http://www.boston.com/dailyglobe2/320/nation/Data_tell_how_but_not_why_plane_broke_apart_and_crashed%2b.shtml

Data tell how, but not why plane broke apart and crashed

By Glen Johnson, Globe Staff, 11/16/2001

WASHINGTON - American Airlines Flight 587 appears to have flown safely through two waves of turbulence it encountered after takeoff, but was then severely jostled and experienced unexplained rudder movements at a time when the pilots appeared to be fighting for control of the aircraft.

In releasing the information yesterday, the National Transportation Safety Board could not explain whether the jostling was caused by a controlled movement of the rudder or whether the rudder moved as the tail began to break apart.

The board also disclosed that the plane's flight data recorder, stored in the tail, quit working while the aircraft was still 2,900 feet in the air, and after the unit had stopped picking up reliable information from the rudder panel.

Board chairwoman Marion C. Blakey stopped short of attributing cause and effect, but the outline she provided suggests that either the turbulence or the pilots' fight for control might have damaged the tail, cutting off power in the area, causing parts to separate, and leaving the plane turning out of control as it fell into a neighborhood in Queens, N.Y., on Monday.

The safety board is looking closely at the carbon-fiber construction of the tail assembly, questioning whether a break in the strong but lightweight plastic may play a role in the first airliner crash of its kind. Use of the material is spreading throughout the aircraft industry, primarily because it promotes fuel efficiency.

''I would point out that the investigation continues, on the basis of all the information we have to this point, to point to an accident,'' Blakey told reporters during a briefing in New York. The crash killed all 260 aboard the flight from New York to Santo Domingo, in the Dominican Republic, as well as five on the ground.

Based on the findings, the Federal Aviation Administration and the NTSB's counterpart in France, where the Airbus A300 that crashed was made, planned to issue a joint airworthiness directive today requiring operators of the A300 and its sister model, the A310, to inspect the tail assembly in a matter of days.

On Wednesday, American Airlines announced that it would voluntarily inspect the tail assemblies on the remaining 34 A300s in its fleet and eventually inspect the same part of all its aircraft. Federal Express has 37 A300s and 46 A310s, while United Parcel Service operates 20 A300s. Worldwide, there are 411 A300s and 213 A310s.

The carbon-fiber tail fin on the plane that crashed, known as the vertical stabilizer, snapped off just above six carbon loops that are used to attach the fin to the fuselage. The loops fit into metal sockets on the fuselage, and the connnection is secured with six titanium metal bolts.

The carbon fiber around one of the six loops on the tail fin was repaired during the plane's manufacturing in 1988, the NTSB said this week.

Investigators said the metal fittings and bolts held through the crash, but the vertical stabilizer snapped off on a line that runs across the carbon loops.

That has not happened before, and the discovery is challenging the 400-person NTSB. The agency has only two experts on the subject of composite components. The FAA has dispatched its own expert to provide assistance in the investigation.

The same aircraft encountered severe turbulence in November 1994 while on a flight from the Caribbean to Boston, injuring 47 people aboard. Blakey said yesterday the plane was inspected after the incident, but ''the mechanics in doing that did not discover anything, in terms of the aircraft, that warranted repair at the time.''

The safety board's afternoon briefing was filled with technical aeronautical jargon.

Blakey said the unit stopped operating about 20 seconds before the cockpit voice recorder. The voice recorder has a battery backup, which would have allowed it to keep operating even after a power disruption.

Both units are located in the tail, but not in the vertical stabilizer or its trailing component, the rudder panel.

The American plane started moving down the runway 1 minute and 45 seconds after a Japan Airlines Boeing 747-400, more than double its weight, began its own takeoff roll on the same runway. The planes flew parallel tracks about three-quarters of a mile apart side to side, with the American plane riding about 800 feet lower but only 90 seconds behind the jumbo jet.

As a rule of thumb, controllers have airplanes wait two minutes between takeoffs, but FAA rules state only that these type of aircraft start at least four miles apart. At the altitude the planes had reached, the guidelines call for such aircraft to be separated by a minimum of 3 miles horizonatally or 1,000 feet vertically.

