In concluding the longest aviation accident investigation in its 32-year history, the National Transportation Safety Board said today that the probable cause of the crash of a Boeing 737 near Pittsburgh was a loss of control of the airplane resulting from the movement of the rudder surface to its blowdown, or maximum available, limit. The rudder surface most likely deflected in a direction opposite to that commanded by the pilots as a result of a jam of the main rudder power control unit servo valve secondary slide to the servo valve housing offset from its neutral position, and overtravel of the primary slide.
On September 8, 1994, USAir (now US Airways) flight 427, a Boeing 737-300, crashed while maneuvering to land at Pittsburgh International Airport. The flight was arriving from Chicago. All 132 persons aboard - 127 passengers and 5 crew - perished in the crash.
Confronted with an extremely fragmented aircraft and a flight data recorder with only 11 parameters of information, the NTSB began an investigation that involved two public hearings and an extremely long series of tests - including computer simulations, kinematic studies, wake vortex flight tests, and examinations of hardware from dozens of aircraft. Before today's final report, the Board had issued 20 recommendations to the FAA during the course of the flight 427 investigation dealing with the 737's rudder and rudder control system, flight data recorders and pilot training. Today's report includes 10 more recommendations on the same topics.
The Safety Board concluded that as flight 427 was approaching Pittsburgh at about 5,000 feet above the ground (6,000 feet above sea level), the aircraft encountered the wake of a preceding aircraft. As a result of the effects of the wake encounter, the Board believes, the pilot made a right-rudder input, but the rudder most likely reversed to a full left position, leading to the loss of control of the aircraft. The Board also concluded that the outer, secondary sleeve of the rudder power control unit's servo valve could have jammed to the valve housing, and when the rudder input was made, the inner, or primary sleeve overtraveled, leading to the reverse rudder movement.
This investigation also included examinations of two other events involving 737s, the 1991 crash of United Airlines flight 585 in Colorado Springs, Colorado, and the 1996 upset incident involving Eastwind Airlines flight 517 near Richmond, Virginia. The Board concluded that the same type of rudder reversal most likely occurred in those two cases, as well. The NTSB acknowledged that these reversals are transient and extremely rare events, and noted that the Boeing 737 has logged 92 million flight hours since its introduction 30 years ago and in that time carried almost the equivalent of the entire population of the world.
Since the crash of flight 427, Boeing has redesigned the PCU's servo valve and is retrofitting it onto all 737s. The Board said that the redesigned servo valve should preclude the rudder reversal failure mode that most likely occurred in these three cases, but that other potential malfunctions of the main PCU actuation system could exist. "[E]ven with these changes," the Board stated, "the 737 series airplanes...remain susceptible to rudder system malfunctions that could be catastrophic." The NTSB noted that a recent rudder event involving a Metrojet 737 with a redesigned servo valve is still under investigation.
The 737 is the only air carrier airplane with two wing-mounted engines that was designed with a single-panel rudder controlled by a single actuator, albeit with a dual-concentric servo valve design. Other rudder designs use multiple rudder surfaces and/or multiple rudder actuators. The Board therefore recommended in today's report that the FAA require that all existing and future 737s have a reliably redundant rudder actuation system. As is the Board's practice, it did not specify how best to achieve that goal, leaving that to the determination of the FAA.
The NTSB also recommended that the FAA convene an engineering test and evaluation board to conduct a failure analysis to identify other potential failure modes in the rudder system. The board should be made up of the FAA, NTSB technical advisors, the Boeing Company, other manufacturers as appropriate, and experts from other government agencies, industry and academia. Its work should be completed by March 2000.
The Safety Board again pressed the FAA to require more parameters on airline flight recorders, especially 737s. Four years ago, the NTSB recommended that the FAA require FDR upgrades on 737s by the end of 1995. The FAA's rule requires fewer parameters than the Board requested, and gives airlines until August 2001 to comply. As an example of how that rule does not encourage or require aggressive retrofit of FDRs, the Board noted that the Metrojet 737 that experienced an unexplained rudder movement last month will be undergoing a major maintenance check this Spring, but will not upgrade its flight recorder for another two years.
"The Safety Board concludes that the FAA's failure to require timely and aggressive action regarding enhanced FDR recording capabilities, especially on 737 airplanes, significantly hampered investigators in the prompt identification of potentially critical safety-of-flight conditions and in the development of recommendations to prevent future catastrophic accidents," the Board stated in its report. It further noted that "737 flight crews continue to report anomalous rudder behaviors, and it is possible that another catastrophic 737 upset-related accident could occur. If such an accident occurs before August 19, 2001, it is likely that the data recorded by the accident airplane's FDR will not be sufficient for investigators to readily identify the events leading to the upset and develop corrective actions to prevent similar accidents."
The Board's report will first be available in a few weeks on its web page, http://www.ntsb.gov under Publications. It will subsequently be available in hard copy for purchase from the National Technical Information Service, 5285 Port Royal Road, Springfield, Virginia 22161, (703) 487-4650. Please specify report number PB99-910401.