ERA15FA221

An engine flame-out due to fuel starvation as the result of an improperly-installed auxiliary fuel pump fitting, and the pilot's failure to promptly feather the propeller following the engine-flame out, which resulted in decreased glide capability and impact with terrain. Contributing to the accident was the pilot's operation of the airplane with no emergency electric propeller feathering system.

The commercial pilot departed on a cross-country flight in the experimental, amateur-built, turboprop airplane. About 3 minutes after takeoff, the pilot initiated a return to the departure airport. About 6.8 nautical miles (nm) from the runway, at an altitude of about 6,500 ft mean sea level, the pilot declared an emergency, stating that he had "lost" the engine. Shortly thereafter, he informed air traffic control that the airplane had experienced a loss of fuel pressure. Witnesses saw the airplane collide with a tree and subsequently impact a pond about 1.2 nm from the runway. They reported no engine sound. Examination of the airplane revealed no evidence of preimpact failures or malfunctions of the flight controls that would have precluded normal operation. The landing gear and flaps were found retracted. There was no evidence of preimpact failure or malfunction of the engine core. The engine-driven fuel pump showed signs of internal cavitation, while the auxiliary fuel pump, which provided fuel under pressure to the engine-driven fuel pump, flowed nearly 48% less than specified by the manufacturer, and had an inconsistent/intermittent laboring sound during testing. Disassembly revealed an improperly-installed fitting, which resulted in the decreased fuel flow rate, a loss of fuel pressure, and a subsequent engine flame-out. Inspection also revealed that a component of the fuel pump had been replaced after manufacture, but there was no record of this maintenance in the airplane's logbooks. While it was not known if the pilot attempted to perform an air restart of the engine, a worn ignitor box and ignitor plug may have precluded a successful restart of the engine. Although the pilot descended the airplane at its published best glide airspeed in his attempt to return to the airport, because the airplane was not equipped with an electrically-operated feathering pump, the propeller blade angle remained where it was set when the engine flame-out occurred, which was likely flat pitch. The flat pitch of the propeller blades significantly decreased the airplane's engine-out glide ratio. Had the pilot promptly feathered the propeller blades to reduce drag following engine flame-out, or had the airplane been equipped with a feathering pump, it is likely the airplane would have been able to reach the intended runway and land uneventfully. Although an enlarged heart was noted during the autopsy and toxicological testing detected quinine, given that the pilot was actively controlling the airplane just before the accident, there is no evidence that a medical condition or use of quinine contributed to the pilot's inability to fly the airplane.

