NTSB Factual Report

HISTORY OF FLIGHT

On July 6, 2005, about 1900 mountain standard time, an Ayres Corporation S2R-G10 (Thrush) restricted category airplane, N440AT, lost engine power and made a forced landing near Wickenburg, Arizona. The airplane was being operated by the State of Arizona as a fire suppression tool to fight forest fires. The commercial pilot, who was the sole occupant, sustained minor injuries. The airplane was substantially damaged. The public-use flight was being conducted under the provisions of 14 CFR Part 91. Visual meteorological conditions prevailed. A company flight plan was filed, and departed Mirana, Arizona, about 1 hour 15 minutes prior to the accident.

According to the pilot, he departed Mirana with a full load of fire retardant and about 4 hours of fuel on board. He dropped the fire retardant over the assigned area (Eagle Fire, south of Wickenburg) and was continuing to Wickenburg at 1,000 feet above ground level (agl), when the engine reportedly lost power. The pilot noticed the engine "lose torque" and "the propeller began to auto-feather." The airplane began to lose altitude immediately and so he made a mayday call. The pilot looked for a suitable landing site, but was unable to find anything level. The airplane touched down on rough and uneven terrain.

PERSONNEL INFORMATION

The commercial pilot had single engine land and instrument airplane ratings. According to the submitted Pilot/Operator Aircraft Accident Report, the pilot accumulated a total of 2,051 hours of flight time, of which 664 hours were accumulated in the same make and model of airplane. However, the pilot had 50 hours of flight time in the accident airplane/engine combination. The pilot received training from the operator regarding the Garrett TPE331 engines prior to flying the firefighting missions since the majority of his experience was obtained flying the Thrushes equipped with Pratt & Whitney PT-6 engines. Prior to the accident date, the pilot had not performed any flying firefighting missions.

AIRCRAFT INFORMATION

The accident airplane was equipped with a Garrett TPE331-10-511N engine (serial number P-36096C), which was rated at 900 horsepower at 1,591 rpm. It underwent its last annual inspection on October 1, 2004, at total time of 3,015.4 hours. The airplane accumulated 229.6 hours since the last annual inspection. At the time of the last annual inspection, the engine accumulated 8,951 hours since new and 1,325 hours since its last overhaul.

A maintenance entry dated October 9, 2003, indicated that the mechanic installed the engine on the accident airplane, and set the T2 bias adjustment one full turn counterclockwise to obtain full power. The engine total time during the adjustment was 8,593.9 hours.

WRECKAGE AND IMPACT INFORMATION

The accident site was located at 33 degrees 54.09 minutes north latitude and 112 degrees 47.02 minutes west longitude. The airplane came to rest within scrub brush and cactus. Photographs taken of the accident site revealed the landing gear were collapsed aft under the airplane and the wings sustained substantial damage. The propeller was separated from its drive shaft, but the four propeller blades remained attached to the hub. The propeller blades sustained chordwise scratches and gouges. The wreckage was recovered to an aircraft salvage facility in Phoenix, Arizona, where the engine was removed and shipped to the manufacturer's facility for further examination.

TESTS AND RESEARCH

On July 11, 2005, a preliminary examination of the wreckage was conducted at Air Transport, Phoenix, Arizona. A Federal Aviation Administrator (FAA) inspector and investigators from Honeywell, a party to the investigation, were present. According to those present for the initial examination, the starter/generator was separated from the gear case, but the input shaft was intact and remained with the starter/generator. All of the propeller studs were pulled out of the propeller mounting flange and remained on the engine propeller shaft. The oil bypass valve was in the retracted position. The oil scavenge pump fitting was removed and the first stage compressor impeller was rotated, which resulted in a corresponding rotation of the scavenge pump. Little or no oil was noted in the torque sensor line, both the "to" and "from" oil cooler lines, the oil line from the unfeather pump to the aft side of oil tank, and the oil line from the unfeather pump to the beta block.

The engine was prepared for shipment to Honeywell's facilities in Phoenix, by removing the tail pipe, nacelle wreckage, and aircraft engine mounts from the engines.

Engine Examination

On August 16 and 17, 2005, the National Transportation Safety Board investigator-in-charge (IIC), and investigators from Honeywell and the Department of Interior's Office of Aviation Safety (also a party to the investigation) examined the engine at the Honeywell examination facility.

