by Bill Adair
“Great,” Demetrio said, still unconvinced. “I understand, just like I did all the other times.”
Goodman abandoned the question and went to the next one. “Do you have any plans to remarry?”
“No,” Brett said.
20. EASTWIND
As Eastwind Airlines Flight 517 neared Richmond, Virginia, the night of June 9, 1996, Captain Brian Bishop felt a bump from the back of the plane. “Did you feel that?” he asked the first officer.
Before Bishop got an answer, the plane’s nose veered right and the right wing dipped toward the ground. Back in the passenger cabin, a flight attendant was tossed into a row of seats. Bishop stomped on the left rudder pedal, but it felt stiff. He turned the wheel to the left and added power to the right engine. That stopped the plane from rolling, but he could not get the wings level. He pressed on the left rudder pedal with all his weight, but could not get it down. The plane was stable, but was flying with the right wing tipped precariously toward the ground. The 737 was heading straight for the lights of downtown Richmond.
Bishop glanced out the window, looking for an area with no lights. If he had to put the plane down, he wanted to do it away from homes and buildings. Suddenly the rudder seemed to release and the wings leveled off. Bishop told the first officer to start the emergency checklist for a sudden roll.
“Autopilot off,” the first officer said.
“Off,” Bishop said.
“Yaw damper off.” One of them reached to the switch on the ceiling.
“Off,” Bishop said.
But then there was another thump, and the right wing dipped again. It seemed as if the pilots were losing control.
“Declare an emergency,” Bishop said. “Tell them we’ve got a flight control problem.”
The first officer relayed that message to the Richmond tower. A controller gave them a new heading to the airport, but Bishop was having trouble turning the airplane. He had the wheel cranked to the left and was putting most of his weight on the left rudder pedal, but he could not get the wings level. He was worried that he might not make it to the airport.
Then the rudder seemed to release again, allowing Bishop to level the wings. His first officer told the controller they could make the turn toward the airport. They were about five or six miles away now. They hurriedly went through the landing checklist. Bishop knew that a plane was more vulnerable to rudder problems when it flew low and slow, so he told the first officer, “I’m going to stay high and fast.” Bishop was afraid that if the strange problem happened again he would not be able to recover.
As they descended toward the runway, fire trucks were waiting with red lights flashing. The plane touched down and rolled almost to the end of the pavement. The first officer told controllers the plane was okay and asked that the fire trucks stay away so passengers would not be alarmed. As they taxied to the gate, Bishop realized that he was so scared his knees were shaking.
He picked up the microphone to make an announcement to the passengers, but then he wondered what he would say. Anything he said would just make it worse. He put the microphone back.
Haueter and Phillips heard about the incident the next day. At first they thought it was a minor malfunction unrelated to the USAir crash. But then they discovered that Bishop had reported earlier rudder problems with the same plane. They headed for Richmond.
Eastwind was a new airline based in Trenton, New Jersey. It had only two planes, both of them 737–200s formerly flown by USAir. The planes had been repainted with Eastwind’s logo, a squiggly line along the windows and, on the tail, a bumblebee wearing sunglasses. The airline had dubbed itself “the Bee Line.”
Three weeks before the Richmond incident, Bishop had felt a slight rudder kick in the same plane when he departed Trenton and leveled off at 10,000 feet. It felt like the copilot had tapped the rudder pedals. He circled back and landed. Mechanics inspected the power control unit and replaced a coupler for the yaw damper, a device that made tiny adjustments to the rudder to keep the plane flying straight. The 737 yaw damper had a reputation for frequent malfunctions, so that seemed to be a logical fix. (The yaw damper could move the rudder only 2 or 3 degrees, so Haueter and Phillips did not believe it could have caused the full 21-degree movement that led to the crash of Flight 427.) Bishop tested the plane the next day, and the rudder pedals felt fine. The plane went back in service.
