An Eastwind Airlines Boeing 737-200 has been grounded for nearly three weeks as officials investigate two recent flights disrupted by rudder problems.
Though no one was hurt, the problem flights closely parallel events surrounding the March 3, 1991, crash of a United Airlines 737-200 in Colorado Springs and the Sept. 8, 1994, crash of a USAir 737-300 in Pittsburgh. Eastwind is a fledgling, low-cost carrier based in Trenton, N.J.
The rudder is the hinged panel on the vertical tail section that controls an aircraft's left-to-right direction of travel. The Eastwind incidents are the latest of hundreds of cases where pilots have reported rudders moving inadvertently in flight, causing the aircraft to swerve or roll briefly off course, including more than 50 reports in the 22 months since the Pittsburgh crash.
Rudder malfunctions are widely suspected of causing the Colorado Springs and Pittsburgh crashes, which remain unsolved by the National Transportation Safety Board. There were no survivors of either crash - a total of 157 people died.
A team of investigators from the NTSB, Boeing, the Federal Aviation Administration, USAir and Eastwind is examining and testing the grounded Eastwind jet at USAir's Greensboro, N.C., maintenance base. USAir is involved because Eastwind, which began operations last August, subcontracts with USAir for maintenance services and crew training.
So far, the investigators are stumped.
"Of course we are aware of the Colorado Springs and Pittsburgh accidents, and we're going to keep this airplane on the ground until we determine what's wrong with it," said Stephen Davis, Eastwind's vice president of operations. "We would be delighted if this turns out to help solve these two unresolved accidents. Maybe this will lead to an answer."
Eastwind flies two 25-year-old Boeing 737-200s leased from General Electric.
On the morning of May 10, aircraft No. N221US took off in clear skies over Trenton with 102 people on board, bound for Providence, R.I. About 10 minutes into the flight, leveling off at 10,000 feet, the pilot felt an "abrupt nose right movement, similar to a jab on the right rudder," according to an NTSB report.
When the rudder jab happened again a few seconds later, the pilot turned off the autopilot computer and an automatic rudder-adjustment mechanism, called the yaw damper, and turned back to Trenton. The errant rudder movements occurred at least twice more as the plane descended into the airport.
According to The Times of Trenton, passengers said the plane seemed to be swinging from side to side for most of the short flight. One passenger cried hysterically and another screamed obscenities, the newspaper reported. The landing was uneventful.
USAir replaced the jet's rudder power-control unit, or PCU, an assembly of valves, linkages and filters that directs pressurized hydraulic fluid to move the rudder. But the replacement PCU failed a functional test and was replaced by yet another PCU, said Julie Beal, an NTSB spokeswoman.
The two PCUs were shipped to the manufacturer, Parker Bertea Aerospace, of Irvine, Calif., for testing, and the plane was returned to service the next day, May 15.
Then, on June 9, the same Eastwind pilot was at the controls of N221US descending into Richmond, Va., at about 5,000 feet when the plane unexpectedly rolled 20 to 30 degrees to the left, according to a report the pilot filed with his superiors. The pilot straightened the jet out, and landed safely.
"He got unexpected rudder movement and rolled a little bit up on the right wing," said Eastwind's Davis.
The plane has been grounded and in the hands of investigators ever since. Beal said the probe has been hampered by a crude flight data recorder, which indicates the plane rolled no more than 10 degrees to the right. She could not explain the discrepancies between the data and the pilot's report.
"We're collecting data, it's an ongoing thing," Beal said. "Our people are looking at these very same issues."
Beal said investigators are examining the possibility that a malfunctioning yaw damper is behind both incidents. The PCU is under close scrutiny as well. After the second incident, the NTSB took custody of the two PCUs that had been sent to Parker Bertea after the May 14 incident.
Boeing, in fact, has long characterized many of the disrupted 737 flights reported by pilots as likely being caused by yaw-damper glitches that should be easily countered by the pilots. By design, the yaw damper under most conditions can move the rudder only a few degrees in either direction.
More recently, the company has said pilots could be mistaking encounters with wingtip turbulence from jets flying ahead, as rudder problems. Boeing declined to comment on the Eastwind incidents, deferring to the NTSB.
However, evidence has emerged since the Pittsburgh crash suggesting that the 737 rudder is susceptible to errant rudder movements that can quickly twist a plane into the kind of nose dives that led to the Colorado Springs and Pittsburgh crashes.
USAir, which trains Eastwind's pilots, has been a leader among airlines changing the way they fly 737s on landing approach to give pilots a better chance to counter an errant rudder. USAir also is among the first airlines to begin teaching pilots special maneuvers to recover from a nose dive caused by an errant rudder.
The grounding of the Eastwind jet also marks a new approach.
Twice in early 1991 - just six days and again three days before the Colorado Springs crash - crews flying the ill-fated United 737 reported uncommanded rudder movements. Each time a yaw-damper part was replaced, and the plane was immediately put back into service.
The Colorado Springs crash occurred on a windy day, as the jet twisted right and dropped straight down, falling several hundred feet in nine seconds. Investigators from Boeing and the NTSB were unable to extract conclusive evidence from the wreckage that the yaw damper or other rudder part failed, causing the crash.
The Pittsburgh crash took place as the USAir jet was descending on a picture-perfect September afternoon. At 6,000 feet, the jet wobbled briefly, then twisted and dived left for 23 seconds. It smashed into a wooded ravine at more that 300 mph. Again, investigators from Boeing and the safety board could not find evidence that convinced them the rudder was at fault.
While Boeing and safety authorities maintain there is no undue safety hazard from the rudder, some aviation safety experts are starting to call for another review.
Bob Besco of Dallas, a former airline pilot who is now an independent aviation consultant, believes rudder problems are being reported frequently enough to warrant a mandatory redesign of the 737 rudder, something authorities have been discussing behind closed doors.
"They're playing Russian roulette with this thing," Besco said. "Even one malfunction a year is too high because you never know when it's going to persist and lock the rudder."