The flight by Tiger Moth VH-AQN was the third in a series of 10-minute joy flights that day from the local airstrip. On board with the pilot were one adult and a 7-year-old child. Both passengers occupied the forward single seat, with the child sitting on the adult's lap. A lap type safety harness attached to the adult harness restrained the child.

The pilot reported that the engine appeared to be slightly harder to start than normal; however, once started, the engine performed normally. At about 150 ft above ground level after takeoff, and without warning, the engine began to run roughly and lose power. The pilot banked the aircraft to the right towards a nearby golf course to avoid the residential area directly ahead. He was able to manoeuvre the aircraft to land on a cleared area of the golf course. However, during the landing roll, the aircraft's left wings came into contact with a tree, which spun it sharply to the left. The aircraft then began to slide sideways and the landing gear collapsed. The aircraft continued to slide until the right wings came into contact with several other trees. After it came to rest, the pilot assisted the passengers from the aircraft. There were no injuries.

The aircraft sustained major structural damage to the upper and lower sections of the wings and the main landing gear separated from the fuselage at the upper attach points.

The aircraft was approved to operate on mogas (automotive gasoline). Examination of the aircraft fuel system indicated that sufficient clean fuel should have been available to power the engine. The engine sustained minimal damage in the accident, which allowed the investigator to conduct a test run of the engine. During the test run, the engine started without difficulty and accelerated normally.

Examination of the carburettor found that the float valve, which was made from natural cork covered with a fuel proof varnish seal, had two large blisters in the varnish. Further examination indicated that the larger of these blisters was binding against the float chamber housing walls. This could have caused either an excessively rich or lean mixture, which would have caused the engine to run rough and stop. When the spark plugs were examined immediately after the accident, they exhibited a slight oil wetness and sooting, which indicated that the power loss was probably due to an excessively rich mixture.

The float was removed and sent for specialist examination. This examination revealed that the blistered lacquer was very fragile, and that it broke easily and peeled in flakes without adhering to the cork. These characteristics were consistent with the use of a non-approved type of nitrocellulose dope, as identified in the Hobson carburettor overhauling and servicing manual. The examination was not able to conclusively determine the cause for the blistering of the varnish. The operator advised that the already varnished float, approved part number CHA31267, was purchased from the UK and was fitted to the Claudel Hobson model A148HIM carburettor approximately 160 flight hours prior to the accident. No other information was available regarding its history, or whether it was recently re-doped or re-lacquered with modern equivalent materials.

A search of the Bureau's database was unable to find any record of a similar event. During the investigation, several experienced Tiger Moth operators were contacted to ascertain their experience with this type of problem. These operators advised that they were aware of several incidents where the varnish surrounding the cork float had cracked and the cork float had then absorbed fuel. However, none had any previous experience of the varnish blistering in this manner.

FACTUAL INFORMATION

The crew of the Learjet registered as VH-SLJ and operating as Spider 14, had flight planned Canberra - Williamtown - Nowra and was preparing to complete the final leg. At the time, the control tower at Williamtown was unmanned and the control zone was operating as a mandatory broadcast zone (MBZ). While parked on the Royal Australian Air Force (RAAF) air movement section tarmac, the crew of SLJ was monitoring the MBZ frequency. They heard traffic information broadcast by the pilot of Navajo, VH-MYG and an Eastern Australia Airlines aircraft that were inbound to Williamtown. The two aircraft were self-separating and MYG elected to land on runway 12 and the Eastern aircraft positioned for a landing on runway 30. The wind was a light easterly from 070 degrees Magnetic.

The pilot of MYG was operating a "bank run" charter and had planned to make a quick turnaround at Williamtown. He landed MYG from a straight-in approach to runway 30 and, while on final approach, had heard Spider 14 make a general taxy broadcast on the MBZ frequency.

The crew of SLJ decided to depart from runway 12, which necessitated a long taxi on taxiway Alpha but less manoeuvring on departure. They made a general broadcast and reported lining up on runway 12. When approaching the threshold, they heard a taxi call broadcast by MYG so they broadcast on the MBZ frequency that their aircraft was rolling. As they commenced their take-off run from runway 12, they sighted MYG entering the runway 30 threshold from taxiway Juliet. The pilot in command considered aborting the takeoff but elected to continue because the aircraft was lightly loaded and would be airborne quickly. The crew saw the Navajo cross the runway and vacate into the operational readiness platform at the threshold of runway 30.

The pilot of MYG reported that during the turnaround at the civil apron, he remained in the aircraft and monitored the Williamtown MBZ frequency 118.3 MHz on the primary radio and the national advisory frequency 127.7 MHz on the secondary radio. After a short stop to offload one bag, he taxied MYG for a runway 30 departure and made a general broadcast to advise that he was taxiing. He reported that prior to entering runway 30, he scanned the runway and checked the final approach path and that on entering the runway, he made an entering and lining up broadcast. After making the broadcast he noticed the Learjet rolling towards him from the other end, and so vacated the runway to the operational readiness platform. The pilot of MYG did not hear crew of SLJ report lining up on runway 12.