Now it was on the ground again, being dissected by engineers from the NTSB, USAir, and Boeing. They discovered some curious things. When mechanics had hooked up the yaw damper, they rigged it wrong. Instead of being limited to 3 degrees either direction, the yaw damper could move the rudder 4.5 degrees right and 1.5 degrees left. (That had not been a problem on Flight 427, however. The PCU had been rigged correctly.)
The Eastwind flight data recorder was rushed back to the NTSB lab to see if the engineers could decipher what had happened. It showed the plane had some strange rudder movements. It had initially gone to 4.5 degrees, presumably because the device on the yaw damper had been rigged wrong. But there was a second movement when it went to about 7 degrees and stayed there for twelve to fourteen seconds. That big a movement could not be explained by any sort of misrigging. It looked like the rudder had gone hardover.
The thirty-nine-year-old Bishop was a wiry former commuter pilot who did not have the polish of the typical airline captain. He had stringy brown hair and was always dashing to airport smoking lounges for a quick cigarette before departure. He had had the misfortune of working for two airlines that went out of business, so he had driven an airport snowplow until he got the job with Eastwind. He had a gritty personality that Haueter liked. His experience in flying small commuter planes came in handy when the rudder kicked. He used asymmetric thrust—putting more power in one engine—to keep the plane flying straight. That was a common approach for a “throttle jockey” flying a small turboprop, but it was rare for a 737 captain.
Boeing officials said Bishop had overreacted. Sure, there had been a malfunction in the yaw damper, the company said, but pilots were notorious for exaggerating their accounts of a sudden roll. Boeing said there were problems with gyros feeding information to the Eastwind flight data recorder that raised doubts about whether the rudder had truly gone to 7 degrees.
Haueter thought that the incident gave him a unique opportunity to test an airplane that may have had a hardover. He knew Eastwind was crippled while the plane was grounded in Richmond—the plane was half of the airline’s fleet—but he couldn’t pass up the opportunity to try flight tests with a crew that might have encountered a hardover.
The parties agreed on two tests. One would measure how a 737 would react to a yaw damper problem. The other would see how a pilot would react to a sudden rudder movement. Bishop would be the guinea pig.
The day before the test, Michael Hewett, the Boeing test pilot, and several NTSB investigators arranged a conference call with Bishop to explain how the test would go. As usual, Hewett was brash—so brash that Haueter thought he was trying to bully Bishop into thinking he hadn’t responded the right way.
“Stop it!” Haueter said. “You’re trying to intimidate this guy.”
Hewett said he was just trying to get Bishop to understand what had happened that night. Hewett seemed to have doubts about Bishop’s competence. “When I put my wife and children on an airliner, I expect the people flying up front to be as good as I am.”
Haueter was furious. He thought that Boeing was trying to influence the test.
Hewett was equally angry. He thought the NTSB was acting like the Gestapo, limiting what questions he could ask. He wasn’t trying to influence Bishop; all he wanted was an accurate story from him.
The first test was done without Bishop. The Eastwind plane was configured exactly the way it had been when Bishop had his scary incident at 4,000 feet, with the same PCU, the same yaw damper. The only change was in the cabin, where seats had been removed so Boeing could install its flight-test computers. The mood was tense a
t a preflight briefing as Hewett and the NTSB went over the test plan. They didn’t know if there was a gremlin in the tail, and no one could be sure that Hewett would be able to control the plane if something went haywire. The route took the plane out to a restricted military area over the Atlantic Ocean so they could try maneuvers away from a populated area. If the rudder went hardover, they wouldn’t wipe out a neighborhood. But Hewett was not nervous at all. He thought the NTSB was overreacting by insisting that the plane fly over the ocean.
Once the plane climbed into the sky, Haueter figured it was safe. The critical time was takeoff and landing. When a plane was flying faster than 200 knots, pilots could easily recover from a hardover.
Hewett and an FAA test pilot flew out over the Atlantic and put the plane through the maneuvers, kicking the rudder right and left. The plane flew poorly because its controls seemed to be out of alignment, but the pilots found no problem with the rudder. They returned to Richmond.
