The Mystery of Flight 427

by Bill Adair

  Haueter began the hearing by reading an account of the five-month investigation: “On September 8, 1994, at about 7:03 Eastern Daylight Time, USAir Flight 427, a Boeing 737–300, registration N513AU, crashed while descending to land at the Pittsburgh International Airport.” He played a video animation of the moments right before the crash. The video showed a plain white 737 (USAir had balked at putting the airline’s logo on the animated plane) that bobbled a bit, rolled smoothly to the left, and then plunged nose down. The video ended while the plane was still at an altitude of 4,000 feet. (A Boeing video played later in the day was more dramatic. It showed the view Emmett and Germano would have seen, with the ground spinning closer and closer until impact.)

  Haueter explained the basics of the investigation—the recovery of the wreckage, the reconstruction in the hangar, the backdrive work in the simulator, and the multitude of tests on the rudder system. He didn’t say they were stumped, but he made it clear that they had not found the answer. “Mr. Chairman, at this time, I am not aware that any party to the investigation or any other persons or organizations have raised avenues of investigation that we have not pursued fully, or are currently examining.”

  One by one, Haueter, Phillips, and other NTSB officials asked the witnesses to tell what the NTSB had found. Haueter and Phillips knew the answers to most questions before they asked them; the point of the big production was not to uncover new facts but to let the public hear the old ones. It was peculiar—NTSB investigators asking outsiders to describe what NTSB investigators had done—but that was how the agency worked. In the safety board’s just-the-facts-ma’am culture, it was preferable to let outsiders take the spotlight. That approach allowed the investigators to be impartial and kept them from speculating publicly about the cause.

  Bill Jackson, the pilot who had ridden in the cockpit on Ship 513’s previous leg, testified how his knee was pressed against the microphone button, which would explain the gurgling sound the passenger heard in first class. An FBI agent testified about the examination of wreckage and how the agency had ruled out the possibility of a bomb. A parade of Boeing engineers explained the NTSB’s efforts to jam the rudder PCU, the simulations in M-Cab, and the tests on hydraulic fluid. All showed that the rudder and the 737 had performed properly.

  McGrew had never testified at an NTSB hearing, but when he walked into the ballroom and saw the families sitting in a special section he realized that the purpose of the hearing was not to advance the investigation. This hearing was just a big show, a way for the NTSB to get publicity. After answering general questions about his education and his job at Boeing, he used the remaining questions to convey his main points—that the plane had passed every test and that Boeing was committed to safety.

  He said his boss had told him “to go out and find the cause and, if it’s anything to do with the airplane, fix it.” He spoke proudly of his plane and its safety record, like he was boasting about his kid’s SAT scores. He wasn’t basing this testimony on some emotional tie he had to the plane, of course. It was all based on hard data. And the data showed his plane was incredibly safe. More than 2,600 737s were now in use in ninety-five countries, he said. The 737 had the best reliability of any airliner and—this was the data talking, mind you—the rate of 737 hull losses was extremely low.

  Unfortunately for Boeing, McGrew did not come across well on the witness stand. His cool reserve about the plane’s record made him appear smug, and much of his testimony sounded rehearsed.

  Toward the end, Phillips tossed him a softball question, asking if there was anything else he wanted to say. McGrew seized the opportunity, pointing out that the PCU from the USAir plane would not reverse, that the fluid was not significantly contaminated, and that there was no evidence of a jam. “That leads us, based on that data, to think that the rudder was doing what it was asked to be doing.”

  In other words, the pilots did it.

  The most unusual event of the week came during the lunch break on the second day, when about twenty family members held a press conference in a tiny meeting room down the hall from the ballroom. They took turns stepping up to the microphones to complain about the shoddy treatment they had received from USAir. “We believe the system for notification of next of kin is deeply flawed,” said Marita Brunner, whose brother-in-law Jeffrey Gingerich was killed in the crash. “It increases the anguish for the families.”

