Turkish airliner crashes at Schiphol

[S.F.L.Y]

For either the 737 or the 777, that should have resulted in a climb and a safe go-around. Approach to satll is something we DO train for.

This is not correct, when you train for that you stop trimming the aircraft well before reaching the stick shaker, which wasn't the case for the Turkish. A lot of people comment on the fact that they didn't use their fully workable engines, which is totally incorrect and shows an obvious lack of understanding of the report. Let's have a look at the official report instead of reading absolutely wrong comments made by so called professionals (the bold characters are my comments):

http://www.onderzoeksraad.nl/docs/ra...ary_EN_def.pdf

According to radar data the localizer of the instrument landing system was intercepted at approximately 5,5 NM from the runway threshold. The standard procedure is that the glide path is approached from below. Because of the shorter line up for the runway at 2000 feet the glide path had to be approached from above. This is a major factor in this incident as on a normal interception thust needs to be reduced while in this particular case thrust had to be added in order to stabilize the glide. Many accidents occurred during similar interceptions from above. Once the localizer was intercepted the crew selected, by means of the vertical speed modus of the automatic pilot, a descent speed of 1400 feet per minute to catch the glide path. The autothrottle system entered the retard mode, and the thrust levers were moved to the idle position and remained in retard modus. In order to conduct the interception from above the crew had to select a vertical speed. While engines were on idle the aircraft was still accelerating due to the Vz. The glide path was intercepted at approximately 1330 feet and the aircraft was now also at the correct altitude for the approach of runway 18R. The aircraft speed had during the time the aircraft was in vertical speed modus increased to 169 knots, and decreased again when the aircraft followed the glide path. Once the crew had intercepted the GS (at 1330'!) and recovered a normal Vz the speed started to decrease towards Vref.


At approximately 900 feet , the flaps were selected to 40 by the crew they most probably had to check the speed for that... and the speed continued to decrease. At approximately 770 feet, the crew set the selected airspeed to 144 knots. At that moment the actual airspeed was 144 knots. As the speed was decreasing toward 144 kts the crew selected it as the approach speed, expecting the AT to adjust thrust to maintain it. This demonstrate that they were monitoring speed a 770 feet on the contrary of what many people wrongly believe. Reduced thrust between 2000 and 770 feet seemed absolutely normal as they were too fast due to the high Vz GS interception from above.The autothrottle system should have maintained the speed selected by the crew but, with the thrust levers at idle, speed continued to decay. Because the automatic pilot wanted to maintain the glide scope, the automatic flight system, in response, commanded increasing nose up pitch and applied nose up stabiliser trim.

The stick shakers activated at approximately 460 feet which is only 310 feet below the selection of the approach speed... meaning that they only had 310 feet (about 30 seconds) to notice that speed was dropping without adjustments from the AT. This is far from the 2000 feet of unmonitored descent described by some "professionals" of this forum. , warning the crew that the angle of attack (AOA) was too high. The data of the digital flight data recorder show that the thrust levers were immediately advanced but moved back to idle.The crew immediately reacted by manual thrust adjustment while the AT immediately neutralized it. When the thrust levers returned to idle, the autothrottle was disengaged. Whether these actions were performed by the crew or automatically is still under investigation. At that moment, the speed was approximately 110 knots, the pitch angle was approximately 11° Aircraft Nose Up (ANU) and the recorded AOA was approximately 20°.


At 420 feet the autopilot was disengaged by the crew and attempts were made to recover the correct flight position by pitching the aircraft. 40 feet below stick shaker activation the crew decided to disconnect the AP and pitch down to preserve speed. At 310 feet a negative nose position was reached of 8° beneath horizon. 8 degrees of nose below horizon shows that the aircraft responded to the pitch down command from the crew. Almost simultaneously the thrust levers were advanced to their most forward position after which the aircraft ascended somewhat and the nose position increased.For the second time in few seconds the crew acted on the thrusts levers which finally gave full thrust as the AT was disconnected. This full thrust adjustment caused the aircraft (which was trimmed to minimum speed) to climb due to the engines momentum. The combination of full thrust and trim position created a pitch-up movement which most probably exceeded the elevator's nose down inputs. At that point pushing on the stick wasn't enough and the attitude increased again. According to the last recorded data of the digital flight data recorder the aircraft was in a 22° ANU and 10° Left Wing Down (LWD) position at the moment of impact. The aircraft stalled as a result of the high induced attitude.


