NTSB Factual Report

HISTORY OF FLIGHT

On December 29, 2000, at 2234 eastern standard time, a Jetstream 4101, N323UE, operated by Atlantic Coast Airlines as flight 331, was substantially damaged when it departed the runway after landing at the Charlottesville/Albermarle Airport (CHO), Charlottesville, Virginia. Two flight crewmembers, 1 flight attendant, and 14 passengers were uninjured. One passenger received minor injuries. Visual meteorological conditions prevailed, and an instrument flight rules flight plan was filed for the scheduled passenger flight. The flight originated at the Washington-Dulles International Airport (IAD), Washington, DC, and was conducted under 14 CFR Part 121.

After landing on runway 21, the airplane rolled out to the end, turned left onto the last taxiway, then continued onto a grass apron, and down a 60-foot embankment.

In a telephone interview about 1 hour after the accident, the captain was asked to describe what happened. He stated:

"I couldn't tell you. It was a textbook landing. We landed in the touchdown zone, on speed. I pulled both throttles to idle and got a bad reverser on the right side. I got a red beta light. It's policy not to use reverse when you get the red light.

I got on the wheel brakes harder, and the brakes didn't slow us down. I just decided to do a high-speed turn off onto the taxiway. I didn't get any other bad indications. I didn't get any with the anti-skid, but the plane just wouldn't slow down."

On December 30, 2000, about 1400, the flight crew was interviewed in person. The captain again provided his recollection of events. He said:

"It was a textbook visual approach and landing in VMC conditions. It was on speed, in the touchdown zone, and on centerline, with the throttles at idle. At touchdown, I was on the brakes with light toe pressure; I pulled the finger lift, and got a red attention-getter. I went to the annunciator, and had a red beta. Coming back up over the stop, the red beta went out. I saw the red beta went out, but I never saw the green. I said, 'We have a bad beta'.

I was hard on the brakes. The airplane veered right, and I was on the tiller to steer back to the centerline. We were not slowing down. The [runway lights] were changing, so I knew we were near the end.

I did a high-speed taxi off the side to avoid all the metal off the end. We got about 90 degrees [through the turn], but we couldn't make it, and went down over the embankment. We egressed, counted everybody, and waited for someone to realize what was happening."

The captain was asked what the computed approach speed was for the flight, and the range of speeds for minimum weight to maximum allowable gross weight. He was also asked what the computed accelerate/stop and landing distances were for the flight. The captain said he could not recall the answers to any of the questions. However, he explained that tabular charts determined all distances and speeds. The charts converted all parameters (wind speed and direction, runway condition, etc.) to a weight value, and these values were added to the total aircraft weight. When the total weight value was below the maximum allowable for the proposed flight, the flight was completed, and the charts provided takeoff, approach, and landing speeds.

The captain was asked if he felt he could explain the accident. He stated:

"I really don't know. When we were on the brakes, it just didn't seem to be slowing down like a normal Jetstream. We were hard on the brakes, but the brakes just didn't seem to be slowing us down."

When interviewed, the first officer explained that it was the responsibility of the pilot-not-flying (PNF) to determine the landing weight of the airplane, based on the tabular data, and choose the card that provided the correct approach and landing speeds. He stated:

"It was a VFR approach to runway 21. I made the 1,000-foot call, and the 500-foot call. I called [speeds] plus 15, plus 10, REF, and we touched down on centerline. I didn't see the red beta light, but the captain said, 'We have a bad beta, we're not gonna use reverse, we're gonna use the brakes.'

I called 70 knots, and moved the condition levers to taxi. I looked out and could see the end of the runway was getting closer. When we were in the final turn, I applied some brake pressure because we were a little fast."

When asked whose responsibility it was to move the condition levers from the flight position (100 percent rpm), to the taxi position (72 percent rpm), the first officer explained that the pilot flying the airplane controls the power levers, and the pilot not flying (PNF) moves the condition levers.

The accident occurred during the hours of darkness, at 38 degrees, 08 minutes north latitude, and 078 degrees, 27 minutes west longitude.

PILOT INFORMATION

The captain held an airline transport pilot certificate with ratings for airplane multi-engine land. The captain's most recent first-class medical certificate was issued September 6, 2000.

The captain reported 4,050 hours of flight experience, all of which were in multi-engine airplanes. He reported 1,425 hours of experience in the Jetstream 4100, 400 hours of which were as captain. The captain also reported he had 45 hours of flight experience in the 30 days prior to the accident.

The first officer held an airline transport pilot certificate with ratings for airplane single engine land, multi-engine land, and instrument airplane. The first officer's most recent first-class medical certificate was issued November 7, 2000.

