Announcement

Collapse
No announcement yet.

Virgin Australia B738, Jan 4th 2013 Incident Report

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

  • #31
    Originally posted by Gabriel View Post
    I don't understand what you mean. In these "pitch controls speed" modes speed is the king. This mode will provide better speed control and assurance than any other. It will go as far as descend when commanded to climb if that's what it takes to hold the airspeed. PROVIDED THAT YOU LET THE MODE FLY THE PLANE FOR GOD'S SAKE!
    As I said already, in FLCH, in descent, if you move the thrust levers to IDLE and hold them there for a couple seconds they will remain there until FLCH reaches its target altitude, and if that target is too low there is nothing to add thrust, there is no speed protection in these modes. The pitch will try to hold speed until the alpha protect overrides it. but nothing will add power IF YOU ARE NOT FLYING THE PLANE.

    Comment


    • #32
      Originally posted by Evan View Post
      As I said already, in FLCH, in descent, if you move the thrust levers to IDLE and hold them there for a couple seconds they will remain there until FLCH reaches its target altitude, and if that target is too low there is nothing to add thrust, there is no speed protection in these modes. The pitch will try to hold speed until the alpha protect overrides it. but nothing will add power IF YOU ARE NOT FLYING THE PLANE.
      Even, look what you have just written. It makes no sense!!!

      The pitch will try and WILL hold speed. There will be no alpha protect because the speed is kept, so the AoA doesn't increase. And that holds true whatever happens with the thrust.

      PITCH -NOT THROTTLE- CONTROLS SPEED IN THIS MODE.

      --- 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


      • #33
        Originally posted by Gabriel View Post
        Even, look what you have just written. It makes no sense!!!

        The pitch will try and WILL hold speed. There will be no alpha protect because the speed is kept, so the AoA doesn't increase. And that holds true whatever happens with the thrust.

        PITCH -NOT THROTTLE- CONTROLS SPEED IN THIS MODE.
        Yes, I get that, if you REMAIN active in FLCH with a selected altitude of zero it will simply fly you into the ground at the selected airspeed. But if you transition to MANUAL flight before then, and pull back relentlessly, thereby reducing airspeed, you still have pitch protection for AoA but no AT 'wake-up' protection for thrust (which I referred to as 'speed protection'). That is how I understand the FLCH trap anyway.

        Comment


        • #34
          Originally posted by Evan View Post
          Yes, I get that, if you REMAIN active in FLCH with a selected altitude of zero it will simply fly you into the ground at the selected airspeed. But if you transition to MANUAL flight before then, and pull back relentlessly, thereby reducing airspeed, you still have pitch protection for AoA but no AT 'wake-up' protection for thrust (which I referred to as 'speed protection'). That is how I understand the FLCH trap anyway.
          Okay, I think you are mixing a couple of things here.
          I'm not sure, because I'm not familiar with the 777's systems, but as far as I understand (not much, so double check) things go more or less like this:

          First of all, and I can't believe we are discussing this, of course, if you say "Otto, you've got the speed", and then kill Otto, of course Otto is not controlling the speed anymore. That is, in no way, a "stealth feature" of the FLCH mode. "Otto" could be the real name of your copilot and this fact will still remain true.
          Other examples: If you tell the AT to keep 250kts, and then kill the AT and manipulate the thrust levers, then the AT will not keep 250kts. If you select the AP in ILS, but disconnect the AP and manually control pitch, then the AP will not follow the glide slope.
          Once you switch an automation off, you lose the right to blame that automation for failing to do anything.

          Point two: What you are talking about is not about autopilot modes, but envelope protections, and in principle they don't have much to do with the automation mode.
          The reason why the thrust didn't automatically increase when reaching a dangerously slow airspeed was not that they had selected a too low altitude, but that they WERE at a too low altitude. The speed protection kicks in increasing thrust when the speed becomes dangerously slow in any AP mode, PROVIDED THAT YOU ARE HIGHER THAN 100FT RA. Because if you are lower, the system assumes that you are in a flare (imagine being in the middle of the flare unable to bleed-of speed as you float along the runway. Now that would be some stealth mode).
          If they had run out of speed at a higher altitude, the speed envelope protection would have kicked in regardless of the selected altitude. They were unlucky to run out of speed when they were already so low.

