WPR16FA176

Total loss of power to the right engine during takeoff due to a fuel contaminant, which blocked the fuel flow transducer and resulted in fuel starvation to the engine.

The airline transport pilot and three passengers were departing in the multi-engine airplane when, during the early phase of takeoff, most likely shortly after rotation, the right engine experienced a total loss of power. The power loss occurred during a critical phase of flight, at a time when the airplane was close to or just below the manufacturer's recommended "safe single-engine speed." The hilly terrain surrounding the airport, particularly in the takeoff direction, left the pilot with few options for a safe climb out. Further hindering the takeoff was the airplane's anemic single-engine climb performance due to the high density altitude conditions and the airplane's weight. As a result, shortly after the loss of power, the airplane rolled right, consistent with it flying slower than its minimum controllable single-engine airspeed, collided with the ground, and sustained substantial damage. The pilot and passengers all sustained serious injuries and could not remember the circumstances of the accident. Examination of the right engine revealed that a clear, gelatinous substance had blocked the inlet port of the fuel flow transducer, leading to fuel starvation. The substance was determined to be silicone, most likely room-temperature-vulcanization silicone sealant. There were no indications that this material had been used anywhere within the fuel system, nor were there any signs of recent maintenance that could have resulted in the introduction of this contaminant. To get to the fuel flow transducer, the silicone would have had to pass through multiple fine mesh filters, which was unlikely based on the material's size when dry. However, silicone becomes semi-soluble once in contact with aviation gasoline, and it is possible that it was extruded through the filters in this state, and eventually coalesced in the flow divider inlet port. Smaller fragments of the silicone were found in the fuel manifold valve, beyond the flow divider, and a fuel injector valve port was partially occluded, possibly with the same material. The specific source of contamination could not be determined. The pilot last purchased fuel for the airplane about one month before the accident. Later on the day of purchase, the fuel system was shut down by the airport management due to metering inaccuracies; however, the problems were all electrical in nature, and did not require the repair or replacement of any components that would have come into contact with fuel.

