Announcement

Collapse
No announcement yet.

Turkish B738 crashlands in Amsterdam.

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

  • Originally posted by Skywatcher View Post
    To me, there are two moments that played a decisive role on the final outcome. First, receiving (false) aural warnings about wrong landing configurations while still flying at high altitude. As far as we know, probably they didn't understood what was going on with those alarms. That was very stressing for them, I'm sure.
    Any pilot certified to fly the airplane should not have been confused by the gear warning. If they were, they should not have been certified. These alarms should be apart of their training.

    Second, the fact that they joined the localizer very high. They were obliged to perform a -1400 feet/sec to join the slope at (only) 1330 feet. So they were only fully established (flaps 40) at 900 feet!! Another very stressing situation. Now, one thing I would like to know is if they had any chance to read the landing checklist until that moment, I'm sure they didn't because the aircraft was not 100% configured. At 770 feet they were still busy setting 144 knots on the box.
    They should have been engaged in a go-around at that point. It's one thing to find yourself struggling to keep up with the airplane but as soon as you fall behind, its go-around time.

    So, now they were finally breathing a bit, maybe they decided to go (finally) on the landing checklist and missed those decisive 20-25 seconds were the speed decayed. They were on a death trap, no way to get out of that.
    In my sincere opinion, most of the things I have been reading here are very unfair to the pilots, those guys were absolutely overwhelmed with huge work load on the final stage of the flight. They did a big mistake (missing that speed decay between 770 and 460 feet) but we have (or we must) to understand what was going on that cockpit.

    My honest five cents.
    Thanks for your honest five cents.

    But these guys allowed the aircraft to crash. I say again, they ALLOWED the aircraft to crash. They made bone head mistakes that disgust most pilots.

    The crash was caused by the flight crew failing to apply basic airmanship while operating an aircraft loaded with passengers. Had they done so, the aircraft would be in service today.

    Had these guys survived, they should have been banned from all form of commercial aviation. IMO this accident was inexcusable.
    Don
    Standard practice for managers around the world:
    Ready - Fire - Aim! DAMN! Missed again!

    Comment


    • Don, I wish I had your patience and eloquence.

      Comment


      • Interesting posts. I find also interesting this concept of stable approach, and wonder if this has any relevance to my latest flights with Ryanair. Normally, on any flight (also in former flights with Ryanair), you feel that the final descent is done softly, as if the plane is let to "glide" and burn its excess energy in the way down. I hope you understand what I'm saying, it's like driving a car and seeing a distant red light, letting go off the gas in advance instead of rushing to the light and then hitting hard the brakes.

        Now, in my four (two roundtrips) latests flights, I've feel like the plane is going down powered, rushing, and instead of a gentle glide the sensation is of a rushed ride, even in good weather, until finally you suddenly are stabilizing and in two minutes you have landed.

        Now, I'm not saying these pilots are doing anything unsafe (I guess that wouldn't be tolerated), but I've observed this just in this airline, and there's a urban legend going around that Ryanair pilots have incentives (don't know which ones exactly) to shorten the flight duration as much as possible. May be totally unfounded, sure. Also I don't think it makes sense economically to burn fuel you don't have to, nor that it would make for more than a few minutes of difference.

        I'm just wondering if someone else has observed the same, or I was just "unlucky" in these four flights. Could this also depend on the airport configuration for approaches?

        Comment


        • The most efficient use of energy, and thus the lowest operating cost, is to retard the throttles at TOD and never touch them until the wheels touch the runway. A stabilized approach requires the engines be spooled up at or about the time the aircraft passes the outer marker on an ILS approach. The stable phase of the approach is generally entered at or about 1000 feet AGL.

          ATC requirements can cause a delay in leaving TOD or cause the descent to begin earlier than necessary. Under these conditions you may hear the engines spooling from time to time or the pilots may need to use the speed brakes. Both procedures burn fuel.

