737 rudder evidence

From:         baco-new@seatimes.com (Byron Acchido)
Organization: The Seattle Times
Date:         18 Sep 95 12:07:05 
View raw article
  or MIME structure

Published Sept. 8, 1995 The Seattle Times pp. A-1

Questions remain in two 737 crashes

    In two crashes of Boeing 737s _ including one a year ago today _ a
mechanism that controls the rudder was immediately suspected. After tests,
investigators said there was no evidence that the mechanism had failed in
either crash. But questions remain whether those tests were done right.

by Byron Acohido
Seattle Times aerospace reporter

     On a blustery spring day four years ago, two United Airlines pilots
banked their Boeing 737 jetliner for final approach into Colorado Springs
when the plane suddenly flipped and slammed nose first into the ground.
All 25 on board were killed.
     Because of the sudden flip, investigators immediately suspected the
rudder _ the large slab on the vertical tail section that controls a jet's
direction _ of playing a role in the disaster. Over the past three
decades, pilots have reported hundreds o f incidents of 737 rudders moving
inadvertently and causing aircraft to swerve temporarily out of control.
      From the wreckage, investigators retrieved a badly mangled power
control unit (PCU), a complex assembly of pistons, levers, cranks, shafts
and valves that directs pressurized hydraulic fluid to move the rudder.
      The PCU's valves were jammed so severely by fire that investigators
from Boeing and the PCU manufacturer, Parker Bertea Aerospace, had to
pound them free with a hammer and chisel-like tool. Inside, they found
moisture, traces of a mysterious white p owder and stringy, bronze-like
chips.
     But no analysis of the contaminants was conducted, despite
manufacturer's records dating to 1965 showing that dirty hydraulic fluid
can jam the PCU and cause severe uncommanded movements of the rudder.
      Greg Phillips, the National Transportation Safety Board's rudder
systems expert, said investigators at the time didn't think the debris was
important.
      Investigators did test the PCU itself, but not before many key parts
were misplaced, replaced or rebuilt to factory-new tolerances. The part
then operated smoothly enough@noCQ@ro for investigators to conclude that
there was no evidence of malfunctio n.
    Twenty-two months later, the National Transportation Safety Board
(NTSB), responsible since 1926 for determining the cause of airplane
accidents, conceded that United Flight 585 was the first crash of a modern
jetliner it could not solve.
      The second might be the crash of USAir Flight 427, still unsolved
after a disastrous end a year ago today.
      Last Sept. 8, two USAir pilots flew a 737 packed with business
travelers, families and vacationers into clear, calm skies over
Pittsburgh.@bo @roFlight 427 was in a routine descent when it suddenly
flipped and screamed nose first into a wooded ravin e. All 132 on board
died instantly.
      Investigators again suspected the rudder, and this time treated the
PCU more methodically, videotaping each step in handling the part. Before
testing the PCU, Parker engineers replaced the main shaft and all external
linkages and levers, because the
 part was too damaged to operate.
      Lab tests were run and the rebuilt part operated smoothly. Only then
did investigators disassemble the PCU. Inside, they found minute particles
of steel and aluminum-nickel-bronze.
      This time, investigators tested to see whether a PCU like the one
used on 737s could flush very high concentrations of debris through its
valves, yet continue to operate smoothly. Tests directed by a Boeing
engineer found that it could.
     However, an independent hydraulics expert now says those tests were
conducted under conditions unlikely to occur in flight.
      Still, based on analysis of the rebuilt PCU and contaminants
recovered from inside the device, NTSB Chairman Jim Hall says there is no
evidence the USAir jet's PCU jammed.
      William Laynor, acting NTSB director of aviation safety, emphasized
that a jammed PCU has yet to be linked to any of the 61 737s that have
crashed since the jet entered commercial service in 1967. He attributed
the hundreds of reports of rogue rudde r movements to routine electrical
malfunctions.
     Twenty-one 737s have crashed in the 1990s, including three since the
Pittsburgh accident. Laynor acknowledges that if dirty fluid jammed both
the PCU's internal valves, ``you're going to have a major controllability
problem.'' But he said no correcti ve measures are under consideration.
     ``We haven't seen that (PCU jams) to be a realistic danger in all the
service history of this airplane,'' Laynor said.
     Boeing spokesman Steve Thieme said the 737 remains statistically
among the safest jet models flying. Thieme contends that rudder system
testing in the investigations of the Colorado Springs and Pittsburgh
crashes legitimately found no evidence of the
 PCU's involvement.
      The stakes are high with the 737, the most widely used jet in the
world, with more than 2,600 in service. In the past two years alone,
pilots have reported more than 200 cases of inadvertent 737 rudder
movements in flight.
       Experts worry that some of those incidents could be examples of
mild PCU jams perhaps foreshadowing a major jam that could occur when the
aircraft is at low altitude, leaving the pilots little room to maneuver.
      Most airlines that fly 737s are considering following United
Airlines' example and providing special pilot training on how to recover
from unusual flying positions. United began ``advanced maneuvers''
training for 737 pilots last year.
     Industry concern has been heightened by a recent Federal Aviation
Administration report outlining several ways the PCU could trigger
uncommanded rudder movements.  The FAA has asked Boeing to analyze the
probability of PCU malfunctions posing a haza rd.
      However, aviation and legal records show safety authorities have
been slow to address 737 rudder concerns partly because they rely almost
exclusively on manufacturers for technical orientation in solving crashes
and tracking emerging safety issues.
      In July 1992, for instance, authorities learned that, under certain
conditions, 737 PCUs could inadvertently reverse in flight. It wasn't
until March 1994 that the FAA ordered airlines to fix the problem, giving
737 operators until March 1999 to do so.
     Dr. Bernard Loeb, director of the safety board's office of research
and engineering, said few independent sources understand complex airplane
systems comprehensively enough to look for signs of trouble.
    ``Absolutely, this is the best way to do this,'' Loeb said.
     However, sources close to the crash investigations and lawyers
pressing claims in Colorado Springs and Pittsburgh characterize the
testing done to date as inadequate. Two attorneys representing families of
passengers from the Colorado Springs crash h ave been particularly
outspoken.
      Philadelphia attorney Art Wolk, who primarily represents families of
air-disaster victims, describes the safety board's investigative technique
as ``ridiculous.''
     ``In other words, you take a part that's broken and is suspected of
causing a problem, then you fix it until it works and then pronounce it
not to have been a factor,'' Wolk said.
       San Francisco attorney Dennis Lods, who specializes in technical
aspects of airplane systems, said he doubts officials will get to the
bottom of the 737's rudder problems. At risk, he says, is another crash.
     ``I'm not at all satisfied with the thoroughness of the
investigations,'' Lods said.  ``There's great deference to the
manufacturer that isn't warranted.''
     Aviation and legal records paint a picture of the role manufacturers
have played in the Colorado Springs and Pittsburgh investigations.
     Two Parker Bertea crash investigators recently testified in cases
pressed by Wolk and Lods that engineers knew from the day the PCU was
designed that dirty hydraulic fluid could restrict movement of its
interval valves.
      While this is characteristic of PCUs in general, designers knew the
737 PCU would have to be in near-perfect adjustment and free of even the
smallest traces of debris to operate properly.
     The PCU's two internal valves were designed to move smoothly back and
forth, opening and closing holes, called ports, which permit pressurized
hydraulic fluid to move through various chambers. The even flow of
pressurized fluid between the chambers c auses the rudder to move.
     But contaminants can collect around the ports and wedge between the
valves, upsetting the flow of hydraulic fluid and causing the rudder to
move inadvertently, according to Karun Nair, a Houston-based independent
hydraulics systems designer.
      Should its valves jam upon receiving a command to move the rudder
slightly, the PCU would swing the rudder all the way over, Parker
engineers Steve Weik and Shihyung Sheng testified.
      The PCU's filtering system is relatively coarse, which experts say
could make it particularly susceptible to varying degrees of jamming.
      The valves are separated by a microscopic distance of 2 microns,
while the filtering system is designed to trap contaminants 10 microns or
larger. The human eye cannot see anything smaller than 40 microns; the
diameter of a human hair is 89 microns.

