CEN14FA278

The pilot's failure to maintain airplane control while operating in instrument flight rules (IFR) conditions, which was due to spatial disorientation resulting from erroneous heading and bank angle information shown on the primary flight display. Contributing to the accident were the pilot's impairment due to diphenhydramine and his improper decision to operate in IFR conditions with the airplane over gross weight and at an aft center of gravity.

The pilot/owner was ferrying the airplane from the United States to Europe, and he had installed an auxiliary fuel bladder in place of the rear seat. Before takeoff, the airplane's fuel tanks were topped off, and 60 gallons of fuel were added to the auxiliary fuel bladder. The estimated weight of the airplane during takeoff was about 509 lbs over its maximum gross weight. The estimated center of gravity (CG) of the airplane was 93.2, which was near the aft limit of the CG range. The flight departed in marginal visual flight rules conditions and, soon after takeoff, climbed into instrument flight rules (IFR) conditions while passing through 1,000 ft above ground level. Air traffic control (ATC) cleared the pilot to fly a northeasterly heading and climb to 12,000 ft, but the pilot did not acknowledge the instruction, and radar track data indicated that the airplane turned right within 1 minute after departure. ATC instructed the pilot to turn back on course, and the pilot complied. The airplane continued on course for about 1.5 minutes, but then it turned right again while still in a climb. ATC instructed the pilot to turn back on course, but the pilot did not respond. The airplane continued to turn right, reached a maximum altitude of about 6,600 ft, and then entered a steep, descending right turn. ATC instructed the pilot to climb immediately, but there was no response, and the airplane continued the steep descending turn and impacted a lake about 5 minutes after departure. A comparison of the radar track data with the flight data recovered from the airplane's primary flight display (PFD) and multifunction display revealed discrepancies between the two data sources regarding airspeed, bank angle, heading, wind speed, and wind direction, indicating that erroneous information was being displayed on the PFD during the flight. Specifically, the flight data indicated periods of straight and level flight when the radar track data indicated the airplane was banking and changing heading. The erroneous information would have made it difficult for the pilot to control the airplane and navigate effectively in IFR conditions. The reason for the erroneous flight data could not be determined. The pilot's toxicology report indicated 0.146 ug/ml diphenhydramine (a sedating antihistamine) in cavity blood, which was above the therapeutic range of 0.0250 to 0.1120 ug/ml. Although diphenhydramine undergoes postmortem redistribution, the postmortem level detected suggests that the pilot likely had impairing levels of diphenhydramine in his system at the time of the accident. To maintain control of the airplane, the pilot would have needed to recognize that the PFD display was faulty and use the information from the standby attitude indicator, turn and bank indicator, and magnetic compass. However, it is likely that diphenhydramine, which impairs cognitive and psychomotor performance, diminished the pilot's ability to recognize and manage the erroneous PFD indications. The pilot's failure to acknowledge the clearance to turn to the northeast and climb to 12,000 feet only a few seconds after he initiated contact with ATC suggests that his attention was diverted for some reason about that time. The pilot verbally acknowledged and responded to a subsequent call to return to course. However, after about 1.5 minutes the airplane again deviated from course and entered a steep descending turn, most likely due to the pilot experiencing spatial disorientation as a result of the erroneous heading and bank angle information on the PFD and his ineffective use of standby flight instruments in restricted visibility conditions. The airplane's aft CG and over gross weight condition would have reduced the airplane's longitudinal stability, and this likely also contributed to the loss of control.