          Point 3: Even if the pilots thought that the speed protection would be active all the way down to the runway, what? Did they rely on an envelope protection, one that will not hold the selected speed but merely and barely prevent a stall? That's crazy!!!
          Any yet, they had speed protection in another shape. You understand very well how speed and AoA are closely related. Keep the AoA below a certain threshold and, no matter what, the speed will not go below certain threshold (except transiently, if you transiently fly at less than 1G which was not the case here).
          And, you also know, in all "normal" and Boeing FBW planes, how the elevator inputs relates with these two parameters. Basically, the deflection on the yoke (and force too, at one given speed) is proportional to the deflection of the elevator, which is proportional to the change in AoA from the "trim" point.
          In this case, the pilot didn't touch the trim, so all the speed reduction from "way too fast" to "way too slow" was achieved by pulling back on the yoke, a constant increasing in deflection and force.
          Not only that, but when approaching the stick shaker and after that, the force needed to keep increasing the AoA and reducing the speed becomes disproportionately high (that's Boeing FBW approach to envelope protection that, unlike the Airbus ones are "soft" and overriddable).
          So yes, there was speed envelope protection, which was overridden by the pilot.

          Finally, I don't know what would happen if you select say 500ft, set FLCH in the AP, disconnect the AP but leave the FD on (the FD is the one that really controls the AP modes, then the AP, if on, just follows the FD command bars) and then manually fly and descend through 500ft. At least in the MD80, when approaching the selected altitude the vertical mode would revert to VS (not very important since it's off) and the AT would revert to "speed" (what you'd call the "wake-up"). So this could have saved the day, but it'd be a strange practice even as a "smart-ass trick".
          But even then, why would any pilot expect this feature to kick in higher if you selected it to happen at zero ft? Would you call it "a stealth feature" that the AT doesn't wake up until the time that you set it to wake up?

          I must be missing something here because other pilots too talk about the "stealth feature" of the FLCH mode and I fail to see what it is.

          --- 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


          • #35
            Originally posted by Gabriel View Post
            First of all, and I can't believe we are discussing this, of course, if you say "Otto, you've got the speed", and then kill Otto, of course Otto is not controlling the speed anymore. That is, in no way, a "stealth feature" of the FLCH mode.
            What we are discussing here is that much debated regime of blended automation and manual flight. MCM is accustomed to Otto or no Otto, but the SOP these days is to use AT from liftoff to touchdown regardless of whether you or Otto is on pitch and roll. If you fail to understand the interactions, you can find yourself in this stealthy middle ground.

            Okay, I think you are mixing a couple of things here.
            I'm not sure, because I'm not familiar with the 777's systems...
            Allow me to shed some light there. On the 777, FLCH is a pitch mode (on the 767 it is an AT mode). When you are in FLCH and the thrust levers reach IDLE and stay there for a moment, the AT mode changes from IDLE to HOLD (THR HLD on the 767 I think). The AT clutches are inhibited; the AT WILL NOT reposition the levers in HOLD mode, and most importantly, the AT WILL NOT activate to provide speed protection in this mode at any altitude. Speed protection is only available in the SPD or THR REF AT modes. In HOLD, in terms of speed protection, it is the same if you are under 100' or at 1000'. There is none. (There is still AoA pitch protection).

            If they had run out of speed at a higher altitude, the speed envelope protection would have kicked in regardless of the selected altitude. They were unlucky to run out of speed when they were already so low.
            Therefore, This is false. Once they went to HOLD mode on AT, they lost speed protection.

            Point 3: Even if the pilots thought that the speed protection would be active all the way down to the runway, what? Did they rely on an envelope protection, one that will not hold the selected speed but merely and barely prevent a stall? That's crazy!!!
            It wasn't that they were relying on speed protection. They were relying on the AT being active and would 'wake-up' when more thrust was needed to maintain the optimum airspeed.

            I suspect that this is a result of many successful applications of FLCH as a 'save' when high on the GS in a PRECISION approach, where the AP mode is going to transition to GS upon intercepting the glideslope and the AT is going to 'wake-up' in SPD or THR mode. Or possibly even on a non-precision approach with VNAV PTH armed where the mode is going to transition upon crossing the VNAV descent path. In the case of Asiana, there was no G/S and VNAV was not armed, so there was nothing to 'wake-up' the AT and no speed protection.