          Your analogy of slowing (coasting) up to a red traffic signal vs waiting and applying hard braking is exactly on target. Conserving energy is the goal.
          Don
          Standard practice for managers around the world:
          Ready - Fire - Aim! DAMN! Missed again!

          Comment


          • Originally posted by Dmmoore View Post
            The most efficient use of energy, and thus the lowest operating cost, is to retard the throttles at TOD and never touch them until the wheels touch the runway.
            Too bad that never happens.

            Comment


            • Regarding stabilized approaches, by all means do read this document:

              http://www.flightsafety.org/alar/ala...blizedappr.pdf

              It mentions overreliance in automation, approaching high and fast, rushing an approach, demanding ATC clearance, being with too much workload during the approach, approaching with idle thrust...

              A good bunch of it looks like it's inspired on this accident, only that it was written eight years ago.

              --- 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


              • This post took me a few days to complete.

                With the new info available, and comparing with what we have been discussing, some things seem to be mitigating, others look even worse.

                For example, we no know that it wasn't 100 seconds, but about 1/4 of that the, time that the speed went below Vref. We also know that, since they were approaching high and fast, idle thrust would have been an expectable thing for the autothrottle to command even if it hadn't been affected by the failure of the radio-altimeter.

                On the other side, we now know that the left side RA didn't just fail at 2000ft. It was failing since long ago during the flight and the failure become evident at 2500ft, much more than 100 seconds before the stickshaker. Also we now know that the crew was "playing" with the flaps and the speed setting well below 1000ft, and that that is against the stabilized approach concept.

                Below you'll find a graphical and written timeline, highlighting (underlined) the many opportunities to brake a link in the chain of events and prevent this accident from happening.

                A standard disclaimer applies: This timeline was prepared with the partial, unconfirmed and incomplete data from the preliminary report and some other data that had become available earlier, and some extrapolation from my part. Other than data, I added opinion too. I tried to keep it neutral, but I'm a subject so don't expect a perfect objectivity. Finally, I am not a professional pilot or a professional accident investigator, so keep everything with a grain of salt.

                That said, enjoy!



                1- Design. The RA and A/T interaction could have been designed with more redundancy and warnings, to make it easier for pilots to catch failures and take appropriate action.

                2- Training / Culture. A proper pilots' selection, training, checking and monitoring, along with the company's SOP and culture, should have been able to brake at least one of the links in the chain of events that led to the accident (see below)

                3- ATC vectored the plane in a way that left it in a high energy situation (too high above the glide path relatively close to the runway), which is potentially dangerous if not properly managed by the pilots (which could include abandoning the approach and going around). The pilots on the other hand accepted the clearances and didn't manage the approach appropriately (see below).

                4- Both RAs detect altitudes AGL up to 8191ft. However, only altitudes from 2500ft and below are displayed in the cockpit. When the plane descended below 8191 ft the right hand RA commenced to record the correct altitude value, but the left RA remained pegged at 8191 ft. At some point during the approach (arbitrarily represented by point 4 in the profile) the plane descended through 2500 ft AGL. AT that point the right RA's display (the one displayed on the FO/PF primary flight display, became alive showing the altitude from 2500 and down. The left RA's display remained dead. Both RAs disagreed for the remaining of the flight. The crew should have detected that disagreement, understood its implications and taken proper measures (according to Boeing documentation, to disengage the A/T)

                5- The plane stabilizes at 2000ft, flaps 15 and gear down are set.

                6- The plane intercepts and starts tracking the localizer. Speed 160 KIAS.