Before Bishop got on the plane for the second flight, Haueter warned him about Hewett. “Look, the purpose of this is I want to know your perceptions. Don’t let anybody talk you out of anything. Take a piece of paper so you can write something down immediately after it happens so it will be fresh. We want to know how similar this is to what you had the night of the ninth, as best as you can recall.”
Bishop took the controls and flew the same as he had on June 9. Without warning, Hewett gave a quick hand signal. An FAA pilot behind them pushed a button that suddenly moved the rudder 4.5 degrees. The plane started to roll.
Bishop quickly stomped on the opposite rudder pedal, which stopped the roll and made the plane roll back toward wings level.
Was that what happened on June 9? Hewett asked.
No, Bishop said, the test was much slower. “This isn’t even a tenth of what we felt that night.”
“Well, it was dark out, you weren’t expecting it,” Hewett said. He seemed to be offering excuses to show that Bishop had exaggerated.
“This wasn’t even close,” Bishop said.
Haueter had studied dozens of suspicious 737 incidents since the crash and had found reasonable explanations for nearly every one. Many were yaw damper problems, some were autopilot malfunctions, and lots were encounters with wake turbulence. But the more he studied Eastwind, the more it matched Flight 427. He believed Bishop’s account, which was corroborated by the first officer, that the rudder pedal would not move. The flight data recorder also verified their accounts and showed that the plane was cross-controlled, meaning that the rudder was turned at the same time the ailerons were going the other direction. The only problem in the pilots’ stories was their confusion about the timing of the events, which was common for all pilots. Haueter concluded that the Eastwind incident was eerily similar to 427, with one important difference. The plane’s speed was 250 knots, well above the crossover point, so Bishop was able to recover before it crashed.
On the two-year anniversary of the crash, by pure coincidence, John Cox took a USAir 737 to Pittsburgh. He was scheduled for his six-month training session at USAir’s simulator center near the Pittsburgh airport. As the plane approached the city, Cox was sitting in Row 8 of coach, studying a pilot handbook about things that can go horribly wrong on a plane—engine fires, takeoff stalls, autopilot failures. The passengers sitting around him might have gotten heartburn if they had seen what he was reading, but to Cox it was like studying for final exams.
The USAir emergency checklist had been improved because of Flight 427. Now there was a procedure for “Uncommanded Yaw or Roll,” which called for the pilot to turn off the autopilot, grab the wheel firmly, and turn off the yaw damper. The cover of the checklist also had changed, listing the general approach for pilots anytime a plane was in trouble:
Maintain aircraft control
Analyze the situation
Identify the emergency
Use the appropriate checklist
Cox wasn’t nervous about his six-month checkup, but he wanted to make sure he aced it. He double-checked the list of maximum fuel temperatures and silently recited the engine fire procedure (“Auto-throttle off, throttle idle, start lever cutoff, fire handle pull. If light doesn’t go out in thirty seconds, rotate handle”). As the flight attendants prepared the 737’s cabin for arrival in Pittsburgh, he was so caught up in preparing for calamities that he didn’t notice the time—7:03 P.M.
Two years to the minute since Flight 427 had crashed into the hill.
The next day in the simulator, he dealt with one crisis after another. He was climbing out of Philadelphia, leveling off at 10,000 feet when bang!—the No. 2 engine seized up. He went through the memory procedure for an engine fire. Once the situation was under control, he pulled out the checklist and started going through it with the first officer. An instructor pounded on the cockpit door, pretending to be a flight attendant.
“What is going on!? What was that bang?” the instructor shouted.
The weather in the simulated Philly was lousy, so Cox diverted the plane to Baltimore and landed with one engine. He then departed for Charlotte and found horrible weather there, too. He had to use the 737’s auto-land system, which can land the plane in thick fog. More mishaps followed. The plane began to stall after takeoff. Another engine problem. And then his beeper went off.
He glanced down at the beeper and saw the number. It was LeGrow, the union’s chairman for the Flight 427 investigation. At the next break Cox walked to the pay phone and called him.