  Joanne Shortley said she started calling USAir the minute that she suspected her husband, Stephen, was on the plane, but all she could get was a busy signal. When she finally got through, a USAir employee took her phone number and said the airline would call back. Her brother called the airline and was told that Stephen was not on the flight. Joanne’s children cheered. But hours went by and Stephen did not come home. Finally, the airline called back at 2:45 A.M. to say that he was on the plane. Like Brett, Joanne thought that her USAir coordinator was poorly trained and unprepared for the family’s grief. “She was not equipped in bereavement,” Joanne said at the news conference. “She was a saleswoman.”

  Judy Lindstrom, whose husband, Gerald, was killed in the crash, complained that USAir had blocked her efforts to get a list of other families. “We had a great need to see each other and be with each other. I was told this was not something the airline would disseminate. We found out later that many of us had made the same request. We had to scramble and scratch to get together.” The families said they were forming the Flight 427 Air Disaster Support League to urge the government to appoint a family advocate who would help families after a crash. “We are demanding that this process be taken away from the airline,” Marita Brunner said.

  Brett agreed with the group’s complaints, but he didn’t attend the news conference. He had kept his distance from the group. He wasn’t much of a joiner, and he did not want to get too wrapped up in the crash. It seemed as if some people in the group wanted to make the crash the centerpiece of their lives. He didn’t. He knew that he had to move on. He spent much of the week reading the docket in his room and exercising in the fitness center.

  Reporters left the news conference and cornered Deborah Thompson, the USAir director of community affairs, on the mezzanine outside the ballroom. Was it true, they asked, that USAir had done a poor job?

  Thompson said the process of identifying the passengers had gone slowly, but that the airline wanted to be sure the list was accurate. “You don’t want to give wrong information,” she said. She acknowledged that the family coordinators had not been trained and said that the airline planned to start training people so they would be better prepared in case of a future crash. “We want to do a better job. But that’s not to say I think we’ve done a bad job.”

  Jim Hall, the NTSB’s new chairman, had just replaced the retiring Carl Vogt. A short man with sandy hair and a mild Tennessee drawl, Hall was a Vietnam veteran, a longtime Democratic activist, and a friend of fellow Tennessean Al Gore Jr. When he was nominated for the job in 1993, Washington Post columnist Al Kamen called Hall “a politically connected white male Democrat whose only transportation experience apparently is a driver’s license.”

  Hall appeared to be a lightweight because he didn’t talk like an engineer. He used folksy Tennessee phrases and often sounded like Andy Griffith on Matlock. While other people were talking about dual concentric servo valves, Hall would be recounting something his mother had taught him. It didn’t help that he had a dog named Trixie in his office.

  A friendly brown-and-white Welsh Corgi, Trixie belonged to Hall’s special assistant, Jamie Finch. Hall was an animal lover, but his three dogs and four cats were back home with his family in Tennessee. So he was glad to have Trixie around, even though she occasionally pooped on his carpet. He would throw tennis balls for her and reach down and pat her during meetings. At Christmas, he and Trixie walked through the office wearing matching Santa hats. She even had her own NTSB badge that said her title was “Safety Dog.”

  The truth was that Hall was not a lightweight at
all. He was a savvy political operator who knew how Washington worked. He was a loyal soldier in the Clinton-Gore crusade to make government more responsive to taxpayers. He was well connected at the White House and in Congress, so he had the clout to get the money and staff that the NTSB needed. But he also had an outsider’s perspective and would fuss at the NTSB engineers to make sure that they explained their findings in everyday English. In the NTSB’s ongoing fights with the FAA, he used his political skills behind the scenes and then, if that didn’t work, he would offer a few choice comments to the media.

  Before Hall arrived, the safety board had not paid much attention to victims’ families. Investigators would politely answer questions, but they thought their job was to find out why planes crashed, not to provide grief counseling. When Hall heard that the families were unhappy and had formed a support group, however, he told the NTSB staff to help them. In his view, helping the families was exactly what the Clinton-Gore approach called for. He asked the NTSB staff to set aside reserved seats in the ballroom for the families and arranged a meeting with them one night after the testimony. He had Haueter come along to answer questions.