                7- The pilots select a descent in vertical speed mode at 1400 ft/min to intercept the glide slope from above. Immediately as the plane starts to descent the left RA recording suddenly changed from 8191 ft to -8 ft. The left RA's display became alive and started displaying -8ft on the capt/PNF primary flight display. As result of that, the gear warning sounds and the A/T changes from "speed hold" mode to "retard" mode. The change of A/T mode is displayed in the flight mode announciators which are part of both primary flight displays and which is one of the top priorities parameters for the pilots to monitor. The thrust levers go to idle when the A/T changes to retard mode, but that could have been expectable even without that failure as the A/T would have reduced the thrust anyway to keep the airspeed when the flight changed from level flight to a 1400 ft/min descent. The A/T remained in "retard" mode and the thrust levers remained at the idle position until after the stick shaker activated. (Perhaps this is as far as they should have gone with the approach, even if the RA had not failed. 5 NM to go, plane not configured, way above the glide slope to the point that a 1400 ft/min descent would be needed, in actual IMC. It looks like a go around at this point would have been wise. On the other hand, hindsight is always 20/20)

                8- At 1330 ft the plane intercepts and starts to track the glide path. By then the speed had increased to 169kts during the steep descent. As the plane starts to follow the shallower glide path, the speed starts to decrease.

                9- At 1000ft the plane crosses what is typically the "minimum stabilizing altitude" in IMC. The stabilized approach criteria is not met. A go around should have been immediately initiated. (stabilized approach procedures vary from company to company, but typically include all of speed at Vref and stabilized -not met-, loc and GS within 1 dot and stabilized (met), plane configured for landing -not met-, landing checklist completed -probably not met-, and thrust above flight idle -not met-)


                10- At 900ft they set full flaps. Speed continues to go down. The fact that they were at 900 ft already and still configuring the plane should have rung a bell inside the pilots' head. On the other hand, it looks almost certain that by then they were monitoring the speed (it's slowing down, still above Vref, good time to set full flaps). I cannot believe that a pilot would set full flaps without looking at the speed first. (But then, I cannot believe that a pilot let the speed go 40kts below Vref before applying thrust). So if they are looking at the primary flight display to check the speed, they are missing the "retard" mode in the flight mode annunciator (in the same primary flight display) that has been there for quite a lot of seconds by then.

                11- At 770ft the plane is at 144kts, exactly Vref. So one pilot stretches his arm, reaches the speed knob in the A/T panel, and selects 144kts. Say again? He changes to 144 kts from what? From the previously selected speed, which certainly was not 144kts, and more than likely was faster (probably the 160kts they were holding before). Why is the plane flying at 144 kts when the selected speed was faster. Now we know. Also, if they knew that a short time before (when they set full flaps) the pane was going faster than Vref, and now it was at Vref, the plane was obviously slowing down, so it's obvious that to stop the slowdown and hold Vref thrust should have to be added, which should give a lot of feedback to the pilots in the form of thrust levers moving, engine parameters increasing, engine noise increasing, and the sensation of acceleration (and, of course, the speed stopping to go down). Nothing of this happened, of course. So, summarizing, the plane was slowing down through Vref and they set the speed to Vref. They should have noted that the plane was not holding the previous speed. They should have also noted that no additional thrust was added and that without it the plane was going to keep slowing down. Everything was saying "hey, there's something odd with the autothrottle". And they don't even check the autothrottle status in the flight mode annunciator? This situation, as a hole, represents a very big clue. And a very big opportunity lost to brake the chain of events

                12- At 500ft the plane crosses what is typically the "minimum stabilizing altitude" in VMC. The stabilized approach criteria is not met (V not at Vref, thrust not above flight idle). A go around should have been immediately initiated.

                13- Immediately after 11- the speed started to drift below Vref and the pitch started to increase to a nose up position which is beyond typical approach pitch. When the plane was at 460ft, some 20 to 25 seconds after point 12- the speed had decayed to 110kts (34 kts below Vref) and the pitch has increased to 11° (typical final approach attitude is about 2°). So for the last several seconds all of the following was happening simultaneously: the flight mode annunciator was displaying "retard" for the autothrottle mode, the speed was well below Vref and decreasing, the pitch was well above final approach pitch and increasing, and the thrust was at flight idle. All of them are unacceptable. Any of them by itself renders the approach unstabilized. Three of them (flight mode annunciator, speed and pitch) are displayed on an instrument that not just by chance has the name of primary flight display, and of which there are two, each just in front of each pilot, and where the main job of one of them is to be looking at it as the number one task and the job of the other one is to monitor it every few seconds. Everything (except the flight path) was unacceptable. The pilot flying should have caught it, the pilot not flying should have caught it, and a go around should have been immediately initiated before this point, when the stickshaker activates.

