The plane, which had a radio call sign of “Boeing 053” instead of a USAir flight number, took off into sunny skies the morning of September 20, 1995. It headed north over the Straits of Juan de Fuca, the spectacular coastline along the Canadian border. Cox, Boeing test pilot Mike Carriker, and USAir pilot Jim Gibbs each had a chance to feel the crossover point for himself.
When Cox got his turn in the cockpit, he pulled the throttle back to slow the airspeed to 190 knots, with the flaps set at “1,” just as they had been for Flight 427. He steadily pushed his foot on the left rudder pedal and simultaneously turned the wheel to the right to keep the plane from rolling, a maneuver known as a steady-heading sideslip.
He flew for miles at that fragile point, balancing the wheel and the rudder pedal. If he slowed to just under 190, he started to lose control. He had to push the stick forward to lower the nose and gain airspeed to recover.
Something is wrong here, Cox thought. The crossover point was way too high. He had expected it to be down around 170 knots, but it was actually at 187. The other two pilots tried the same maneuver and had slightly different interpretations of the speed, but they agreed on an important fact: Flight 427 was right at the crossover point when it flipped out of the sky.
Haueter, who had stayed in Washington, got an excited call about the discovery from Tom Jacky, the NTSB engineer who headed the performance group.
“They ran out of roll authority,” said Jacky.
“You’re kidding,” said Haueter.
“If you slow the airplane up, you don’t have enough wheel to stop the roll.”
“Holy shit!” Haueter said.
With the Seattle tests complete, Cox and the other pilots flew the plane across the country to Atlantic City for the wake turbulence tests. The FAA had agreed to loan its 727, a plane used primarily for research, to play the part of the Delta plane that was four miles ahead of Flight 427. The FAA jet had been equipped with special smoke generators on its wings.
Wakes were usually invisible, but the 727’s smoke generators made them show up as spinning white tubes. They twisted across the sky like an abstract painting. The dual wakes would keep spinning for minutes, and depending on the winds, they would turn, curl, and bounce off each other. Standing near the runway when the FAA plane flew past, Cox could hear the whooosh from the spinning tube of air as it lingered long after the plane was gone.
The next day, Cox got to see the phenomenon from the air as his 737 flew in and out of the 727’s wakes. They were stronger and closer together than Cox and the other pilots had expected. They marveled at the way the lines twisted and soared in the sky. It was as if they could finally see the pothole that they had been running over all these years.
“Oh, that was neat,” Cox said to Boeing pilot Mike Carriker in the cockpit after they watched the wakes twirl back and forth. “I liked that.”
For a week they tried 160 different conditions behind the 727—flying across the wakes, going up and down through them, and holding the wings, tail, or engines in them. For several tests, they took their hands and feet off the controls and rode the wakes like a roller coaster.
As Haueter had expected, there were no breakthroughs. The pilots could easily recover from a wake, no matter the angle at which they flew into it. When Phillips asked Cox whether he was startled by the wake, Cox said he was not. Reacting to the wake, he said, “is as natural as breathing.”
But the engineers learned a lot about wakes. They learned how long they lasted and how they could roll a plane. Haueter and Cox said afterward that the tests were well worth the $1 million price because the data could be used by researchers for years.
The biggest breakthrough of the Atlantic City tests involved the mysterious thumps that had been heard on Flight 427’s cockpit tape. Cash, the shy sound expert, had rigged the voice recorder in the 737 so it was identical to the one on Flight 427, and he had Carriker fly in and out of the wakes.
Carriker and the other pilots heard whooshing sounds when they flew in the wakes, but the sounds were not consistent. When Carriker landed, he told Cash the sounds didn’t match the thumps from 427.
Still, Cash believed he might find a match. Noises often sounded different on the tape because the cockpit microphone picked up sounds through the airframe. Cash took the recording back to his hotel room, put headphones on, and queued up the tape.
Bingo! Exactly the way it sounded on Flight 427! The thump was caused by the Delta plane’s wake. He could finally rule out the theories about explosions and birds. He called Haueter the next morning. “We’ve got a really good match,” he said.
The sparring between Boeing and ALPA had remained behind the scenes, confined to phone calls and closed-door meetings. The angry exchange of letters about pilot error had not been released to the public. But as the two rivals prepared for the second hearing in Springfield, Virginia, in November 1995, it appeared that the conflict would go public.
It was highly unusual for the NTSB to have a second hearing, but Chairman Jim Hall wanted it to release findings of the flight tests and give an update on the investigation. Hall wanted to reassure people that the crash had not been forgotten. He often said that the NTSB was working on three accidents: “Pittsburgh, Pittsburgh, and Pittsburgh.” The safety board chose suburban Springfield for the hearing because it was the closest place to Washington that had an available ballroom.
Boeing and ALPA came to Springfield with sharply different goals. The union wanted to keep the focus on the plane and emphasize the crossover point, but it was walking a thin line. Its leaders did not believe the 737 was unsafe—Cox and hundreds of other union members still flew it every day—but they wanted to emphasize the plane’s problems and show that the 737 needed safety fixes.
The union leaders expected Boeing to come out with guns blazing, raising questions about whether Emmett and Germano had made mistakes in the cockpit. So ALPA prepared two strategies to respond through the news media. A normal, subdued approach would be used if the debate was civil. But if the hearing turned into nuclear war, the union would use tougher words. Cox was not on the initial list to testify, but two days before the hearing ALPA leaders decided that he should. They wanted to give the union’s account of the flight test to counter the presentations from several Boeing witnesses. The union was wary of John Purvis, Boeing’s chief accident investigator, who could cross-examine witnesses like a crafty trial lawyer. The ALPA leaders were afraid that Purvis might try to back Cox into a corner by asking hypothetical questions and challenging Cox for not being a test pilot.
Boeing’s goal was just the opposite from ALPA’s: The company wanted to keep the focus on the possibility of pilot error. McGrew was convinced that Emmett and Germano had been startled by the wake turbulence, had made a crucial mistake by pulling back on the stick, and had failed to turn the wheel right and keep it there. Boeing witnesses would remind everyone that there still was no evidence that the plane had malfunctioned, emphasize that the wake turbulence was “an initiating event,” and raise the possibility that one or both pilots had stomped on the rudder pedal. The word “startled” would be used a lot.
The hearing in the Hilton ballroom was again set up like a giant trial, with Boeing attorneys and engineers clustered around one table, the ALPA team at another. The other parties in the investigation—USAir, the machinists union, the FAA, the valve manufacturer Parker Hannifin, and the flight attendants union—also had tables. But this would primarily be a showdown between Boeing and ALPA.