Announcement

Collapse
No announcement yet.

Virgin Australia B738, Jan 4th 2013 Incident Report

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

  • Virgin Australia B738, Jan 4th 2013 Incident Report

    This was on AV Herald today:

    Originally posted by AV Herald
    On May 26th 2014 the ATSB released their final report without a formal conclusion but a safety message:

    This occurrence highlights the importance of consistent attention to auto-flight system modes and aircraft energy state. Operation of the auto-flight system in an unintended or inappropriate mode can lead to an undesirable energy state, or in other cases, unintended operational or procedural non-compliance.

    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. When the aircraft climbed through FL260 the autopilot consequently switched to Mach 0.62. While approaching FL350 about 20 minutes into the flight the first officer noticed a "buffet alert" caution in the scratch pad of the CDU, the captain at the same time noticed the autopilot made a small not noticeable reduction of attitude. Both pilots detected that the indicated airspeed was just on top of the lower amber speed band maintaining minimum maneouvering speed. The crew reduced the pitch attitude further permitting the aircraft to descend until the speed had increased sufficiently that the aircraft was able to continue a safe climb to the intended flight level.

    The ATSB reported that usually VNAV would have been selected soon after takeoff, the first officer could not recall why LVL CHG had been selected, the captain reported they were expediting the climb probably through a cloud layer and/or turbulence.

    According to the recorded data the aircraft reached a lowest speed of 201 KIAS, 6 knots below minimum maneouvering speed.
    I really get the feeling this A/P mode is not well explained to pilots.

  • #2
    Actually, in this particular case, I think that the problem was not a lack of understanding but lack of attention. And, a feature of the automation that I thought was taken into account in the design but apparently it isn't.

    When not using VNAV, LVL CHG is the standard vertical mode for a climb: You select the climb airspeed, select the climb thrust, and the AP will adjust the pitch to keep the selected airspeed, whatever the resultant climb rate is.

    Compare this to selecting vertical speed for the vertical mode and airspeed for the autothrottle. As the plane climbs and the engine looses power, the autothrust will first increase the thrust beyond the desired climb thrust to keep the speed. When it has advanced the thrust all that it can, the speed will start to decay. By doing so the plane will tend to reduce the climb rate below the set value, so the AP will pitch up to keep it at the desired point, what will only make the speed decay faster. We know how these things end.

    So, the LVL CHG mode is designed specifically for these climbs or descents where you want to hold a given speed and an optimum power setting and you don't care much about the precision of the resulting vertical profile, and a climb to cruise altitude fits perfectly in this category.

    The problem here was, again, that nobody was flying the plane.
    They selected 250 kts in the vertical mode. That's the speed to hold below 10,000ft, not because it's the best speed for the plane, but because that's the "speed limit" below 10,000ft for all airplanes (except those where 250kts in clean config would be below acceptable). When crossing 10,000ft, you have to select the optimum climb speed, which typically is about 280kts. When you do that in LVL CHG, the plane will reduce the pitch, accelerate to 280, and then pitch up again as needed to hold 280. The engines are left at climb thrust.

    But, alas, these pilots forgot to do it. Not just for a few seconds. They got to FL260 in that condition. It must have taken some good 5 minutes to go from FL100 to FL260. So my question is not just how they forgot to adjust the speed for so long, but rather, didn't it call their attention that they were above FL100 and the speed was 250? Or (more likely I'm afraid) they didn't monitor the speed for 5 minutes? (you know, Otto is taking care of it, so why bother?).

