Birgenair 301

As regular readers of my blog know, I have a macabre interest in airline crashes. This probably stems for a healthy diet of such stories in my youth. Growing up as the son of a Boeing Project Manager and the grandson of a Pan Am Pilot will do that to you.

I saw a documentary about the sad case of Birgenair 301 the other day. The Boeing 757-225 had spent almost a month sitting on the Tarmac at the airport in Puerto Plata before being called into service in 1996. The Turkish crew was glad to finally be returning home, which may have been a factor in the Captain's decision not to abort takeoff when he and the co-pilot noticed that their airspeed indicators were disagreeing. Later, the investigation would determine that the pitot tube on the captain's side had been blocked and was giving false readings as a result. The co-pilot's system was correct all along.

At the point the two pilots noticed the problem, they still had room to abort, but instead the captain decided to proceed to acceleration and the take-off roll. Once in the air, he continued a string of mistakes that lead to the loss of the aircraft and 189 people. First, he activated the autopilot despite the faulty airspeed readings. By default the autopilot computer uses input from the captain's side pitot tube and therefore was misled. As the plane gained altitude, the pressure outside the blocked pitot tube went down, creating a differential between the inside and outside of the tube that was interpreted as data for the ASI (Air Speed Indicator). As they gained altitude the "speed" increased. The autopilot attempted to correct the problem by pulling the nose up and reducing power to the maximum of its authority.

In reality, the plane's speed was dropping precariously, but the pilots were being bombarded with warnings that they were traveling too fast. The pilots, confused about the source of the warnings, wasted precious seconds resetting the circuit breakers. This would only temporarily silence the alarms. Not trusting either ASI, the captain then made the critical mistake of pulling back on the throttles still further. This brought them to the edge of a stall. The computer sensed the stall and automatically disconnected the autopilot and made the pilot's control column shake as a warning of immanent stall.

Confused by the contradictory warnings (overspeed vs. stall) the captain failed to take action to avoid the stall. The co-pilot and a relief pilot in a jump seat behind the captain suggested that he maneuver to level flight. However, he maintained the nose-up attitude and the airplane began to enter the stall. Getting desperate, the pilot pushed the throttles to full power, but by that time the airspeed over the wing was too slow to support the setting and one engine failed. The asymmetrical power introduced yawl which exacerbated the stall. The airplane fell over on one wing and plummeted into the ocean. Everyone on-board died.

Several failures caused the accident. For one thing, the pitot tubes should have been covered to protect them on the ground. For another, the captain should have aborted take-off. Once in the air, he should have maintained awareness of what the aircraft was doing or delegate that task. Instead, he became preoccupied with the faulty instrumentation and failed to notice that the autopilot was taking them to the edge of the flight envelope. Then, once they were about the enter the stall, he ignored the advice of the two junior pilots. At that point, American crews are trained for the co-pilot to take over control. He should have taken hold of the control column and pushed forward. But remember, this was a turkish crew, and it is not culturally appropriate for a junior to usurp authority in this manner. These days, crews are trained in something called CRM (Crew Resource Management), which recognizes that crew members need to be smart about who pays attention to what as a crises looms. Too many people have died because pilots have been overwhelmed by the complexity of their aircraft.

To be fair to the captain, several crews put through a simulator failed this test. The contradictory warnings have a way of unnerving people. As a result, Boeing made changes in the warning systems and pilots are now required to go through at least one blocked-pitot tube scenario in simulator training.

That makes me feel better.

-t