, the automation actually pulled the thrust levers back and keep adding nose-up trim turning the very visible and noisy trim wheel; the fact that this was not enough for these pilots these days doesn't mean that it is not better than doing the same actions "stealthily").


                                There were some reports on AVherald of AF buses going into Alpha Floor or Alpha Prot on approach. But there are cases when the crew did avoided a stall on their own, before any protections were engaged. Any many cases of UAS where no stall occurred.

                                As for feedback, I think procedures and basic airmanship are more important. ECAM messages can confuse, but the pilots have a display with the plane's attitude, speed, altitude/VS, and a separate display wit the engine parameters (so Turkish and Asian shouldn't have happened either). I think this is the most important feedback. So train pilots to use that, or if you're going to redesign something, redesign those basic flight parameter displays and engine parameter displays to account for human factors.

                                Sully kept saying the alpha protections were not engaged during his ditching. But they were, and it should have been visible on the speed tape (as well as the below optimum speed). Now, if a great pilot like Sully did not take notice because of the stressful situation (and if possibly he remembers things a bit differently from how they actually happened), what can we say about more average pilots? Humans are humans. Therefore we need to go more into human factors. Study them and account for them better. But not necessarily add physical feedback, even if it would be better than not having any. I think the basic flight parameter idnications are more important.

                                Comment

                                Working...
                                X