    At FL260, the autopilot automatically changed the vertical mode from "speed" to "Mach". This is normal and standard, because as the plane climbs at constant airspeed, the true airspeed is ever faster, and the speed of sound is ever lower (because the temperature descends as you climb and the speed of sound is a function of the square root of the temperature), and the result of a higher TAS and a lower speed of sound is an ever increasing Mach number, and if the Mach number becomes too high the drag starts to increase and the climb performance worsens. So a typical climb profile can be "hold 280 KIAS nad let the Mach number increase, as the Mach number reaches M 0.72, hold M 0.72 and let the airspeed decay". The altitude at which the "hold KIAS" and the "hold mach" figures coincide is called changeover. That altitude is not exact, it just happens where it happens, but it happens in a quite narrow margin of a few thousand feet. So the LVL CHG mode of the AP is designed to automatically switch from "hold IAS" to "hold Mach" when crossing through what we could call a fixed "average changeover altitude". The system will alert the pilot when doing so (I don't remember exactly how, maybe a chime and and the word MCH replacing IAS in the FMA and blinking for some seconds). When doing so, the selected Mach number to hold will be the Mach number that the plane was doing at that time. This might be not exactly (but very close to) the Mach number that the climb profile said to hold, so if there is any slight difference the pilot would make the required minor adjustment in the setting.

    The problem here is that, because they were holding 250 instead of something like 280, when the changeover happened, the resulting Mach happened to be 0.62 instead of something like 0.72.

    So, when the changeover happened, and they AP alerted the pilots of that, did't the pilots alert? Didn't they notice that the new selected Mach was way below what it was supposed to be? No, they didn't.

    So... the pilotless plane kept climbing holding 0.62 as designed. And the IAS started to decay as expected. And so things went until they reached FL350, which is maybe another good 5 minutes. Again, didn't the M 0.62 call their attention? Didn't an airspeed that was approaching 200kts and the amber zone call their attention? DIDN'T EITHER OF THEM WATCH THE AIRSPEED INDICATOR EVEN ONCE IN THE LAST 10 MINUTES, the approximate time that it must have taken to go from FL100 to FL350? Did they need to get a "buffet alert" and the AP lowering the nose overriding the selected airspeed to prevent a stall to, as the report says, "direct their attention to the primary flight display" which has the airspeed, Mach number, and flight mode data? (and which, as I always say, have the words PRIMARY and FLIGHT in its name for a reason).

    Yes, I know, they were having a meal, and pilots NEED meals not just to live and satisfy their appetite, but also to perform well in the cockpit. But isnt' there a meals SOP that ensures that at least one pilot is flying the plane at all times? It looks that, in the case, the pilots could have had the meal in a business class seat, because no one was flying the plane anyway.

    All all this pilot-bashing done, I fully subscribe to this comment by the ATSB:
    The ATSB commends the flight crew and operator for submitting a report in relation to this occurrence, noting that the occurrence was not reportable under the definitions provided in Transport Safety Investigation Regulations 2003. The report was submitted in the interest of aviation safety and to provide an opportunity for others to learn from the occurrence.
    Basically, they could have kept silent and get away with it, but they chose to expose their stupidity so others would learn what they learnt the hard way.

    And finally, what the pilots said:

    The crew commented that the effectiveness of their instrument scan and auto-flight system mode awareness was probably compromised to some degree by the distractions sometimes associated with relatively short sectors. While the aircraft was climbing, each pilot consumed a meal before commencing preparations for their arrival into Melbourne. While the crew were broadly scanning aircraft instruments throughout the climb, nothing specifically drew their attention to the unintended auto-flight climb mode or the gradually reducing airspeed (beyond the changeover altitude). The crew commented that the occurrence provided a salient reminder regarding the importance of maintaining auto-flight system mode awareness.
    Hmmm... What about an airspeed indicator, no, wait, three of them? What is "broadly scanning aircraft instruments throughout the climb" if it doesn't include even one look at the airspeed indicator by either the flying pilot or the monitoring pilot in 10 minutes?

    ATSB report here:

    --- 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


    • #3
      Evan, this wasn't poor mode knowledge, simply not realising that it was the mode they were still using.

      Generally a good summation Gabriel. A couple of points:

      There is no notification to the pilot that this changeover has happened other than the number in the speed window on the MCP (which is not something you would necessarily be looking at) changes (well not on the 767 or 747 anyway, so I imagine it is the same for the 737).