The engine did not display any external catastrophic failures. The propeller separated from the propeller flange. The flange was bent aft on one end, but the propeller bolts remained intact. The fuel control unit (FCU) controls were attached. The reversing function operated when manually manipulated. The engine's internal components rotated freely when manually rotated through the exhaust via the turbine blades or through the inlet via the compressor impeller. The engine was placed on an engine mount. The reduction gearbox (RGB) chip detector was removed. Ferrous-type metal shavings were noted over the magnet. About 1.5 liters of oil were removed from the RGB and a sample was captured.

The nose cone was removed from the RGB. The nose cone remained intact and the propeller shaft was intact and in place; however, the propeller drive coupling was cracked in the same area as the propeller flange damage. The crack was irregular in shape and originated in the middle of a flat area between splines (not in a radius). It also displayed 45-degree shear lips in the fracture surface. The crack propagated through the majority of the coupling and the lip that remained intact was warped. The inside diameter spines were scraped and metal was shaved off of the splines. The propeller shaft displayed rotational scoring and gouging near the aft end of the shaft on its flats, and near its forward end on the fractured propeller drive coupling.

The planetary gears were removed from the ring gear. They were intact and did not display any gouging or galling, and their bearings were smooth. All bearings in the RGB were smooth and free to rotate. No discoloration, galling, or gouging was noted.

The engine/propeller installation was designed so that in the event of an engine failure, in which the propeller drives the power section, the engines NTS (negative torque sensing) system will modulate the propeller blades at a minimum drag condition. In order for the propeller to go into the feather position, the pilot must position the emergency cutoff handle from the normal position to the cutoff position (aft). Doing this will close the fuel shutoff valve and position the propeller feathering valve to the feathered position (dumping oil pressure to the propeller piston allowing the feather spring and counterweights to drive the propeller to the feathered position). The fuel cutoff switch was found in the open position and the feathering valve was not in the feathered position. The feathering valve operated when manually manipulated.

The torque sensor metering valve gear was suck in an aft position and could not be moved forward on its helical splines. According to Honeywell personnel, this is not an unusual finding during a disassembly following a propeller strike. During terrain impact the gear will sometimes be pushed aft past its helical spline limit and get hung up. Functional tests of the component in this condition would only provide a torque reading at the stuck position; therefore, the Safety Board IIC elected not to functionally test the unit in this position.

Examination of the RGB diaphragm assembly revealed that the locating shoulders for the ring gear were offset and the corresponding holes in the ring gear were elongated.

The turbine oil pump (transfers oil from the rear turbine bearing to the front of the engine) was removed. The pump shaft was intact and the splines were in good condition. Upon removal of the pump, oil drained from the pump area. The turbine bearing displayed no evidence of galling or overheating.

The turbine wheels and stators were removed and examined. The turbine wheel blades displayed burned organic material on the leading edges and faces of the blades. None of the turbine wheel airfoils displayed evidence of FOD or excessive erosion. The number 1 turbine nozzle displayed five blades with eroded (burned through) trailing edges. None of the blades were fractured. According to Honeywell, this is not an unusual finding in engines that come in for overhaul. Sometimes cracks develop and the hot gases erode the material away from the crack. It could also be caused by fuel nozzle streaking, but the combustor can did not display any burn through or hot spots. The number 2 stator nozzle also displayed six blades with eroded trailing edges and green discoloration and bubbling. There was also a significant amount of fine dust/dirt in the turbine section, and in some areas it was piled up to two inches thick. Microscopic examination of the first stage of turbine blades revealed shinny resolidified metal splatter on the leading edges. The honeycomb shroud, for the turbine wheel's labyrinth seal, displayed corresponding scores and polished edges.

The compressor section was disassembled. The first stage impeller displayed worn edges and a sandblasted appearance on its leading edge. One of the blades was bent aft opposite the direction of rotation. The compressor shroud displayed rotational scoring and gouging noticeable both visually and tactually around 270 degrees of the inside diameter.

