Texas

Approximately 8 months prior to the accident, during a cross country flight, the owner shutdown the left engine due to low oil pressure and diverted from his intended destination to a nearby airport. During descent, the right alternator failed, and the owner performed the emergency gear extension procedure. Following an emergency gear extension, the landing gear of this model airplane cannot be retracted until the system has been ground serviced. A mechanic reported that about 7 months prior to the accident, with the owner present, he removed the oil filter from the left engine, found it packed with metal shavings and told the owner that the engine needed overhaul. Two other mechanics reported that approximately three weeks before the accident, they installed an oil filter on the left engine, changed the oil, and cleaned the oil pressure regulator. They ground ran both engines with no discrepancies noted. One of the mechanics reported that following the engine run, the left engine oil filter was removed, examined, and no metal was found. The landing gear was not serviced. According to the owner, the pilot was "hired" by one of the two mechanics to ferry the airplane with the gear extended to a location where the gear could be serviced. While en route, the pilot reported a loss of power on the left engine, that he was having trouble feathering the engine, that the airplane would not maintain altitude and he was looking for a place to land. Witnesses observed the airplane flying low, wheels down and losing altitude. They further observed it roll into a steep left bank, hit trees and a fence, catch fire, come to rest inverted on a road and burn. Post accident examination of the left engine revealed a hole in the right crankcase half over the #3 cylinder attach point. Disassembly of the left engine revealed that the #3 connecting rod was separated from the crankshaft, and the rod bolts, rod cap, and top of the rod were deformed. The #5 piston pin had one cap missing. Scoring was noted on the crankshaft journals, and the main bearings exhibited discoloration and deformation consistent with oil starvation. The cylinders exhibited deformation, scoring in the barrels, and deposits on the domes. The camshaft exhibited discoloration and scoring on the camshaft lobes. Disassembly of the left propeller revealed that it was in the vicinity of low pitch/latch position and not rotating at impact. The disassembly of the right engine and propeller did not reveal any discrepancies that would have precluded operation prior to impact. Estimates of the airplane's climb performance indicated that with the landing gear down and the left propeller stopped, it was not capable of sustained flight.

The flight was approaching a private airport (elevation 983 feet msl), that did not have an instrument approach system, during instrument meteorological conditions. The pilot informed the air traffic controller that he had the airport in sight, and cancelled his instrument flight plan. The twin turbofan airplane impacted a tree approximately 4,000 feet northeast of the airport in an upright position. The airplane then impacted the ground in an inverted position approximately 200 yards from the initial impact with the tree. The weather observation facility located 16 miles northeast of the accident site was reporting an overcast ceiling at 400 feet agl, and visibility 4 statute miles in mist. The weather observation facility elevation was 541 feet msl. Local residents in the vicinity of the accident site stated that there was heavy fog and drizzle at the time of the accident. The pilot had filed an alternate airport (with a precision instrument approach); however, he elected not to divert to the alternate airport. Examination of the wreckage did not reveal any evidence of pre-impact anomalies that would have prevented operation of the airplane.

