Castle Aviation

During an instrument approach at night in a twin-engine turboprop airplane, the pilot reported an engine failure, but did not specify which engine. About 9 seconds later, the airplane impacted terrain about ¼-mile short of the runway and a postcrash fire consumed a majority of the wreckage. During that last 9-second period of the flight, the airplane’s groundspeed slowed from 99 kts to 88 kts, as it descended about 400 ft in a slight left turn to impact (the airplane’s minimum controllable airspeed was 92 kts). The slowing left turn, in conjunction with left wing low impact signatures observed at the accident site were consistent with a loss of control just prior to impact. Postaccident teardown examination of the left engine revealed that the 1st stage turbine rotor had one blade separated at the midspan. The blade fracture surface had varying levels of oxidation and the investigation could not determine if the 1st stage turbine blade separation occurred during the accident flight or a prior flight. The 2nd stage turbine was operating at temperatures higher than the 1st stage turbine, which was consistent with engine degradation over a period of time. Additionally, the 2nd stage turbine stator assembly was missing vane material from the 6 to 12 o’clock positions, consistent with thermal damage. All of these findings would have resulted in reduced performance of the left engine, but not a total loss of left engine power. The teardown examination of the right engine did not reveal evidence of any preimpact anomalies that would have precluded normal operation. Examination of both propellers revealed that all blade angles were mid-range and exhibited evidence of little to no powered rotation. Neither propeller was in a feathered position, as instructed by the pilot operating handbook for an engine failure. If the pilot had perceived that the left engine had failed, and had he secured the engine and feathered its propeller (both being accomplished by pulling the red Engine Stop and Feather Control handle) and increased power on the right engine, the airplane’s performance should have been sufficient for the pilot to complete the landing on the runway.

The cargo flight was departing on its fourth flight leg of a five-leg flight in night instrument conditions, which included a surface observation of light snow and a broken ceiling at 500 feet above ground level (agl). One pilot who departed just prior to the accident flight indicated that moderate snow was falling and that he entered the clouds about 200 feet agl. The accident airplane's wings and tail were de-iced prior to departure. Radar track data indicated the accident flight was about 45 seconds in duration. An aircraft performance radar study indicated that the airplane reached an altitude of about 1,130 feet mean sea level (msl), or about 400 feet above ground level, about 114 knots with a left bank angle of about 29 degrees. The airplane descended and impacted the terrain at an airspeed of about 155 knots, a pitch angle of -16 degrees, a left roll angle of 22 degrees, and a descent rate of 4,600 feet per minute. The study indicated that the engine power produced by the airplane approximately matched the engine power values represented in the pilot's operating handbook. The study indicated that the required elevator deflections were within the available elevator deflection range, and that the center-of-gravity (CG) position did not adversely affect the controllability of the airplane. The study indicated that the load factor vectors, the forces felt by the pilot, could have produced the illusion of a climb, even when the airplane was in a descent. The inspection of the airframe and engine revealed no anomalies that would have precluded normal operation.

The pilot departed on a night multiple-leg flight series with full tanks. On approach for the second landing, the green landing gear lights extinguished. The pilot discovered that if the landing gear was extended the circuit breaker would stay in and if the landing gear was retracted the circuit breaker would pop. The circuit breaker controlled the landing gear lights, the engine oil pressure, and both fuel quantity gauges. The pilot continued on four additional flights in that condition. The fourth flight terminated 5 miles from the runway with a double power loss. Examination revealed the fuel tanks were empty. The pilot had flown the airplane 4.3 hours including 6 takeoffs, one missed approach, and a 20 minute ground run prior to departure with both engines running on the final flight.