The National Transportation Safety Board concluded an unusual three-day hearing last week urging The Boeing Co. and the Federal Aviation Administration to take prompt action to improve the 737's flight controls.
Aviation officials questioned whether the 737's flight-control system should be redesigned - an astronomical expense considering the complexity of the system and the more than 2,600 planes in service.
Yet the safety board is no closer to explaining why USAir Flight 427,a Boeing 737-300, flipped out of calm, clear skies on Sept. 8, 1994, on approach to landing in Pittsburgh. Nor can authorities say why a United 737-200 crashed under similar circumstances in Colorado Springs, Colo., four years ago.
Pilots have reported several hundred cases of 737 flight controls moving inadvertently in flight - more than 45 in the last 14 months. "How many more?" Relatives of the 132 people killed on Flight 427 are pressing for answers. At the hearing, many of them were briefed by safety-boardchairman Jim Hall and lead investigator Tom Haueter. Jon Hamley, whose wife, Sarah Slocum-Hamley, was Flight 427's lead flight attendant, has closely tracked the investigation.
He noted that 737s have crashed in England, Turkey, El Salvador and Nigeria since Flight 427 went down. "How many more 737s have to go down before they do something?" he asked. "I'm no expert. . . . but 28 years' worth of incidents? Four crashes since (Flight) 427?" In Boeing's view, deceased pilots Peter Germano and Charles Emmett could be to blame for the crash of Flight 427. Jim Kerrigan, principal 737 engineer, testified that a re-creation of the crash shows the USAir jet's rudder could have swung to the extreme left, causing the fatal dive.
But Kerrigan said Boeing has been unable to pinpoint "any particular failure mode." The only other possibility is pilot error, he said.
Boeing and safety officials still call the 737 one of the safest jets. But other testimony and records shed light on lingering questions: -- Rudder hardover: Officials are concerned the rudder, the part of the vertical tail piece that controls the jet's left/right heading, could deflect to an acute angle in flight.
In certifying the 737 as safe in 1967, the FAA never required Boeing to prove the jet could safely recover from a rudder hardover; it was deemed improbable.
But officials now know of many ways the rudder can, in theory at least, swing hardover when dirty hydraulic fluid moves through the flight-control system. NTSB chief technical adviser William Laynor said there's been no action because there's no concrete evidence of any rudder hardover in flight.
Boeing is studying the probability of rudder hardovers in flight, and the FAA has called for a probe into how sensitive the 737's flight controls are to dirty fluid.
But in Seattle Friday, Boeing President Phil Condit said he doubtedthat dirty hydraulic fluid caused 737 flight controls to malfunction. "I don't see any information that would tell me that," Condit said."None. We have done lots of testing. Boy, I don't see anything."
-- Rudder reversal: One type of hardover can occur when an out-of-adjustment rudder valve jams and moves too far, causing the rudder to deflect in the wrong direction. The FAA has ordered airlines to replace an internal valve part and verify proper adjustment of the valve to eliminate the possibility of rudder reversal.
But planes with upgraded valves still have problems. NTSB records show the rudder of an America West 737 moved inadvertently in late August and again in early September near Phoenix. The jet's valve was upgraded in response to the FAA order last May.
After the second incident, the upgraded valve was removed and sent to its manufacturer, Parker Hannifin, for examination, records show. Haueter, the safety board's chief investigator, said he was unaware of the statusof that valve.
-- Yaw damper: This computer makes thousands of small rudder adjustments during flight. It has been blamed for hundreds of inadvertent rudder movements over the years. But officials say there's no danger because the yaw damper can only command comparatively small rudder movements, and pilots can easily switch the yaw damper off.
But Boeing flight-controls engineer Richard Kullberg testified that the yaw damper can move the rudder more than twice its normal limit, should a stand-by rudder valve jam.
Levers on the stand-by valve of both the United jet and the USAir jet were found "galled," or worn from rubbing too close to each other, indicating they may have jammed.
Kullberg said Boeing is working on a "service bulletin" advising airlines to install a special bushing to make the stand-by valve less susceptible to jamming, but the improvement is voluntary. He said Boeing does not consider the matter a safety issue.
Jean McGrew, chief 737 project engineer, testified that Boeing is evaluating improvements to the yaw damper but wouldn't describe them. Though Boeing contends "the yaw damper is not involved in the accident in any way," the company does believe the reliability of the device needs improvement, McGrew said.