      Just because you look at the airspeed indicator doesn't mean you would necessarily notice anything was wrong. Remember VNAV handles the speed bug most of the time, and these guys do a LOT of sectors at varying weights where you expect to see different things. As long as the speed bug was within the manoeuvre margins, you may not necessarily twig that the speed was slower than normal, particularly above about 20,000ft (as it may not really have been that much slower than normal). So, its only really in the last little while when the speed got very close to the amber band that you would have a huge cue that something wasn't right. You say your optimum climb speed might be 280kts, but it isn't that fixed. It changes a fair bit flight to flight - on the 767 by up to about 60kts.

      Yes, absolute 100% paranoid attention should make you notice this - but it doesn't work that way in real life.

      Let the Boeing pilot who has never forgotten that he was in a LVL CHG mode and not VNAV on a climb cast the first stone.

      These guys made an error, and then corrected it when realised. They didn't stall the aeroplane, nor go into stall warning territory. They corrected the flight path as soon as the aircraft got to the manoeuvre margin (yes, the aircraft assisted). They were not completely absent, and reacted appropriately. So its not all bad.

      Comment


      • #4
        MCM,
        Some things that you are saying are really unexpected to me.

        A speed of 250kts at FL130, 180, 220 would't call your attention? Or a Mach number of 0.62 at FL300? Is the target climb speed ever as low as 250 or even 260 kts above FL100 and below the changeover?

        What you say about the bug is interesting. The "actual" speed is no very obvious indication in modern displays. Just the relative speed regarding some references. At a glance you can see if the speed is near the speed bug or near an amber zone, red zone, etc. But you have to actually read the number to know the speed. Maybe this event would not have happened with the old steam gauges where, with a glance, you can see where the needle is (even without paying attention to the exact value). Maybe the pilot would have said "hey, this is not where the needle tends to be in the climbs, let's double check everything".

        Finally:
        "Let the Boeing pilot who has never forgotten that he was in a LVL CHG mode and not VNAV on a climb cast the first stone."

        The problem is not so much to be climbing in LVL CHG and not in VNAV. In fact, this was the standard case when the planes didn't have VNAV.

        The problem is to keep climbing at 250 above 10000ft, something that can happen even if you are intentionally climbing at LVL CHG and forget to adjust the speed at 10000ft.

        So, given all the mitigating factors that you mention, what is or should be the line of defense against this? Sure that it's easy to forget that you had set LVL CHG and keep this mode unintentionally on. And sure that if the AP gives no warning that its changing over from IAS to Mach, it will be hard to catch it. If you say that the speeds they encountered were with the realm of a normal climb all the way from FL100 to FL350 until just before the amber zone, then the speed or Mach would not be a hint either. The MCP speed is not something that one monitors. The FMA showing the wrong mode..., well, that one should have been caught at least. The FMA should be one of the things most frequently monitored.

        But what else? Don't the pilots know what speed they SHOULD be flying as to be able to crosscheck with the ASI? What's the point of monitoring the instruments if you don't know what they SHOULD be showing? It shows 210? Ok. It shows 320? Ok too. Both values are possible and within the flight envelope. That's it?

        I don't feel comfortable with "Otto will provide". Both in terms of "he knows what speed to fly" as in "he will alert us or even save us if the speed approaches an undesirable limit". Note that these pilot only noticed that they were approaching the minimum maneuvering speed (and actually went 6kts below that) when they got a buffet warning and the AP started to lower the nose to accelerate above the speed that he was supposed to hold (which was too slow). Is that acceptable?

        I have a few proposals to begin with:

        - That the FMA shows not only the mode but the selected value that it's tracking, in two lines (nothing new, it's what the MD-80 shows). For example, the vertical mode in LVL CHG would read:

        LVL CHG
        IAS 280

        LVL CHG
        M 0.75

        Whenever there is a change, there is a chime and the new thing blinks for 10 seconds.