            Finally, I don't know what would happen if you select say 500ft, set FLCH in the AP, disconnect the AP but leave the FD on (the FD is the one that really controls the AP modes, then the AP, if on, just follows the FD command bars) and then manually fly and descend through 500ft. At least in the MD80, when approaching the selected altitude the vertical mode would revert to VS (not very important since it's off) and the AT would revert to "speed" (what you'd call the "wake-up"). So this could have saved the day, but it'd be a strange practice even as a "smart-ass trick".
            Exactly! Which is why they would set it at 0'. Setting it at 0 is essentially turning off that feature. Probably the original intention was to drop into the glideslope or descent path at whatever altitude that might occur and transition to the active AT mode. Then, as MCM says, things got pear shaped and they reverted to manual flight, either unaware or forgetting that the AT was in HOLD.

            I must be missing something here because other pilots too talk about the "stealth feature" of the FLCH mode and I fail to see what it is.
            Do you see it now? Because, when used correctly (and hopefully not on final) with a sensible altitude setting and subsequent modes armed, the AT will 'wake-up'... this creates the impression—in the mind of a pilot too ignorant to understand the aircraft he is entrusted with—that there is always this AP-AT interaction. And then there is a scenario where there is no AT interaction at all. And in an age of blended automation, this is all the more unexpected.

            Again, not the automation at fault. A failure to train pilots on the automation and system interactions and a failure to impress upon pilots the importance of adherence to standard procedures rather than homegrown workarounds.

            Or, to put it in a Gabrielesque equation... FLCH + IDLE x 1.2 secs * (ARMED MODE x 0) * (GLIDESLOPE x 0) * (LVL OFF ALT=0) * (SPEED PROTECTION x 0) * (INATTENTION TO BASIC FLIGHT PARAMETERS + FMA) = DEATH x 3 + B777 / 2

            (I flunked algebra).

            Comment


            • #36
              Originally posted by Evan View Post
              Do you see it now? Because, when used correctly (and hopefully not on final) with a sensible altitude setting and subsequent modes armed, the AT will 'wake-up'... this creates the impression—in the mind of a pilot too ignorant to understand the aircraft he is entrusted with—that there is always this AP-AT interaction. And then there is a scenario where there is no AT interaction at all. And in an age of blended automation, this is all the more unexpected.

              Again, not the automation at fault. A failure to train pilots on the automation and system interactions and a failure to impress upon pilots the importance of adherence to standard procedures rather than homegrown workarounds.

              Or, to put it in a Gabrielesque equation... FLCH + IDLE x 1.2 secs * (ARMED MODE x 0) * (GLIDESLOPE x 0) * (LVL OFF ALT=0) * (SPEED PROTECTION x 0) * (INATTENTION TO BASIC FLIGHT PARAMETERS + FMA) = DEATH x 3 + B777 / 2

              (I flunked algebra).
              Yes, I see it now:

              1- They tell the AT to forget the speed (by now at least) and the AP "speed is yours now, adjust pitch to hold it". (i.e. they select FLCH).
              2- They turn off the AP, which is the one they told to hold the speed, and take manual control of the pitch.
              3- They set a clock-alarm to wake up the AT (so it goes back to controlling the speed). The wake-up time they set is zero ft.
              4- They expect the AT to, somehow magically, control the speed after it was sent to sleep and before it was time to wake up.

              And that's a stealth feature of the FLCH mode???? I'd say it's a stealth failure of the human brain (and I'm not starting to mention their inability to control the most basic parameter in a visual approach and the commendable effort made by the flying pilot to keep pulling increasingly up with increasing force along about a couple of minutes).

              By the way, and just as side questions:
              The FLCH mode is a pitch mode in the 767????
              What would have happened if they had not kept idle for as long as 1.2 seconds?

              --- 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


              • #37
                FLCH is still a 'speed on elevator' mode on the 767. Its just that the system has options for thrust that is not N1 or IDLE for SMALL level changes. It still goes to THR HLD once it is at the idle stop. Same for the 777.

                There is no 'wake up' because there doesn't need to be. Thats the very heart of the mode. When the aircraft reaches the assigned level, it goes into a different mode - an altitude capture mode (ALT CAP on the 767).

                This is a low/medium level of automation mode - it does what it says on the tin. It climbs or descends to the level in the altitude window at the speed you have set in the speed window. It will adjust the climb/descent rate to achieve what you have asked for.

                Why would the A/T ever 'wake up' on a descent? The mode is doing EXACTLY what it is designed to do, and what you have asked it to do. There is nothing 'stealthy' in that.

                The reason the A/T works on the 777 to touch down is because it is flown in modes that are designed to achieve that. FLCH is most definitely not a mode for approach (Boeing say so), and so you can't expect it to do something it wasn't designed for (nor at its very heart appropriate for). How can a mode such as this EVER decide when it should ditch its actual primary function in life, and start doing something else strange?