-- Redesigning flight controls: A rudder hardover is especially worrisome when the 737 is slowing and descending on approach to landing. At lower speeds, the pilots lose the ability to counter the effects of the rudder by turning the control wheel in the opposite direction, which extends wing panels called ailerons.
Retired USAir Capt. Herb LeGrow, representing the Airline Pilots Association, questioned Boeing engineers about the possibility of redesigning the ailerons and/or rudder to give pilots more ability to recover from a rudder hardover at lower speeds.
One way to do that would be to restrict the rudder's range of movement once it becomes airborne. Pilots only need an acute ruddermovement in a specific emergency, when one engine shuts down during takeoff. In that case, deflecting the rudder hardover helps keep the plane flying straight ahead.
Or the ailerons could be made bigger or redesigned to extend at more acute angles. But redesigning the 737's flight controls would likely require the plane to be largely recertified, an ALPA source said.
Art Wolk, an attorney representing families of victims from the Colorado Springs and Pittsburgh crashes, said authorities have more than enough evidence to order major 737 improvements. "Here's an airplane that's telegraphing it's going to crash, and nobody is willing to ground it and make Boeing redesign the rudder," Wolksaid. "This is like the Simpson jury wanting a video tape of the crime in order to convict Simpson."
SPRINGFIELD, Va. - If USAir Flight 427 flew into turbulence swirling from the wingtips of another jet flying four miles ahead, the encounter was entirely routine. That's what Federal Aviation Administration test pilot Les Berven concluded after repeatedly flying a Boeing 737 into the "wake vortexes" of another jet as part of a $1 million experiment last September.
Berven's testimony, given yesterday at a special National Transportation Safety Board hearing, seemed to undermine speculation that a wake vortex encounter disoriented pilots Peter Germano and Charles Emmett, causing them to fly a fully-loaded 737-300 jet into the ground.
The NTSB has been unable to explain why the USAir jet piloted by Germano and Emmett suddenly flipped out of calm, clear skies the evening of Sept. 8, 1994, in Pittsburgh. Nor can authorities say why a United Airlines 737-200 crashed under similar circumstances in Colorado Springs four years ago.
Berven said he purposely flew more than 200 times into wake vortexes and proved what many experts and pilots have long believed: Wake vortex encounters are inconsequential for big jets. "I think we covered just about everything that could be done. We went up and down and crossways," Berven testified. "You can pull up in there with absolute impunity. You can fly in any way you want. You can hit anyone or both of the vortexes without a controllability problem."
Despite Berven's testimony, The Boeing Co. used the same wake vortex test results to lay more groundwork for the startled-pilot theory. Boeing test pilot Mike Carriker, who participated in the experiment, testified that a wake vortex encounter could take airline pilots by surprise and confuse them. Carriker said most U.S. pilots have "average" flying skills.
Carriker showed a videotape of one test flight piloted by himself and USAir Capt. John Cox in which the test aircraft hit a wake vortex and rolled sharply to the left. His testimony appeared to contradict Berven's assertions. But then Cox took the witness stand and pointed out that he and Carriker had purposely let the airplane drift into the vortex at an unusual angle - with their hands and feet off the controls and theautopilot turned off. "We literally had our feet flat on the floor and our hands straight up," Cox said.
Commercial pilots routinely encounter wake vortexes and instinctively make the small control adjustments needed to smooth out flight within seconds, Cox said, adding that he has never become disoriented while encountering wake vortexes on actual flights. Another theory for both the Colorado Springs and Pittsburgh crashes holds that the rudder system jammed and moved the rudder sharply without it being commanded to do so.
Flight 427's flight data recorder, which charts the aircraft's trajectory, shows the jet fell from the sky in a way that can only be explained by a sudden, acute movement of the rudder. Moreover, the cockpit voice recorder captured the sound of the engines revving, consistent with the rise in engine noise that occurs whena rudder is deflected in flight.
SPRINGFIELD, Va. - Aviation safety authorities are being urged totake a harder look at whether dirty hydraulic fluid may be causing theflight controls of Boeing 737 jetliners to regularly malfunction inflight.
But a Boeing Co. hydraulics expert, Richard Kullberg, testified yesterday he could not think of any detective work that hasn't already been done.
