ERA17LA246

The fatigue failure of an exhaust valve spring retainer due to air trapped in the lubrication system, which resulted in a total loss of engine power.

**This report was modified on December 4, 2022. Please see the public docket for this accident to view the original report.** While in the airport traffic pattern for landing at the conclusion of a cross-country flight, the airplane experienced a total loss of engine power, and the pilot performed a forced landing during which the airplane sustained substantial damage. The airplane had recently been purchased and the Rotax 912 ULS engine had 13.2 hours total operating time. Review of onboard data indicated that the fuel pressure, cylinder head temperature, and oil temperature remained relatively steady until the loss of power occurred, which indicated that the engine failure likely did not involve the fuel system, cooling system, or lubrication system. Examination of the engine revealed that there was no oil in the oil line between the oil thermostat and oil pump. The oil pump drive pin also displayed excessive wear in relation to the operating hours of the engine, and the magnetic plug was covered in metallic particles, although the oil filter was clean. Further examination of the engine revealed that the No. 1 cylinder was substantially damaged, and evidence of bluing was present. The cylinder’s exhaust valve spring retainer was fractured in half, and one half of the cotter was fractured. A small ridge could be felt on the exhaust valve spring retainer and galling (a rough surface) was visible on the exhaust valve bore in the cylinder head. The hydraulic lifter for the exhaust valve displayed a small indentation on the edge of the lifter, and when the hydraulic lifters were manually depressed, the lifter for the exhaust valve was easier to depress than the lifter for the intake valve. The pushrod for the exhaust valve was straight, but displayed a ridge on the rocker arm side of the pushrod, and the rocker arm displayed impact damage on the valve connection face. The exhaust valve was found in the combustion chamber. It was chipped, and bent, and deformed into an “S” shape. A hole was visible in the piston as the result of the piston face striking the exhaust valve after it dropped into the cylinder. A small amount of oil captured from the hydraulic tappets indicated that the oil contained significantly elevated levels of nickel, which could have come from manganese-containing alloys, as they occur in high-alloy hardened steels, e.g. for camshafts, valves or valve shafts. Examination of the fractured surface on the exhaust valve spring retainer revealed the presence of fatigue with pronounced vibration stripes when viewed with an electron microscope; however, the heat treatment corresponded to the target specifications, as did the statistical process control value. Between 2 and 3 years after this accident, four more cases of broken valve spring retainers on Rotax 900 engine series occurred in the United States. All the engines had differing hours of operation. Extensive metallurgical examination of the engine components from these four engines revealed that they met their specifications, and the fractured surfaces on the valve spring retainers revealed the presence of fatigue with pronounced vibration stripes, which was the same pattern that was observed on the valve spring retainer from this accident. Review of the engine manufacturer’s published guidance revealed that air could be introduced into the oil lubrication system through several means, including exceedance of the maximum bank angle of 40°, poorly or insufficiently vented hydraulic valve tappets, lack of proper oil system purging, spinning the propeller in the reverse direction from normal rotation, or opening portions of the oil system during maintenance or servicing. Testing of an exemplar engine with air introduced into the lubrication system revealed that with air trapped in the hydraulic tappets, it took about 6.5 minutes of engine operation at 2,538 rpm for air to be purged from the tappets, allowing them to work as designed. This indicated that with air trapped in the hydraulic tappets, the valve train could be overloaded, which could lead to a fatigue crack and breakage of a valve spring retainer; this was likely the reason for the fatigue cracking of the valve spring retainers in this accident and in the other four failures identified.

