Just after liftoff, the pilot noticed a vibration throughout the aircraft then heard a bang and number one engine over-speed. The engine was immediately shut down and the pilot then saw that the number one propeller was missing. He made a MAYDAY call and landed immediately on the crossing runway. Inspection of the failed propeller revealed the barrel assembly, which retains the blades, had failed due to cracking, liberating both blades. During the investigation, the remaining propellers were removed and inspected but no abnormalities were observed. After repairs were completed, the aircraft was released for service. However, after 17 hours of operation, the number three propeller was removed due to an oil leak.

Inspection of the removed propeller revealed cracking in the barrel in the same area as that found on the original failure. The remaining propellers were removed, and fretting was observed on all propeller cone surfaces indicating the propellers had been operating in a loose condition.

Analysis

The manufacturers maintenance data for the propeller requires that, after installation, the propeller nut torque should be checked again after the first flight and then at 150 hourly intervals. Additionally, the barrel should be inspected for cracking every 25 hours. Review of the aircraft history revealed that the propellers had originally been fitted in the United Kingdom and the nuts were re-torqued after the first test flight. The aircraft was subsequently flown to Australia where it operated for 176 hours prior to the failure.

There is no evidence in the aircraft maintenance records that propeller nut re-torque was performed at 150 hours subsequent to its installation, or that the barrel was inspected for cracking at 25 hourly intervals. The manufacturers installation data for the propeller requires that the propeller nut be torqued to 600 lbs/ft and, whilst maintaining that torque, the wrench is given two taps with a soft mallet in the direction of nut rotation to ensure the cones are seated. Maintenance personnel involved with the propeller installation advised they had not used this method but simply torqued the nut to the required value with a torque wrench. Experiments were carried out on applying torque to a propeller nut.

It was found that when the required torque was obtained and the wrench was tapped twice with a soft mallet, the nut rotated a further 1/4 to 1/2 turn. This indicated that the nut was not sufficiently tight to seat the cones when only the initial torque was applied.

The propeller manufacturers Technical News Sheet number 26, dated January 1963, states that investigations of bracket type propellers which have suffered failure of major components have directly attributed the failures to operation of the propeller in a loose condition due to:

1. Incorrect installation.

2. Failure to check tighten or to effectively check tighten due to incorrect procedures.

Findings

1. The initial failure of the number one propeller was consistent with operation of the propeller in a loose condition because propeller nut re-torque was not carried at the recommended time periods.

2. The failure may have been averted had inspections of the barrel at the recommended intervals detected cracking.

3. The cracking of the number three propeller barrel and fretting of the other propeller cones was consistent with operation of the propeller in a loose condition as a result of incorrect nut torquing procedure.

Significant Factors

Maintenance personnel involved in the maintenance of the propellers failed to observe the requirements of the manufacturer's maintenance data.

Safety Action

The safety deficiencies identified during this investigation were corrected as they were identified. Consequently, no safety recommendations have been raised.

On approach into Melbourne the landing gear was selected to the down position for landing. The usual thump of the nose gear uplock unlatching and subsequent air noise was not evident. The nose landing gear had not unlocked and only two green indicator lights illuminated for the main gears. The aircraft was vectored away from the circuit where a manual gear extension procedure succeeded in lowering the nose gear.

The aircraft returned for a normal landing.

Investigation disclosed that a bush on the nose landing gear drag brace had seized. The aircraft's normal landing gear extension system was unable to overcome the drag created by the seized bush. It is suspected that inadequate lubrication or excessive washing out of the lubricant was responsible for the bush seizing.

A fleet check was carried out, however, all other bushes were found to be satisfactory.

During an aerial agricultural flight, the engine began to lose power. The pilot observed a low oil pressure indication before the engine suddenly seized. During the subsequent forced landing, the pilot attempted to extend the glide in order to avoid trees. However, the aircraft stalled at a height of about 30ft and landed heavily on the left main landing gear, which then collapsed

An inspection revealed that two top cylinders of the radial engine had failed, and there was no significant oil left in the engine. The engine had accumulated only 10 minutes flight time since major maintenance had been carried out, during which a top cylinder had been changed. It was later found that the oil sump plug was loose. It was apparent that the engine oil had leaked from the engine, which had then seized. The LAME who performed the engine repair was apparently distracted by other events and had failed to tighten the sump plug.

The pilot had flown the helicopter from Darwin to Jabiru, carried out some local flying at Jabiru and then returned to Darwin on the day of the accident. As the helicopter was approaching Darwin the pilot was instructed by an air traffic controller to hold position in the Palmerston area and await further clearance. Shortly afterwards the engine stopped, and the pilot was forced to complete an autorotational landing. The pilot misjudged the approach, and the aircraft landed heavily.

The pilot did not complete a flight plan prior to the flight and no evidence was found to indicate she used any form of formal fuel management to ensure that fuel available met the required reserves.

The aircrafts total fuel capacity of 200 L provided for an endurance of approximately 2.6 hours at a normal fuel consumption rate of 75 L/hour. The flight to Jabiru took 2.5 hours. A total of 194 L of fuel was required to fill the tanks following this flight. The tanks were also refilled following the local flying. Although almost all the fuel had been exhausted on the flight to Jabiru the pilot departed for the 2.5 hour flight to Darwin without having made any alternative fuel arrangements. The fact that the aircraft had almost run out of fuel on the flight to Jabiru did not register.