The torsion shaft, which drives the RGB through the high-speed pinion, was sheared at the aft splined end. Approximately 1.25 inches of the aft shaft was sheered off. The splines in the area of the sheer failure displayed torsional deformation. The splines were offset by at least one spline width. According to the Honeywell investigators, the compressor impellers and turbine wheels are assembled onto the main shaft. A nut at the aft end of the main shaft provides the clamping force to hold these together. The torsion shaft is splined to the aft end, inner diameter, of the main shaft. There is a coupler shaft, which couples the high-speed pinion assembly (in the RGB) to the rotating group via the front end of the torsion shaft. During a propeller strike, the propeller is stopped, which stops the gearbox and the front end of the main shaft. There is a significant force being produced by the rotating group (main shaft and associated components), which continues to rotate or drive the aft end of the torsion shaft. This provides the relative torsion, which results in the torsional deformation and sheer failure of the torsion shaft.

All of the internal engine bearings appeared to be in good condition with no evidence of galling or overheating. The oil reservoir displayed a crack adjacent to a weld line near the drain fitting. This area of the reservoir sustained impact damage and deformation.

Accessory Examinations

The FCU shaft was intact and the splines appeared to be in good condition. The propeller governor was removed and its shaft was intact. The FCU was placed in Honeywell's test cell after removing the Inlet Temperature Probe (P2T2 probe; a micrometer is utilized in flow tests to simulate various temperature conditions) and the filter cap, as the filter cap fitting was impact damaged and displaced to an extent where the fitting threads were visible. The FCU ran throughout all test parameters but with excessive fuel flow pressures. According to Honeywell investigators, the FCU is normally reset by mechanic on the airplane, and test cell only provides a comparison to a calibrated performance. In no condition did the flow pressures fall below calibrated limits. Examination of the P2T2 probe bellows revealed that they were offset slightly, but remained flexible.

The flow divider was placed on test cell and fuel was applied to the unit. The unit allowed flow through both the primary and secondary ports, but when greater fuel pressure was applied, fuel began leaking from a sheared screw noted on the cover plate. The plate was offset and the threaded section of the screw remained in the unit.

The fuel shutoff valve (electric/manual) was removed and taken to the test facility. The valve unit was connected to the test stand and electrical power was applied to the unit. The technician heard the valve's electric microswitches click once upon power application. Following the power application and subsequent "click" the valve was closed. Fluid could not be passed through the valve. Electrical power was applied in an attempt to reopen the valve, but it would not open. The valve switch cover was impact damaged and dented inward on one face and on one corner. The cover was removed so the internal components could be examined. Power was again applied to the valve and the switches and valve operated normally. When the internal electrical wires were pressed with finger pressure in the same area as the cover face indentation, the switch throws were blocked by the impinged wire. When the wires were not pressed, the unit functioned without anomaly. The manual operation of the valve (the valve can only be manually positioned to the closed position) displayed no anomalies.

The propeller governor was placed on test cell and operated throughout the normal test range. The unit displayed no anomalies.

Propeller Examination

On October 20, 2005, the propeller was examined at McCauley Propeller Systems, Wichita, Kansas, under the supervision of a Safety Board investigator. According to the report provided by McCauley, the propeller damage was "the type associated with impact forces, with gross deflections... There were no indications of any type of fatigue failure of any of the components. Several hub sockets contained blade counterweight impact marks indicating blade angle at impact in the normal operating range. No impact marks were found at or near the feather position." In addition, the type and extent of blade bending and twisting "indicates high power at impact." The propeller blades contained significant chordwise scratches and gouges "indicating rotation at high speed."

Additional FCU Examination

On March 28, 2006, the FCU was examined at the manufacturer's facility (Woodward Governor Company) in Rockton, Illinois, under the supervision of an FAA airworthiness inspector. The FCU, with fuel pump still attached, was mounted to a test stand. The unit passed all standard day acceleration, 35,000-foot, hot day, and cold day schedule test points. One test point on the 15,000-foot schedule was above the upper tolerance, providing more fuel flow than required. The underspeed governor minimum was higher than that prescribed. The power lever schedule test showed all flows were high. All of the test parameters were similar to those found during the testing at Honeywell's facility. No anomalies were noted that would have resulted in the reported loss of engine power.

ADDITIONAL INFORMATION

The wreckage was released to the owner's representative upon completion of the examinations. No parts or pieces were retained.

NTSB Probable Cause

the partial loss of engine power for undetermined reasons.