                14- Altitude 460ft. Speed 110kt (34 below Vref). Pitch 11° nose-up. Thrust at flight idle. The stickshaker (stall proximity warning) activates. Throttles are advanced (probably manually) but autothrottles are not disengaged. Throttles move back to idle (probably commanded by the autothrottles still engaged in "retard" mode). Throttles should have been kept at the forward mechanical limit. The autothrottle should have been disengaged.

                15- Altitude 420ft. The autopilot (but not the autothrottle) is disengaged. Nose down is commanded.

                17- Altitude 310ft. The plane has been pitched down to -8° (remember it had started with +11°, so it was a remarkable push down of almost 20°). The throttles are firewalled (almost sure manually) but it's too late already. The plane starts to pitch up again

                18- Pitch is 20° nose up. The plane is descending fast vertically but slow horizontally. There is no altitude left. Some say "the plane pancaked". It crashed fully stalled.

                --- 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


                • Originally posted by Gabriel View Post
                  This post took me a few days to complete.

                  With the new info available, and comparing with what we have been discussing, some things seem to be mitigating, others look even worse.

                  For example, we no know that it wasn't 100 seconds, but about 1/4 of that the, time that the speed went below Vref. We also know that, since they were approaching high and fast, idle thrust would have been an expectable thing for the autothrottle to command even if it hadn't been affected by the failure of the radio-altimeter.

                  On the other side, we now know that the left side RA didn't just fail at 2000ft. It was failing since long ago during the flight and the failure become evident at 2500ft, much more than 100 seconds before the stickshaker. Also we now know that the crew was "playing" with the flaps and the speed setting well below 1000ft, and that that is against the stabilized approach concept.

                  Below you'll find a graphical and written timeline, highlighting (underlined) the many opportunities to brake a link in the chain of events and prevent this accident from happening.

                  A standard disclaimer applies: This timeline was prepared with the partial, unconfirmed and incomplete data from the preliminary report and some other data that had become available earlier, and some extrapolation from my part. Other than data, I added opinion too. I tried to keep it neutral, but I'm a subject so don't expect a perfect objectivity. Finally, I am not a professional pilot or a professional accident investigator, so keep everything with a grain of salt.

                  That said, enjoy!



                  1- Design. The RA and A/T interaction could have been designed with more redundancy and warnings, to make it easier for pilots to catch failures and take appropriate action.

                  2- Training / Culture. A proper pilots' selection, training, checking and monitoring, along with the company's SOP and culture, should have been able to brake at least one of the links in the chain of events that led to the accident (see below)

                  3- ATC vectored the plane in a way that left it in a high energy situation (too high above the glide path relatively close to the runway), which is potentially dangerous if not properly managed by the pilots (which could include abandoning the approach and going around). The pilots on the other hand accepted the clearances and didn't manage the approach appropriately (see below).

                  4- Both RAs detect altitudes AGL up to 8191ft. However, only altitudes from 2500ft and below are displayed in the cockpit. When the plane descended below 8191 ft the right hand RA commenced to record the correct altitude value, but the left RA remained pegged at 8191 ft. At some point during the approach (arbitrarily represented by point 4 in the profile) the plane descended through 2500 ft AGL. AT that point the right RA's display (the one displayed on the FO/PF primary flight display, became alive showing the altitude from 2500 and down. The left RA's display remained dead. Both RAs disagreed for the remaining of the flight. The crew should have detected that disagreement, understood its implications and taken proper measures (according to Boeing documentation, to disengage the A/T)

                  5- The plane stabilizes at 2000ft, flaps 15 and gear down are set.

                  6- The plane intercepts and starts tracking the localizer. Speed 160 KIAS.