        - The vertical mode or the autothrottle (depending which one is controlling the speed, according to the mode), regardless of whether it's tracking a IAS or a Mach, doesn't accept neither IAS nor Mach values that are beyond a "normal" limit (minimum maneuver speed, max flap or gear speed, Vmo, Mmo). For example, if the selected Mach would result in a IAS below minimum maneuver, the AP or AT will track the minimum maneuver (assuming that this is the first limit that you'll find, if the buffet warning comes first for the present condition, then it would track the buffet warning speed). The MCP speed would keep showing the originally selected value, but in the second line the FMA would show a yellow blinking message like LOW LIM or something like that, indicating that the AP or AT will not track the MCP value because it's below the low limit.

        - Crossing 10000, cruise climb check:
        FMA ==> Desired mode
        Speed/Mach ==> Check that matches the target (of course you need to know the target)
        Engines ==> Green arc
        Sterle cockpit ==> Over
        Meal ==> Eat, one pilot at a time, the other pilot assumes PF role.

        If all of the above doesn't work:
        VNAV ==> Eliminate.
        LVL CHG ==> Eliminate.
        Autothrottle ==> Eliminate.
        This will force the pilots to pay attention to the flight parameters and not rely on the automation (because there would be barely any automation left).
        Manually set climb thrust, climb in V/S mode and adjust the selected V/S every minute or so to keep the desired IAS/Mach. Like in the ole times.

        I'm not serious about this last point, but come on guys. Too much workload and you can't track all parameters because of task saturation. Too little workload and you can't track all parameters due to boredom and complacency. There's no pleasing you!!!

        --- 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


        • #5
          Thanks, all very informative.

          I was getting at something else though, not that LVL CH is the wrong mode (at least to begin with) but that I wonder if pilots truly understand the stealth implications of this AP mode? This is just based on a number of recent incidents where LVL CH has caught up with the pilots not flying.

          In other words, I would like to know that pilots are using this mode cautiously and are aware of the danger involved and the absolute need to monitor the airspeed rather than the ravioli and change modes at the correct stage.

          But you have to actually read the number to know the speed.
          Wow, how hard is that?

          Comment


          • #6
            Originally posted by Evan View Post
            Thanks, all very informative.

            I was getting at something else though, not that LVL CH is the wrong mode (at least to begin with) but that I wonder if pilots truly understand the stealth implications of this AP mode? This is just based on a number of recent incidents where LVL CH has caught up with the pilots not flying.

            In other words, I would like to know that pilots are using this mode cautiously and are aware of the danger involved and the absolute need to monitor the airspeed rather than the ravioli and change modes at the correct stage.
            Well, what you said applies not to LVL CHG, but any "classic" flight mode (that is, everything but full AP/AT VNAV, where the FMS manages all the parameters at the same time).

            In fact, the LVL CHG mode is, in my opinion, the least steal mode because it has speed as its priority. A climb in vertical speed, or even an altitude hold on the wrong side of the power curve, and even a GS mode (glide slope) are much more dangerous because the autopilot doesn't manage the speed and doesn't interact with the autothrottle, so it will gladly stall the airplane (except if there are envelope protections) in it's attempt to hold the required vertical profile with insufficient power to do it.

            This incident had nothing to do with stealth characteristics that are specific of the LVL CHG vertical mode.

            Wow, how hard is that?
            Hard enough that it requires conscious and intentional action to read the number, which is not hard really. But with the steam gauges, you can be looking at the panel in general and suddenly and unconsciously something calls your attention to the ASI. You don't know what it is but there is something strange in the picture. When you turn your attention to the ASI, you see that the needle is not where it typically is in these climbs. Without knowing it, your peripheral vision had detected that and a red light went off in your brain. Now that you diverted the attention to the instrument and identified that the needle is sort of odd, you read the number.

            Let me give you an example. Say that by mistake you select 230kts instead of 320kts. In the "vertical tape" indicator (like the ones found in the EFIS PFDs) the picture doesn't look too different. There is a speed bug and, if the actual speed is close to the speed bug, the speed bug will be close to the speed index, in the middle of the tape. In a round display, it will be very obvious with an unintentional split-second glance that neither the speed bug nor the actual speed are anywhere close to where the are supposed to be.

