CEN19FA088

The pilot's conduct of a departure into instrument meteorological conditions (IMC), which resulted in spatial disorientation and subsequent loss of airplane control. Contributing to the accident was the pilot's poor instrument flying skills and his decision to depart into IMC with an unresolved autopilot maintenance issue.

The instrument-rated private pilot and passenger departed into instrument meteorological conditions with a 600-ft cloud ceiling in an airplane that was about 550 lbs over gross weight. Air traffic control data showed the airplane in a climbing left turn that continued beyond the assigned heading. After reaching 1,400 ft msl, the airplane continued turning left and its altitude and speed began to vary. The airplane continued in a left spiral, completing more than two full circles, then decelerated in a right turn and rapidly descended until impact with terrain. Examination of the flight control system revealed no evidence of mechanical malfunctions and downloaded engine data indicated normal engine operation. Downloaded data from the autopilot system revealed three in-flight error codes. The first error code, which likely occurred about 1 minute after takeoff, would have resulted in the autopilot, if it was engaged at the time, disengaging. The subsequent error codes likely occurred during the erratic flight profile, with the autopilot disengaged. Before the accident flight, the pilot had informed a mechanic, who is also a pilot, of intermittent issues with the autopilot system and that these issues were unresolved. The mechanic had flown with the accident pilot previously and assessed his instrument flying skills as weak. The flight instructor who provided initial flight training for the turbine engine transition stated the pilot's instrument flying proficiency was poor when he was hand flying the airplane. Toxicology testing revealed that the pilot had used marijuana, and his girlfriend stated the pilot would take a marijuana gummy before bedtime to sleep more soundly. However, given that no psychoactive compounds were found in blood specimens, it is unlikely that the pilot was impaired at the time of the accident. The instrument conditions at the time of the accident, the airplane's erratic flightpath, and the pilot's reported lack of instrument proficiency when flying by hand support the likelihood that the pilot experienced spatial disorientation sometime after takeoff. In addition, given the reports of the intermittently malfunctioning autopilot that had not been fixed, it is likely the pilot experienced an increased workload during a critical phase of flight that, in combination with spatial disorientation, led to the pilot's loss of airplane control.