                Nothing stealthy about it - anyone in the right mind knows that FLCH won't magically control the thrust on final. Its a fundamental core of its existence. The Asiana crew were just in a state of mind/overload/unfamiliarity that they didn't even realise it was in FLCH and what the consequences were. They panicked, and reverted to what they were used to - expecting the G/S mode to look after them.

                Comment


                • #38
                  Great answer MCM.

                  Regarding this:
                  expecting the G/S mode to look after them
                  Wich of course makes no sense. G/S mode a visual approach without a G/S? Not until we put cameras for the AP to look at the PAPI or VASI.

                  I'm not saying that what you say makes no sense. I'm saying that what these pilots did and expected makes no sense, whatever they did and expected. Be it that the AP (that was turned off) would capture an nonexistent G/S, that the AT would magically keep the speed, or whatever.

                  --- 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


                  • #39
                    Originally posted by MCM View Post
                    Nothing stealthy about it - anyone in the right mind knows that FLCH won't magically control the thrust on final. Its a fundamental core of its existence. The Asiana crew were just in a state of mind/overload/unfamiliarity that they didn't even realise it was in FLCH and what the consequences were. They panicked, and reverted to what they were used to - expecting the G/S mode to look after them.
                    Com'on guys, I feel like I'm talking into my hat here.

                    Yes, there is nothing stealthy about the system when used as designed by a crew that fully understands it. Yes. A thousand times yes. Indeed and concur.

                    The stealth factor in descent I refer to is found in human nature. A crew (or an entire pilot culture perhaps) that takes to using FLCH for something it is not intended to do, specifically because it works very well in fact, seemingly every time, without fully inderstanding the far-reaching implications of doing so, will be caught off gaurd when, in that rare instance where circumstances converge against them (G/S n/a, approach not well planned and briefed, proper modes not armed, get-there-itis), the automation fails to do what it did all those other times.

                    I don't believe this was simply a case of mind/overload/unfamiliarity and panic. I don't believe they didn't realize it was in FLCH. You have to push the FLCH button to get it there. I believe they were following an established practice created by pilots as opposed to procedural engineers. I believe that they didn't understand fully the interaction between the AP mode FLCH and the autothrottle mode HOLD. (Again, on the 767 FLCH is an autothrottle mode, so perhaps it's more apparent).

                    The stealth factor in climb I refer to is also found in human nature. Since this mode is not supposed to require monitoring to stay on speed, I think occasionally pilots might be more inattentive while using it. I also wonder if a crew that does not foresee the possibility of the scenario at the top of this thread might intentionally choose to remain in FLCH longer than it is intended for, because it's dinner time. And there it is, going from kts to mach, a hidden pitfall...

                    For the nth time. I see nothing wrong with the aircraft systems. I see a weakness in training and discipline focusing on certain stealth factors of automation. The reason I see this is because it keeps happening.

                    Originally posted by Gabriel
                    Wich of course makes no sense. G/S mode a visual approach without a G/S? Not until we put cameras for the AP to look at the PAPI or VASI.
                    Gabriel, I think what MCM is getting at is that they were used to doing this on a precision approach and were mentally conditioned to expect the AT to perform in FLCH as it does when intercepting the G/S. No, it doesn't make sense in a rational state of mind but under stress and confusion and the force of a bad habit it just might seem logical.

                    Comment


                    • #40
                      Originally posted by Evan View Post
                      Com'on guys, I feel like I'm talking into my hat here.
                      Yes, there is nothing stealthy about the system when used as designed by a crew that fully understands it. Yes. A thousand times yes. Indeed and concur. The stealth factor in descent I refer to is found in human nature.
                      Yes, yes. Roger that Evan, but there are a couple things that I don't get.

                      - That they expected the AT to "wake up" and "keep the speed" as it "always do".
                      - That they set zero ft.
                      - That FLCH is "tricky" in climbs.

                      Let me explain:

                      - You say "they were used to doing this on a precision approach and were mentally conditioned to expect the AT to perform in FLCH as it does when intercepting the G/S".
                      And they were used to FLCH intercepting the GS with the AP turned off too? Ok, leave that at a side.
                      How is that they expected FLCH to intercept an nonexistent GS?
                      They briefed for a visual approach. The flying pilot said he was nervous because he had to hand-fly the visual approach and that he considered it difficult and dangerous (that alone disqualifies him as a pilot, but oh well).
                      You say "under stress and confusion and the force of a bad habit it just might seem logical". I don't buy it. I'm not buying that they expected that the AT would "wake up and keep the speed" when the "FLCH intercepted some eerie GS". I think that either they were unaware that they were in FLCH (because they selected it by mistake or because they forgot that they had set it) or they they selected a mode (FLCH) of which they had no clue whatsoever of how it works. And I'm not talking of the nuances of the system. I'm talking of the core of it: Pitch controls speed.