            --- 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


            • #7
              Originally posted by Gabriel View Post

              In fact, the LVL CHG mode is, in my opinion, the least steal mode because it has speed as its priority. A climb in vertical speed, or even an altitude hold on the wrong side of the power curve, and even a GS mode (glide slope) are much more dangerous because the autopilot doesn't manage the speed and doesn't interact with the autothrottle...
              Wait a sec, as I understand it using LVL CH in climb requires the pilot to set the optimum power setting manually, then the AP maintains the MCP target speed with pitch using that power setting. The A/T is not active and there is no A/T 'wake-up' protection (there is a still pitch-reduction/pitch trim protection). Am I wrong about that? Where is the autothrottle interaction?

              Comment


              • #8
                Originally posted by Evan View Post
                Wait a sec, as I understand it using LVL CH in climb requires the pilot to set the optimum power setting manually, then the AP maintains the MCP target speed with pitch using that power setting. The A/T is not active and there is no A/T 'wake-up' protection (there is a still pitch-reduction/pitch trim protection). Am I wrong about that? Where is the autothrottle interaction?
                No. You can set the AT to keep one N1 or EPR (depending on the engine).
                I don't know in the 737, but in the MD80 you can set the AT to keep an EPR defined by the TRC (thrust rate computer), so you select "CLB" in the TRC and a window will display the climb EPR (which varies as you climb).

                In the MD80, before the FMSs with VNAV, the SOP was more or less:
                Set TO/GA or TO-FLEX in the TRC.
                Set the clean-up altitude as target in the vertical mode. (say 3000ft)
                Set 250kts in the speed window of the AT (note that in this plane the speed for the vertical mode of the flight director / AP is set in a different window, in the same one that you would set the vertical speed, where you are setting a vertical speed or an airspeed depended on the vertical mode selected).
                Advance the throttles to somehow the take-off EPR shown in the TRC window.
                Engage the AT in TRC mode.
                Rotate, positive climb, gear up.
                Follow the FD bar to keep V2+10 / V2+20 during the first part of the climb.
                Engage the AP in "speed" mode (what would be LVL CHG).
                The AT will keep the TOGA or TO FLEX EPR, and the AP will hold the speed.
                Approaching 3000ft (at 2700ft to be precise) the AP will change autmatically to VS, the window of the vertical mode that was showing the speed will show the current climb rate, and the AT will revert to speed hold and start to go for the target of 250kts, but since it's already at max thrust it will do nothing by now.
                The AP (in fact the FD) will start to reduce the target VS in the window, so it will start to lower the nose, and the plane, still at max truths will start to accelerate.
                As you cross the different flap speeds, you clean up the plane. By when you finish the clean up chances are that the plane will be already established and holding 3000ft and the last retraction to "flaps zero" happens in 200/220kts range, so the plane will still be accelerating towards the 250kts.
                Select the cruise altitude (the plane will do nothing yet, waiting for "instructions" about how you want to get there), select CLB in the TRC, select TRC mode in the AT, select "speed" mode in the AP, select 250kts in the speed window of the vertical mode, and up we go with the AT holding the CLB thrust and the AP holding 250kts.
                Set the speed in the AT to Mach and to whatever will be your cruise Mach. The AT will do nothing with that by now because it's set to hold the EPR, not the speed.
                When crossing 10000ft, select say 280 (or whatever the climb airspeed is) in the window of the vertical mode.
                In the high 20s, the vertical mode AP will automatically change from "speed" to "Mach" (setting the current Mach), and it will hold it. The AT is still keeping the CLB thrust defined by the TRC. At that time, pay attention in case you need a small adjustment in that Mach number.
                700ft before the cruise altitude, the AP will revert to VS (automatically selecting the current VS) and the AT to Mach (remember you had already selected this Mach many minutes earlier, just after clean-up).
                As the plane approaches to the cruise altitude, the VS shown in the window will diminish until the plane reaches, and the AP tracks, the cruise altitude.

