Guest columnist
A scary pattern of malfunction with Airbus rudder system

By Lee Gaillard
Special to The Times

AP An investigator climbs onto wreckage of American Airlines flight 587 Nov. 14, 2001, in Rockaway Beach, New York.

E-mail this article Print this article Search archive Most read articles Most e-mailed articles

Remember Hal, that sinister computer that tried to take over the spaceship in the classic sci-fi film "2001: A Space Odyssey"? We now face the same threat on some commercial jetliners, where dangerous, uncommanded — not pilot-initiated — rudder movements have disrupted flights.

Just such an event may have ripped the tail fin off an Airbus A300-600R (American Airlines Flight 587 from New York) in November 2001, killing 265 people.

Last Tuesday, however, the National Transportation Safety Board laid primary blame for that disaster on "unnecessary and excessive" pilot actions.

Yes, First Officer Sten Molin was at the controls as the A300 entered mild wake turbulence from an earlier 747 departure. But did Molin himself command ensuing violent rudder movements? Or were they the latest in the dangerous series of more than 30 major uncommanded rudder deflections and faulty autopilot inputs initiated autonomously by A300 computerized flight-control systems?

Take this near-disaster ignored by the NTSB: In January 2002, a Caracas-bound A300 "experienced significant uncommanded rudder inputs" as it climbed through 10,000 feet. "Accelerating through 290 knots [about 333 mph], the pilots experienced smooth, uncommanded yawing," severe enough to force buckling and popping sounds from passenger exit doors, according to the American Airlines A300-600 Pilots' Report on A300 uncommanded rudder incidents. The pilots made an emergency landing back at Miami.

They were lucky. Given a takeoff profile closely matching Flight 587's (with no dissipating wake encounter), this A300's system-induced rudder malfunctions could have replicated the Flight 587 catastrophe. During its almost three-year investigation, the NTSB ignored such examples meticulously documented by American Airlines A300 pilots.

Despite requests from representatives of pilots flying roughly 115 A300s for freight and passenger airlines in the United States, why did the Federal Aviation Administration not ground these aircraft for thorough checks of tail fins, wiring and flight-control system software?

No wonder American Airlines A300 pilots requested assignment transfer to non-Airbus aircraft after such incidents as:

• January 1990: An American Airlines A300 diverted to Bermuda following "continuous uncontrollable rudder deflections" and multiple system failures; on landing, a severe uncommanded yaw almost forced it off the runway. After takeoff from John F. Kennedy International Airport in December 2002, this same aircraft suffered uncommanded "sharp and abrupt" rudder movements and was ordered to return to the airport.

• January 2002: Over a three-day period, one American Airlines A300 underwent successive uncommanded rudder jolts (one "moved the whole aircraft 5 or 10 feet from side to side") and experienced a yaw damper (which controls the rudder) that tripped and would not reset. Two months later, "at least one major uncommanded rudder movement" rocked this aircraft during final approach to Miami.

With such "in-flight anomalies" recalling the National Aeronautics and Space Administration's negligence concerning space shuttle foam strikes that doomed the Columbia and its crew, one engineer asserted that the A300's many malfunctions constitute "a remarkable testament to a troubled rudder system." Indeed, Airbus admitted that by October 2002, it had quietly modified its control-system software for "passenger comfort"; uncommanded rudder events decreased thereafter. Why did the NTSB not connect the dots and demand meticulous analysis of earlier software versions?

Furthermore, why did we only just learn of an Airbus engineer's June 1997 memo warning that "rudder movement from left limit to right limit will produce loads on fin/rear fuselage above ultimate design load" — information not shared with airlines and pilots four years before the Flight 587 disaster? (And why, antithetically, did Airbus A300 operating manuals say that to eliminate "unlocked landing gear unsafe indication," pilots should "perform alternating side slips in an attempt to lock the gear" — maneuvers not possible without cycling the rudder from left to right?)

Scary.

Preparing for departure that November morning, Flight 587's pilots reported yaw damper and pitch-trim malfunctions — "a common problem on A300 aircraft," according to one technician. It was the plane's 11th computer-linked pitch-trim problem within a year. Maintenance personnel reset a circuit breaker. Underlying causes remained undiscovered and unresolved.

Although the previous day's aircraft logbook recorded a significant problem with its computerized flight-management system, Airbus denies that "possible flight-control system failures" could have caused the accident.

We know the rest.

"The pilots of Flight 587 were my friends," one American Airlines first officer told me. They deserve better — and they deserved a more complete answer from the NTSB. So did families of crew members, of passengers and of victims on the ground. So did 47 airlines still flying the A300 worldwide.

Lee Gaillard is a freelance writer based in Philadelphia who writes frequently on defense and aviation issues.