  About a hundred family members came to the meeting room on the mezzanine level. Hall opened the session by explaining the NTSB’s role and the purpose of the hearing. Haueter updated the families on the investigation and went through the plans for the remainder of the hearing. Then it was time for questions. The 737 sure seemed to have lots of rudder problems, someone said. Why not ground it?

  That wasn’t up to the NTSB, Haueter said. His agency would only make recommendations to the FAA, and it needed solid evidence before it made such a serious request. So far, he hadn’t found the evidence.

  Why have Boeing people up there? someone else asked. “They’re just going to lie to you.”

  Haueter and Hall explained the party system. “These people know the plane’s systems best,” Haueter said.

  Someone else was amazed about the lack of measurements on the flight data recorder. “Why aren’t you doing anything about that?”

  “Look, you’re preaching to the choir. The safety board has been saying that for years,” replied Haueter. Hall vowed to bring up the issue with the FAA.

  The families had a litany of complaints about USAir—the long delays in confirming who was on the plane, the poorly trained employees, and the airline’s refusal to share the names of other families.

  Hall listened and came away convinced that the families had been mistreated. The things they wanted were reasonable. USAir didn’t have to pamper them, but the company should have shown a little common decency. Hall would have preferred that the airlines do it on their own, but they had shown they did not care enough to do it right. It was time for the government to get involved.

  The hearing revealed growing tension between the NTSB and Boeing. The company had fought some of the NTSB’s requests for data, had stamped “PROPRIETARY” on several items and said the board could not release them to the public because of Boeing’s confidentiality agreements with the airlines. Days before the hearing, the NTSB learned about several 737 incidents that Boeing did not include on a list that Phillips had requested.

  Haueter was surprised that the company was so disorganized. He didn’t think Boeing was deliberately hiding anything, but he was disappointed that it could not keep track of important data. Chairman Hall was furious. He reminded a Boeing official of the request for all incidents and asked, “Is that simple enough?”

  Another Boeing official acknowledged that the company should have included a rudder incident involving an Air France 737 on the list, but the report did not sound serious to Boeing employees when they first heard about it. When the testimony concluded, Hall told McGrew he was concerned that the company’s list of 737 incidents was incomplete. “When we end up in a situation, Mr. McGrew, just to be straight with you, that we request information and then another party sends us information that is pertinent that we didn’t get from you, it causes concern.” Hall said he knew that things fell through the cracks, but he told McGrew to “go back and examine every crack so we don’t have any question that there’s been any incident with this rudder or any of these systems that might assist us.” Hall then apologized for being so harsh. “I’m from Tennessee and I don’t know to express myself any more than just that way.”

  McGrew was angry. He felt like he had been used so Hall could get his name in the press. The NTSB staff knew all the details about the Air France rudder kick. Either they didn’t tell Hall or Hall conveniently ignored it so he could land a few punches on Boeing. McGrew did not like the fact that politicians were getting involved in plane crash investigations. The NTSB was becoming a political beast, under enormous pressure to come up with an answer. And McGrew was not sure they would come up with the correct answer.

  As the hearing concluded, reporters gathered around Haueter and Hall to get their reactions to the week. Haueter had consistently told the press that he was confident he could solve the mystery. A reporter asked how he could be so confident with so little good evidence.

  “It’s based on experience and the data available,” Haueter said. “I’ve seen two other accidents I’ve worked on where I had much less data than this, much less help than this, and we determined probable cause very definitively. I can’t say I identified any new alleys [to investigate] this week… but there are a lot of avenues available that have to be fully explored.”

  As McGrew returned to his office in Renton, he was more confident than ever that nothing was wrong with his airplane. It bothered him that critics were saying Boeing might be covering up a hidden flaw in the plane. McGrew said he felt no pressure from lawyers or anyone else to protect the plane. If the 737 had a flaw, Boeing wanted to find it.