The airplane was equipped with digital instruments that recorded engine parameters, including compressor speed (N1), propeller speed (N2), Torque, inter-turbine temperature (ITT), Oil Pressure/Temperature, and Fuel Level. The instruments were retained and submitted to the NTSB's Vehicle Recorder Division located in Washington, D.C. The instruments contained data from the last power cycle, and parameters were shown in relation to elapsed time since the application of power to the instrument. Three recorded values were noted for oil pressure and oil temperature at 236 seconds, 472 seconds, and 708 seconds. The oil pressure values were 17 pounds per square inch (psi), 34 psi, and 31 psi; the oil temperature values were 15° C, 49° C, and 57° C, respectively. Torque values at 218 seconds, 436 seconds, and 654 seconds were 14%, 78%, and 17%, respectively. Three values for ITT were recorded at 217 seconds, 434 seconds, and 651 seconds: 485° C, 620° C, and 410° C, respectively. There was only one recorded value for N1(58.4%) and N2 (1950 rpm), which was logged at 163 seconds. The pilot's first contact with ground control before taxi occurred at 0909:37, and the airplane was last visually spotted by ATC personnel at 0921:07, resulting in an elapsed time of 690 seconds. However, no correlation to real time could be made, because the time between engine start and the pilot's first contact with ground control could not be determined. According to the engine manufacturer representative, comparing the recorded data with established limits of the engine revealed that all recorded readings for oil pressure and the last two readings for oil temperature were within specified limits. The last two recorded ITT and torque readings were consistent with an engine at flight idle. Examination and operational testing of the auxiliary fuel pump were performed at the manufacturer's facility. According to the report from the manufacturer, examination of the pump revealed that one of the two motor brush caps was red in color, indicating installation of parts by someone other than the manufacturer or factory repair station. For operational testing, the inlet and outlet hoses and their respective fittings remained attached. With a 28-volt electrical supply, the fuel boost pump outlet fuel flow was 55 gallons per hour (gph), which is below the factory requirement of 105 gph. Further examination revealed that the pump outlet adapter fitting was installed deeper than that allowed by the AND10064 specification for this interface, which impeded the maximum open position of the internal no-return poppet valve, thereby restricting the outlet fuel flow. After the incorrectly-installed adapter fitting was removed, the pump performance was 109 gph, an acceptable factory test pressure. During subsequent vacuum testing, the motor had an inconsistent/intermittent laboring sound. Review of the airplane maintenance records revealed no record of removal, replacement, or repair of the auxiliary fuel pump. Examination of the annunciator panel was performed by the NTSB Materials Laboratory. The results indicated all bulb filaments were intact, and none exhibited evidence of stretching. Review of the Pilot Information Handbook revealed an emergency checklist titled, "Engine Flame-out In Flight." The first step specified to move the propeller control lever (PCL) to the feather position. The checklist stated to maintain a minimum airspeed of 130 knots indicated if at a low altitude. According to FAA Order 8130.2H, Airworthiness Certification of Products and Articles, aircraft inspection guidelines for issuance of a special airworthiness certificate specify, in part, that the flight control system should operate properly and the engine(s), propeller(s), and associated instruments operate in accordance with the manufacturer's instructions. There was no mention for a Designated Airworthiness Representative to determine compatibility of airframe, engine, and propeller systems related to emergency systems. A review of FAA Advisory Circulars (AC) AC 20-27G titled, "Certification and Operation of Amateur-Built Aircraft" issued September 30, 2009, and AC 90-89B titled, "Amateur-Built Aircraft and Ultralight Flight Testing Handbook" issued April 27, 2015, revealed no guidance for experimental aircraft builders regarding turbine engine installation considerations. Performance Study An NTSB Performance Study indicated that the pilot maintained the airplane's best glide speed of 130 knots following the emergency declaration and during the attempted return to the runway. The airplane's angle of attack (AOA) was between 1 and 2° before the pilot reported the loss of fuel pressure and between 5 and 10° after. Based on the radar data, the accident airplane's glide ratio was estimated to be 7.5. The glide ratio with a feathered propeller would have been 12.0, resulting in a power-off glide distance of about 12.8 nm from an initial altitude of 6,500 ft. A postmortem examination of the pilot was performed by the Newberry Pathology Associates, P.A. The cause of death was listed as multiple blunt force injuries. The autopsy report further indicated that the heart weighed 560 grams. Forensic toxicology was performed on specimens of the pilot by the FAA Bioaeronautical Sciences Research Laboratory, Oklahoma City, Oklahoma. The toxicology report indicated the results were negative for carbon monoxide and volatiles, and unquantified amounts of Quinine were detected in the submitted urine and iliac blood specimens. The airplane crashed in a pond located adjacent to houses in a residential area. The wreckage was recovered and transferred to a hangar located at Lexington County Airport at Pelion (6J0), Pelion, South Carolina. According to first responders and FAA personnel, the empennage was partially submerged in the water and was in an inverted position, with the airplane's nose on a magnetic heading of 032°. The wreckage location was about 1.24 nm and 279° from the approach end of runway 11 and about 600 ft south of the runway's final approach path. Further examination of the accident site area revealed no damage to unmarked powerlines located adjacent to the pond. A strong smell of jet fuel and fuel sheen were noted on the west end of the pond. Damage to the top of a pine tree was noted about 70 ft above ground level; the top of the tree was observed in the water adjacent to the tree. Pieces of curved acrylic material were at the base of the tree, and pieces of composite material were in the water at the west end of the pond. There was no evidence of any tree limbs cut by the propeller blades. Examination of the wreckage following recovery accounted for all primary and secondary flight controls. The one-piece wing was separated from the impact-damaged fuselage. The vertical and horizontal stabilizers remained attached, and the primary and secondary flight control surfaces remained attached to their respective attach points. With the rudder placed in the neutral position, the rudder trim tab was trailing edge left (tail left); with the elevator in the neutral position, the trim tab was positioned trailing edge up (tail up). Damage to the right horizontal stabilizer was noted on the outboard portion. The engine remained attached to the airframe by the engine mount and the propeller was separated from the engine. There was no evidence of fire. Examination of the wing revealed extensive impact damage, including fracture of the forward and aft spars of the left wing about 5 ft outboard of the landing gear attachment, and fracture of the aft spar of the right wing. Impact damage was also noted to the left and right ailerons and flaps. The right main landing gear was retracted, and the left main landing gear was extended, although the FAA inspector reported it was retracted when the airplane was recovered. The center section of the wing revealed the flap actuator remained attached