When the pilot was instructed to hold position in the Palmerston area and await a further clearance she advised the controller that the aircraft was low on fuel but she did not communicate any urgency about the situation. Ten minutes later, whilst still in a holding pattern, the engine stopped as a result of fuel exhaustion. The helicopter had been airborne for 2.6 hours when the engine stopped.

The pilots actions indicate she was never sufficiently aware of the aircraft's fuel situation. The lack of planning and the failure to use a formal management procedure probably exacerbated the situation.

The pilot initiated normal gear selection after arriving in the circuit area at Manilya. The gear extension process stopped after the nosewheel and right main gear had extended. The pilot checked the circuit breakers and recycled the gear twice more with the same result.

The pilot decided to return to Carnarvon where the gear could be inspected visually. Observations by a pilot and engineer confirmed that the left main gear was up. Use of the manual extension system and the introduction of 'g' forces were unsuccessful in causing the left main gear to extend.

After discussion with engineers in Carnarvon and Perth the pilot elected to carry out a wheels up landing on the grass flight strip beside sealed runway 18. The landing was successful.

An inspection of the landing gear systems disclosed that the hinge bolt, from the front hinge of the left landing gear door, was missing. The missing bolt allowed the landing gear door to twist, in the airflow, jamming the door closed and preventing the left gear from extending.

The missing bolt was not recovered. A number of landing gear hinge bolt nuts, including one that was adjacent to the hole left by the missing bolt, were recovered from the channel surrounding the landing gear bays on both sides of the aircraft. It was not possible to determine if any of the nuts came from the missing bolt. There was no evidence of stress, in the area of the missing hinge bolt, to indicate that the bolt had been working prior to its loss.

The landing gear doors had been removed, inspected for cracking, and replaced eight flying hours prior to the accident. The gear had been extended and retracted a number of times, without apparent problems, after the replacement and prior to the final flight. The failure sequence indicates that the bolt was probably in place for a period following the maintenance and fell out on the accident flight. No history of problems with the hinge bolts was disclosed during the investigation.

It was not possible to determine whether the bolt failed and fell out, as a result of fatigue or overload, or if it was incorrectly fitted following the maintenance conducted shortly before the accident flight. There was some evidence that the engineer responsible for re-fitting the suspect bolt was distracted by a personal problem during that process however he did not recall any specific event that may have contributed to the loss of the bolt.

The pilot reported that during the approach the airspeed was too high. He decided not to go around as the tall trees surrounding the strip made it look restricted. He flew the aircraft straight and level to wash off airspeed and felt he could still land and stop before the end of the runway.

On touchdown the aircraft ballooned. The pilot recovered to fly the aircraft parallel to the runway but did not add power, to overcome the ballooning effect, as he felt he would run short of runway. On the second touchdown the aircraft bounced and on the third touchdown the nosewheel collapsed. The aircraft skidded to a halt towards the left side of the runway.

The accident sequence was typical of other accidents where the runway perspective appears different to that previously experienced by the pilot. The perspective can be altered by a narrower than usual runway for its length and/or by the surrounding terrain or foliage.

The pilot was landing on the westerly strip at McCaffrey Field with 10 knots of wind from the north (all cross wind). During the round out for landing, the aircraft ballooned, and the pilot pushed the nose down and at the same time retracted the air brakes. The aircraft struck the ground in a nose down attitude and bounced. It then veered left under the influence of the cross wind and bounced again, before running into the boundary fence.

During climb to cruise, passing 7,500 ft, the ground proximity warning system sounded as the aircraft encountered moderate rain and light hail. There were storm cells on either side of track.

Post flight inspection revealed hail damage to all leading edges, right landing light cover, and radome.

The ground proximity warning system often gives a false alarm when it encounters precipitation (heavy rain and/or hail). This is a known deficiency against which there is no technological defence.

The pilot had remained clear of the storm cells by using his airborne radar. It is a known phenomena that hail can fall well outside its associated storm cells.

Sequence of Events

The instructor pilot said that his student had completed four dual circuits with very satisfactory landings off each approach. During the fifth approach, the student flared the aircraft too high and before the instructor could take control, it stalled and struck the ground in a heavy landing.

The left main gear collapsed, and the aircraft came to rest on its bent nose gear and left-wing tip.

Analysis

The instructor was unprepared for the high round out following the student's good landing performance on the previous occasions. He did not take control in time to prevent a heavy landing.

Whilst on descent through about FL240 with C autopilot engaged, the aircraft rolled rapidly to the right but immediately corrected itself. Speed brakes were extended at the time.

Flight Data Recorder information did not reveal any unusual flight control inputs, however individual spoiler positions are not recorded, only the speed brake control handle position is recorded. A ground functional test of the flight control system was normal, but a loose electrical connector at number 6 spoiler (left wing) power control actuator was found. It is considered that the loose connection may have commanded the spoiler to retract, causing a roll to the right which was corrected by the autopilot.

The operator has instigated a fleetwide check of spoiler electrical connections.