What strikes me in this report is that the crew had to conduct a GS interception from above with excessive speed in IMC (interception from above at 1330' with 169 kts). There is a major ATC factor in my opinion. Following that interception the crew obviously monitored speed as low as 770' since 144 kts was selected during the deceleration. While the crew expected the AT to adjust speed kept decreasing until stick shaker, which occurred 310' lower. This didn't give them much time to realize the AT wasn't working properly and this is far from the 2000' of unmonitored speed commented by "professionals" who obviously didn't even read the report. As the stick shaker was activated the crew applied full thrust manually (which came back to idle), disconnected the AP and pitched down. Realizing the levers went back to idle the captain re-adjusted the thrust to the maximum and the aircraft entered an uncontrolled pitch up movement leading to the final stall. In any case they crashed without using fully serviceable engines, they crashed with engines at full power which basically caused the stall due to the extreme pitch up trim position. Engines were set to full power for the last 300 feet, it wasn't a matter of a basic go around. If the BA guys had recovered full thrust after stick shaker activation they would also have enjoyed a roller coaster trip.

In regard to the BA038 they basically achieved the worst gliding performances they could. Not only their aircraft flew a wrong path aiming way too far beyond the threshold, it also wasted all of its energy to fly at an extremely low speed which is definitely giving the poorest gliding performances, especially with headwind. Reducing to minimum speed is basically how you achieve the shortest gliding distance. You don't need to be an astronaut to understand that (even if the space shuttle guys are doing it for real). Ask any helicopter pilot familiar with autorotations, it's exactly the same thing. If you add to that the fact that they still had little available thrust, flying at best glide speed would have made the most effective use of this available thrust while flying slower increases minimum required thrust. Again, this is very basic.

Again, for those who are little bit slow to understand, I'm not judging any crew, but I think it's excessive to fully blame the Turkish crew without proper understanding of the accident while calling the BA guys heroes as they achieved the worse gliding performances they could...

[Denti]

This is not correct, when you train for that you stop trimming the aircraft well before reaching the stick shaker, which wasn't the case for the Turkish. A lot of people comment on the fact that they didn't use their fully workable engines, which is totally incorrect and shows an obvious lack of understanding of the report. Let's have a look at the official report instead of reading absolutely wrong comments made by so called professionals

Your training must be pretty crap in that case. Dunno about others, but in the outfit i fly i had to do exactly this case every six months in the simulator. Autopilot flying the ILS, thrust idle, recover on stick shaker, the autopilot keeps trimming until stick shaker onset. It is not rocket science, it is mandatory training. And yes, it was trained also to push the thrust levers forward and click the autothrust disconnect at the same time, because to land in that situation something must have gone wrong, and it can well be the autothrust computer for some reason, best to remove it there. Even more important though is that the PM has to actively monitor thrust application and reset go-around thrust if not set correctly as well. Again, 6 months training exercise. Yes, the 737 trimmed full backwards is a handfull to fly, but it is really nothing spectacular either, it is simply basic flying. Even more basic is to monitor the speed during flight so as not to end up in that situation in the first case.

[lexxie747]

there is one magic statement made above: only 30 seconds........to notice decay of airspeed.
THEY DIDNOT NOTICE IT FOR 30 SECONDS! what were they doing during that time? my left eye doesnt leave the speedindication till flare. 30 second is a very long time, ask my wife.....

[BOAC]

Denti - I think that at the time of the crash your company was unusual in conducting a stick-shaker/fully trimmed stall recovery 'every six months'.

The prelim report does change my understanding from the initial reports in as much as I understood the autothrottle had remained engaged during recovery. It appears that whatever the route INTO the stall was, this crew experienced a classic under-slung engines pitch event along the lines of the BOH 737, the ?CPH? BA757, the ?LHR? BA 747 and at least one other I cannot recall somehwere in Scandiland. I have always understood that the only way to recover this event is either to reduce power or to apply significant bank, both of which are not 'natural' reactions. On a quick read-through I also did not see any statement on the history of any reporting of the radalt 'defect' other than DFDR data.

It would seem that Denti's company is leading the way in what I think should now become a mandatory recurrent item in underslung types.

[S.F.L.Y]

Here's a good publication about the Thomson incident:
http://flightsafety.org/asw/aug09/asw_aug09_p28-31.pdf
The incident sequences are more than similar to the Turkish. The aircraft experienced a severe pitch-up momentum, beyond pitch control authority, following the application of thrust. The aircraft stalled nose up and speed went as low as 82 kts. Fortunately the crew had enough height to recover.

