N6094H

On April 21, 2000, at 0930 hours Hawaiian standard time, an American Eurocopter AS-350 BA, N6094H, experienced a loss of engine power during cruise. The pilot made a hard landing in rough mountainous terrain in Iao Valley, 7 miles southwest of Kahului airport, Maui, Hawaii. The helicopter was operated by Sunshine Helicopters, Inc. as a nonscheduled, domestic, passenger, sightseeing flight under 14 CFR Part 135, and sustained substantial damage. The commercial instrument helicopter pilot, and five passengers were not injured. Visual meteorological conditions existed for the 30-minute tour of the west Maui Mountains. A company visual flight rules (VFR) flight plan had been filed. The flight originated at 0915, and was scheduled to terminate at the Kahului airport.

According to the pilot's written statement to the Safety Board, approximately 20 minutes into the flight he informed the passengers that he was experiencing a loss of engine power and that he was going to make a precautionary landing. He stated that without prior indication of a problem, the engine oil pressure light illuminated followed by the generator light and the low rotor rpm audio warning. He lowered the collective to stop the low rotor rpm warning from sounding, to regain rpm, and to enter an autorotation, then made an attempt to restart the engine. He simultaneously briefed his passengers on emergency landing procedures. He noted that the N1 was at 10-12 percent and that the generator light was still not illuminated. He assumed the engine had restarted and "pulled a little collective" and "instantly noticed by 'sound' the rotor dropped a little." He knew he had experienced an engine failure and continued towards his preselected landing zone. In order to make the landing zone, the pilot had to clear several obstacles, then flare, and make a "run on landing." He flared the helicopter 2-3 times to clear obstacles, and then a final deceleration as he attempted the run-on landing. Approximately 10 feet short of the landing zone, he knew he was going to land hard and instructed his passengers to sit straight. After landing, the helicopter bounced 10 feet forward, coming to a sudden stop after a skid caught in a ditch.

Two passengers seated on the left side in the back, stated that about 15 seconds before the engine out low rpm audio sounded they smelled fuel fumes. One of the passengers started to feel sick and was about to tell the pilot when the warning sounded. None of the other passengers or the pilot recalled smelling fuel.

PERSONNEL INFORMATION

A Safety Board investigator reviewed the pilot's training records. The pilot held a commercial pilot certificate with ratings for rotorcraft and instrument helicopter. There were no restrictions noted. The pilot received his commercial certificate on October 19, 1993.

A review of the pilot's medical certificate revealed that he held a second-class medical issued on October 8, 1999, with no limitations.

According to the pilot's training records he had a total time of 6,755.10 hours. In the last calendar quarter (January - March 2000), he had accrued 273.80 hours of total turbine time. Total time in the AS 350 for that same time period was 273.80 hours. In the month of April 2000, he had accrued 41.2 hours of flight time. The pilot's training records indicated that he completed an airman competency/proficiency check FAR 135.293, in an AS-350 BA on March 20, 2000. No discrepancies were noted during the check.

AIRCRAFT INFORMATION

The helicopter was inspected at the operator's facilities by a Federal Aviation Administration (FAA) inspector. During the inspection of the fuel system, the FAA inspector noted that a T-fitting had separated from the ignition solenoid valve housing. The inspector noted that the safety wire did not appear to be a field installation. Maintenance personnel from the operator indicated that they had not worked on this component since the last overhaul.

The inspector further noted that when he rotated the N1 section manually it turned freely.

The inspector also reviewed the helicopter airframe and engine logbooks, which showed a total airframe time as 10,066.3. The last annual inspection had been conducted on January 26, 2000. The last 100-hour inspection had been conducted on April 5, 2000.

On June 6, 1999, Turbomeca's Repair Station in Grand Prairie, Texas, overhauled the engine. Total engine time since new was 6,705.6 hours. Total engine time since overhaul was 525.1 hours.

TEST AND RESEARCH

On May 10, 2000, an engine examination took place at the Turbomeca overhaul facility in Grand Prairie, under the supervision of the Safety Board. Parties to the investigation in attendance at the examination were American Eurocopter, Sunshine Helicopter, Inc., and Turbomeca Engine Corporation.

EXTERNAL EXAMINATION

Technicians removed the ignition solenoid valve, T-fitting, and the start drain valve from the engine for inspection. The T-fitting remained attached to both the ignition solenoid valve and the start drain valve; however, it had separated from the ignition solenoid housing. Prior to removal of the unit, Turbomeca representatives confirmed that the safety wire, a swedged unit, had been installed at the Grand Prairie facility during the last overhaul.

ENGINE RUN

Prior to the engine run, the Fuel Control Unit (FCU) fuel screen and the oil filter were removed and inspected. The FCU fuel screen was free of debris. The oil filter appeared to have been recently installed. Carbon particles were found in the oil filter; however, the manufacturer considered this normal.

A new ignition solenoid valve, T-fitting, and start drain valve were installed on the engine. The engine was placed in a test cell for an engine acceptance test run. The engine started and ran for approximately 20 minutes while a range of acceptance tests were conducted. The engine ran at 97 percent N1 through high and low rpm/torque ranges with no deviations. The throttle was advanced to maximum N1 with no discrepancies noted. During this operation it was also noted that the bleed air valve opened and closed within the normal range.

A fuel leak was introduced into the system and the engine ran at 97 percent N1. A variety of tests were conducted until the engine quit. Another engine run with the fuel leak was conducted with the bleed air valve manually opened. When that element was introduced, the engine flamed out immediately.

INTERNAL EXAMINATION

The T-fitting was removed from the ignition solenoid valve. The T-fitting was covered with a black substance and metallic shavings were stuck to the threads. The threads of the unit inside the valve were also covered with a black substance. Metallic shavings were also found inside the valve. The O-ring, which was seated between the T-fitting and the ignition solenoid valve, was flattened and had material missing from it. The unit was sent to the Safety Board's metallurgical laboratory in Washington, D.C., for further examination.

METALLURGICAL EXAMINATION

The metallurgist indicated that the T-fitting was made of steel, which is then threaded into the ignition solenoid valve housing that is made of an aluminum alloy. An O-ring is placed in the groove that allows sealing between the threaded portions of the T-fitting and ignition solenoid valve housing.

Energy dispersive x-ray spectroscopy (EDS) was conducted on the black substance found on the T-fitting threads and inside the ignition solenoid aluminum alloy housing. The EDS spectra in those areas had peaks of aluminum. However, no aluminum peaks were observed on the T-fitting away from the threaded ends where there was no black substance observed. The threads inside the aluminum alloy housing exhibited deterioration that was consistent with fretting.

The metallurgist stated that the O-ring was flattened and that material was missing. This damage extended approximately 135 degrees in circumference. On one side of the center circumferential plane of the O-ring a corresponding impression was observed on the opposite side that would be consistent with a pinched O-ring.

The resulting examination by the Safety Board metallurgist of the chamfer area revealed that the chamfer area was outside the manufacturer's design specifications. The chamfer is between the outer surface and the recessed area in the aluminum alloy housing. It is designed to guide the O-ring into position and to prevent it from being pinched between the T-fitting and ignition solenoid valve as it is being installed. Insufficient clamping forces due to flattening of the O-ring may have allowed the lock nut to loosen.