Propeller/rotor malfunction

The pilot was carrying out a series of aerobatic manoeuvres when the propeller separated from the airframe. A forced landing was carried out in a suitable paddock. The pilot then noted that one wing had sustained damage from the departing propeller. The evidence available indicates that the propeller fell from the aircraft when the remaining five of the six propeller retaining bolts failed as a result of fatigue cracking and overload. The sixth bolt was not found but lack of marks on the crankshaft flange indicate that it had fallen out prior to failure of the other five bolts. All bolts were in place and secured with lock wire prior to take off. CAA Airworthiness Directive AD/PEP/1 AMDT 2 addresses a known problem which relates to propeller wood shrinkage and loosening of the retaining bolts. The directive recommends special checks on the propeller bolt torque in certain circumstances. The propeller was fitted to the aircraft in Victoria, in a cool moist environment, and the aircraft was then transferred to Perth and operated in a warm dry environment. This change in environment is one of the circumstances set out in the airworthiness directive. No evidence was found to indicate recent compliance with the directive. It is probable that shrinkage of the propeller wood caused loosening of the propeller retaining bolts. The loose bolts were than able to "work" eventually resulting in the failure of the lock wire and loss of one bolt and the failure of the other five bolts.

The pilot reported that while the aircraft was in cruising flight it suddenly began to vibrate heavily. The pilot closed the throttle but the violent vibration continued. The surrounding terrain was generally unsuitable for a forced landing, and in the latter stages of an approach towards a small paddock the right wing struck a tree. The aircraft then dived into the ground and was destroyed. It was subsequently determined that more than half of one of the two propeller blades had separated in flight. The propeller had failed as a result of fatigue cracking. A similar crack was discovered in the other blade. The propeller had only operated for 42 hours total time before the failure occurred. During the approach to land, the pilot's vision was affected by the violent vibration caused by the propeller imbalance. As a result, he was unable to accurately judge his flight path between two trees on the edge of the selected paddock.

During the take-off run the pilot sensed a loss of power being transmitted to the propeller. He assessed that he could not stop in the strip length remaining and reduced the throttle setting in the hope that the reduction drive would transmit the reduced power to the propeller. The aircraft cleared the fence at the end of the strip, but the pilot was forced to turn through about 90 degrees to avoid a power line and to line up with the contours of the ploughed field he had selected for a forced landing. During this turn the airspeed reduced to the point where a high sink rate developed, and the pilot had insufficient height or power to effect a recovery before the aircraft mushed into the ground. On impact, the left mainwheel was torn off and the aircraft overturned some five metres beyond the initial touchdown point. Specialist investigation showed that disc springs had been incorrectly replaced during a modification to the reduction gearbox. This allowed the dog gear to ride up and out of the dog hub, under the reduced stiffness of the axial spring assembly, and reduce the transmission of power to the propeller. Investigation also revealed that there was some ambiguity in the assembly instructions for the gearbox modification.

After application of full throttle during the takeoff run, a severe vibration developed which rapidly stopped the engine from rotating. When the aircraft came to rest the pilot noted that one of the two propeller blades had cleanly separated from its hub. The hub failed as a result of fatigue cracking which extended around approximately half of its circumference. The cracking originated in the root of the blade retention thread and grew progressively over about 850 flights. The cracking was not visible during normal periodic inspections because the hub had been painted to help prevent corrosion.

Shortly after lift-off, both propeller blades separated from the aircraft, which then touched down and subsequently overturned. One of the propeller blades was located 80 metres to the left of the strip and the other, with hub and constant speed unit attached, was located 40 metres to the right of the strip. The propeller hub contained a fracture which was caused by the propagation of a fatigue crack originating at the root of the blade retention thread. The engine had been fitted with two fifth order crankshaft damper weights. This was contrary to the manufacturer's requirement, which was for one fifth order and one sixth order counterweight to be fitted. It is considered likely that fitment of the incorrect counterweights initiated and contributed to the propagation of the fatigue crack.