                  7- The pilots select a descent in vertical speed mode at 1400 ft/min to intercept the glide slope from above. Immediately as the plane starts to descent the left RA recording suddenly changed from 8191 ft to -8 ft. The left RA's display became alive and started displaying -8ft on the capt/PNF primary flight display. As result of that, the gear warning sounds and the A/T changes from "speed hold" mode to "retard" mode. The change of A/T mode is displayed in the flight mode announciators which are part of both primary flight displays and which is one of the top priorities parameters for the pilots to monitor. The thrust levers go to idle when the A/T changes to retard mode, but that could have been expectable even without that failure as the A/T would have reduced the thrust anyway to keep the airspeed when the flight changed from level flight to a 1400 ft/min descent. The A/T remained in "retard" mode and the thrust levers remained at the idle position until after the stick shaker activated. (Perhaps this is as far as they should have gone with the approach, even if the RA had not failed. 5 NM to go, plane not configured, way above the glide slope to the point that a 1400 ft/min descent would be needed, in actual IMC. It looks like a go around at this point would have been wise. On the other hand, hindsight is always 20/20)

                  8- At 1330 ft the plane intercepts and starts to track the glide path. By then the speed had increased to 169kts during the steep descent. As the plane starts to follow the shallower glide path, the speed starts to decrease.

                  9- At 1000ft the plane crosses what is typically the "minimum stabilizing altitude" in IMC. The stabilized approach criteria is not met. A go around should have been immediately initiated. (stabilized approach procedures vary from company to company, but typically include all of speed at Vref and stabilized -not met-, loc and GS within 1 dot and stabilized (met), plane configured for landing -not met-, landing checklist completed -probably not met-, and thrust above flight idle -not met-)


                  10- At 900ft they set full flaps. Speed continues to go down. The fact that they were at 900 ft already and still configuring the plane should have rung a bell inside the pilots' head. On the other hand, it looks almost certain that by then they were monitoring the speed (it's slowing down, still above Vref, good time to set full flaps). I cannot believe that a pilot would set full flaps without looking at the speed first. (But then, I cannot believe that a pilot let the speed go 40kts below Vref before applying thrust). So if they are looking at the primary flight display to check the speed, they are missing the "retard" mode in the flight mode annunciator (in the same primary flight display) that has been there for quite a lot of seconds by then.

                  11- At 770ft the plane is at 144kts, exactly Vref. So one pilot stretches his arm, reaches the speed knob in the A/T panel, and selects 144kts. Say again? He changes to 144 kts from what? From the previously selected speed, which certainly was not 144kts, and more than likely was faster (probably the 160kts they were holding before). Why is the plane flying at 144 kts when the selected speed was faster. Now we know. Also, if they knew that a short time before (when they set full flaps) the pane was going faster than Vref, and now it was at Vref, the plane was obviously slowing down, so it's obvious that to stop the slowdown and hold Vref thrust should have to be added, which should give a lot of feedback to the pilots in the form of thrust levers moving, engine parameters increasing, engine noise increasing, and the sensation of acceleration (and, of course, the speed stopping to go down). Nothing of this happened, of course. So, summarizing, the plane was slowing down through Vref and they set the speed to Vref. They should have noted that the plane was not holding the previous speed. They should have also noted that no additional thrust was added and that without it the plane was going to keep slowing down. Everything was saying "hey, there's something odd with the autothrottle". And they don't even check the autothrottle status in the flight mode annunciator? This situation, as a hole, represents a very big clue. And a very big opportunity lost to brake the chain of events

                  12- At 500ft the plane crosses what is typically the "minimum stabilizing altitude" in VMC. The stabilized approach criteria is not met (V not at Vref, thrust not above flight idle). A go around should have been immediately initiated.

