Dutch Safety Board cites 'convergence of circumstances' in 2009 THY 737 crash

By Cathy Buyck | May 7, 2010

The Feb. 25, 2009, crash of a Turkish Airlines 737-800 on approach to Amsterdam Schiphol was caused by a malfunctioning radio altimeter, a non-stabilized approach, poor reactions from the pilots and insufficient directions from air traffic control, the Dutch Safety Board concluded in its final report released yesterday.

Nine persons were killed and a further 120 injured after the aircraft crashed short of the runway during final approach. "That the accident could happen was the result of a convergence of circumstances," the board said. "These circumstances could only have resulted in the accident happening because of their mutual interaction."

While executing the approach by means of the ILS with right autopilot engaged, the 737's left radio altimeter showed an incorrect height of minus 8 ft. on the left primary flight display. This resulted in activation of the autothrottle's "retard flare" mode whereby the thrust of both engines was reduced to a minimal value (approach idle) in preparation for the last phase of the landing. Owing to the approach heading and altitude provided to the crew by ATC, the localizer signal was intercepted 5.5 nm. from the runway threshold rather than at least 6.2 nm., with the result that the glideslope had to be intercepted from above.

"This obscured the fact that the autothrottle had entered the retard flare mode. In addition, it increased the crew's workload," investigators noted. When the aircraft descended to 1,000 ft., the approach was not stabilized "so the crew should have initiated a go-around," the board said. The right autopilot (using data from the right radio altimeter) followed the glideslope signal. As the airspeed continued to drop, the aircraft's pitch attitude kept increasing. "The crew failed to recognize the airspeed decay and the pitch increase until the moment the stick shaker was activated," according to the report. "Subsequently the approach to stall recovery procedure was not executed properly, causing the aircraft to stall and crash."

The board said its investigation did not uncover a reason for the altimeter malfunction but stated that "the failure of radio altimeter systems in 737-800 aircraft has a long history." Boeing "reasonably" could have realized that the problem, particularly the effect on the autothrottle, could have had an impact on safety, it said. It is recommending that the manufacturer improve the reliability of the radio altimeter system and that it review its "approach to stall" procedures with regard to the use of autopilot and autothrottle.

The report also was critical of AMS controllers, who "must observe the procedures as they are described in the rules" and not engage in individual interpretation. "The given heading instructions did not enable the aircraft to be in level flight on the final approach track before intercepting the glidepath from below," the board said. "This is not in line with the ICAO guidelines for this type of approach."

THY said it disagrees with "certain aspects" in the DSB report, in particular two conclusions: "Approach stabilization is not a factor in the causation of this accident and it is claimed by the report that the crew could have recovered the aircraft after the stall warning was received. However, even though the crew promptly reacted, autothrottle kicked back unexpectedly. The second attempt by the crew, after disengaging the autothrottle, to advance thrust levers was successful but too late."

Discuss this news 8

07 May15:51

If they were intercepting the

By Syed Mohammad Husain

If they were intercepting the glideslope from above with the vectoring provided by ATC, the autothrottles would be in idle position, but later, speed will only decay with increase of pitch attitude if the altitude is hold is on or a misreading of an altitude of minus 8 feet provided by the radio altimeter indicating flare mode.
The crew could have been deceived as mentioned in the report about the radio altimeter malfunction.

08 May00:35

Can someone explain to me why

By Matt K.

Can someone explain to me why a value of -8 feet from a radio altimeter is not rejected by the autothrottle or flight control computer as "implausible", and thus ignored? It seems to me that a little bit of code checking the reasonableness of incoming data could have helped prevent this.

09 May01:47

Much of Holland is below sea

By Jaime Vazquez

Much of Holland is below sea level, and in particular Schipol airport is at -11 feel level (i.e., below sea level). So an altitude reading of -8 feet would indicate the aircraft was 3 feet (1 meter) above ground, which is possible, although apparently incorrect in this case.