Shortly after the pilot applied full power and commenced the take-off roll there was a loud bang and the aircraft yawed violently. The pilot closed the throttle and stopped the aircraft. An inspection disclosed that one complete propeller blade was missing. A detailed inspection of the propeller blade indicated that the blade had failed in the area of the threads at the root of the blade. The fracture surfaces indicated that a fatigue crack had started approximately 500 hours prior to the final flight and that the blade had failed because the remaining unaffected metal could no longer carry the loads involved. The inspection was unable to find any reason for the origin of the fatigue crack. The cracking occurred in an area of the blade which is not visible when the blade is fitted to an aircraft, consequently the cracking could not be disclosed during either preflight or periodic maintenance inspections carried out in accordance with the maintenance system in force at the time of the failure. The blades are normally inspected for cracks at each overhaul. An overhaul is carried out after the blades have been in service for 1500 hours. The blade that failed had been in service for 1052 hours.

The pilot reported that shortly after establishing cruise at 300 feet above ground level the aircraft suddenly began to shake as if something had been thrown into the propeller. The engine stopped shortly afterward and, after establishing a glide and making a radio call, the pilot attempted to restart the engine. The engine would not turn over. A landing amongst scattered scrub and clumps of grass was made. Examination of the propeller disclosed that one blade had failed due to a large fatigue crack adjacent to the mounting holes in the blade root. The blade had then severed a bracing wire running between the wing trailing edge and the fuselage behind the propeller. The wire had become entangled in the other propeller blades. The propeller blades were constructed of fibreglass strands in a resin matrix covered with a layer of carbon fibre cloth. Visual and x-ray examination of the blades indicated that the strands were not uniform in direction in the area of the blade roots, some sections had no strands at all, and there were voids in the matrix. These faults reduced the strength of the blades in that area. All blades showed evidence of fatigue cracking near the mounting holes. The holes had been made by multiple drilling. The blade manufacturer advised that he was aware of the faults and had altered the production method. He had advised all owners of similar blades of the deficiencies. This accident was not the subject of a formal on-scene investigation.

At the start of the take-off run, one propeller blade separated from the hub. The resulting propeller imbalance caused damage to the engine and airframe. The fracture was caused by the growth of a fatigue crack from the second blade retention thread. No material abnormalities or stress concentrators contributed to fatigue initiation. It was estimated that fatigue growth had occurred over a period of approximately 2100 take-off/landing cycles. The Civil Aviation Authority subsequently amended the Airworthiness Directives relating to this type of McCauley propeller, when fitted to agricultural aircraft. These amendments include a reduction in hours between overhauls, and more detailed inspections of the threaded section of the blades.

At about 300 feet agl on climb out, the pilot heard a loud bang and saw that the propeller had separated from the aircraft. He lowered the nose to regain airspeed but the aircraft descended steeply and landed heavily in a rice paddy. The engine crankshaft counterweights were found to be worn and detuned. The resulting vibrations were transferred to the propeller hub, which failed and released one of the blades. This type of detuning usually occurs as a result of large and rapid power variations, and as this aircraft was used for agricultural operations it is likely that the work requirements precipitated the failure.

Circumstances:

Seven days prior to the accident a new propeller was fitted to the aircraft after which two brief test flights were made. The aircraft was not flown again until the accident flight. Witnesses reported that on the day of the accident the pilot had been flying the aircraft for about 20 minutes. He was proceeding in a westerly direction at an altitude of about 100 feet above the ground when the rotor blades appeared to disintegrate. The propeller had in fact come off its mounting and hit one of the rotor blades. The rotor blade was destroyed and the aircraft dived to the ground. Subsequent investigation showed that the six bolts securing the propeller to the hub assembly had failed. The propeller was secured to the hub assembly with a compression plate on either side of the propeller hub to distribute the compression loads from bolt tension over a large area of the hub. The plates had been coated with black paint which had allowed slippage to occur between the propeller and the hub plates. This resulted in the inability of the hub assembly to achieve successful engine torque transfer through friction. The torque transfer was achieved through the bolts only which failed in bending.

Significant Factors:

The following factors were considered relevant to the development of the accident:

1. Inadequate engineering design in that the propeller was attached to its mounting in a manner that resulted in the drive being transferred to the mounting bolts.

2. Failure of the mounting bolts, which led to the separation of the propeller.

Recommendations:

1. It is recommended that the Australian Ultralight Federation and the Australian Sport Rotorcraft Association advise their members of the CAA Airworthiness Advisory Circular - May 1987, Item 187-1 addressing the basic principles and conditions of engine torque transmission to wooden propellers.