                As you can see, it's quite a workload until just after the clean-up (and we are dealing with the VNAV part only, we have not messed with the SID, but LNAV existed in the FMSs since much earlier than VNAV, airplanes have not been flying raw navigation for quite some time, so most likely the AP will be following all the lateral navigation from the FMS so it will not add a lot of workload).

                But after that, you only need to do 2 things:
                At 10000ft, change the speed window of the vertical mode from 250 to the climb speed.
                At changeover (high 20s), check the resulting Mach and adjust it if necessary, in the window of the vertical mode again.
                So that's just turning one knob once or twice.
                You don't even need to press a button anywhere from just after clean-up to the start of the descent (unless you need to change the speed or the flight level along the cruise part).

                And the thrust levers? You have not touched them since advancing them for take-off, and chances are that you will not touch them until the flare (you know, the part where an Airbus would insult you: "Retard!, Retard!")

                And we are talking about a 1980 plane without VNAV.

                --- 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


                • #9
                  Can I suggest you guys be a lot more specific about the aircraft type.

                  The way the MD80's version works is not the same as the 737. And the 737 is not even the same as the 767 - they're slightly different modes.

                  Wait a sec, as I understand it using LVL CH in climb requires the pilot to set the optimum power setting manually, then the AP maintains the MCP target speed with pitch using that power setting. The A/T is not active
                  Incorrect. If engaged to accomplish a climb, LVL CH engages the A/T in N1 mode (and gives climb N1). On descent, it goes to IDLE thrust and then remains armed.

                  Comment


                  • #10
                    Originally posted by MCM View Post
                    Can I suggest you guys be a lot more specific about the aircraft type.
                    They're being specific, it's just that they're being specific about different aircraft, which strikes me as kind of pointless.

                    Comment


                    • #11
                      Originally posted by MCM View Post
                      Can I suggest you guys be a lot more specific about the aircraft type.

                      The way the MD80's version works is not the same as the 737. And the 737 is not even the same as the 767 - they're slightly different modes.
                      Well, I was referring to Boeing aircraft with LVL CH or FLCH modes. I would also ask about Airbus OPEN modes. My impression is that there are some stealth aspects to these modes (such as the one referenced in this incident) that might not be clearly understood by some pilots and that some supplemental training might be a good idea (particularly on when to use these modes and when not to).

                      But MCM, you have reassured me that there is no extraordinary risk in climb with these modes, outside of the universal risk of pilot inattention. It seems to be just that, as Gabriel points out, a few checks and adjustments may be required that are not required of VNAV or V/S (true?), so closer attention might be needed to avoid the above scenario. (for instance, because there is no need to modulate power settings, do pilot see this as a 'set and forget' mode in climb, perfect for an extended dinner break?)

                      Descent, especially final, is another story. As we have seen. I think FLCH should be forbidden beyond the FAP (why would you ever need it on a stable approach?).

                      And, while it might not be the same animal at all, thanks for the MD-80 tutorial Gabriel.

                      Comment


                      • #12
                        LVL CHG and FLCH are different modes and operate slightly differently. FLCH has more control over thrust.

                        You are correct that there is no particular risk associated with them.

                        LVL CHG requires the pilot to set the speed, whereas VNAV sets the speed for you. As such, its one thing that can be forgotten. 98% of the time the climb (certainly in the upper stages) is done in VNAV, and so it would be easy to miss the fact that you have to control the speed because its not something you are used to doing. The mode is obvious to use - the catch is that the crew forgot they were in it to begin with. Thats the bigger problem.

                        All the modes have their catches. V/S is a mode that needs very careful attention, particularly in climb. Whereas LVL CHG will eventually stop climbing if you run out of thrust, V/S will keep pitching up and get you into a far nastier position. V/S is an extremely useful mode lower down, where you aren't going to be thrust limited, and you need to be more precise with rates of climb or descent - conversely LVL CHG has uses in the terminal area, but not for approaches.

