CEN12FA250 · Sikorsky S-76B

Four days before the accident flight, the accident helicopter experienced an intermittent loss of power to the No. 2 engine that lasted about 10 to 20 seconds while it was in cruise flight. The pilot described the event as a "rollback" but was uncertain as to how much power the engine actually lost during the event. He reported that the No. 2 engine reestablished power by itself. When he arrived at his destination, he shut down the helicopter and checked for any fault codes generated by the electronic engine control (EEC) units. No faults were recorded. He then performed ground and hover checks, and the helicopter appeared to be operating normally. He continued to fly and completed another two flights to oil platforms that same day. The next day the pilot flew another three flights without incident. A mechanic checked for fault codes, but none were recorded. The helicopter was released for flight with no further ground or flight checks. Because the pilot continued to fly the helicopter without determining the reason for the intermittent loss of engine power, the risk of another engine power loss remained high.

On the day of the accident, the pilot departed for an oil platform with six passengers and full fuel. The postaccident calculated takeoff weight was about 515 pounds over the maximum gross weight. The postaccident calculated weight of the helicopter during the approach to the oil platform was about 55 pounds below the maximum gross weight. Although the helicopter was within weight limits during the approach, its higher gross weight as compared to what it would have weighed if the pilot had loaded it within limits for the departure decreased its performance capability. The closest surface weather station, located 21 miles away, indicated that the wind was from 070 degrees at 5 knots at the time of the accident, but the pilot believed his last GPS reading of the wind was from 220 degrees at 5 to 6 knots. The pilot flew a visual approach to the oil platform on a 190-degree heading, which limited the go-around potential since it was on a direct course for the oil platform's super structure. The pilot reported that the helicopter was about 60 feet from and 15 to 20 feet above the landing pad with a nose-high attitude in the flare when there was a loss of engine power. The pilot was unsure which engine had the loss of power. With the loss of power, the pilot realized that the trajectory of the helicopter placed it short of the landing pad and that the helicopter was going to hit the platform. He pulled collective pitch and moved the cyclic control aft and to the left to clear the platform. Once clear of the platform, he attempted to lower the collective and gain airspeed, but the helicopter was in a high rate of descent with low airspeed. He pulled collective pitch and flared the helicopter before water impact. The pilot reported that it was about 3 to 4 seconds from the time he maneuvered to avoid hitting the platform to water impact. The helicopter remained on top of the surface as the pilot kept power on the helicopter to keep it from sinking. He deployed the emergency floatation bags and attempted to water taxi toward the oil platform, but there was no directional control since the tail boom was partially separated from the fuselage. All personnel were rescued without injury.

The examination of the wreckage revealed an anomaly within the stepper motor for the No. 2 engine's fuel control. The examination revealed that the end of the output shaft of the stepper motor had overstress fractures and that the shaft was bent. During the overhaul of the stepper motor 8 years before the accident, the pin that attaches the flapper valve lever to the output shaft was pressed onto the output shaft. The force applied to the external lever was sufficient to both crack and bend the output shaft. This condition eventually resulted in a "stuck" stepper motor which limited the fuel flow to the engine and resulted in an intermittent loss of engine power. The EECs did not monitor the performance of the output shaft or the flapper valve lever; therefore, no fault codes were generated by the EECs.

Verbatim NTSB analysis and probable cause from the NTSB dataset