The first officer reported 4,818 hours of flight experience, 950 hours of which were in multi-engine airplanes. He reported 68 hours of flight experience in the Jetstream 4100, 32 hours of which were in the 30 days prior to the accident. He reported 4 hours of experience in the 24 hours prior to the accident.

AIRCRAFT INFORMATION

The airplane was a 1995 Jetstream Aircraft Limited Model 4101, and was registered in the transport category. It was powered by two Allied Signal TPE-331 turbo-shaft engines.

The airplane had accrued 14,456 hours of total flight time. It was on a continuous airworthiness inspection program, and its most recent inspection was completed on December 29, 2000.

METEOROLOGICAL INFORMATION

The weather reported at Charlottesville included calm winds with a broken cloud layer at 12,000 feet. The temperature was 24 degrees Fahrenheit and the dewpoint was 15 degrees Fahrenheit.

AERODROME INFORMATION

The Charlottesville-Albermarle Airport was at a field elevation of 639 feet mean sea level (msl), and had a single grooved runway. Runway 03/21 was 6,001 feet long and 150 feet wide.

Six perpendicular taxiways, designated A thru F, connected the runway to the parallel taxiway, designated north/south.

The runway surface, which was inspected approximately 4 hours after the accident, was clean and dry.

FLIGHT RECORDERS

On January 4, 2001, data from the airplane's Flight Data Recorder (FDR) was retrieved and graphed by a Safety Board recorder specialist. According to the specialist's factual report:

"For the accident landing, the increased activity in the vertical acceleration FDR data, at approximately 11,307.55 seconds, indicated initial touchdown occurred. At the time of this initial touchdown, the magnetic heading was approximately 210°, the indicated airspeed was approximately 98.5 knots, the pitch attitude was 2.58° nose up, and the roll attitude was 1.25° right wing down.

Approximately 2.2 seconds later, the "Air-Ground Nose discrete" changed from "flight" to "ground" (at 11,309.77 seconds). Less than a second later, at 11,309.88 seconds, the pitch attitude had decreased to -2.11 degrees nose down. One second after the Air-Ground Nose discrete changed from "flight" to "ground", the Air-Ground Left Main discrete changed from "flight" to "ground" (at 11,310.77 seconds) and then one second later, the Air-Ground Right Main discrete also changed from "flight" to "ground" (at 11,311.77 seconds).

Approximately 12.8 seconds [later], at 11,324.55 seconds, with an indicated airspeed of 69.5 knots, the left and right spoiler discretes changed from "stow" to "deploy" and then 5 seconds later, with an indicated airspeed of 57.2 knots, the left and right spoiler discretes changed back to "stow" (at 11,329.55 seconds), where it remained for the rest of the recorded FDR data.

Approximately 21.2 seconds later, at 11,350.77 seconds, the Air-Ground Nose discrete changed from "ground" to "flight" where it remained for the rest of the recorded FDR data. One second later, at 11,351.77, the Air-Ground Left Main discrete changed from "ground" to "flight" and then 3 seconds later, at 11,353.77, changed back to "ground" where it remained for the rest of the FDR data. Similarly, one second after the Air-Ground Left Main discrete had changed back to "flight", the Air-Ground Right Main discrete changed from "ground" to "flight" (at 11,352.77 seconds), where it remained for the rest of the FDR data. Approximately 29 seconds later, at 11,382 seconds, the recorded FDR data ended."

Examination of the FDR data revealed that about 13 seconds after touchdown, and approximately 70 knots, engine power decreased and the spoilers deployed. Five seconds later, engine power increased and the spoilers stowed.

A review of the FDR data from the two landings previous to the accident flight revealed that engine rpm stabilized around 72 percent, and engine torque stabilized around 6 percent, during the landing roll.

A review of the accident landing revealed that engine rpm remained between 80 and 100 percent, and engine torque values were between 22 and 30 percent, until the airplane went over the embankment.

WRECKAGE INFORMATION

The airplane was examined on December 30, 2000, and all major components were accounted for at the scene. A set of three skid marks, which corresponded with the positions of the airplane's landing gear, began about 2,500 feet prior to the departure end of the landing runway.

The skid marks angled slightly to the right side of the runway, then arced 90 degrees to the left, and onto taxiway alpha, at the very end of the runway. The skid marks continued about 400 feet on taxiway alpha, and then departed the pavement into the grass, at the intersection of alpha and the parallel taxiway, where they continued over the embankment and down to the wreckage.