A replay of the audio recording of the MBZ frequency indicates that the crew of SLJ did report lining up on runway 12 but that the pilot of MYG overtransmitted the last portion of the transmission. The crew of SLJ did not hear MYG report lining up on runway 30. Military staff informed the investigation team that the tower radio frequency at Williamtown (used as the MBZ frequency) had some gaps in its coverage.

Neither the crew of SLJ nor the pilot of MYG submitted an air safety incident report, which was required in accordance with AIP Australia ENR 1.14, paragraph 1.3.2.

ANALYSIS

The wind was a light easterly from 070 degrees magnetic and did not favour any particular runway. Mandatory radio broadcasts made on the MBZ frequency by the crews of both aircraft were appropriate and timely.

The pilot of MYG did not hear the "lining up" or the "rolling" broadcast made by the crew of SLJ due to limitations in the MBZ radio coverage. The crew of SLJ also did not hear the pilot of MYG report entering and lining up on runway 30 due to limitations in the MBZ radio coverage. The basis of this analysis is the tape recording of the event, where the pilot of SLJ had been transmitting for sometime before that transmission was overridden by the taxying broadcast made by the pilot of MYG. As a result of the limitations of the radio coverage of the Williamtown MBZ frequency, the crews of both aircraft were prevented from developing accurate situational awareness.

The pilot of MYG did not adequately scan the runway prior to entry for take-off.

SIGNIFICANT FACTORS

Neither crew heard the transmission of the other crew reporting lining up on the reciprocal runway.

The radio coverage of the Williamtown MBZ frequency was limited: the tower radio frequency at Williamtown (used as the MBZ frequency) had some gaps in its coverage.

The pilot of MYG did not adequately scan runway 12/30 prior to entry.

SAFETY ACTION

Local Safety Action

Williamtown ATS authorities issued a Notice to Airmen on 10 December 1998, alerting pilots of the unreliability of the MBZ frequency for ground-to-ground communications due to possible shielding.

Bureau of Air Safety Investigation Safety Action

As a result of this occurrence, the Bureau of Air Safety Investigation is investigating a perceived safety deficiency relating to limitations of the MBZ frequency.

Any safety output issued as a result of this analysis will be published in the Bureau's Quarterly Safety Deficiency Report.

As the aircraft descended through 7,500 ft, the controller informed the pilot that radar services were terminated. Below 8,500 ft the aircraft was in `Class G demonstration airspace' and the pilot was responsible for separation from other aircraft. To assist him in this task, he was advised of radar-observed traffic by the controller, within the limits of radar coverage and as the controller's workload permitted. The pilot had expected that such assistance would not be terminated at 7,500 ft but would be provided to a much lower altitude. Recorded data from three radar sites covering the area was examined during the investigation. 

Returns from the aircraft to one site were reliable down to 7,500 ft, while returns from the other two sites were broken for varying periods of time on a number of occasions during the descent. The radar system software utilised a built-in matrix to determine which radar data was used in specific geographical areas. The matrix had recently been changed as technicians attempted to obtain the best possible coverage for the area. The investigation found that radar coverage in the area the aircraft was operating was unreliable below about 7,500 ft.

The pilot reported that the aircraft appeared to be slow to respond to aileron control inputs during the previous flight. While carrying out a flight control check prior to departure for the next flight, he noticed that the right aileron did not respond to cockpit control column inputs. The pilot cancelled the flight and advised company maintenance personnel of the apparent problem. Examination revealed that the control rod for the right aileron had fractured, separating the aileron from the aileron control system.

Specialist examination of the control rod confirmed that the fracture originated in the internally threaded tube end fitting at the point where the fitting was pinned to the centre tube section of the rod. The features of the fracture were consistent with the application of excessive stress. There was no evidence of pre-existing cracking or other defects. The spherical bearing in the rod end fitting had seized and there was evidence of sliding contact between the outer edge of the inner race of this bearing and the aileron attachment fitting. This suggested that the rod probably failed as a result of aileron control input loads applied after the rod end bearing had seized.

It could not be determined when the control rod had been fitted to the aircraft. A scheduled 100 hourly maintenance inspection of the aircraft had been completed on the day prior to the incident and no abnormality concerning the aileron control system was noted. The maintenance organisation considered that during routine maintenance activities it was difficult to inspect the complete aileron assembly in the area where the failure occurred. Following advice concerning this incident, the Civil Aviation Safety Authority commenced a review of the maintenance organisation's inspection and certification procedures.

