On March 3, 1991, United Airlines Flight 585 departed Denver’s Stapleton International Airport at 9:23 a.m. for a routine flight to Colorado Springs Municipal Airport.With 10,000 flight hours, Capt. Harold Green, 52, was in command of the Boeing 737. The co-pilot was Patricia Eidson, 42. Three flight attendants and 20 passengers were also on board. It was a sunny morning, with winds at 13 knots, gusting to 23 knots.Flight 585 turned for a visual approach from the south. The tower cleared the flight for landing on runway 35 at 9:43 a.m. The plane then suddenly rolled sharply to the right and pitched nose down until it crashed in Widefield Community Park less than 4 miles from the runway, digging a crater 10 feet deep. There were no survivors.Stretched thin from a fire at a retirement home that killed 10 elderly women just hours before, emergency responders and volunteers combed the debris field. When found, the flight data recorder showed that 45 seconds before the crash, the plane’s airspeed was consistent with a normal final approach at 1,000 feet. Green was at the controls with Eidson handling communications. When the roll started, Green attempted a go-round by setting the flaps at 15 degrees and increasing thrust to 4Gs. The plane continued to dive, traveling at 245 mph.The National Transportation Safety Board’s investigation started immediately. Two theories emerged:A freak gust of wind flipped the plane into a nose dive. The swirling wind would have had to strike the plane head on and turn earthward at just the right moment to drive it into the ground. Boeing provided data and a computer simulation explaining just how a rotor wind might have caused the crash.A problem with the plane’s rudder power control unit: The rudder is the hinged part of the tail that controls a plane’s horizontal movement. Just a week before the crash, the plane had turned inexplicably to the right while in flight. The flight crew switched off a control called the yaw damper, and the plane returned to normal flight. Mechanics replaced the yaw-damper coupler and returned the plane to service. Two days later, another flight crew reported the plane’s rudder again moving to the right for no reason. This time, mechanics replaced a valve in the yaw damper and returned the plane to service. Four days later, the plane crashed.The NTSB relied on representatives from Boeing, which designed the rudder, and Bertea Corp., a division of Parker Hannifin, which manufactured it, to test Flight 585ís mangled rudder parts. The parts were shipped to United Airline’s San Francisco maintenance center.The yaw damper was crushed and untestable. The rudder PCU was severely burned, its main rod bent, and its internal parts stuck.When they took the rudder PCU apart, the investigators focused on the servo valve, a cylindrical metal slide positioned inside a slightly larger slide. The slides work together to direct the proper flow of pressurized hydraulic fluid that moves the rudder.The team decided to test the servo valve at Bertea’s Irvine, California, facility. The parts were packed, and a Bertea engineer hand-carried them there. When the package was opened, a spring, spring guide and end cap were missing. The three parts are crucial for channeling just the right amount of pressurized hydraulic fluid, at just the right moment, to move the rudder in the direction and to the degree commanded by the pilot or the automatic yaw damper.The team decided to smooth the servo valve’s interior housing walls and the exterior of the outer slide, creating like-new surfaces. They replaced the missing parts with new parts. Although the reassembled servo valve didn’t work perfectly, the investigators concluded there was no evidence showing the rudder PCU had caused Flight 585 to crash.The NTSB’s report, issued Dec. 8, 1992, said it ìcould not identify conclusive evidence to explain the loss of United Airlines Flight 585.î It was one of only four unsolved cases in the NTSB’s 27-year history.Then, on Sept. 8, 1994, another Boeing 737, USAir Flight 427, crashed on approach to Pittsburgh International Airport. All of the 132 people on board were killed.This time, the plane rolled to the left, stalled and rolled upside down. It then rolled upright but continued to roll while pitched nose-down. Traveling at 300 mph, it slammed into the ground. Boeing argued that Flight 427ís pilot had overreacted to wing turbulence from the plane landing ahead of it, or perhaps had suffered a seizure that locked his foot on a rudder pedal.When tested, Flight 427ís rudder PCU performed perfectly.Frustrated, in August 1996, NTSB investigators gathered top hydraulic engineers to test Flight 427ís rudder PCU and a new one. They would try to cause a jam by testing at gradually colder temperatures and injecting 170-degree fluid at each test point to simulate flying conditions.With both PCUs passing the higher temperature tests, engineers started to pack up and head home ñ- until Flight 427ís PCU jammed at 40 degrees below zero. The NTSB had proven the PCU could jam.Weeks later, a Boeing engineer was reviewing the hydraulic engineers’ PCU test data when he noticed unexpected fluctuations in hydraulic flow, and concluded that if the PCU jammed, the rudder would reverse. That is, if the pilot pushed on the right rudder pedal, the rudder would go left ñ- a catastrophic event if it were to occur at low altitude while on approach.Boeing tested a new 737. In the first tests, the rudder performed as it should, but when the test pilot aggressively stomped on a rudder pedal as fast as he could, the rudder swung in the wrong direction.Boeing presented the results directly to the Federal Aviation Administration, proposing to issue a bulletin requiring airlines to test rudders for jams, train pilots on how to deal with rudder problems and retrofit existing aircraft. In the long term, Boeing would redesign the servo valve. The FAA agreed.The NTSB’s final report on the crash of Flight 585 in Colorado Springs cites the same probable cause for the crash of Flight 427 in Pittsburgh: ìa loss of control of the airplane resulting from the movement of the rudder surface to its blowdown 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.îAt Widefield Community Park, the names of Flight 585ís crew and passengers are inscribed on a plaque that rests under a gazebo in honor of those who lost their lives that day in March 1991.Sources: Seattle Times and St. Petersburg Times