The captain was the flying pilot for the night landing on runway 01 in instrument meteorological conditions (IMC), with a right cross wind from 110 degrees at 14 gusting 18 knots, drizzle, and a wet runway. Prior to starting the approach, the flightcrew determined that the landing approach speed (Vref) and the approach speed (Vapp) were 122 and 128 knots, respectively. DFDR data showed the airplane flying on autopilot as it passed the middle marker at 200 feet AGL at 130 knots on the ILS approach. Approximately 3 seconds after the first officer called "runway in sight twelve o'clock," the captain disconnected the autopilot, while at a radio altitude of 132 feet and on a heading of approximately 18 degrees. Within approximately 11 seconds after the autopilot disconnect, the glideslope and localizer deviation increased. The first officer called "runway over there." Approximately 5 seconds before touchdown, the airplane rolled right, then left, then right. DFDR data-based performance calculations showed the airplane crossed the threshold at an altitude of 35 feet and 130 knots. The airplane touched down 2,802 feet from the approach end of the 5,495-foot runway (844- foot displaced threshold) at 125 knots on a heading of 10 degrees. The airplane overran the runway and struck a ditch 175 feet beyond the departure end of the runway. Landing roll calculations showed a ground roll of 2,693 feet after touchdown, consisting of 1,016 feet ground roll before braking was initiated and 1,677 feet ground roll after braking was initiated until the airplane exited the pavement. According to Saab, for a wet runway, the aircraft would have needed 1,989 feet from the time of braking initiation to come to a complete stop. The American Eagle Airlines, Inc., FAA approved aircraft operating manual (AOM), states in part: Stabilized approaches are essential when landing on contaminated runways. Landing under adverse weather conditions, the desired touch-down point is still 1,000 feet from the approach end of the runway. Touchdown at the planned point. Cross the threshold at Vapp, then bleed off speed to land approximately Vref -5. Use reverse, if needed. To achieve maximum braking effect on wet runway, apply maximum and steady brake pressure. In 1992, the City of Killeen submitted a proposal that included extending the north end of runway 01 by 194 feet. The FAA originally disapproved the proposal, in part, because the runway extension decreased the length of the runway safety area (RSA) which was already shorter than the recommended 1,000 feet for a 14 CFR Part 139 certificated airport. The proposal was subsequently approved and a drainage ditch was installed in the north RSA, perpendicular to the runway and approximately 175 feet north of the departure end of runway 01. In 1993, the airport received FAA Part 139 certification. The 1998 and 1999, FAA airport certification inspection reports noted the inadequate RSA; however, neither letter of correction, sent from the FAA to the City of Killeen following the inspections, mentioned the RSA. Following this accident, the ILS runway 01 was flight checked by the FAA and all components were found to be operating within prescribed tolerances. Examination of the airplane found no anomalies that would have prevented it from operating per design prior to departing the runway and encountering the ditch.

Freezing rain, mist, and ice pellets were forecast for the destination airport with temperatures 34 to 32 degrees F. During the daylight IMC descent and vectors for the approach, the airplane began to accumulate moderate clear ice, and a master warning light illumination in the cockpit indicated that the horizontal stabilizer heat had failed. The airplane was configured at 120 knots and 10 degrees flaps in accordance with the flight manual abnormal procedures checklist; however, the crew did not activate the horizontal stabilizer deice backup system. The aircraft touched down 1,500 ft down the runway, which was contaminated with slush, and did not have any braking action or antiskid for 3,000 ft on the 7,753-ft runway. Therefore, 3,253 ft of runway remained for stopping the aircraft, which was 192 feet short of the 3,445 ft required for a dry runway landing. Upon observing a down hill embankment and support poles beyond the runway, the captain forced the airplane to depart the right side of the runway to avoid the poles. After the airplane started down the embankment, the nose landing gear collapsed, and the airplane came to a stop.

Witnesses reported that during the airplane's takeoff roll they heard a heard a series of repeated sounds, which they described as similar to a "backfire" or "compressor stall." Several witnesses reported seeing the airplane's right propeller "stopped." One witness reported that as the airplane lifted off the ground, he heard "a loud cracking sound followed by an immediate prop wind down into feather." He continued to watch the airplane, as the gear was retracted and the airplane entered a climb and right turn. Subsequently, the airplane pitched up, entered a "Vmc roll-over," followed by a 360-degree turn, and then impacted the ground. Radar data indicated the airplane took off and climbed on runway heading to a maximum altitude of about 200 feet agl. The airplane than entered a right turn and began to lose altitude. A radar study revealed that the airplane's calibrated airspeed was 97 knots when the last radar return was recorded. According to the flight manual, minimum controllable airspeed (Vmc) was 93 knots. Examination of the accident site revealed that the airplane impacted the ground in a near vertical attitude. A post-crash fire erupted, which destroyed all cockpit instruments and switches. Examination of the propellers revealed that neither of the
propellers were in the feathered position at the time of impact. Examination of the left engine revealed signatures consistent with operation at the time of impact. Examination of the right engine revealed that the second stage impeller shroud exhibited static witness marks indicating that the engine was not operating at the time of impact. However, rotational scoring was also observed through the entire circumference of the impeller shroud. The static witness marks were on top of the rotational marks. Examination of the right engine revealed no anomalies that would have precluded normal operation. The left seat pilot had accumulated a total flight time of about 950 hours of which 16.9 hours were in an MU-2 flight simulator and 4.5 hours were in the accident airplane. Although he had started an MU-2 Pilot-Initial training course, he did not complete the course. The right seat pilot had accumulated a total flight time of about 2,000 hours of which 20.0 hours were in an MU-2 flight simulator and 20.6 hours were in the accident airplane. He had successfully completed an MU-2 Pilot-Initial training course one month prior to the accident.