The aircraft was on climb passing through 1200 ft, when the crew observed a fire warning on the right outboard (No. 4) engine. Following completion of the checklist items, the engine was shut down. The aircraft returned to Adelaide and conducted an overweight landing. Engineering inspection of the No 4 engine, found that an "O" ring in the combustion chamber fuel manifold at the No 7 fuel nozzle, had split internally.

The resultant small fuel leak and fire had triggered the fire warning. The detector wire is approximately 3 inches downstream of the failed ring. As the engine was approaching a 5,000-cycle heavy inspection, an engine change was carried out and the aircraft returned to service. The fuel nozzle "O" ring has a finite life of 5,000 cycles. The failed ring had completed 4673 cycles and is the subject of a Defect Report submission to CASA.

While conducting spraying operations, the aircraft hit a single wire earth return (SWER) power line. The power line contacted the propeller and proceeded to wrap about the propeller hub. The power line was pulled taut by the aircraft, the pilot lost control, and the aircraft impacted the ground 1,000 metres beyond the point of wire contact. The lateness of the day combined with low overcast is believed to have contributed to the pilot's inability to detect the wire in his flightpath.

During the impact sequence, the underfloor framework with the attachment points for the seat and seat belt appear to have distorted altering their geometry. This in turn may have allowed sufficient movement of the pilot against the loosened restraints for his head to contact with the instrument panel. As he was wearing a helmet at the time, the sequence resulted in severe concussion and not a more serious head trauma.

The pilot landed the Piper Aerostar with the landing gear in the retracted position. Investigation of the landing gear retraction and extension system, by the operator, failed to identify a problem. The aircraft sustained damage to the propellers, lower fuselage, landing gear doors and lower rudder. Repairs to these areas have been carried out and full retraction tests performed without fault. The aircraft was then returned to service.

During the investigation, the pilot reported that when the landing gear was selected down, it did not extend. The pilot later reported to the company that she did not hear the landing gear warning horn, did not confirm that the landing gear was down prior to the landing and did not complete the pre-landing checklist. The company, in consultation with the Civil Aviation Safety Authority, has instituted a revised check program for its flight crews in response to this accident.

Whilst in cruise at FL250 the aircraft de-pressurised and the oxygen masks deployed. The crew declared a PAN, executed an emergency descent, and performed the Rapid Depressurisation and Emergency Descent drills. During the descent it was determined that the selector switch for the cabin pressure controller was in the DITCH position which had caused all cabin pressure to be lost. The crew re-selected the switch to its normal position and regained control of the pressurisation. The descent was terminated at FL160 and the flight continued to its destination. It was determined that the incident occurred when the co-pilot's shoulder contacted the switch while he was returning to his seat.

On this and one other aircraft in the operator's fleet the selection to DITCH is achieved by pushing in and rotating the switch in an anticlockwise direction. All other aircraft in the fleet are fitted with a selector where the selection of DITCH is achieved by pulling out the selector switch and then rotating in an anticlockwise direction. The operator removed the pressure control panel and returned it to the vendor for a check of the switch depression and spring tension settings. Pending the results of these tests the operator is examining the replacement of the push actuated switches with the pull actuated switches. An Operational Notice has been issued to all pilots reminding them of the position and operation of the selector, alerting them to take particular care when getting in and out of the right pilot's seat in flight.

The airship was tasked to fly at night over the city of Melbourne to broadcast a live link to a TV variety show. The pilot telephoned the Bureau of Meteorology aviation forecasters twice before 1830 for a briefing on weather conditions. He was advised by the forecasters that conditions would be windy. The area forecast issued at 1415 indicated that winds at 2,000 ft would be from the north at 35kt.

Moderate turbulence was forecast at all levels with occasional severe turbulence in the lee of the ranges, situated to the north of Melbourne. At 1626 the Bureau issued an amended forecast for Essendon with winds from 350 degrees at 18 kt gusting to 35 kt with moderate turbulence below 5,000 ft. The airship departed Essendon at 2105 and was cleared to operate in the Albert Park Lake area at 1,000 ft. At 2125 the air traffic controller in the Essendon tower noticed that the airship was operating below 1,000 ft.

At the same time the controller and police began to receive telephone calls from the public reporting that the airship was flying very low and erratically. The pilot advised ATC that he was operating at a lower level due to winds, otherwise operations were normal and no emergency conditions existed. When ATC observed a radar altitude of 400 ft AMSL an altitude alert was issued to the pilot. Concerned members of the public alerted the police and emergency services, both of which followed the airship during its return to Essendon.

After arrival at Essendon the airship landed and was secured without further incident. Video recordings of the flight taken both on board and from the ground showed the airship pitching, rolling and yawing while being buffeted by the wind. The extent of the disruption to normal flight could not be accurately assessed because some of the images had been increased in speed when broadcast by television news outlets.