CEN14FA046

The pilot's loss of airplane control during a known one-engine-inoperative condition. The reasons for the loss of control and engine shutdown could not be determined because the airplane was not equipped with a crash-resistant recorder and postaccident examination and testing did not reveal evidence of any malfunction that would have precluded normal operation.

Radar and air traffic control communications indicated that the Mitsubishi MU-2B-25 was operating normally and flew a nominal flightpath from takeoff through the beginning of the approach until the airplane overshot the extended centerline of the landing runway, tracking to the east and left of course by about 0.2 nautical mile then briefly tracking back toward the centerline. The airplane then entered a 360-degree turn to the left, east of the centerline and at an altitude far below what would be expected for a nominal flightpath and intentional maneuvering flight given the airplane's distance from the airport, which was about 5 miles. As the airplane was in its sustained left turn tracking away from the airport, the controller queried the pilot, who stated that he had a "control problem" and subsequently stated he had a "left engine shutdown." This was the last communication received from the pilot. Witnesses saw the airplane spiral toward the ground and disappear from view. Examination of the wreckage revealed that the landing gear was in the extended position, the flaps were extended 20 degrees, and the left engine propeller blades were in the feathered position. Examination of the left engine showed the fuel shutoff valve was in the closed position, consistent with the engine being in an inoperative condition. As examined, the airplane was not configured in accordance with the airplane flight manual engine shutdown and single-engine landing procedures, which state that the airplane should remain in a clean configuration with flaps set to 5 degrees at the beginning of the final approach descent and the landing gear retracted until landing is assured. Thermal damage to the cockpit instrumentation precluded determining the preimpact position of fuel control and engine switches. The investigation found that the airplane was properly certified, equipped, and maintained in accordance with federal regulations and that the recovered airplane components showed no evidence of any preimpact structural, engine, or system failures. The investigation also determined that the pilot was properly certificated and qualified in accordance with applicable federal regulations, including Special Federal Aviation Regulation (SFAR) No. 108, which is required for MU-2B pilots and adequate for the operation of MU-2B series airplanes. The pilot had recently completed the SFAR No. 108 training in Kansas and was returning to Tulsa. At the time of the accident, he had about 12 hours total time in the airplane make and model, and the flight was the first time he operated the airplane as a solo pilot. The investigation found no evidence indicating any preexisting medical or behavioral conditions that might have adversely affected the pilot's performance on the day of the accident. Based on aircraft performance calculations, the airplane should have been flyable in a one-engine-inoperative condition; the day visual meteorological conditions at the time of the accident do not support a loss of control due to spatial disorientation. Therefore, the available evidence indicates that the pilot did not appropriately manage a one-engine-inoperative condition, leading to a loss of control from which he did not recover. The airplane was not equipped, and was not required to be equipped, with any type of crash-resistant recorder. Although radar data and air traffic control voice communications were available during the investigation to determine the airplane's altitude and flightpath and estimate its motions (pitch, bank, yaw attitudes), the exact movements and trim state of the airplane are unknown, and other details of the airplane's performance (such as power settings) can only be estimated. In addition, because the airplane was not equipped with any type of recording device, the pilot's control and system inputs and other actions are unknown. The lack of available data significantly increased the difficulty of determining the specific causes that led to this accident, and it was not possible to determine the reasons for the left engine shutdown or evaluate the pilot's recognition of and response to an engine problem. Recorded video images from the accident flight would possibly have shown where the pilot's attention was directed during the reported problems, his interaction with the airplane controls and systems, and the status of many cockpit switches and instruments. Recorded flight data would have provided information about the engines' operating parameters and the airplane's motions. Previous NTSB recommendations have addressed the need for recording information on airplane types such as the one involved in this accident. Recorders can help investigators identify safety issues that might otherwise be undetectable, which is critical to the prevention of future accidents.