On August 19, 2014, the engine and propeller were examined under NTSB oversight in Prague, Czech Republic. The engine examination revealed that the power turbine was intact; however the hub was rotationally scored on both faces. The leading edges of all the blades were rotationally scored and bent aft. The gas generator turbine was intact and the blade tips were circumferentially scored with metal transfer evident on the convex sides of the tip, consistent with contact against the gas generator turbine shroud. Additionally, there were randomly distributed bright shiny flakes deposited on the convex side of some of the blades, which is consistent with a metal spray condition. The compressor rotor, consisting of two axial compressors and one centrifugal impeller was intact. The axial compressor blades were intact and the tips were circumferentially scored consistent with contact against their respective shroud elements. The impeller vanes were rotationally scored, consistent with contact against the impeller shroud. The compressor and impeller shrouds exhibited rotational scoring. The propeller examination revealed that the Nos. 1 and 2 blades were bent aft at the mid-span to a bend angle of about 90 degrees, with no evidence of blade twisting. The No. 3 blade had a slight bend with no evidence of blade twisting deformation. The No. 1 piston guide was slightly dented at a location which indicated that the blade pitch at impact was 15 degrees, a low angle corresponding to the hydraulic low pitch stop. This, in turn, corresponded to a low power setting of the engine. The NTSB Materials laboratory examined the rudder bellcrank, the rudder spherical bearings, and a section of the lower rudder and vertical stabilizer bulkhead. The examination revealed that the rudder bellcrank was comprised of a left arm and a right arm. On each arm, there were attachment points for a control cable input and a rod output. There were two holes in the vertical stabilizer bulkhead that allowed the rods to connect to the forward rudder spar via an attachment fitting. The right and left rods were fractured. Pieces of the right and left rods were attached to the rudder and a piece of the right rod was attached to the bellcrank. By contrast, there was no corresponding piece of the left rod attached to the bellcrank. A closer examination of the right rod fracture surfaces revealed that they did not match; indicating that the right rod had fractured in two or more locations and an intermediate section had been separated and was not recovered. An examination of the bellcrank revealed a deformation mark on the forward portion of the right arm in the vicinity of the bellcrank stop. The left and right rod fracture surfaces were visually examined using a stereomicroscope. The fracture surface on the left rod piece attached to the rudder consisted of inclined slant fractures, and no apparent out-of-plane deformation, consistent with a tensile overstress fracture. The right rod had collapsed near each fracture. The initially circular tube cross sections had deformed by elongating in one direction and collapsing in the other direction. The rod end at the forward end of the right rod (attached to the bellcrank via a rod end bearing) was bent. The features were consistent with overload by compressive buckling. The NTSB Vehicle Recorders laboratory examined the Avidyne FlightMax Entegra MFD removable compact flash card. While the compact flash card successfully read, it did not contain any recorded information. Avidyne confirmed that the Entegra does not record any information when installed in an experimental, turbine aircraft. The NTSB Vehicle Records laboratory examined the accident pilot's Nokia C5 cell phone that was found in the airplane wreckage. The Nokia cell phone turned on, but the screen was damaged and no further recovery attempts were attempted. Two Chelton IDUs were recovered from the accident aircraft and sent the NTSB Vehicle Records laboratory for examination. The units sustained minor impact and water damage. The units integrate multiple primary flight instruments including airspeed, altitude, electronic compass, turn rate, bank angle, pitch angle, vertical speed, and an optional slip/skid ball. They can also function as a navigation and engine display. They may integrate with external components, including a GPS/Air Data/AHRS. Units are typically installed in pairs, providing PFD and MFD capabilities. The units are capable of recording a log of aircraft parameters at a rate of 1 sample per second to an internal PCMCIA card. The parameters recorded depend upon installation and include primary flight instrument data, GPS position data, AHRS data, and engine data. The current log file "LOG00.DAT," was retrieved from the download of each Chelton IDU unit. The "LOG00.DAT" file from the unit with serial number 292 contained recorded data on June 7, 2014 between 16:00:16 universal coordinated time (UTC) and 16:20:53 UTC. The file recorded primary flight instrument data, GPS position data, AHRS data, and engine data. There were about 20 additional log files of prior flights, four of which were recorded on June 6, 2014. For the accident flight, 5 hours were subtracted from UTC to convert to CDT. The data showed that the aircraft departed from KDLH, climbed towards Lake Superior with intermediate level offs, exhibited fluctuations in pitch, speed, and roll, and then descended rapidly and crashed into Lake Superior. The maximum altitude attained was 6,607 ft with an indicated airspeed (IAS) of 94 kts, which was the minimum recorded IAS after the initial departure climb. Thereafter, the IAS increased and the aircraft descended, reaching a maximum recorded IAS of 262 kts about 6 seconds before the end of the recording. On the prior flight from Bend, Oregon, to Duluth, Minnesota, the day before the accident, the roll, heading, and course each oscillated about +/-10 degrees for about an hour in cruise flight. See the NTSB Vehicle Recorder Division's report "Electronic Devices" in the docket material associated with this investigation for further details. The NTSB's Vehicle Performance Division conducted an airplane performance study which described the accident airplane ground track, altitude, and speed, as well as the timing of select radio communication between ATC and N86NW, including estimates of airplane pitch, roll, and heading derived from radar, as well as airplane and engine data recovered from the Chelton IDUs. The study compared the data derived from radar and the Chelton IDU's log data. The Chelton log data is the data being displayed to the pilot on the PFD. The study indicated that during two periods about 30 seconds in length during the accident flight (the first centered around 16:17:40 and the second at16:19:50) the radar-derived airspeed exceeded the airspeed recorded in the log file by 10 to 20 kts. During approximately the same time frame, the radar-derived bank angle exceeded the bank recorded in the log file by as much as 25 degrees (i.e., more right-wing-down). Additionally, the recorded log data bank angle is different than the heading shown by the radar data. During the first 30 second period centered at 16:17:40, the log data shows little or no bank, while the heading derived from radar data showed the airplane turning to the south/right. The other notable difference in the comparison was between the heading derived from radar and that recorded in the log file. For nearly two minutes, early in the flight, the log file heading was 20 degrees to 25 degrees more airplane-nose-left than that estimated from radar. From 16:19:20 until the end of the data the log file heading was upwards of 45 degrees more airplane-nose-right than that estimated from radar. The Chelton log files recorded the groundspeed, wind speed, and wind direction during the flight. During the climb to 6,600 ft msl, the wind speed varied between 5 and 88 kt