                        FLCH is slightly smarter in that the speed changeover is at 300/.80 (on the 767), which is more 'middle of the road', and modulates thrust for small altitude changes.

                        Its about horses for courses - and remembering which horse you got on.

                        I think FLCH should be forbidden beyond the FAP (why would you ever need it on a stable approach?)
                        'Non-ILS approaches are normally flown using VNAV or V/S pitch modes. The use of FLCH is not recommended after the FAF.'

                        I have never seen anyone attempt to use it - it is widely regarded as unsuitable and ineffective. The only times I've ever read about it being used it has been in error (still engaged after descent, where it was used appropriately, and the crew have not transitioned to an appropriate mode).

                        Comment


                        • #13
                          EDIT: I've posted this before reading MCM's post just above
                          Originally posted by Evan View Post
                          And, while it might not be the same animal at all, thanks for the MD-80 tutorial Gabriel.
                          Well, yes, they are different animals: One is a tiger and the other a leopard.

                          I don't know exactly what are all the differences, but some of them, from what I understood from MCMs explanation:
                          - The 737 has only one MCP speed window, where you select the speed to track by whatever will controllit, the AP or the AT. In the MD80 you have 2 windows.
                          - The 737 sets climb N1 or idle, depending on whether you set LVL CHG to climb or descend. In the MD80, when you select "speed" mode the AP will automatically select the current speed in the vertical mode window, and the AT will remain engaged in "clamp" (manual) leaving the throttle levers in their current position. So basically the plane will keep doing what it was doing, waiting for "further instructions" from the pilot, who can manually set a higher or lower thrust to climb or descent (including "idle"), or set the AT to hold a computed thrust (like CLB). In any event, the AT remains engaged and will autmaticaly revert to "hold speed" (and the AT to "VS" and then "alt hold") when the altitude approaches the selected target.

                          But, basically, the core concept is the same:

                          In any of these modes, elevator controls speed, throttle controls vertical speed. In contrast with the AT in speed and the AP in VS or alt hold, where the throttle controls speed and the elevator controls vertical speed).

                          One important result of that is than in LVL CHG type of modes speed has priority (the plane will pitch up or down as needed to hold the speed). The plane will gladly descend if needed to hold the speed even if you selected a climb. (or will gladly climb even if you selected a descent).

                          In VS, ALT or GS the vertical path (in terms of VS or glide slope) has priority. If you are not with the AT in speed mode, or if you are but max thrust or idle are not enough to keep the selected speed, you will depart from the selected speed. And the AP will gladly stall or overspeed in its attempt to enforce the vertical path (unless stopped by an envelope protection). We have countless of accidents and incidents like this (Turkish and Pinnacle comes to mind).

                          I agree that LVL CHG (or similar) makes no sense in a stable approach.
                          But the BIG BIG mistake in Asiana was not so much using LVL CHG in the approach, but expecting that the automation would still control the speed when you just commanded the AP "now you control the pitch to enforce the speed" and, immediately after, disengaged it and start to manually control the pitch (and hence the speed).

                          --- 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


                          • #14
                            MCM,
                            Does one given plane have both LVL CHG and FLCH?
                            Or they are two "equivalent" modes (although with some differences) in different planes?

                            --- 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


                            • #15
                              Originally posted by Gabriel View Post

                              I agree that LVL CHG (or similar) makes no sense in a stable approach.
                              But the BIG BIG mistake in Asiana was not so much using LVL CHG in the approach, but expecting that the automation would still control the speed when you just commanded the AP "now you control the pitch to enforce the speed" and, immediately after, disengaged it and start to manually control the pitch (and hence the speed).
                              ...which was the result of using FLCH and apparently (though not yet confirmed) failing to set a minimum safe altitude (where VNAV should have been armed) on the MCP... ?

                              Comment

                              Working...
                              X