The airplane was not equipped, and was not required to be equipped, with a cockpit voice recorder, flight data recorder, or cockpit image recorder. No problems with communications equipment were reported. The Tulsa International Airport, TUL, is a public, controlled airport located about 5 miles northeast of Tulsa, Oklahoma, at a surveyed elevation of 677.5 feet. The airport features two concrete runways, runway 18L/36R, which is 9,999 feet by 150 feet, and runway 8/26, which is 7,376 feet by 150 feet. Runway 18R/36L is asphalt and 6,101 feet by 150 feet. The runway 18L threshold is at an elevation of 626.5 feet, and the runway slopes upward at a 0.4-percent gradient. The runway contains a four-light precision approach path indicator (PAPI) on its left side with a 2.75 -degree glidepath. The listed obstruction to the runway is a 41-foot tree, which is located 1,894 feet from the runway and requires a 41:1 slope to clear. According to the instrument landing system (ILS) approach plate for runway 18L, the outer marker, identified as OWASO, is 5.6 nautical miles (nm) from the end of runway 18L. The glideslope/glidepath crossing altitude at OWASO is 2,346 feet msl. Aircraft Performance Radar Study The aircraft performance radar study used Airport Surveillance Radar (ASR) data to calculate the position and orientation (pitch, yaw, and roll angles) of the airplane in the minutes preceding the accident. This information was then used to estimate various performance parameters of interest, including horizontal and vertical speeds, terrain clearance, AOA and proximity to stall, and required engine power. The ASR-9 radar at TUL received returns from the airplane starting at 15:30:13, when the airplane was 60 nm northwest of the TUL runway 18L threshold and descending through 14,500 feet msl. The TUL ASR continued to track the airplane throughout its approach to TUL, including the last minutes of the flight. From 15:43:37 to 15:44:04, as the airplane was on final approach between 6.1 and 5.1 nm from runway 18L, there was a 27-second gap in the ASR secondary data (corresponding to five missing radar returns). However, during this time, three primary returns consistent with the position of the airplane were received by the ASR. In addition, the Chelsea/Afton, Oklahoma ATCBI-6 (QAF) radar received secondary returns from the airplane. The QAF secondary returns indicate that the gap in the TUL secondary returns was not due to a problem with the airplane's transponder but to some other unknown cause. The data indicated that at 15:42:50, the airplane was descending on a southeasterly heading through 2,500 feet msl about 8 nm north of the TUL runway 18L threshold. About 15:43:00, the airplane leveled off briefly at 2,200 feet msl and started a right turn toward the runway. At this time, the airplane was already below the PAPI glidepath for TUL runway 18L. The airplane crossed the extended runway centerline at 15:43:19 and continued the right turn to correct back to the centerline. When the airplane resumed its descent at 15:43:25, the PAPI would have displayed four red lights, indicating that the airplane was below the PAPI centerline. About 15:44:00, as the airplane was descending through 1,400 feet msl, it started a left turn that continued to the end of the radar data, with the airplane almost completing a full 360-degree turn. During this turn, the pilot reported to air traffic control (ATC) that "I've got a control problem" (at 15:44:51) and that "<unintelligible> I've got a left engine shutdown" (at 15:45:06). The airplane crashed less than 0.05 nm southwest of the last radar return, about 5 nm north of the runway threshold, and about 0.05 nm left (east) of the extended runway centerline. The airspeed data presented in the study indicated that the airplane was operating close to the 20-degrees flaps, one-engine inoperative minimum controllable airspeed (Vmc, 20) of 93 knots calibrated airspeed (KCAS) during the time that the pilot reported control and engine problems. In addition, the calculations indicated that shortly before the end of the radar data, the airplane's lift coefficient (CL) reached the maximum CL (CLmax) for the flaps 20 configuration, which suggested that the final descent of the airplane into the ground followed an aerodynamic stall of the wing. This finding was consistent with the condition of the wreckage, its location very close to the last radar point, and witness statements. The CL and airspeed data computed from the radar returns indicated that the flaps must have been deployed at some time before 15:44:18, because the airplane's calculated CL exceeded the CLmax for 0-degrees flaps beyond this time, but the airplane continued flying. Consideration of the power requirements computed from the radar data suggested that the flaps may have been deployed to 20 degrees around 15:43:30, shortly after the airplane decelerated below the speed at which flaps 20 could be extended (140 KCAS). The roll angle computed from the radar data indicated that the airplane required about 13 degrees of roll during the right turn between 15:42:55 and 15:43:50, as the airplane maneuvered to line up with the extended runway centerline. The roll angle required during the final 360 -degree left turn was about 15 to 25 degrees, with the roll angle increasing from about 5 degrees left to 22 degrees left between 15:44:20 and 15:44:30. This increase in roll angle corresponded to the time that the airplane arrested its deceleration and descent (that is, decay in energy) and leveled at about 95 KCAS and 1,100 feet msl, or about 400 feet agl. Associated with this level-off was an increase in required horsepower. The power increased to about the maximum available from one engine for the corresponding flight conditions. Because, as suggested by the pilot's reports of an engine problem, the increase in horsepower was only available from one engine, then any thrust asymmetry between the two engines would also increase and would increase the rudder deflection and/or sideslip angle required to compensate for the asymmetry. Consequently, the increase in roll angle at this time may reflect these changing parameters affecting the trim of the airplane. Also, by 15:44:15, the airspeed had already decayed to around 95 KCAS, close to the Vmc, 20 of 93 KCAS. Consequently, with full power on the operating engine, and at this speed, the airplane was close to the limit of controllability. The airplane may have been easier to control at lower power settings on the operating engine but may still have presented a challenging situation to the pilot, given the low energy state of the airplane and its proximity to the ground. During the final 360-degree left turn, the highest priority to ensure the safety of the flight would have been to increase the control margin by increasing the airspeed further above the 93 KCAS Vmc, 20 speed. However, to increase the speed, a pilot would have to increase power on the operating engine (thereby exacerbating the thrust asymmetry and control problem at low speed, even if additional power were available), trade altitude for airspeed (which a pilot may be reluctant to do if the airplane is already at a low altitude), or perform some combination of these actions. A pilot could also increase the speed and margin from Vmc, 20, by retracting the landing gear, thereby lowering the airplane's drag. Hence, at the time the power was increased between 15:44:10 and 15:44:30, the airplane was already in a difficult situation because of the combination of low altitude, low airspeed, and the reported problem with the left engine. Engine Examination The engines were disassembled at Honeywell's facilities in Phoenix, Arizona, under the supervision of the NTSB. Disassembly and examination of the engines did not reveal evidence of preimpact malfunctions. Disassembly of the right engine revealed the compressor section 1st stage