CEN16FA307

The pilot's failure to maintain airspeed following an engine anomaly during the initial climb after takeoff, which resulted in the airplane exceeding its critical angle of attack and experiencing an aerodynamic stall. Contributing to the accident was an engine anomaly, the reason for which could not be identified during postaccident examination.

The airline transport pilot departed from a 13,500 ft long runway with a right quartering headwind present on a local test flight in an experimental, amateur-built, twin-engine airplane, which was a replica of a 1930s-era air racer. During the initial climb after takeoff , the airplane entered a right bank followed by a left bank. The left bank increased, and the airplane entered a nose-low descent that continued until it impacted terrain in an inverted attitude. The airplane was equipped with several onboard cameras that captured video footage of the accident flight. The video revealed that the left/forward engine began surging after liftoff and reached its maximum operating speed (red line) twice during the short flight. Although the pilot attempted to control the forward engine rpm with the throttle lever, the throttle inputs had no apparent effect. Based on the design of the propeller drive train, it is a possibility that the forward engine clutch was slipping. The airplane's airspeed decreased below its design stall speed and the angle of attack increased; the airplane then rolled left and subsequently impacted the ground. A postaccident examination of the airplane did not reveal any preimpact flight control anomalies. Examination of the engines did not identify a reason for the surging of the forward engine or slipping of the clutch. The sequence of events as described by witness statements and the onboard video was consistent with a loss of airspeed following an engine anomaly and a subsequent aerodynamic stall. Toxicological testing revealed alcohols were present in samples taken during the autopsy. Given the putrefaction of the samples, it is likely that all detected alcohols were the result of postmortem production. Terrain from the accident site to one quarter mile north of the accident site was suitable for an emergency landing there.