                  13- Immediately after 11- the speed started to drift below Vref and the pitch started to increase to a nose up position which is beyond typical approach pitch. When the plane was at 460ft, some 20 to 25 seconds after point 12- the speed had decayed to 110kts (34 kts below Vref) and the pitch has increased to 11° (typical final approach attitude is about 2°). So for the last several seconds all of the following was happening simultaneously: the flight mode annunciator was displaying "retard" for the autothrottle mode, the speed was well below Vref and decreasing, the pitch was well above final approach pitch and increasing, and the thrust was at flight idle. All of them are unacceptable. Any of them by itself renders the approach unstabilized. Three of them (flight mode annunciator, speed and pitch) are displayed on an instrument that not just by chance has the name of primary flight display, and of which there are two, each just in front of each pilot, and where the main job of one of them is to be looking at it as the number one task and the job of the other one is to monitor it every few seconds. Everything (except the flight path) was unacceptable. The pilot flying should have caught it, the pilot not flying should have caught it, and a go around should have been immediately initiated before this point, when the stickshaker activates.

                  14- Altitude 460ft. Speed 110kt (34 below Vref). Pitch 11° nose-up. Thrust at flight idle. The stickshaker (stall proximity warning) activates. Throttles are advanced (probably manually) but autothrottles are not disengaged. Throttles move back to idle (probably commanded by the autothrottles still engaged in "retard" mode). Throttles should have been kept at the forward mechanical limit. The autothrottle should have been disengaged.

                  15- Altitude 420ft. The autopilot (but not the autothrottle) is disengaged. Nose down is commanded.

                  17- Altitude 310ft. The plane has been pitched down to -8° (remember it had started with +11°, so it was a remarkable push down of almost 20°). The throttles are firewalled (almost sure manually) but it's too late already. The plane starts to pitch up again

                  18- Pitch is 20° nose up. The plane is descending fast vertically but slow horizontally. There is no altitude left. Some say "the plane pancaked". It crashed fully stalled.
                  Could you reduce the above to some kind of a concise, basic point?

                  Comment


                  • Originally posted by MCM View Post
                    Just re-iterating what I said in a previous post, many major airlines use the concept of stable approaches, and if you aren't at a reasonable speed, fully configured, and thrust at normal approach settings by 1000ft in IMC, then you must go around.

                    That concept might have saved this crew.
                    Does anyone know whether Turkish Airlines mandate use of the stable approach?

                    Comment


                    • Gabriel,

                      That was great stuff.

                      Comment


                      • Originally posted by TheKiecker View Post
                        Gabriel,

                        That was great stuff.
                        Yea, but where's the sketch I'd attached to that post, that now only shows like a red X.

                        Not only it was crucial for the post and to understand the chain of events, it also took me a good effort and a good while of my time to produce it.

                        Mods? Admins?

                        --- 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


                        • Originally posted by Spad13 View Post
                          Could you reduce the above to some kind of a concise, basic point?
                          The pilots stuffed up pretty much sums it up. The only correct thing they did with the approach was having the correct directions. They were too high which is why they dropped the throttle yet still kept the autopilot on till the very end almost. Also because they were configuring the plane to land at under 1000 feet should tell the pilots to abort the landing which they did not do. Both these pilots failed big time.

                          Comment


                          • Report issued

                            Originally posted by kaiser_chief View Post
                            The pilots stuffed up pretty much sums it up. The only correct thing they did with the approach was having the correct directions. They were too high which is why they dropped the throttle yet still kept the autopilot on till the very end almost. Also because they were configuring the plane to land at under 1000 feet should tell the pilots to abort the landing which they did not do. Both these pilots failed big time.
                            Seems like the Dutch investigators agreed with you. http://www.news24.com/World/News/Alt...crash-20100506

                            Comment


                            • Turkish Airlines Flight 1951 Accident Animation

                              Below is a link to an animation prepared by the Dutch Safety Board. It's a very interesting video which certainly adds perspective to the events leading up to the accident.

                              Auf YouTube findest du die angesagtesten Videos und Tracks. Außerdem kannst du eigene Inhalte hochladen und mit Freunden oder gleich der ganzen Welt teilen.

                              Comment


                              • Maybe I missed it earlier but here is the investigation report:

                                Comment

                                Working...
                                X