Occurrence Summary
On a visual meteorological day with calm winds and a temperature of 33°C, flight BLS363, a Pilatus PC-12/45 single-engine turbine aircraft, was ready for takeoff on Runway 26 at Red Lake, Ontario, Airport. Two pilots and three passengers were on board. When the condition lever was moved from ground idle (65% Ng) to flight idle (75% Ng), the Pratt & Whitney PT6A-67B engine flamed out. A restart attempt was unsuccessful, and the aircraft was towed off the runway.
Investigation
Both pilots held valid commercial pilot licences and were properly trained. The aircraft had been fueled the previous night at Sioux Lookout Airport, and a run-up at that time found no anomalies. Fuel sumps were clean. On the day of the occurrence, engine start was performed per standard operating procedures using the aircraft battery, but no run-up was conducted. After ground idle was achieved, the air conditioning, avionics, and vent blowers were turned on. Taxiing revealed no anomalies.
About 90 to 120 seconds after ground idle, the condition lever was advanced to flight idle. During the transition, the battery ammeter indicated a draw of +300 amps, and the engine flamed out. The restart attempt was terminated because battery voltage dropped to 19 volts and Ng reached only 8.5%. The aircraft manufacturer stated that after a battery start, the normal recharge load is about 190 amps, tapering off; the flight manual specifies that battery charging current must not exceed 15 amps before takeoff, requiring a delay of 5 to 8 minutes. At the time of the draw, electrical equipment—air conditioning, navigation lights, beacon, and vent blowers—collectively drew 110 amps, combining with battery charging to produce the 300-amp load.
After towing, company maintenance staff ran the engine using a ground power unit but could not duplicate the power loss or high amperage draw. Fuel filters and sumps were inspected and found free of contamination. The oil filter contained particles: wood, synthetic and mineralogical fibers, and carbon or paint. The electrical system showed no anomalies. The No. 1 generator, fuel control unit (FCU), and engine-driven fuel pump were replaced.
An oil sample from the engine had a magnesium count of 39.9 ppm. The PT6A maintenance manual does not prescribe magnesium limits, but an internal Pratt & Whitney guideline of 35 ppm triggers borescope inspections. A borescope inspection found wear on the second-stage splines of the reduction gearbox (RGB). The engine was removed and sent to Pratt & Whitney Canada for teardown. Disassembly confirmed substantial fretting of the RGB second-stage mating splines and ring gear. Pratt & Whitney Service Bulletin No. 14267 R1 addresses replacement of the RGB case when wear limits are reached; about 20 of the 200 PT6A-67B engines in service had complied, and no catastrophic failures have occurred.
Gas generator disassembly showed no anomalies. The No. 2 bearing had been modified per Service Bulletin No. 14320 R2 and was clean, but black carbonaceous and magnesium deposits were found in the bearing oil housing. The power section thermal blanket, power turbine shaft housing, and No. 3 bearing cover were covered with soot. The power turbine rotor air seal had reduced hole diameters due to carbonaceous residue.
Bench testing of the FCU, fuel pump, and flow divider revealed only that specified governing air pressures produced insufficient fuel flow, which did not contribute to the occurrence. When installed on another engine in a test cell, no anomalies were found during acceleration, deceleration, and various power settings. The flameout could not be duplicated. Disassembly of these units revealed no anomalies. The intake of water or air can interrupt fuel scheduling without leaving traces. The starter generator was bench tested and disassembled, with no findings that would explain the 300-amp draw.
Canadian Aviation Regulations require a mean time between failure of 0.01/1000 or less for single-engine turbine aircraft used in passenger transport; fuel supply interruptions are not considered engine failures for this standard.
Findings
The TSB concluded that the engine power loss most likely resulted from an interruption in fuel supply, though this could not be determined with certainty. The electrical draw was from battery charging and selected equipment and did not contribute to the power loss. Fretting between the second-stage ring gear and RGB splines released magnesium particles into the oil, which became coated with carbonaceous material and were trapped in the oil filter. Carbon particles originated from the No. 2 bearing area prior to the service bulletin modification.