[Wizofoz]

What strikes me in this report is that the crew had to conduct a GS interception from above with excessive speed in IMC (interception from above at 1330' with 169 kts). There is a major ATC factor in my opinion. Following that interception the crew obviously monitored speed as low as 770' since 144 kts was selected during the deceleration. While the crew expected the AT to adjust speed kept decreasing until stick shaker, which occurred 310' lower. This didn't give them much time to realize the AT wasn't working properly and this is far from the 2000' of unmonitored speed commented by "professionals" who obviously didn't even read the report. As the stick shaker was activated the crew applied full thrust manually (which came back to idle), disconnected the AP and pitched down. Realizing the levers went back to idle the captain re-adjusted the thrust to the maximum and the aircraft entered an uncontrolled pitch up movement leading to the final stall. In any case they crashed without using fully serviceable engines, they crashed with engines at full power which basically caused the stall due to the extreme pitch up trim position. Engines were set to full power for the last 300 feet, it wasn't a matter of a basic go around. If the BA guys had recovered full thrust after stick shaker activation they would also have enjoyed a roller coaster trip.

They didn't HAVE to intercept glids slope from above, nor continue with the approach. They could and almost certainly should have initiated a go-around well before things got to that stage.

I can't speak for Turkish, but what you describe would constitute an unstable approach by any SOPS I've operated to, and a go-around would have been manditory from 1000ft. If nothing else, all SOPs I've operated to mandate thrust not be at idle by 1000'- clearly not the case here.

You make much of the fact that they set 144kts at 770 ft. Firstly, the need to be changing the bug at that height means the approach was unstable. Secondly, at that or any height the anticipated thrust increase should have been manually backed up and over-riden if it did NOT increase as expected (Interestingly, I'm just thinking back to My F/Os approach into DXB not two hous ago. With a FUINCTIONAL a/t system, he still needed to manually adjust thrust during a bumpy approach as the A/t wasn't reacting fast enough.)

If they did indeed have the SA to be expecting a thrust increase as the approach stabilised, they should have had the CRM to adjust for it when it didn't.

As to the attempted recovery, firstly approach to stall mandates disconnection of the A/t.Secondly, once applying max thrust, your hand should not come off the T/Ls until you are in a safe flight condition.

[S.F.L.Y]

They didn't HAVE to intercept glids slope from above, nor continue with the approach. They could and almost certainly should have initiated a go-around well before things got to that stage.

I can't speak for Turkish, but what you describe would constitute an unstable approach by any SOPS I've operated to, and a go-around would have been manditory from 1000ft. If nothing else, all SOPs I've operated to mandate thrust not be at idle by 1000'- clearly not the case here.

Don't you think it's also ATC's responsibility to send aircraft on an ILS interception from above?? It happens very often and as you said this shouldn't be accepted. However, many crews, and not only Turkish, prefer to intercept than going around.

The accident isn't the result of a single cause as many posters seems to believe. Here are some of the main contributions:

• ATC cleared the aircraft for intercepting the ILS beyond the interception point and above GS.
• The rad-alt failure set the AT to retard mode
• Despite being fully established at 1330', the high Vz interception increased the airspeed which couldn't be stabilized at 1,000'
• Vref was reached at 770' and crew failed to notice the lack of thrust adjustment
• Trim was set to max nose up as speed decaying during the next 300'
• Manual full thrust input combined to the high nose up trim setting caused the aircraft to pitch up and stall at 300' resulting in a low speed/high attitude impact.

From what I've read, I wouldn't say it's only a matter of 3 guys sleeping for 100 seconds and crashing without using their fully serviceable engines.

I think CVR would be very useful to comment on crew's actions, be it in the BA or in the Turkish. It's amazing to see professionals celebrating or accusing colleagues without any knowledge of what actually happened in the cockpit.

While I find totally stupid to call the BA guys heroes (who did nothing but letting the AP fly the wrong way) I find even more inappropriate to accuse a dead crew without having properly read the report. Is this what you consider as a professional attitude?

[Wizofoz]

I accuse the Turkish crew of nothing more than exactly what you ascribe to them above.