A review of ForeFlight and FlightAware data revealed that the pilot typically flew about three flights per month during the 12 months before the accident. The pilot completed initial turbine transition ground and flight training in February 2016 and accomplished recurrent ground and flight training in May 2018. The flight instructor for the pilot's initial flight training stated the pilot flew well with the autopilot turned on but that his hand-flying skills during instrument training were weak; the flight instructor also indicated that he stressed to the pilot the importance of getting "recurrent [training] often" for hand flying. The airplane underwent a JetProp DLX conversion in 2016 and was equipped with a Pratt & Whitney PT6-35 engine, and a composite MTV-16 propeller. Avionics on the airplane included a Garmin G500 primary flight display (PFD) and multifunction display (MFD), a dual Garmin GTN 750/650 GPS, an Avidyne EX600 MFD, L3 ESI-500 backup instrument system, and a King KFC 225 autopilot. Based on pilot and passenger weights provided by the coroner, a fuel load of 120 gallons, and baggage weights from the wreckage examination, the airplane's ramp weight was estimated to be 4,906 lbs and the center of gravity was about 145.41 inches. The maximum ramp weight is 4,358 lbs and the allowable center of gravity range was 143.3 to 147.2 inches. The pilot had been informed by Rocket Engineering Corp. of ongoing efforts that were underway with the Federal Aviation Administration (FAA) to increase the maximum ramp weight to 4,518 lbs. According to the Louisiana State University Health Sciences Center, Department of Pathology, Shreveport, Louisiana, Medical Examiner autopsy report, the pilot's cause of death was multiple blunt force injuries. The FAA Forensic Sciences Laboratory toxicology testing detected tetrahydrocannabinol's (THC) psychoactive metabolite (11-hydroxy-delta-9-THC) in the pilot's urine but not in his blood. THC's non-psychoactive metabolite, carboxy delta-9 THC (THC-COOH), was detected in the pilot's urine and blood. However, THC was not detected in the pilot's blood or urine. The metabolite 11-hydroxy-delta-9-THC is as psychoactive as marijuana but is rapidly metabolized to THC-COOH, which can be found in urine days to weeks after the last use of the drug. Thus, positive test results for THC metabolites in both blood and urine do not reflect recent marijuana use and cannot be used to determine that the user was under the influence of the drug at the time of testing. Toxicology testing also detected tetrahydrozoline, an over-the-counter eye drop medication commonly marketed as Visine, in the pilot's urine but not in his blood. It is not considered to be impairing. According to the pilot's girlfriend, the pilot was physically and mentally healthy. He routinely played active sports and had no issues with dizziness or vertigo. She reported he had been using a continuous positive airway pressure machine over the last year to help him sleep soundly and that, over the last several months, he had been sleeping well. She thought he was well rested for the flight. Other than his cholesterol medication, the pilot reportedly took a marijuana gummy in the evening to help him sleep. The airplane came to rest inverted about 17 ft below the surface of the Red River. A post recovery examination revealed the left and right wings were fractured and remained attached to the fuselage, which was deformed with aft buckling. The vertical and horizontal stabilizers were attached to the empennage, and the elevator and trim tab were attached to the vertical stabilizer. The rudder, left aileron, left flap, portions of the right flap, and nose baggage door were separated from the airplane and not recovered. Examination of the flight control system found the aileron cables remained attached to the control sectors, and cable continuity was established to the cabin controls. Rudder and elevator cable continuity was established through the cabin to their respective control surfaces and the elevator trim cables showed broom-straw separations aft of the pressure bulkhead. The pitch trim indicator pointed to the full nose-down position and the pitch trim barrel was in a position correlating to full nose down. A vacuum-operated attitude indicator (AI) was mounted on the far-right side of the instrument panel. The AI was removed from the panel and disassembled; the rotor and housing showed no obvious rotational scoring. The fuel selector was positioned between the left and right tanks, the landing gear selector was down, the flap selector was up, and the standby vacuum pump was on. The throttle, speed, and condition levers remained attached. The throttle lever was set at "INCR," the speed lever was fully forward, and the condition lever was located at a mid-travel position. The engine remained attached to the firewall, and the exhaust section was crushed. The propeller, the propeller flange, and part of the reduction gearbox were separated and not recovered. An Electronics International MVP-50T engine data monitor (EDM); Appareo Stratus GPS receiver; autopilot; and the pitch, roll, and yaw servos were recovered from the wreckage for further examination. EDM data captured included torque, fuel flow, fuel/oil pressures, and engine rotational speeds. No anomalies were observed during the taxi, takeoff, climb, or turning oscillations. Postaccident examination did not reveal any preaccident malfunctions or failures with the airplane or engine that would have precluded normal operation. On February 28, 2019, about 1040 central standard time, a Piper PA46-350P airplane, N428CD, was substantially damaged when it was involved in an accident near Shreveport Downtown Airport (DTN), Shreveport, Louisiana. The private pilot and passenger were fatally injured. The airplane was operated as a Title 14 Code of Federal Regulations Part 91 personal flight. The pilot and passenger were flying to Vernon, Texas, for a hunting trip. According to air traffic control (ATC) information, the airplane departed runway 32 about 1037. After climbing to 600 ft mean sea level (msl), the pilot was instructed to turn left to a heading of 270° and climb to 12,000 ft msl. According to recorded weather information, the sky condition was overcast clouds at 600 ft above ground level (agl). Radar data indicated the airplane entered a climbing left turn that continued beyond the assigned heading of 270°. The controller issued a revised heading of 140°, and the pilot responded "standby." After climbing to 1,400 ft msl, the airplane's altitude and airspeed began to fluctuate as it continued in a tightening left spiral (see figure 1). About 35 seconds after transmitting "standby," the pilot stated, "we're in trouble up here." The airplane subsequently made a decelerating right turn then rapidly descended. The last ATC data recorded for the flight indicated a groundspeed of 31 knots at 575 ft msl. Onboard data showed the airplane's altitude, vertical velocity, and speed oscillated several times and the turn rate increased during the last minute of the flight. From 10:39:25 to 10:39:58, the airplane's altitude varied from 1,987 to 822 to 1,980 ft msl and the vertical velocity fluctuated from -6,001 to +5,834 ft per minute (fpm). The last data captured at 10:40:07 was 1217 ft msl and -6818 fpm. Figure 1 - Air Traffic Control Flight Track of N428CD The autopilot system was examined at the manufacturer's facility. Testing of the three servos and mounts revealed out-of-tolerance measurements; however, they should not have affected proper function. Three error codes were logged during the accident flight: manual electric trim (MET) fail, roll monitor fail, and yaw fail. The MET fail code indicated that a possible latent failure of the manual electric trim switches had been detected during flight. According to the maintenance manual, when a MET fail is detected, m