      Visible debris was found in the PCUs of the United and USAir jets.
      According to testimony by another Parker engineer, Wally Walz,
investigators found an unknown, white, powdery substance, traces of water
and stringy, bronze-like chips inside the United jet's PCU.
     ``It was a white substance none of us understood,'' said Walz, adding
that it was ``unusual to see the bronze-appearing material.''
     Phillips, the safety board's rudder expert, said investigators at the
time ``hadn't seen any service history'' indicating such contaminants
might be a problem. Phillips said he assumed the PCU's filters prevented
anything dangerous from getting into its internal parts.
     But when particles of steel and aluminum-nickel-bronze turned up in
the USAir jet's PCU, Phillips asked Boeing engineer John Curulla to
conduct some tests. Curulla took a PCU similar to that used on 737s and
set it up to receive constant movement com mands from an
electrically-powered actuator. Curulla added sand, steel and
aluminum-nickel-bronze to the PCU, which operated smoothly even with very
high concentrations of debris.
     Phillips said Curulla's tests helped convince him ``that we've proven
that contamination wasn't a factor'' in the Pittsburgh crash.
      But hydraulics expert Nair said Curulla's tests flushed debris
through the valves in a way unlikely to occur in flight.
      ``Their conclusion is correct for what they did,'' Nair said. ``But
it is hard to imagine that somebody is going to shake the rudder in flight
back and forth all the time.''
      Jams typically come and go, varying in severity and often leaving no
trace, as contaminants work their way free in the constantly moving fluid,
Nair said.
      A more representative test would be to monitor the USAir jet's PCU
operating with dirty fluid over a period of time, with only occasional
commands from manual levers that match the pilot's rudder pedals, Nair
said.
      ``You have to do a waiting game, and eventually, it will happen,''
Nair said.
      After failing to find enough evidence to implicate the PCU in
Colorado Springs, the safety board extensively examined _ but did not
accept _ a Boeing theory that a freak gust of wind, bouncing like a
horizontal tornado off nearby foothills, caused t he crash.
      This month, the safety board will explore a Boeing theory that
another kind of horizontal tornado caused the USAir jet to crash in
Pittsburgh.
      Elaborate flight tests, estimated to cost $2 million, will examine
whether spiraling turbulence from the wings of a jet 4 miles ahead might
have tossed USAir Flight 427 into a prolonged twisting motion.
* * *