09 May02:34

Dear Much Of Holland, The

By Anonymous

Dear Much Of Holland,

The altimeter is a radio altimeter, therefore it will read the height below the aircraft irreveant of the airport elevation.

09 May02:34

Dear Much Of Holland, The

By Anonymous

Dear Much Of Holland,

The altimeter is a radio altimeter, therefore it will read the height below the aircraft irreveant of the airport elevation.

09 May12:30

For Syed who wrote; "speed

By Mark

For Syed who wrote; "speed will only decay with increase of pitch attitude if the altitude is hold is on..." Speed will also deteriorate (as was the case here) with glide-slope capture and the thrust still at idle. To not noticed increasing pitch, deteriorating airspeed (with tell-tale decreasing wind noise) is poor situational awareness - regardless of a failed radio alt. Couple that with a poorly executed stall recovery (no AT's disconnect until the second attempt) and this was the tragic result. ATC and radio altimeters played a very small part imo.

11 May10:10

As Mark said, the primary

By David Connolly

As Mark said, the primary cause of this THY accident was a general lack of tactile and situational awareness of the flight crew. This was doubly ironic with the presence of a "Safety Pilot" in between the operating crew. It is also ironic that this B-738 NG has a lot of OG/Old Generation technology, thanks to Boeing's shortsighted bowing to one customer-operator, namely Southwest, who are only just begining to use autothrottles and autobrakes. Quite a Texan "my bull-way or the highway" attitude, with predictable eventual results,like Chicago Midway on December 8 2005.
THY lost a B-737-400 on April 7 1999 after Take-off from Adana, Turkey in convective conditions with the Pitot heat off. Every Boeing since the B-757/767 has had switch selectable critical functions, automated, once the first engine was started. It is therefore foolproof. Irony struck again with the AF-447 loss of June 1 2009. This A-330 was also thus foolproof configured. The better fool that came along, as they do, after several similiar cascading failures and a subsequent Northwest/Delta event near Tokyo on June 23 was the Thales pitot probe's heating element. The relic B-738NG bias to the Captain's default position of the Captain's RadAlt must be right on the left is a dangerous tradition of the historic Captain God. In compliance with this tradition, the nose gear tiller is only operable from the captain's side. Tradition of economy has an eventual price to pay. What if the captain slumps dead on taxi ?, a lot of awkward shuffling on the F/O's part, at best. A pilot is singular, a crew is plural. The crew should be greater than the sum of it's parts.
The latest THY accident illustrates, what I call "Latent Failures" of tradition carried forward. Tradition is the spoilt child of bad habit.
The sooner Boeing adapts the B-787 attitude carried from the B-777 and B-747-400 (which I fly), the better. Incest is best, in Family Guy tradition of absurd sanity, in the cross fertilisation of the B-747-400 and B-777. "NO OPERATIONAL LIVES LOST ON BOTH TYPES SINCE 1989", the B-747-400's debut operational year with Northwest. This illustrates a true evolutionary symbiosis of both models. The B-787 is the pinnacle of human perfection-evolving, in that all options were added with a Co-Command cockpit structure from the B-744/777 tradition, evolved with all options not optional. If backward luddite operators like Southwest, chose to fly in a fog of ignorance, that is their option and customers, Boeing has no longer made it their option by combining a Henry Ford Model T mantra with economics of mass production.
Human pilot operators fail every way in every day and mostly through CRM, trap those errors safely. But the Old- New Gen B-737 is the plane that time forgot in the forward "Flare-Retard" PFD A/T mode NOT, Retard-Flare(Very heavy landing guaranteed) and overhead backward panel will keep the error trap open to catch the unwary as it did in Amsterdam on Feb 25 2009. Hopefully Boeing's B-737 Next Generation, will actually be a new generation born of the B-787. Hope springs eternal, like intercepting the Glideslope from below.

13 May02:33

As a computer person, I

By James

As a computer person, I wonder why don't they introduce some checks so that if the Right and Left Altimeter are producing major differences, it warns the crew that something isn't right.