Accepting an unstable approach, continuing with a Vs that meant speed did not stabilize, not noticing that the speed bled back below Vref whilst the thrust stayed at idle and allowing the aircraft to trim full nose up are all included above and are actions by or omissions from the crew.

I also do not anoint the BA Crew with the Hero tag. I think that, dealing with a MAJOR malfunction, they did a perfectly reasonable job and got a very nearly optimum outcome. I doubt that a nicely flared landing on the piano keys was possible under any circumstances, so zero fatalities is a pretty good result.

Yes, the Turkish guys were at the end of a chain of events, not all of their making, but the fact remains that the aircraft was serviceable apart from a minor glitch, the dealing with of which is what we're paid for.

SFLY, I'm actually sympathetic to wards the Turkish crew. They made some fundamental errors, but there but for the grace of God....

What got me involved in this thread and is earning you so much rancor is your attempt to equate them with the 38 crew. They did not make ANY errors in their handling of the situation. They were not (and neither am I or you) trained to get the optimum gliding performance out of the aircraft (which, by the way, would have been impossible to know, as the presence of residual thrust would have changed the required speed.) and did a good job with the little time and knowledge available too them.

We ARE trained not to let a simple automatic glitch end in an accident.

We should certainly look deeper into training, culture, complacency and over-reliance on automatics to try and find and correct the root cause of the problem and hey, maybe giving us an AofA meter and information on how best to use it would benefit the next crew who find themselves gliding a jet.

Your support of the Turkish guys is admirable. Your attempt to say they were no more in error then the BA38 crew is misguided.

[MathFox]

From what I've read, I wouldn't say it's only a matter of 3 guys sleeping for 100 seconds and crashing without using their fully serviceable engines.

It was three guys trying to do a good job; ATC trying to achieve efficient traffic flow and some piece of automation, installed to make landing safer, that failed them all. All indications are that the failure of the AT was unnoticeable till 770' and when the crew noticed (soon after) they took the appropriate decision to GA, losing control of the plane after selecting full power.

First question that pops up: Why did the PF lose control during the GA? Lack of specific training?
An important point to notice is that the technical failure crept upon the pilots... had they known about it at a higher altitude, they could have disabled the system and hand-flown the plane. While the fault in itself is pretty trivial, the chain of events lead to a fatal result. I think this accident raises interesting questions, looking at the man-machine interface... "Why did the throttle retract after a manual override?" is one of them.

[BOAC]

Originally Posted by MathFox

"Why did the throttle retract after a manual override?" is one of them.

- easily answered I think? With the A/T engaged, levers would return to idle if opened manually due to faulty radalt signal. When the A/T was disengaged, they did not.

[S.F.L.Y]

I think this accident raises interesting questions, looking at the man-machine interface... "Why did the throttle retract after a manual override?" is one of them.

At least it retracted, in an airbus it wouldn't. Look at the Sao Paulo accident where the crew didn't notice that one engine was in TOGA while the other was in reverse... Certainly a lot to say about human/machine interface.

BOAC, while it's logic that the levers moved back to idle, it is questionable to decide whether or not the AT should prevail following contradictory manual inputs.

Wiz, I'm not judging the crews of both BA & Turkish. Of course BA guys obtained excellent results and didn't make any fault leading to the accident, while the Turkish crew actively created the context of their accident. The point I'm trying to make isn't there. I'm looking at the situational awareness and the acceptance of an unusual situations by the crews, leading to rapid decisions. The BA and Turkish guys didn't have much time to understand and accept their situation in order to act accordingly and basically they took the actions which seemed to be the most appropriate. In the Turkish case, applying full thrust in their configuration sealed their fate while BA038 achieved a safe landing despite their flight profile.

[MathFox]

"Why did the throttle retract after a manual override?" is one of them.

easily answered I think? With the A/T engaged, levers would return to idle if opened manually due to faulty radalt signal. When the A/T was disengaged, they did not.

I look at the issue from an "interaction designer" point of view. (I am involved in software design for a living.) The design issue I see is that the human override only had a temporary effect; Boeing might have designed the system so that autothrottle would disengage when the pilot moved the throttle manually.
I know each of those choices creates its own issues; an accidental disengagement of AT creates problems too. The difference for the pilot would be:

• move thrust lever and press button to engage AT again (if desired)
• press button to disengage AT before moving thrust lever

What would be safer/more natural for a pilot?