  McGrew approached that mission with vigor. He spent hours reading reports on the crash and led daily meetings of Boeing engineers. Ninety-five Boeing employees had worked 42,000 hours on the investigation, with twenty-three of them dedicated to it full-time. It had cost Boeing $1.5 million. McGrew had become obsessed with the crash. Occasionally, he awoke at 3 A.M. with questions about a theory, crawled out of bed, put on his bathrobe, and went downstairs to his home computer. He would pull up a spreadsheet of 427’s flight data and try adjusting ratios and parameters to see what effect they would have. Then he would go into the office that morning and ask one of his engineers to try a new computer run.

  He kept a list of theories titled “Items Under Consideration.” It had eighty-five possibilities, including everything from bird strikes to thrust reversers. All but eleven had been ruled out. Many of those eleven were long-shot theories that were likely to be disproved by future tests. The investigation had largely become a tug-of-war between two theories—a rudder system malfunction and a mistake by the pilots.

  From the start, McGrew had been open to the possibility that something was wrong with the 737. That was the classic engineer’s approach, to consider every possible failure mode of your creation. That’s why redundancy is such an important concept in aircraft design. If something fails, there is at least one backup, often two. The rudder system had lots of redundancy. Every lever inside the power control unit had a second lever that moved in concert, in case one should break. In the valve itself there were two slides in case one should jam. It was powered by two hydraulic systems in case one should fail. And there was a standby actuator with its own valve in case the main PCU stopped working.

  But McGrew’s plane had come up clean. Test after test failed to find anything wrong with the 737 rudder system. The PCU would not reverse, the valve didn’t leak, and there were no marks to indicate that it had jammed. The hydraulic fluid from the USAir plane had been dirtier than it was supposed to be, but tests found that made no difference. You could make the fluid as thick as Dijon mustard and the valve could still do the job. The bottom line, as summed up by Phillips in his report, was that the PCU was “capable of performing its intended functions.” Besides, the worldwide fleet of
737s had an extraordinary twenty-seven-year record, with 65 million flight hours and an extremely low crash rate. If there was a flaw in the rudder system, it would have shown up years earlier. McGrew felt it was time to take a closer look at the human part of the equation—the pilots.

  The NTSB had spent weeks studying and testing the PCU, but McGrew felt the investigation had barely looked at Emmett and Germano. What kind of training had they had? What did previous incidents tell about how they responded to wake turbulence? The NTSB had not explored these areas very much.

  McGrew kept using the word “startled.” There was plenty of evidence that the pilots had been surprised by something—maybe a bird, the wake turbulence, or another plane—at the moment that Germano said “Sheeez!” The surprise could have caused Emmett to jerk the wheel too far to the right. When the plane quickly rolled in that direction, Emmett might have tried to stop the roll by slamming his foot on the left pedal and continuing to press on it without realizing what he was doing. There also was no question that the pilots had made a huge blunder by pulling back on the control column. That stalled the airplane and gave them virtually no chance to recover. Yet that mistake got virtually no attention from the safety board.

  On February 15, 1995, Boeing went on the offensive and faxed Haueter a seven-page letter that said the NTSB should take a closer look at the pilots. It cited a British report that attributed many military accidents to “overarousal,” when pilots were so surprised by an alarming situation that they could not recover. The letter said a perceived emergency was all it took to startle the pilots (a subtle reference to the moment when Flight 427 was jostled by the wake turbulence). The Boeing letter also cited an FAA study that said a pilot might take up to ten seconds to respond after being startled. “The NTSB should explore, from a review of the literature and all available databases and records, whether the Flight 427 flight crew could have responded to the unexpected and startling encounter with significant wake turbulence by (1) making an inadvertent application of left rudder, or (2) having an accidental or cognitive failure that led to an application of left rudder,” the letter said.