Airbus Helicopters INC AS350 accident near Gustavus, Alaska, September 28, 2018

The NTSB's narrative final · quoted from the NTSB record

The private pilot had just purchased the helicopter and was returning it to his home airport accompanied by a commercial-certificated safety pilot and two passengers. Video from the cockpit image recorder indicated that the pilot, seated in the right seat, was manipulating the flight controls from takeoff until the accident occurred. The surviving passenger stated that, while en route, the pilot indicated that they would be landing on a beach in order to stretch their legs. The recovered data showed the helicopter flying about 500 to 700 ft over water when the pilots began conversing and pointing toward the shore. The safety pilot, whose hands were on his lap, then raised his right hand in a manner that appeared that he was guarding the cyclic control in anticipation that his assistance or intervention might be required. Shortly thereafter, the pilot twisted the collective twist grip throttle from FLIGHT to IDLE; data from the engine data recorder (EDR) indicated that, at this time, the engine fuel control was set to the idle power setting. Such action is consistent with a practice autorotation. The collective control was adjusted downward slightly, but the main rotor speed (Nr) gauge showed Nr decay from the normal "green" operating range of 375 to 405 rpm to the "yellow" cautionary range of 320 to 375 rpm about 5 seconds after the twist grip was set to IDLE. An Nr value of less than 360 rpm would have resulted in a continuous aural tone to alert the pilots of the low rotor speed. About 7 seconds after the twist grip was set to IDLE, the pilot reached toward the center console, after which the "HORN" caution light illuminated, consistent with the pilot muting the aural tone signaling the low Nr condition. After the horn was muted, Nr continued to decay to a low of 254 rpm. Several small cyclic inputs were observed, then the recorded data ended. The helicopter impacted the water and was destroyed. The four occupants were ejected from the helicopter; one of the passengers survived the accident and swam to shore. Postaccident examination of the helicopter and the recovered data did not reveal any helicopter malfunctions or failures that would have precluded normal operation. Since recorded cockpit imagery did not capture microphone audio, the reason for the pilot's decision to roll the throttle to IDLE could not be determined, and what conversations the pilots may have had before and during the maneuver could not be determined. The passengers' headphones were muted from the pilots' conversation, and the surviving passenger did not know if they planned to practice an autorotation to the beach. Cockpit video recordings from previous flights indicated that the safety pilot did not guard the flight controls when the pilot was performing takeoffs or landings. It is plausible that the pilot's intention was to practice or perform an autorotation to the beach, with the safety pilot guarding the controls. However, the beach was not the most suitable location for a practice autorotation to landing procedure and was out of the ordinary from their previous flights, as observed on the recorded cockpit imagery. When the twist grip throttle is set to IDLE position during normal flight, Nr will decay unless the pilot takes action to maintain it, such as a reduction in main rotor collective pitch or returning the twist grip throttle to the FLIGHT position. However, after the throttle was set to IDLE, the pilot failed to reduce main rotor collective pitch, which resulted in low rotor rpm. The pilot's action of muting the low rotor rpm alert horn indicated that he was aware of the annunciation to an anomalous condition, but it could not be determined if the pilot understood the nature of why the horn annunciated or the criticality of the situation.

The complete narrative as the NTSB published it. The NTSB's docket holds the report as a PDF and any photographs, statements and other documents from the investigation.