Those developments came yesterday as the National Transportation Safety Board opened an unusual hearing to publicly air its investigation of two unsolved 737 crashes in four years - and more than 45 recent cases of flight controls moving inadvertently on 737 flights.
A USAir 737-300 inexplicably flipped out of the sky Sept. 8, 1994, in Pittsburgh, four years after a United 737-200 crashed under similar circumstances in Colorado Springs, Colo. The two crashes killed 156 in all. Despite extensive investigations, the safety board has been unable to solve either accident.
Investigators revealed new evidence yesterday indicating the USAir jet's rudder swung to an extreme position, twisting the fully-loaded Boeing 737-300 jetliner into a fatal dive.
Recent analysis of sounds captured on the cockpit voice recorderindicate the engines changed pitch as the rudder moved acutely to the left.
But the central mystery remains whether pilot error or a mechanical failure caused the rudder to move so drastically. Boeing engineer Kullberg, who also serves as a designated Federal Aviation Administration representative in charge of certifying 737 and 757 flight-control systems, reiterated the company stance that there is no evidence of equipment failure in either the Colorado Springs or Pittsburgh crashes.
Kullberg characterized the 737s long-running flight control problems as posing no safety hazards. Even so, he said Boeing recently responded in writing to 12 of 27 concerns outlined in a special "critical design review" of the 737's flight control system completed by the FAA last May.
Among other things, the FAA called for tests to determine how sensitive the 737's flight controls are to dirty hydraulic fluid as well as for an investi gation into all the ways the 737's wing and tail panels,including the rudder, can move inadvertently in flight.
In response to questioning from safety board Chairman Jim Hall, Kullberg said Boeing has responded to the FAA's suggestions with extensive written analysis, though he admitted none of the analysis is supported by any actual testing.
The FAA is now reviewing Boeing's response. But Mike Zielinski, a member of the FAA review team, said no process exists to formally implement the recommendations outlined in the agency's special review.
"Personally I have a concern of a lack of closure and continuing discussion with no real resolution," Zielinski said. Kullberg testified, "We still are talking about possibly running someother tests," but said he was "at a loss to come up with anything that would make sense that hasn't already been done." Other witnesses suggested a couple of possibly fruitful areas that Boeing and safety authorities thus far have left unexplored.
FAA hydraulics expert Werner Koch called for someone to conduct tests to establish how "silting" of hydraulic fluid contaminants affects the 737's rudder over a long period.
Heavy concentrations of contaminants were found in both the United and USAir jets that crashed. The NTSB never analyzed the contaminants retrieved from the wreckage in Colorado Springs, and it permitted Boeing to conduct comparatively crude tests of contaminants like those recovered in Pittsburgh. (The tests were conducted under conditions not likely to exist in flight.)
Koch said the possible effects of contamination silting now deserve closer scrutiny. "As a result of lack of a smoking gun, that might be a logical place to look next," Koch said, broadly describing the kind of testing that investigators could do next. The safety board's Hall then turned to NTSB flight controls investigator Greg Phillips and said, "Mr. Phillips, is that something we can do?" "Yes, it is," Phillips responded. "Well, let's do it," Hall said.
Paul Knerr, a member of an engineering-standards group, said clarification of how many of the world's 2,600 737s may be flying with highly contaminated fluid needs to be addressed as well. After investigators retrieved highly contaminated fluid from the USAir jet, they took more than 100 random fluid samples from 21 other 737s and found 22 percent of the samples were highly contaminated.
That finding shocked the Society of Automotive Engineers (SAE), a group that establishes fluid standards for auto, factory, aircraft, weapons and space systems. Knerr, vice president of engineering for Canyon Engineers, a hydraulics component manufacturer based in Valencia, Calif., said the SAE had assumed jets were flying with cleaner fluid, and he noted that Boeing has no established contamination limit for fluid used on its airplanes.
Yet it is well-known that debris can readily contaminate hydraulic components in the manufacturing process and when they are being used and serviced on airplanes. The SAE is now wor king with the FAA to create a contamination-level standard for commercial jets, Knerr said. In the meantime, Knerr asked Hall to support an effort to collect and analyze fluid samples drawn during flights of an older 737.
These samples would show whether debris tends to circulate in higher concentrations when the airplane is shaking. Such information could help investigators sort out whether dirty fluid should be considered more of a potential threat during takeoffs and landings, when the aircraft is vibrating heavily. It would also help parts makers, like Knerr, design better components. "For a component manufacturer to know that the (contamination) level varies widely is very important in our design," Knerr said.
