|Part Four: Rules of engagement
The Greatest Minds in Hydraulics had arrived in Washington.
That was what NTSB Chairman Jim Hall called the special panel he invited to review Tom Haueter and Greg Phillips' work. It had been almost two years since USAir Flight 427 went down, and still they couldn't prove that their best suspect, the rudder valve, was responsible.
Haueter, the lead investigator, considered the panel a waste of time. Why did they need a bunch of so-called experts? Phillips, his hydraulics expert, knew more about the 737 rudder system than just about anybody on the planet.
Actually, Phillips had been one of the people behind Hall's idea. He thought it would help to have an impartial panel validate the investigators' work. Maybe fresh eyes would help.
Ralph Vick, a quiet, serious man who was now semi-retired, was on the panel because he had designed dozens of valves in his long career. He had never forgotten the day 30 years earlier when hot met cold and the 747 valve stuck. Maybe the rudder valve on the USAir plane had stuck the same way.
In his room at the L'Enfant Plaza Hotel, he sketched how the test might go. The valve would be frozen to 40 below zero the coldest air Flight 427 would have encountered then pumped with 170-degree hydraulic fluid.
No one expected a breakthrough. Every day, hundreds of 737s withstood huge temperature swings without problems. Jean McGrew, Boeing's chief engineer for the plane, said there wouldn't be a thermal jam until 737s could fly to the moon.
Still, Haueter and Phillips were willing to give the test a try. Why not? "We've tried the obvious and the remote," Haueter said. "We're now open to the bizarre."
Two months later, in August 1996, the safety board investigators and the Greatest Minds in Hydraulics gathered at Canyon Engineering, a tiny hydraulics company in Valencia, Calif., that looked more like a garage than a modern test facility. They would test two rudder valves one from the USAir crash and a new one straight from the factory.
The PCUs, the heavy steel contraptions that contained the valves, were placed in a big, white Coleman cooler and frozen with nitrogen gas. The room filled with a steady rhythm of clicks and hisses as the PCUs were put through their calisthenics. Click, hiss, click, hiss.
Both PCUs were passing every test. Things were going so smoothly that the Greatest Minds started to pack up and say goodbye. Another theory ruled out.
They had reached the most extreme condition, when the frigid PCU from the USAir plane was pumped with 170-degree fluid. Click, hiss, click, hiss.
But then: Click, hisssssssssss.
The valve had jammed.
* * *
It took six weeks before anyone believed the jam was real.
The Valencia tests had been conducted under such sloppy conditions that no one could be sure the results were valid. But then it happened again in a sophisticated Boeing lab: The valve from Flight 427 jammed; the valve straight from the factory did not. Not only had the tests proved that the valve could jam, they had shown the Flight 427 valve was more prone to jam.
Still, nobody was shouting that the mystery was solved. There was no evidence of a thermal shock on the USAir plane. Nor were there any scratches to indicate the valve had jammed before the crash.
But then a young Boeing engineer named Ed Kikta noticed something strange in the test results. Poring over charts and numbers at his desk, Kikta discovered odd dips in the hydraulic flow that shouldn't have been there.
If he was reading the data right, it would mean that when the valve jammed in the thermal shock test, the rudder would not have gone to neutral, the way it should. It would have reversed.
A pilot could push on the right pedal, but the rudder would go left.
That would mean catastrophe.
Could this be right? Kikta double-checked the data. Sure enough, there it was. A reversal.
He looked up and saw his boss, Jim Draxler, putting on his coat to leave. Kikta stopped him.
"We might have a problem here," he said.
The beauty of the valve-within-a-valve system was that it was designed to work even if one part jammed. But Kikta's discovery meant a single jam could cripple a plane and even cause a crash.
Boeing sprang into action. For six days, engineers worked feverishly to confirm Kikta's discovery. They concluded that levers were flexing slightly, which allowed hydraulic fluid to flow through the wrong set of holes in the valve.
They scrambled to get a 737 to check their theory. Boeing didn't own one that would be right for the tests, so the engineers took one right off the assembly line, before it was delivered to an airline. That usually takes two weeks to arrange, but the Boeing engineers were so worried that they got one within 24 hours.
The plane landed at Boeing Field in a cold Seattle rain on Oct. 29, 1996. Michael Hewett, the Boeing test pilot, climbed into the cockpit; Kikta stood on a platform on the tail of the plane, watching the rudder.
The first two tests went smoothly. Then came a more aggressive test. Hewett stomped on the pedal as fast as he could.
The rudder swung the wrong direction.
* * *
|Page 1||Page 3|
© Copyright 2006 St. Petersburg Times. All rights reserved.