Faulty Altimeter Downed Boeing 737, Investigators Say

[Barb-SAN]

http://www.avweb.com/eletter/archives/avflash/1326-full.html#199905 When a Turkish Airlines 737-800 crashed short of the runway in Amsterdam last week, killing nine people, it was one of those relatively rare accidents where no apparent cause was readily discernible. But on Wednesday, Dutch Safety Board investigators said (http://www.onderzoeksraad.nl/en/index.php/onderzoeken/onderzoeksraad-start-onderzoek-crash-turkish-airlines-op-schiphol/) a faulty radio altimeter fed misinformation to the autopilot. The altimeter registered that the airplane was approaching ground level when in fact it was still at 1,950 feet. The autopilot initiated a power-down of the engines in preparation for landing and the airplane slowed to near stall speed. The pilots responded to the situation too late, and the airplane hit the ground. The 737 data recorder showed that problems with the altimeter had occurred twice before, investigators said.

Boeing has issued a statement to 737 operators reminding pilots to carefully monitor instruments during critical phases of flight. The Dutch Safety Board said in its report that it is "of the opinion that extra attention is needed for the role of the radio altimeter when using the automatic pilot and the automatic throttle system." The board asked Boeing to pay extra attention to a part of a manual for the Boeing 737 in which is stated that in case of malfunction of the radio altimeter(s), the automatic pilot and throttle system that are connected to it may not be used for approach and landing. The board would like Boeing to consider an investigation into whether this procedure is also applicable during other phases of flight.

Any thoughts on this, Capt. Hutch, as a 737 pilot?

[MathFox]

:coffee: (Unconfirmed) rumours in Dutch media (http://www.nu.nl/algemeen/1928312/boeing-waarschuwde-luchtvaartmaatschappijen.html) say that Boeing sent out a bulletin on Feb 11th recommending replacement of a connector in the radar altimeter that is susceptible of corrosion in moist conditions.:shocked:

Design and maintenance issues... *check*
High cockpit workload.... *check*

That's two of the main ingredients for an accident.

[Barb-SAN]

Here's another article with more details (in English ;), though not about the corrosion issue).
http://seattlepi.nwsource.com/business/402229_plane05.html

[Captain Hutch]

Barbsan,

This is an interesting occurrence, though I'm sure that we don't have all the data yet as far as to what other systems might have failed or not. However, from a pilot's standpoint I can highlight a few areas. In almost all cases of low visibility or low clouds or both the autopilot is engaged and performs the approach. In cases of very low visibility, i.e. down in the 1000 foot range and lower, the autopilot performs an autoland if the aircraft, airport, and pilots are so trained and equipped. Using the autopilot greatly reduces the workload so that basically all the pilots have to do is monitor the aircraft so that all systems perform normally. NOTE: recall the thread somewhere in here that discusses pilotless aircraft. This is the type of insidious failure that required the pilots to act, although it appears they were too slow, though we don't know the reasons yet. However, during my own autopilot approaches I keep my hands on the throttles, as required by my own airline. At the same time the copilot also keeps his hands near the throttles so as to detect any abnormal movement. Once during an approach the throttles moved abnormally and I performed a manual go-around. My main crosscheck during the auto approach is on basic instruments like airspeed, altitude and rate of descent. Why these pilots missed what seems to be several clues, I can't really explain right now, but the cockpit recorder hopefully will reveal a few more clues as to what went wrong.

Thanks,

Hutch :tiphat: