History of Flight
The aircraft, kept uncovered at a farm strip but fitted with covers, was prepared by the pilot. He drained about one litre of fuel from each wing tank via a common drain valve at the fuselage's lowest point, then added fuel to slightly over half in each tank. After starting, the engine was run at 2,500 rpm for 5-10 minutes. The takeoff was uneventful, with the aircraft climbing to about 600 feet above ground level, flying with the right fuel tank selected. Approximately three minutes into the flight, the engine faltered and RPM reduced, then stopped completely 3-5 seconds later. The pilot chose a field for forced landing, judging altitude insufficient for a restart and noting the stoppage seemed final.
The selected field had an initial up-slope leading to a level area, with a rough grass pasture surface. The landing direction was approximately into wind. The approach appeared satisfactory except for high speed. Initial touchdown occurred on the nose and left main wheel, followed by a short bounce, then a three-point landing. Groundmarks showed the aircraft traveled 30-40 metres before the nosewheel entered a pothole, damaging the spat. After another 18 metres, the nose leg collapsed, allowing the nose underside to contact the ground. The aircraft veered right and came to a halt 15 metres further. Occupants were uninjured and exited via doors. The pilot later drained much fuel from tanks, indicating adequate fuel was available.
Examination of Aircraft
The aircraft was recovered to a maintenance company for repair. Examination focused on the engine and fuel system. The fuel system comprised a tank in each wing, with delivery lines running down behind cabin pillars. ON-OFF selectors allowed running from left, right, or both tanks. Fuel lines met at a T-piece in the fuselage centre, with a single line to the engine. The drain valve was located just downstream of the T-piece. In the engine compartment, fuel passed through a gascolator, a fuel flowmeter, and an engine-driven pump to twin carburettors. The engine had a dual electronic ignition and liquid cooling.
No defects were found in the engine. During a ground run, fuel poured from the right carburettor overflow, possibly due to a jammed float valve from impact shock. After reseating the float valve, the carburettor operated normally. The overflow indicated the fuel pump was working, but even without it, gravity from the high-wing design would supply fuel. Fuel flow checks from both tanks were normal, and the system was exceptionally clean with no contamination in the gascolator. A boroscope inspection of fuel tanks found no debris blocking the coarse filters.
In a ground run, fuel was shut off at about 4,400 RPM (takeoff RPM 5,200 stationary). The engine ran normally for about 30 seconds, then irregularly for a few seconds before stopping, supporting a fuel starvation hypothesis rather than ignition failure. Fuel vapour lock was considered unlikely due to the relatively cool day.
Additional Testing
After repairs, the pilot flew the aircraft back to his private strip. He selected both tanks ON. The right tank was less than a quarter full, left about half full. After a 15-minute flight, the right tank was nearly half full and left less than a quarter full. On the ground, levels equalised over time.
Further tests were conducted with the aircraft tethered on level ground. With both tanks initially empty, 18-20 litres were added to each 35-litre tank. When both selectors were ON, no fuel appeared at the drain valve until suction was applied, a phenomenon noted during construction. The vent lines were forward-facing tubes on the wing undersides, pressurised by ram air and propeller wash. Tests at 4,400 RPM showed a pressure difference: left vent gave 60 mph indicated airspeed, right 45 mph, suggesting differential pressurisation due to propeller corkscrew effect. This could explain the fuel transfer observed earlier.
Flow rates were measured: from drain valve, right tank 20 seconds per litre, left 18 seconds; from fuel flowmeter outlet, right 50 seconds, left 46 seconds. The flowmeter was from a different manufacturer than some Rotax 912 installations known for seizure, but the manufacturer stated seizure would not impede flow. Pressurising each vent to simulate 90 mph increased flow by about 1 second per litre. A test with left tank pressurised to 30 mph (twice the observed difference) caused significant fuel transfer from left to right tank, but not enough to empty it.
Both fuel tanks were removed for detailed boroscope inspection; tight baffling prevented viewing the fuel outlet area. Debris of terracotta-coloured particles was found, analysed as an alumino-silicate and aliphatic ester agglomeration, possibly from fuel containers or hand pump.
Summary
No conclusion was reached regarding the engine failure cause despite thorough examination of the engine and fuel system. Observations from the post-repair flight suggested a possible restriction in the right tank fuel line, but the tests did not confirm this.