Blakey said Wednesday the positioning could leave the Airbus in the unsettled air trailing the JAL plane, subjecting it to mini-tornadoes known as wake vortices. Yesterday, she said the flight data recorder provided information consistent with the American plane's hitting two wakes.

The first came about 28 seconds before the recorder data stopped flowing. It jostled the Airbus from side to side, but the plane's direction and altitude did not change, nor did the pilots appear to provide any response.

A few seconds later, the cockpit voice recording picked up an airframe rattle, which would be expected with a wake encounter.

About 20 seconds later, just eight seconds before the data stopped flowing, the flight data recorder shows the Airbus hit a second wave of turbulence of about the same intensity as the first.

The cockpit recorder picked up the sound of the captain, Edward States, commenting about a ''wake encounter'' a few seconds later.

According to the data recorder, in the first few seconds after hitting the new wave of turbulence, the plane appeared to respond to flight-control moves made by the pilots.

About that time, the copilot, Sten Molin, who was flying the plane, called for maximum power, which pilots would do to recover from a wind sheer, another, more severe type of turbulence.

A few seconds later, however, the plane was subjected to stronger forces, two sideway pushes of 0.3 and 0.4 times the force of gravity in one direction, and one in the opposite direction of 0.3 times the force of gravity.

''The point to take away from this is this is a very significant lateral acceleration we're talking about here,'' Blakey said.

Neither she nor Thomas A. Haueter, the NTSB's deputy aviation safety director, would say if the deflections were caused by the pilots or inputs from the autopilot or other automatic control mechanisms on the aircraft.

In that time frame, the cockpit voice recorder picked up comments from the pilots suggesting a loss of control.

Haueter said at a briefing last night: ''The wake vortices are essentially over by the time we have these increases in lateral accelerations.'' Both pilots had received special simulator training from American, including a scenario in which their plane was upset by the wake of a Boeing 747.

About two-and-a-half seconds before the flight data recorder stopped, information about the rudder became unreliable, Blakey said. The panel has a sensor that is wired to the recorder.

Finally, the plane was subjected to even stronger forces, which also remain unexplained. Blakey said the sideways accelerations increased from about 0.3 times the force of gravity to about 0.8 times, the direction the nose began to move about 10 degrees per second, and the bank angle - the measure of how level the wings are on the horizon - increased past 25 degrees to the left, indicating the left wing was dropping sharply.

At the time, the pilots' control yoke was turned to the right, the opposite direction from the wing, Haueter said.

The final two pieces of data provided by the recorder showed the plane's nose was pointed 30 degrees down, while the aircraft was subjected to a load exceeding twice the force of gravity.

If the plane continued spinning at the rate it started, it would have turned about 200 degrees, or more than halfway around from the direction it had been intending to fly.

Airplanes are designed to withstand any force that might batter them in the sky, from rough air to lightning strikes or a sudden movement to avoid a crash. Haueter said the initial forces that struck the Airbus were ''slightly higher that you would see,'' but forces exceeding 0.4 sideways are not common.

That leads investigators back to the tail assembly itself and questions about whether there is a flaw in the design, or whether the aircraft had some undetected flaw that was exacerbated when it hit the wake of the JAL airplane.

The aircraft model that crashed, the A300-600R, uses carbon-fiber reinforced plastic, akin to the fiberglass in boat hulls. It is located in the tail fin, the rudder panel, engine casings known as nacelles, the doors that close over the landing gear, as well as the nose cone and the flaps.

It also has plastic reinforced with glass fibers on the edges of the fin and on parts of the wings.

Airbus started using the materials on the A300 in the early 1980s and has steadily increased their use in newer models. The manufacturer is planning widespread use of composites in the A380, a two-deck behemoth that will be bigger than the Boeing 747 and is currently on the drawing board.

In spreading the technology, Airbus has also worked with its buyers to make sure airlines are trained in inspecting and maintaining the materials.

-- Anonymous, November 16, 2001