Review of a chase helicopter's video showed that there was no inflight fire and that the accident airplane's fire started after the ground impact. CSM was a public, towered airport, which was owned by the Oklahoma Space Industry Development Authority/State of Oklahoma. It was located about 2 miles west of Burns Flat, Oklahoma. The airport had an estimated elevation of 1,922.1 ft above mean sea level. Two runways, 17R/35L and 17L/35R serviced the airport. Runway 17R/35L was a 13,503 ft by 75 ft runway with a concrete surface. Runway 17L/35R was a 5,193 ft by 75 ft runway with a concrete surface. Airport operations personnel examined the runway after the accident and no liberated airplane parts were found. The airplane was fitted with GoPro cameras for the flight. Six of these cameras were found in the area of the wreckage and were sent to the National Transportation Safety Board (NTSB) Recorder Laboratory. The airplane wreckage was released and subsequent to the release, a mechanical engineer in the recorder laboratory examined the cameras, convened a Video Group as its Chairman, and subsequently produced an Onboard Image Recorder Factual Report. The Onboard Image Recorder Factual Report stated that the cameras exhibited witness marks consistent with various levels of impact damage. The cameras recorded video data on micro secure data (microSD) cards. Five of the six microSD cards contained retrievable video data for the entire flight and one microSD card contained retrievable data for a portion of the flight before impact. The report, in part, described the timing and correlation of the cameras' data and the group's observations of the accident flight recorded video and a previous flight's recorded video. The description of the accident flight, in part, indicated that the pilot was in a conscious state during the recording. No pilot or ground crew conversations pertinent to the investigation were captured. All preflight activities appeared to be consistent with known procedures. The pilot was seated and belted during the recording. He moved the left/forward ignition master switch to its "on" position and depressed the starter button. Then a sound consistent with a running engine was heard and the front propeller rotated counter-clockwise. The pilot depressed the right/rear starter button. No additional engine sound was heard and the pilot moved the right/rear ignition master switch to its "on" position. The pilot then depressed the starter button again, the rear propeller spun clockwise, and the sound consistent with a running engine was heard. The pilot appeared to manipulate the area consistent with the location of the engine clutch engagement lever and the front propeller began to spin counter-clockwise. The pilot movements were consistent with flight control check. The engine and gearbox gauge indications, which included engine oil temperature, engine oil pressure, fuel pressure, water temperature, volts, gearbox oil temperature, and gearbox oil pressure for both engines were within their respective green ranges at the start of the taxi to runway 35L and through the remainder of the recording. The airplane crossed the runway edge marking for runway 35L, the pilot added power, and the airplane tracked the right side of the runway centerline. The pilot added power and the airspeed indication became alive during the takeoff roll. The airspeed was about 60 knots during the roll abeam taxiway E. The airspeed indicated 80 knots after the airplane passed abeam taxiway D. The pilot applied backpressure to the control stick when the indicated airspeed was above 80 knots. The airplane crossed abeam taxiway C and it became airborne. The left/forward throttle lever was about 3/4 knob-width behind the right/rear throttle lever. The airplane laterally transitioned from the right side of the runway centerline to the left side of the centerline. The pilot moved the gear selector switch to the "up" position, a red light nearby illuminated, and the light extinguished about five seconds later. The runway centerline was visible below and to the right of the airplane. A change in pitch was heard in the ambient engine sounds. The rpm indication for the left/forward engine began to climb and the right rear engine appeared to remain stabilized. The pilot looked downward in the cockpit area near the hydraulic valve lever. The end of runway 35L became visible and the airplane was left of runway centerline. The pilot's right arm appeared to reach in the direction of the hydraulic valve lever. The left forward throttle lever appeared to be a knob and a half width distance from the right/rear throttle lever. The left/forward rpm indications trended upward, the pilot returned his left hand to the throttles, and his right hand to the control stick. The airplane entered an uncommanded slight left roll. The left/forward engine rpm indication reached about 10,000 rpm and the pilot pulled back the left/forward throttle lever near the closed position. Engine sounds decreased, the left/forward rpm indication decreased, and the airspeed was around the start of the green arc about 70 knots. The ambient engine sound surged. The pilot appeared to have pushed the right/rear throttle forward. The left/forward engine rpm indicated an increase in rpm near its redline. The left/forward throttle lever was positioned near its closed position. The airplane exhibited an uncommanded right roll and some flutter was observed on the left aileron. The airspeed was below the green arc about 65 knots. The right roll was arrested and the airplane appeared level. About a second later, the airplane entered an uncommanded left roll. The airspeed indication was about 65 knots. The control stick was in a neutral position. The left/forward rpm indication was near redline and the right/rear engine indication was about 4.500 rpm. As the airplane rolled through 90° of left bank, the pilot placed both hands on the control stick and commanded a right roll with a positive pitch attitude. The airplane continued to roll left, the nose dropped, and a green field came into view out of the front of the windscreen. The airplane rolled inverted and the recording continued until the subsequent ground impact. The altimeter during the recording did not exhibit an increase in altitude. However, an estimate from a chase helicopter video showed that airplane reached a maximum altitude between 80 and 100 ft above ground level. Additionally, a plot of observed parameters during the accident flight video was produced. The Onboard Image Recorder Factual Report is appended to the docket associated with this investigation. An NTSB aerospace engineer, who was a member of the video group, reviewed the video recordings, assisted in observed video documentation, and produced an Airplane Performance Study. The performance study, in part, reviewed instrument readings as a function of camera elapsed time. The readings included indicated airspeed (VIAS), indicated angle-of-attack (a), left/forward and right/rear engine throttle lever angles (TLA), and the corresponding engine speeds (rpm). A plot of the tabulated TLA's, rpm's, and VIAS's as a function of camera elapsed time was produced and the data showed that the engine speed for the forward engine began increasing from 6,000 rpm about 7 seconds elapsed time without any apparent TLA input from the pilot. The pilot responded by reducing TLA for the forward engine at 31 seconds elapsed time, about two seconds before the forward engine reached its maximum operating speed (red line) of 9,500 rpm. The pilot continued to reduce TLA to a minimum of about 40° for the forward engine until, about 38 seconds elapsed time, he increased the forward TLA by 10°. The airplane's airspeed was observed decaying. The forward engine reached red line for a second time about 42 seconds elapsed time. The input TLA and engine rpm for the right/rear engine appeared more consistent than for th