                      - Even then, why set zero feet? Even trying to be smart-stupid, I can't think of a good reason. The only "trick" that I can think of to set zero ft is if I wanted to fly all the way to the ground either in VS (using the VS knob to adjust the descent along the way while the AT keeps the speed) or FLCH (using the throttles to adjust the descent along the way while the pitch mode keeps the speed). If I expect the system to "wake up" in another mode before zero ft, why set zero ft?

                      - "I also wonder if a crew that does not foresee the possibility of the scenario at the top of this thread might intentionally choose to remain in FLCH longer than it is intended for, because it's dinner time."
                      Longer than intended? This is the mode to use for long climbs and descents if, for any reason, you are not using VNAV. You can use it all the way from just after lift-off to TOC.
                      Yes, you have to adjust the climb speed when crossing the procedural speed restriction limit of 10000ft, but then again this is the limit of the sterile cockpit policy where activities are restricted only to those necessary for the safe operation of the flight. Dinner doesn't qualify. So 10000ft, adjust speed, bring the dinner.
                      "And there it is, going from kts to mach, a hidden pitfall...". The other option, that it doesn't automatically go from IAS to mach, would cause that an inattentive crew that fails to make that change manually would find themselves exceeding Mmo. So any option would be a "hidden pitfall". Note that if you did the adjustment at 10000ft, then the automatic changeover will set a mach that could be a bit off the optimum one, but will be very well within the operating envelope.

                      --- 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


                      • #41
                        Originally posted by Gabriel View Post
                        Yes, yes. Roger that Evan, but there are a couple things that I don't get.
                        - That they expected the AT to "wake up" and "keep the speed" as it "always do".
                        From not understanding that FLCH leaves the AT in HOLD upon AP disconnect rather than transitioning to SPD or THR mode as it does when intercepting the glideslope on a precision approach. They might have thought that when the mode is cancelled, so is the AT HOLD. And keep in mind that they rarely fly non-precision approaches.

                        - That they set zero ft.
                        That is speculation but comes from various accounts I've read in the legitimate aviation media about the 'FLCH trap'. The technique involves setting either 0 or an altitude far lower than the one where you expect the mode transition to occur. 0 seems to be the easiest way to do this.
                        - That FLCH is "tricky" in climbs.
                        No, just the opposite. It seems to me that FLCH is the easiest in climbs and therefore might breed inattentiveness in certain pilots IF they fail to understand the SPD/MACH issue associated with FLCH.

                        And they were used to FLCH intercepting the GS with the AP turned off too?
                        How is that they expected FLCH to intercept an nonexistent GS?
                        So, as I said above, it is the same autothrottle behavior that they probably expected due to conditioning when using this technique in precision approaches. Remember the issue I'm addressing in blended automation. I don't think they expected to intercept a nonexistent glideslope. They just assumed the AT would behave the same way when taking manual control of the glidepath.

                        Longer than intended? This is the mode to use for long climbs and descents if, for any reason, you are not using VNAV. You can use it all the way from just after lift-off to TOC.
                        From the report:
                        Originally posted by ATSB Report
                        The ATSB reported the first officer, pilot flying, had selected LVL CHG mode while climbing through 3900 feet at a speed of 250 KIAS with the intention to later switch to VNAV. Both crew subsequently overlooked that the autopilot was still in LVL CHG mode at 250 KIAS.
                        "And there it is, going from kts to mach, a hidden pitfall...". The other option, that it doesn't automatically go from IAS to mach, would cause that an inattentive crew that fails to make that change manually would find themselves exceeding Mmo. So any option would be a "hidden pitfall".
                        Agreed. Either way a potential pitfall that is somewhat stealth and must be clearly understood in pilot training. That's my concern.

                        Comment


                        • #42
                          Much like the flight that landed into the mountain on landing last month, I dont get why its so complicated. Nor why all of a sudden pilots need to start flipping AP and AT modes back and forth while ignoring their heading or elevation. Just fly the tube.

                          Foolishness. FLCH this .

                          Comment

                          Working...
                          X