On July 15, 2017, about 1615 eastern daylight time, a Costruzioni Aeronautiche Tecnam P92 airplane, N561TU, was substantially damaged when it was involved in an accident near Stevensville, Maryland. The two private pilots were not injured. The airplane was operated as a Title 14 Code of Federal Regulations Part 91 personal flight. The airplane had recently been purchased by the owner and placed on a lease-back operation with the operator. Two days before the accident, the owner, along with the pilot who was in the right seat during the accident flight, took delivery of the airplane in Apopka, Florida, and flew it to Bay Bridge Airport (W29), Stevensville, Maryland. On the day of the accident, the airplane was fueled to about 16 gallons (8 gallons per side) for a roundtrip flight to Shoestring Aviation Airfield (OP2), Stewartstown, Pennsylvania. Before departing on the return flight, the left seat pilot checked the oil, coolant, and fuel. The oil and coolant levels were normal, and the airplane contained about 12 gallons of fuel. Upon arrival in the area of W29, the pilots obtained the current conditions from the automated weather observation and entered the traffic pattern for runway 29 on the crosswind leg. They observed no other traffic in the pattern at the time, and due to noise-abatement rules, they conducted the runway 29 downwind leg about 2 miles south of the airport. The left seat pilot, who was flying the airplane, reduced engine power and began to configure the airplane for landing abeam "the 29 numbers." Several seconds after the power reduction, the engine abruptly started to run rough. At this time, the right seat pilot took the controls. Both pilots scanned the engine indications but did not observe any anomalous readings. The right seat pilot turned onto the base leg of the traffic pattern, but did not turn directly toward the runway out of concern for arriving too high at the threshold and a flightpath that would have resulted in overflight of a densely-populated townhouse community. The pilots increased the flaps setting to correct for the high glidepath, and about 20 seconds later, the engine abruptly stopped. The right seat pilot turned directly toward the runway threshold, and both pilots determined that the airplane would not reach the runway. After considering their forced landing options, the right seat pilot turned the airplane toward a cleared but rough area of open ground about 45° left of their flightpath. The airplane "firmly" glanced off the top of an earthen berm and settled onto the rough ground beyond it. During the landing roll, about 150 ft from the touchdown point, the airplane struck a second berm, the right main landing gear and nose gear separated from their mounting points, and the airplane came to rest about 20 to 30 ft beyond the second berm. The pilots shut off both fuel valves and the master switch and then egressed. The airplane was equipped with a Garmin G3X electronic flight instrument system (EFIS), which provided full primary flight display attitude and directional guidance along with electronic engine information. Review of data downloaded from the G3X indicated that fuel pressure, cylinder head temperature, and oil temperature all remained relatively steady until the loss of power occurred. On September 9, 2017 and April 12, 2018, the airplane and engine were examined by the NTSB. The airframe was substantially damaged. During the impact sequence, the nose landing gear separated from its mounting location, the right main landing gear bent back and toward the left main landing gear, and the left main landing gear was damaged. One blade of the two-bladed propeller was broken off, the engine had been pushed back toward the firewall, the engine mounts were bent, the firewall was buckled, and the fuselage and wings displayed multiple areas of crush and compression damage. External examination of the engine revealed that the air filter was clean and the exhaust system was damaged, but no anomalies were noted. The cooling system was intact. The oil line between the oil cooler and oil thermostat was kinked during the impact sequence, and the Nos. 2/4 (left side) carburetor had been displaced from its intake socket. The propeller gearbox rotated smoothly with no binding noted. The sparkplug electrodes appeared normal and the spark plug gaps were all 0.19 inches. Both the Nos. 2/4 (left side) and Nos. 1/3 (right side) carburetor float bowls contained automotive gasoline. No anomalies were noted with the carburetors. Both the mechanical and electric fuel pump were functional. No oil was found in the oil line between the oil thermostat and oil pump. The oil pump drive pin displayed excessive wear in relation to the operating hours of the engine. The oil cooler appeared to be undamaged. The magnetic plug was covered in metallic particles. The oil filter was clean. Cylinder No. 1 displayed substantial damage and evidence of bluing was present. The exhaust valve spring retainer was fractured in half, and one half of the cotter was fractured. A small ridge could be felt on the exhaust valve spring retainer, and galling (a rough surface) was visible on the exhaust valve bore in the cylinder head. The hydraulic lifter (tappet) for the exhaust valve displayed a small indentation on the edge of the lifter, and when the hydraulic lifters were manually depressed, the lifter for the exhaust valve was easier to depress than the lifter for the intake valve. The pushrod for the exhaust valve was straight, but displayed a ridge on the rocker arm side of the pushrod, and the rocker arm displayed impact damage on the valve connection face. The exhaust valve was found in the combustion chamber. It was chipped, and bent, and deformed into an “S” shape. A hole was visible in the piston. No unusual marks were seen on the intake valve rocker arm, valve keeper retainer, valve or valve stem, or valve keepers. Cylinder Nos. 2, 3, and 4 did not display any anomalies. The crankshaft was twisted and would not rotate within the crankcase; the camshaft displayed no visible anomalies. The internal configurations of the oil system thermostat and associated oil system hoses were documented using radiographic images, and there were no indications of blockages, broken components, or hose breaches. On November 13, 2018 in the presence of the Austrian Federal Safety Investigation Authority (BMK), the No. 1 cylinder head assembly, cylinder, oil pump assembly, oil tank assembly, and oil cooler were examined at Rotax Aircraft Engines in Gunskirchen, Austria. Examination of the fractured surface on the valve spring retainer by electron microscope revealed the presence of fatigue with pronounced vibration stripes. The heat treatment, however, corresponded to the target specifications, as did the statistical process control value. Examination of the cylinder head revealed that the shim from the intake valve and exhaust valve showed unusual wear on the spring contact surface, indicative of increased spring movement. A hardness test of the shim indicated that it corresponded to the drawing specifications. No deviation from the drawing specifications was discovered. Examination of both hydraulic valve tappets revealed that the oil control plates showed noticeable wear. Examination of the oil pump showed no indication of malfunction. The housing, suction inner and outer rotor showed no abnormalities. Examination of the oil tank showed no abnormalities. No indication of a malfunction was visible. Examination of the oil cooler did not reveal any abnormalities and a leak check revealed no indication of a leak or visible malfunction. A small amount of oil from the hydraulic tappets was captured and sent to an independent laboratory for analysis. According to the analysis report: The lead content in this sample is an indication of the usage of leaded fuel. Nickel is significantly elevated. Could come f