Wickenburg Airport is at an elevation of 2,378 ft, and is composed of a single 6,101-ft-long asphalt runway, designated 5/23. Runway 23 is on a 1.2% uphill gradient. Terrain 1.5 miles beyond the departure end of runway 23 rises to a peak about 300 ft above runway elevation. Highway 60, which is offset about 30° right of the runway centerline, follows the foothills of the rising terrain, about 200 ft below the peak. The only fueling facility at the airport was a self-serve pump, managed by the City of Wickenburg. During the period July 16 through 30, the pilot serviced the airplane twice at Wickenburg, and then three times at different airports in Kansas and Wisconsin. The last fuel purchased for the accident aircraft before the accident was from the Wickenburg pump on July 30, 2016. He then flew to Payson, Arizona a few days later. The airport operations manager provided the certificate of analysis for the fuel delivered to the tank farm during that period, and the sample met the tested specifications for ASTM 5191 (vapor pressure), ASTM D86 (distillation), and IP 559 (density). Additionally, daily fuel system facility checks for the month of July and August, did not reveal any anomalies, and no pilots reported issues with fuel. According to the operations manager, the fuel system experienced a failure on July 12, 2016, attributed to a lightning strike, and as a result, the system's computer motherboard was replaced. Metering problems persisted, and on July 30, the same day that the accident pilot purchased fuel, the system was shut down for a week because the delivery meter did not read correctly. The meter's pulse transmitter was replaced; however, anomalies persisted, and in early December, the entire fuel island was shut down for redesign. The operations manager stated that the problems were all electrical in nature, and did not require repair or replacement of any components that would have come into contact with fuel. Examination of the engine control quadrant at the accident site revealed that both mixture controls were in the full rich position, the propeller controls were 1-inch short of full forward, and the throttle controls had bent to the right and over the quadrant about midrange. Both the flap actuator and landing gear assemblies were in positions consistent with retraction. Left Engine The left engine had partially separated from the firewall during the impact sequence, sustaining damage to the throttle body and rocker covers, and exposing the valve springs and rocker assemblies for all cylinders except number 4. The propeller blades and hub assembly remained attached to the crankshaft. Both blades exhibited a 15o bend about 12 inches from the hub, along with multiple nicks and chordwise scratches to their leading edges. The fuel lines along with both the engine and propeller controls were intact, and the spark plug electrodes exhibited normal service life wear signatures, and dark grey coloration. The magneto-to-engine timing was correct, and "thumb" compression was confirmed at all cylinders, along with drive train continuity to all valves and accessories. The fuel lines from the metering unit through to the fuel flow transducer and the fuel manifold valve were free of obstruction, and the internal impellor of the transducer could be heard spinning when low-pressure air was applied to the inlet. Disassembly of the fuel manifold valve, engine driven fuel pump, and throttle body metering unit revealed no mechanical anomalies, and residual fuel was observed within the cavity of the manifold valve. The fuel inlet screen was found clear and free from obstructions. During disassembly of the metering unit, debris was observed on the spring side of the mixture control cam. The debris appeared to be a combination of dried grease and ferrous material. Right Engine The right engine sustained similar impact damage, with the propeller hub assembly remaining attached to the crankshaft. Both blades had detached from the hub, and both were straight, with neither exhibiting any damage signatures associated with rotation such as leading edge nicks or chordwise scratches. Both blades displayed blue streak marks, similar in color to the paint on the refuse trailer which was struck during impact. The engine exhibited comparable magneto-to-engine timing, cylinder compression, and spark plug characteristics as the left engine. Disassembly of the fuel manifold valve, engine driven fuel pump, and throttle body metering unit revealed no mechanical anomalies. However, about 1/8 of one side of the surface of the throttle body inlet screen was covered in lint material, and the fuel injector nozzle for cylinder 3 was partially occluded and coated with a solid glaze. No fuel was observed within the cavity of the manifold valve or the fuel line between the fuel flow transducer and the fuel manifold valve. Disassembly of the fuel lines revealed that a clear gelatinous substance had completely blocked the fuel flow transducer inlet port (metering orifice) (Photo 1). The material was removed, and had a slimy wet texture. After one hour of exposure to air, the material had hardened and took on a texture similar to room-temperature-vulcanization (RTV) silicone. Six fragments were recovered, which, after drying for 24 hours, ranged in size from 1 to 3 mm. Further examination of the fuel manifold valve revealed a similar fragment of the material within the manifold cavity on the pre-filtered side of its screen. Photo 1 - Fuel Flow Transducer Inlet Port Fuel System The airplane was equipped with a 20-gallon auxiliary fuel tank in each wing, and a 50-gallon main tank at each wingtip. The auxiliary tanks fed the system through gravity, and an electrically driven submerged fuel pump was housed in each tip tank for use during priming and starting, and for backup operation to the engine-driven fuel pump. Each wing housed a combination fuel selector valve/strainer, which was controlled by a selector lever in the cabin via a set of cables. The mesh size of the strainer was 104 microns. The fuel flowed from the strainer to the engine driven fuel pump, and onward to the inlet port of the fuel metering unit, which was protected by a 210-micron mesh filter. Downstream of the metering unit, the fuel passed through the fuel flow transducer, and into the fuel manifold valve, which contained a 210-micron mesh filter. Maintenance records revealed that a Shadin 910502 fuel flow indicating system was installed in 1982, in accordance with supplemental type certificate SA573GL and SE552GL. The fuel flow transducer installed at the time of the accident was a FloScan 201 series (p/n 680501), which according to the engine logbook, had been installed in 1995 as a replacement for the original unit. The transducer inlet and outlet ports used 1/4-inch NPT threads, and the inlet metering orifice was about 0.115 inches (2.92mm) in diameter. Both outboard tip tanks had been breached, exposing their inner surfaces. No debris was observed within the tanks. The auxiliary tanks were intact, and no debris was observed when examined through the respective fuel filler necks. The airframe and engine fuel lines, filter plugs, fittings, and gaskets were examined to determine if RTV sealant material had been used as a sealing medium. No traces of such material were observed. Additionally, the fuel lines within the engine compartment were stiff, almost brittle, and exhibited significant chaffing damage. The owner stated that he had never used RTV silicon to seal any components within the fuel system. Data provided by Continental Engines indicated that the fuel pressure from the engine driven fuel pump to the metering unit was between 28.8 and 31.0 psi when the engine was operating at 2,625 RPM, and 6.5 to 7.5 psi at 600 RPM. The metered fuel pressure at 2,625 RPM varies between 17.8 and 18.8 psi. Material Examination The rubber-like mate