SPRINGFIELD, Va. - Investigators today revealed evidence indicating that USAir Flight 427's rudder swung to an extreme position, twisting thefully loaded Boeing 737-300 jetliner into a fatal dive outside Pittsburgh 14 months ago. Moments before the rudder moved, the jetliner encounteredwingtip turbulence from a 727 trijet flying four miles ahead, recent testing shows. Those findings emerged today at an unusual hearing of the National Transportation Safety Board to publicly air details of its so-far futile attempt to solve the Sept. 8, 1994, crash.
The safety board's acoustic expert, James Cash, described how investigators analyzed three "thumps" captured on the cockpit voice recorder moments before the jet ran into trouble. Similar thumps turned up on flight tests in which a 737 was purposely flown back and forth into the wingtip turbulence of a 727 trijet flying a few miles ahead. Cash said sounds recreated on the flight tests showed "the source of the thumps . . . was most probably an encounter with the wake turbulence of the preceding 727 aircraft."
The tests, which cost $1 million, also revealed that 737 engines produce a heightened roar when the rudder is deflected to an acute angle in flight. Cash testified that Flight 427's cockpit recorder captured an engine roar - matching the roar produced in the tests - just moments after the thumps, indicating an acute rudder movement.
The central mystery remains whether the wake vortex encounter, normally a routine event, caused veteran pilots Peter Germano and Charles Emmett to panic and stomp on the rudder control pedals, which would clearThe Boeing Co. of liability in the accident.
Representatives from the Air Line Pilots Association are supporting the other leading crash theory: that the rudder moved sharply on its own, due to a malfunction, just after the plane encountered the turbulence. NTSB Chairman Jim Hall opened the hearing acknowledging that despite 50,000 investigative staff hours and millions of dollars worth of lab and flight tests, investigators have been unable to explain the crash of Flight 427, which killed all 132 people on board.
Nor has it been able to solve the similar crash of a United Airlines 737 four years ago in Colorado Springs, Colo. All 2 5 people on board died in that accident. Moreover, independent aviation experts have recently begun questioning whether there may be a correlation between the two crashes and more than 45 reported incidents of flight controls moving inadvertently on 737 flights in the past 14 months.
Few expect the hearing, scheduled to run through Friday, to produce any breakthroughs. At best, insiders said, they might reveal which way investigators are leaning on the two leading theories: that the 737's rudder system drastically malfunctioned or that the USAir pilots overreacted to the wingtip turbulence.
Among the witnesses scheduled to testify is Malcolm Cohen, a spatial disorientation expert with the National Aeronautics and Space Administration's Ames Research Center. Cohen is expected to address the issue of whether Germano and Emmett may have been so relaxed flying a routine approach in clear, calm air that they were startled by unexpected turbulence, causing them to animatedly adjust the flight controls.
The startled-pilot theory has been gaining popularity. A member of the investigation team, who was not identified, recently told The Washington Post that co-pilot Emmett was so relaxed he used a fake French accent to acknowledge a corporate jet flying by: "Oh yaa, I seee zee Jetstream," Emmett is reported to have said.
Oddly, previously released transcripts of Flight 427's cockpit voice recorder tape do not describe Emmett using a French accent when making that statement. Later in the hearing, the safety board is expected to re-examine the possibility that dirty hydraulic fluid caused the rudder to move on its own, putting Flight 427 into its fatal dive.
Investigators have acknowledged that dirty hydraulic fluid can, in theory, cause a 737's rudder to suddenly reverse. Authorities also know dirty fluid can cause the rudder to move to an extreme position, in response to a command for slight movement from thepilot or the yaw damper, a computer that makes minor rudder adjustments during flight. But investigators say exhaustive testing and analysis have turned up no evidence of such a malfunction occurring on either the USAir or United jets that crashed.
Last May, the Federal Aviation Administration completed a special "critical design review" of the 737's flight control systems. The FAA called for a new combined investigation of the Colorado Springs and Pittsburgh crashes.
The agency also said an investigation was needed to determine how sensitive the 737's flight controls are to dirty fluid. Mike Zielinski, a member of the FAA review team, testified today that Boeing has only recently responded to recommended safety improvements outlined in the review. Zielinski said Boeing's response was under review, but under questioning he acknowledged that no formal process exists to implement the agency's suggestions.