“Hey, Herbie, what’s going on?” Cox asked.
“John Boy, have you checked your Aspens?” LeGrow asked, referring to the ALPA voice mailbox system.
“No.”
“You’re going to like it,” LeGrow said. “Kenny has the safety recommendations. They couldn’t be any better. They are just what we wanted.”
“Yes!” Cox said, punching his fist in the air like a pitcher who just struck out a .300 hitter.
He walked back to the simulator and told the instructor, “Four-twenty-seven is drawing to a close. And we’re going to like the results.”
Greg Phillips’s recommendations for the 737 had come back to life because Laynor, the safety board official who had blocked them, had retired. They found a much more receptive audience in Bernard Loeb, the new head of aviation safety, and his deputy, Ron Schleede. That broke the logjam and meant that Haueter and Phillips could try again to get approval. The recommendations that were going to the board members were essentially the same list from Phillips’s memo eighteen months earlier. They called for sweeping changes to the 737 and the way it was flown. They did not specify what Boeing should do—“We don’t want to be junior engineers telling them how to do it,” Haueter said—but they essentially meant that Boeing would have to install a limiter on the rudder or change the ailerons so pilots would have more roll control. The list also called for fixes to the yaw damper, a better method for pilots or mechanics to detect a jam in the valve, a new cockpit indicator to show pilots when the rudder moved, and new procedures telling pilots how to respond to a hardover.
Cox was ecstatic when he returned to St. Petersburg the next day, calling the recommendations a watershed event that would prevent future crashes. “I’m so happy, I am doing double back flips,” he exulted. “We’ve now got the mechanism in these recommendations to fix the airplane.”
McGrew was feeling burned out. He was overdue for a vacation and felt drained. His son said he looked like he had aged five years since the crash. It had been only two. He had lost enthusiasm for the investigation and thought that others at Boeing were in the same rut. In the first two years there had been a high level of energy throughout the company to solve the mystery, but nearly everyone had since moved on to other projects and now when they worked on 427 it was hard to get them fired up again.
To make matters worse, he had been removed from his job as the 737 chief engineer and given a different position, overseeing new Boeing models. It was a lateral move, but he did not want
to go. His bosses didn’t mention burnout as a reason for his reassignment, but McGrew figured that was one of the factors. He stayed involved in the Flight 427 effort, but he found less time each week to work on the investigation.
He was frustrated at the NTSB’s lack of effort in studying the pilots. Sure, the human factors team had pursued several leads that Boeing wanted, but McGrew had heard that Brenner, the NTSB human factors investigator, believed the pilots had their feet on the floor and never knew when the rudder went in. How could Brenner say such a thing? There was no conclusive proof of that, and in McGrew’s view, there was evidence to the contrary. McGrew and most people at Boeing still believed that Emmett or Germano (most likely Emmett, since he was the flying pilot) had mistakenly slammed his foot on the pedal and then pulled back on the control column, stalling the airplane and causing the crash.
McGrew and other Boeing officials had been traveling the world to reassure airlines about the safety of the plane. They were under tremendous scrutiny. A series of stories in the Seattle Times said Boeing had not responded to the rudder problems, despite many incidents. McGrew found that allegation preposterous. He said the company had thoroughly investigated the incidents and had found no systemic problem.
He said he felt no pressure from Boeing management to defend the plane and had done so only because there was no evidence that the rudder system had malfunctioned. Likewise, he said, the company’s costs for lawsuits had no effect on what he did. “If it’s a ton of money, that’s too bad,” he said one day while driving up Interstate 405 to a meeting. “If there’s something wrong, you’ve got to fix it.” He was convinced that Emmett and Germano just got into a situation that was over their heads.
McGrew and other officials thought it was time to throw the equivalent of a Hail Mary pass. They would go over Haueter’s head directly to the board members, the five political appointees who would vote on the probable cause. Rick Howes, the Flight 427 coordinator in Boeing’s air safety investigation office, made a courtesy call to Haueter. Howes said Boeing was going to be “aggressive” in informing the board.