UPDATE : Faulty Rudder Control Suspected in Crash of USAir Flight 427 : Experts say it’s the only system that could have caused the jet to crash as it did in September. But parts tests find no defects.

In solving a crime, prosecutors must not only identify the suspect, they must explain how he did it.

In figuring out what caused the crash of USAir’s Flight 427 last September, federal investigators must do the same thing: not only identify their prime suspect--now revealed as the Boeing 737’s rudder--but explain what might have caused the rudder to deflect fully to the left, sending the plane into a spiraling dive that killed all 132 people on board.

After a weeklong National Transportation Safety Board hearing here that included more than 40 hours of expert testimony from USAir, Boeing, the Federal Aviation Administration and others, there seems to be general agreement on one point:

The only control system on the jetliner that could have caused it to swivel, roll and dive as it did was the rudder--the hinged back half of the vertical part of the tail that helps the plane turn right or left by shoving the tail in the opposite direction.

Assuming the rudder is involved--and because of limited data, the evidence is still purely circumstantial--that leaves two possibilities: Either the cockpit crew deliberately stomped on the left rudder pedal, or the rudder system malfunctioned, grossly distorting the pedal commands and slamming the rudder to the left all by itself.

It seems unlikely that the captain, Peter Germano, 45, or the first officer, Charles B. Emmett III, 39, suddenly pushed down hard on the left pedal.

Both were experienced pilots with good performance ratings. Few think either of them would have deliberately initiated the sort of violent maneuver that would throw the plane out of control.

So the experts looked at the complex workings of the rudder system--a Rube Goldberg tangle of cables, pulleys, rods, yokes and torque tubes, all hooked to servo units and hydraulically powered actuators.

Although tests, computer simulations and studies of previous mishaps revealed several scenarios that might explain how the rudder system caused the crash, evidence from the wreckage seems to show that none of the scenarios fits.

But the Boeing 737 rudder system is beset with problems.

Boeing records include at least 187 instances in which a 737 rudder system has malfunctioned, often causing the rudder to move on its own. While most of these incidents were minor, some caused planes to roll, yaw or dip violently.

Although no rudder system malfunction has ever been blamed conclusively for the crash of a Boeing 737, some experts think a rudder glitch may have downed United Airlines Flight 585, a 737 that crashed in Colorado Springs, Colo., in 1991. The cause of that crash has never been determined officially, but like Flight 427, Flight 585 suddenly swiveled to one side, rolled over and slammed nose-first into the ground.

NTSB investigators in Colorado looked carefully at the shattered wreckage of Flight 585’s rudder control system. There are three principal components: the main power control unit, the standby power control unit and the yaw damper.

Responding to commands from the rudder pedals that are relayed back by the linkage system, a servo valve in the main power control unit releases pressurized hydraulic fluid, driving a piston that deflects the rudder left and right, a maximum of 26 degrees in either direction. The rudder movement is relayed back through the linkage system, shutting off the servo so the rudder will not continue on past the point commanded by the rudder pedals.

If the main power control unit fails, the cockpit crew can switch over to the standby unit, which works the same way.

Because a jetliner tends to fishtail slightly when cruising at high speed, the Boeing 737 is also equipped with a yaw damper. The damper, responding automatically to commands from a gyroscope, eliminates the fishtailing by correcting it with rudder deflections of up to 3 degrees in either direction.

All three components are interconnected through the complex linkage system. NTSB investigators found that when one of them jams, that affects the way the others operate.

Inspection of Flight 585’s standby rudder system showed that its input shaft was binding. NTSB investigators said that because of the interconnecting linkage system, this binding impaired normal shut-off feedback from the rudder.

This condition, they said, could cause the rudder to deflect fully, even though the pilot was pushing the rudder pedal only slightly. But they said that if that did happen, the pilot should have been able to correct for the deflection by using the plane’s other controls.

Several months after the crash of Flight 585, the rudder of a Boeing 737 taxiing at Chicago’s O’Hare International Airport jammed. Inspection of the main power control unit showed that parts had been machined improperly.

Tests showed that under some circumstances, the Chicago unit could reverse on its own. That raised the terrifying prospect that a pilot trying desperately to apply right rudder in an effort to correct a dive to the left would instead be applying left rudder that intensified the problem.

Similar tests showed that a unit from a Sahara India 737 could reverse because it was repaired with improper parts. Tests of other malfunctioning units showed that severely contaminated hydraulic fluid could cause those units to jam, leading to excessive rudder deflection. And tests of binding yaw dampers from other planes showed they could cause the rudder to deflect beyond 3 degrees.

But examination of the rudder control components from Flight 427 showed they were machined properly, contained the proper parts, were free of severe contamination and did not bind, Boeing said.

So the mystery remains.

Solving it has implications far beyond the crashes in Colorado and Pennsylvania. The 2,600 Boeing 737s aloft have carried more than 4.2 billion passengers--the equivalent of half the world’s population--and if there is something fundamentally wrong, it could force the grounding of the most widely used air transport plane in history.

“We want as badly as anyone else to find out if there is something wrong with that plane,” Boeing spokesman Jack Gamble said. “We need to get it fixed, and get it fixed now.”

(BEGIN TEXT OF INFOBOX / INFOGRAPHIC)

The Prime Suspect

The rudder pedal sends a complex linkage system into motion, eventually moving the rudder left or right. If the main power control unit fails, the cockpit crew can switch to a standby unit. The third component in the linkage system, a yaw damper, eliminates the normal fishtailing that occurs during high speeds.

HOW IT MAY HAVE HAPPENED

1) Rudder apparently deflects to the left . . .

2) . . . rolling the plane left and down . . .

3) . . . and into at least one revolution . . .

4) . . . before the plane hits the ground nose-first.

Source: National Transportation Safety Board