[captplaystation]

I don't think you can equate setting the speed of 144kts on the MCP as being confirmation that they were monitoring speed, and if they were, they gave up at this point as they didn't notice the subsequent very substantial reduction in speed until the stick shaker woke them up.

Why ?

Well, without the benefit of hearing the CVR we are only guessing, but as the approach was "rushed" I wonder if the landing checks had been read & responded to, and completed with ," holding at flaps", as many companies do.

OR , and this scenario would fit the outcome very nicely, immediately after 144kts was set ,PNF started reading & PF responding to, the landing checklist.
This would be about the right time scale (30 seconds or so ? ) for the period that no-one appears to have been monitoring the airspeed trend.
This leaves another Q of course, what exactly was the "safety" pilot on the jump seat doing at this stage to enhance safety?

As a previous poster has said, BASIC situational awareness would lead you to expect the thrust levers to advance as speed stabilised, and it feels awfully wrong & unnatural if they don't , IMHO very difficult to miss. Which leads me also to suspect they had buried themselves in completing the landing checklist, the ability to prioritise, seems to have been lost in that case. AVIATE NAVIGATE COMMUNICATE.

[S.F.L.Y]

BASIC situational awareness would lead you to expect the thrust levers to advance as speed stabilised, and it feels awfully wrong & unnatural if they don't

While I'm totally with you, Airbus seems to have another point of view when it comes to aircraft feedback (be it levers or artificial sensations). This "new" philosophy might well infiltrate Boeing's cockpit as I'm not sure all pilots have their hands on the levers at all time.

[chuks]

Some questions for S.F.L.Y.:

1. Do you hold any sort of pilot's licence/license and if so what is your level of experience? For instance, have you flown large aircraft into Schiphol?

2. What do you mean by the term "Vz" which you use repeatedly? Vx and Vy I am vaguely familiar with but my aeronautical knowledge does not extend beyond those to Vz, hence my question. Perhaps it is something that has come into use while I was distracted renovating my shed so please enlighten me and your other many readers here.

[BOAC]

z is the vertical axis about which yaw takes place.

Courtesy of Wiki
Coordinate systems

The position (and hence motion) of an aircraft is generally defined relative to one of 3 sets of co-ordinate systems:

• Wind Axes
  • X Axis - Positive in the direction of the oncoming air (relative wind)
  • Y Axis - Positive to Right of X Axis, perpendicular to X Axis
  • Z Axis - Positive downwards, perpendicular to X-Y plane

• Inertial Axes (or Body Axes) - based about aircraft CG
  • X Axis - Positive forward, through nose of aircraft
  • Y Axis - Positive to Right of X Axis, perpendicular to X Axis
  • Z Axis - Positive downwards, perpendicular to X-Y plane

• Earth Axes
  • X Axis - Positive in the direction of North
  • Y Axis - Positive in the direction of East (perpendicular to X Axis)
  • Z Axis - Positive towards the center of Earth (perpendicular to X-Y Plane)

For flight dynamics applications the Earth Axes are generally of minimal use, and hence will be ignored.

[ZimmerFly]

S.F.L.Y.

I suggest you read the full report (all 64 pages) of the Thomson "incident"


http://www.aaib.gov.uk/sites/aaib/publications/formal_reports/3_2009_g_thof.cfm

[S.F.L.Y]

Some answers for Chucks:

1 Yes I do hold some kind of professional licenses, I'm even above level 4. Still, I'm certainly not as professional and experienced as many of you since my medical class 1 doesn't include a prostate inspection.

2 Vz is not an indicated airspeed, it means vertical speed since it is related to the vertical axis kindly explained by BOAC. The Z axis is usually used by those eventually evolving in the third dimension.

[chuks]

It's chuks, actually...

But never mind that now.

What sort of AME do you use, anyway? None of mine ever insisted on a prostate exam and I can see how that might just put someone off a career in aviation. Well either that or else give you a good preview of how you shall be treated by management!

Oh, okay, Vz... Hmm, but wouldn't that put a whole other meaning on Vx and Vy, then? Perhaps we should use "rate of climb/descent" instead so that we all know what is under discussion here.

I look forward to reading the full accident report, which should be out soon, yes? Until then!

[S.F.L.Y]

Sorry for the misspelling Chuks...
In regard to the rate of descent (ROD), I guess that vertical speed (Vz) is widely used to be understood by the majority or at least self explanatory in the specific context it was used.

ZimmerFly thanks for the report, very interesting and so relevant!