|Part Two: The soda can
Dozens of people had seen the 737 twist toward the ground, and most agreed it had rolled left and then plunged nose-down. They agreed on little else. Some heard a growling sound, others heard nothing.
A USAir utility worker saw smoke coming from the front of a wing and somebody else said one of the engines looked like it was cocked to the side. But most people didn't see smoke and said the engine looked normal.
Rumors spread about a passenger named Paul Olson, a cocaine dealer who had just met with prosecutors in Chicago about testifying against a drug ring.
"WITNESS ON DOOMED PLANE TIED TO COLOMBIAN DRUG TRADE" said the Gannett News Service. Gannett quoted lawyer F. Lee Bailey: "I believe a bomb caused it, particularly after I found out what went on in the cockpit. In my view, in the totality of what we know so far, the scenario is consistent with a bomb. I think two bombs might have gone off in sequence."
But the FBI told Haueter that Olson was a bit player who would not be the target of a hit. "He's a nobody," an FBI agent said.
The wreckage told the same story: no bomb.
The only pieces of the plane found away from the hill were light objects, such as insulation, that were carried away in the hot plume of smoke. The wreckage, including the 4,000 copies of BusinessWeek, was concentrated in a tight area on the hill. If a bomb had gone off, BusinessWeeks would have rained from the sky. They'd still be finding them in Cleveland.
To be absolutely sure, Haueter consulted Edward Kittel, an FAA specialist in airplane explosions. Kittel said hot gas from a bomb would leave unmistakable marks. The wreckage would be pitted with tiny craters and black streaks that would not wipe away with a finger. But a fire after a plane hit the ground would burn cooler and slower.
Kittel inspected thousands of pieces that came off the hill, searching for the telltale pockmarks. He found none.
With the bomb scenario put to rest, Haueter could focus on his best theory -- that something went wrong with the rudder of USAir 427.
A rudder problem would explain why the plane suddenly rolled left. In fact, it sounded suspiciously like what happened to the United 737 that crashed in Colorado Springs.
Everyone at the NTSB knew about Colorado Springs, one of only four unsolved crashes in the safety board's 27-year history. Haueter vowed that the big words "FOR UNDETERMINED REASONS" would never appear on his report. He pictured himself not only solving USAir 427, but also untangling United 585 in one fell swoop. He wanted to be the guy who solved the riddle of the 737.
* * *
For the USAir crash there was stronger evidence from the flight recorder pointing to the rudder, so Haueter's investigators had quickly focused on a device the size of a soda can that moved the big panel.
The soda can gadget was an ingenious device that had simplified the 737's rudder system, trimming at least 50 pounds off the plane's weight. That saved the airlines thousands of dollars in fuel.
Its real name was impossibly dull the dual concentric servo valve but engineers used an amusing hand gesture to show how it worked. They curled fingers on one hand to create a hole, then stuck their index finger in and out, like they were demonstrating sex.
The valve was housed inside a heavy steel contraption about 3 feet long and 2 feet wide called a power control unit, or PCU. When a pilot stepped on a rudder pedal, the PCU moved a tube in and out of the soda can-sized valve. That allowed hydraulic fluid under pressure to pass through the holes in the valve. Depending on which holes the fluid passed through, the rudder would turn right or left.
In the arcane world of hydraulics, the valve was revolutionary because it had two tubes, one inside the other. The valve-within-a-valve design provided an important safety feature: If one tube jammed, fluid could move through the other and neutralize the rudder.
Haueter assigned his colleague and close friend Greg Phillips to lead the investigation of the plane's mechanical systems, including the rudder. A frizzy-haired NTSB engineer who had worked on the Colorado Springs crash, Phillips knew 737 rudders better than just about anyone. Ever since Colorado Springs, he kept a running list of suspicious 737 incidents, like a detective tracking a killer.
Two weeks after the Hopewell crash, Phillips and his team of investigators flew to Seattle to test the key pieces at a Boeing lab.
The wreckage arrived in crates that reeked of Clorox. They were stored in a locked room at Boeing to guarantee that nobody could tamper with the evidence. The NTSB's "party system" made Boeing, USAir and the pilots union equal partners in the investigation but trust had its limits.
Phillips believed in the NTSB culture of caution. They had to consider every possible theory. "We can't just stay focused on the rudder," he told his colleagues.
Dozens of tests on cockpit gauges, switches and other items turned up nothing out of the ordinary. This was the drudgery of the NTSB's methodical approach. A hundred tests might produce nothing, but crossing each one off the list brought them closer to solving the mystery.
They saved the rudder tests for last, flying to Irvine, Calif., home of Parker Hannifin, the company that made the PCUs and the soda can valves.
There was tension in the air when Phillips' group arrived. To Parker employees, it was like a visit from the IRS. The NTSB team was there to see if Parker's product was responsible for the disaster.
Some investigators thought this test might solve everything. They suspected a jam in the soda can valve had made the rudder suddenly turn, flipping the plane out of the sky.
Technicians pulled the PCU from the crate and hooked it to a pump. A technician flipped a switch.
No one spoke. Everyone was listening for a hiss or gurgle or some other awful sound the valve might have made before it threw 132 people to their deaths. It seemed like everyone was holding their breath waiting for the big moment.
The valve didn't budge. They could just as well be staring at a rock garden.
A technician moved a lever to simulate a pilot stepping on a rudder pedal. The tubes inside the valve slid smoothly back and forth.
The technician pushed the lever harder. He tried to get it to reverse, slamming it as hard as possible.
"No piston reversal," a Parker employee wrote on the test records.
Test after test, the device worked perfectly.
Hundreds of tests and they had found nothing wrong with the plane or the soda can valve. Haueter was back where he started.
* * *
© Copyright 2006 St. Petersburg Times. All rights reserved.