The airplane came to rest upright, and faced downhill on a 47-degree slope, about 120 degrees magnetic. The airplane rested on its nose at the bottom of the slope where the terrain leveled off. The drop from the landing surface to the bottom of the embankment measured approximately 60 feet.

The nose gear was separated from the airframe and was on the ground, connected only by a cable.

The right main landing gear was twisted and collapsed, and the right wing rested on the ground. All five propeller blades on the right engine displayed twisting, bending, leading edge gouging, and chordwise scratching.

The left main landing gear was down and locked, and supported the left wing. The left propeller appeared undamaged. The inboard tire on the left main landing gear was shaved flat, and exhibited a hole approximately 2 inches in diameter.

Both wings and the wing box structure were displaced aft as a single unit. The fuselage structure was not twisted or wrinkled.

According to the cabin attendant, the main cabin door, the aft cabin door, the two over-wing exits, and the emergency exit lighting had all operated effectively.

TESTS AND RESEARCH

At the request of the flight crew, several components from the airplane's braking and anti-skid systems were functionally tested. The crew also requested that the flight idle time delay baulk relay, which prevents inadvertent movement of the power levers aft of the flight idle position in flight, and the right beta switch be tested.

The anti-skid control box, four wheel-position transducers, and the inboard and outboard anti-skid control valves were operational and tested within maintenance inspections limits.

All wheel brake assemblies were pressurized and found operational, with no leaks noted. The brake pistons retracted, and running clearance was within maintenance inspection limits.

The flight idle time delay baulk relay switch was bench tested with no defects noted.

The right beta switch was tested, and operated within the operational test parameters. However, the technician stated that the switch would not be returned to service due to occasional intermittent operation.

In addition, the propeller pitch control, right-hand weight on wheels switch, right engine fuel control, and right engine governor control unit were all functionally tested and found operational.

Both the captain and the first officer stated that, other than the red beta light indication, there were no mechanical deficiencies with the airplane.

Tests were conducted in both the Jetstream 4101 airplane, and the Jetstream 4101 flight simulator. The purpose of the tests was to duplicate the power control positions, flight control positions, and speeds from the accident flight as represented by the FDR data, and observe the effect on the performance, handling, and braking of the airplane.

Tests in the airplane were conducted on February 20, 2002. The fleet manager of turbo-props for Atlantic Coast Airlines flew the airplane from Dulles International Airport to the Charlottesville-Albermarle airport to conduct the tests. He flew the first approach and landing to determine the approximate touchdown point on the runway. According to the fleet manager's report:

"The first approach was flown straight in following approximately the same course as that flown by the accident aircraft. The aircraft was flown using the VASI for vertical guidance until such time as it was necessary to flare the aircraft for landing. The aircraft touched down approximately parallel to the 4000 foot remaining marker. Normal deceleration using reverse and a very light application of braking just prior to initiating the turn off the runway at approximately 40 knots was more than adequate to allow the aircraft to comfortably stop by the second to the last intersection."

After a second approach and landing were completed with similar results, a high-speed taxi test was conducted. According to the fleet manager's report:

"[The purpose was] to determine the approximate speed the aircraft would achieve on the ground roll with the condition levers at the full forward or flight position and the power levers at the flight idle position, which is just ahead of the gate. Secondly to determine the same with the condition levers at the full aft or taxi position.

This was accomplished by performing a high-speed taxi from a standing start using the Static Takeoff Profile from the Atlantic Coast Airlines J41 Flight Standards Manual. It consists of holding the brakes and advancing the power levers to a setting of approximately 30-40 percent torque before brake release. After brake release the power is advanced to 100 percent torque or the limiting temperature. The aircraft was allowed to accelerate to approximately 70 knots with the power at 100 percent before the power levers were brought back against the flight-idle gate.

The aircraft continued to accelerate to a speed of approximately 81 knots before it stabilized. When the condition levers were brought back to the full aft or taxi position the aircraft started to accelerate once again. It was still accelerating through 85 knots when the high-speed taxi was discontinued."

On February 27, 2002, flight tests were performed in the Jetstream 41 simulator in Dulles, Virginia, under Safety Board supervision. The flight test was conducted with the supervisor of flight standards for the J41 and the director of safety for Atlantic Coast Airlines, and the Air Line Pilots Association (ALPA) representative.

The simulator was programmed to approximate the environmental conditions and the airplane's configuration on the night of the accident.

Several approaches and landings were flown to a runway 6,000 feet long, with various combination

NTSB Probable Cause

The captain's improper application of power after responding to a beta warning light during landing rollout, which resulted in an excessive rollout speed and an inability to stop the airplane before it reached the end of the runway.