During the take-off run, just before reaching V1, the crew heard two bangs in quick succession. Both the No 3 engine EPR and EGT indications were seen to fluctuate, so the take-off was immediately rejected. Whilst decelerating, the crew requested a frequency to communicate directly with the Rescue and Fire Fighting Service (RFFS) but were given the ground control frequency. That frequency was cluttered with other traffic, which made direct communication to RFFS difficult. They were finally referred to a different frequency. The aircraft vacated the runway onto taxiway Alpha 5 where a small brake fire was extinguished by the RFFS. The aircraft was subsequently towed back to the departure gate, where the passengers and crew disembarked.

Ground inspection revealed damage to several compressor stages of No.3 engine. The engine was changed, and the aircraft released for service.

Safety Action

The Bureau of Air Safety Investigation is evaluating aspects of a discrete air traffic services frequency on which aircrews and rescue firefighting services at airports can be transferred during an emergency situation. Any forthcoming recommendations will be published in the Quarterly Safety Deficiency Report.

FACTUAL INFORMATION

Sequence of events

The flight was to conduct photographic work around the Kununurra area. The crew consisted of the pilot in command and a second pilot in the right front seat to assist with look-out, radio communications and communicating with the photographer. The photographer operated from the rear seat without a headset.

After the crew had completed part of the task, the pilot started to climb the aircraft to a higher altitude. As the aircraft passed about 600 ft above ground level, the engine stopped suddenly without warning. The pilot transmitted a Mayday message and carried out a forced landing into 30-foot trees. During the impact, the aircraft turned over and the stitching of the harness of both pilots came undone, releasing the shoulder harnesses.

The pilot in command and the photographer suffered minor injuries. The right-seat pilot was seriously injured.

Wreckage

The aircraft came to rest inverted with the left wing torn off. The propeller blades were broken but showed no signs of rotation during impact. The Bureau was advised that, four days after the accident, an inspection of the fuel tanks indicated that the right tank was about three-quarters full, while the left tank contained only about 20 litres. Fuel was leaking from the impact-damaged left tank-to-engine feed lines. Consequently, the fuel quantity remaining in the aircraft at the time of the engine failure could not be accurately determined. Subsequent specialist inspection of the fuel-system components found that the engine-driven fuel pump outlet port was blocked by grit wrapped in fibrous material. The system filters were also contaminated.

The aircraft was normally refuelled at Kununurra, but had been refuelled from drums when operating from another strip in the area.

Aircraft maintenance

Further inspection of the engine found that the mixture control was excessively worn and that three spark plugs could be unscrewed by hand. The time at which contamination of the fuel system occurred was determined. The last periodic inspection was conducted 94 hours prior to the accident.

Emergency locator transmitter

A portable Pointer 3000 emergency locator transmitter (ELT) was carried in the baggage compartment. The ELT was turned on by the pilot after the accident and its signal was detected by the Brisbane Search and Rescue Co-ordination Centre.

ANALYSIS

The reported blockage of the engine-driven fuel pump outlet port is consistent with the circumstances of the engine failure. Contamination could have been introduced to the system if appropriate filtration was not used during refuelling from drums. Although there were indications that the aircraft maintenance was deficient, there was insufficient evidence to suggest that inadequate maintenance was implicated in the circumstances of the accident.

SIGNIFICANT FACTOR

It is likely that engine power was lost due to fuel starvation caused by system contamination.

The pilot was tracking for left base, runway 22 (about 290 degrees), and was instructed to report abeam the Queen Elizabeth 2 stadium. A twin- engined aircraft, VH-DRB, was descending to 1500 feet from CTA to the south east of Archerfield. The pilot of this aircraft had initially requested a landing on runway 28 but the controller had difficulty in facilitating this request. The controller ultimately instructed VH-DRB to make a visual approach via a wide left base for runway 22, and to follow a third aircraft on final approach.

In VH-JVV, the pilots were monitoring the progress of VH-DRB by radio. The trainee instructor alerted the pilot in command to the position of the other aircraft. An immediate climbing turn was initiated in order to avoid a collision. The controller attempted to confirm that JVV had VH-DRB in sight at about the same time.

The pilot of VH-DRB saw the other aircraft during its avoiding turn. His aircraft was still descending, and the other aircraft was probably hidden from view by his aircrafts nose section, the instrument combing or the right windscreen pillar. He did not take any avoiding action.

The aerodrome controller had not passed traffic information about each other to the two aircraft involved. The warning transmission from the controller was made at the same time as the pilot of VH-JVV reported at QE 2 stadium.

The pilot of VH-DRB originally requested runway 28, and the controller had initially attempted to facilitate this request. In doing so he had not considered the possible track of that aircraft when it was instructed to track for a wide left base for runway 22.

The weather at the time was fine, with a clear sky and unrestricted visibility.

At 1025 the pilot of VH-NJY contacted Mackay Tower at 30 NM inbound on the 155-degree radial on descent to 6,000 ft. VH-NJY was then cleared to descend to 4,600ft and track via a 5 NM arc to the west of track and then to track overhead for a NDB approach. At 1027, VH-EWS departed Mackay for Brisbane tracking on the 137-degree radial on climb to 10,000 ft. Mackay tower controllers commenced trying to visually observe VH-NJY inbound. This was done by the use of standard issue binoculars. The inbound track of the aircraft is situated over a prominent topographical feature known by tower staff, however, VH-NJY could not be visually acquired at this stage. At 1029, VH-NJY was observed visually at approximately 12 NM and 12-13 degrees east of track. VH-EWS was 6 NM on the 137-degree radial. VH-NJY was instructed to turn left to commence the diversion for the NDB approach.

A traffic alert was not issued to the aircraft, as it was the judgement of both controllers on duty that although the aircraft were found to be closer than normal, a collision risk did not exist. They believed there was approximately a mile laterally between both aircraft, and they were diverging. Subsequent replay of the radar tape indicated that VH-NJY and VH-EWS passed each other within approximately 0.7 NM laterally and 800 ft vertically. The Aeronautical Information Publication Australia (AIP) Air Traffic Rules and Services (RAC) 44.4.3 requires aircraft to be navigated by reference to the aid which provides the most precise track guidance, which in Mackay is the Mackay VHF Omni Range (VOR). The pilot in command is required at all times to take positive action to regain track as soon as a deviation from the correct track is recognised. The VOR tracking tolerance at 10 NM is approximately plus or minus 0.9 NM VH-NJY indicated on radar as being approximately 2.6 NM off track at 10 NM which is 1.7 NM outside the approved tolerances. Had this fact been known to air traffic controllers, then corrective action to ensure separation could have been taken. Mackay tower is not equipped with radar. Radar data from Swampy Ridge radar sensor is processed through the technical facilities at Mackay, enroute to Brisbane centre, but no radar display is provided for tower controllers. VH-NJY was under radar control during the flight until the aircraft was transferred to Mackay tower at 30 NM when radar services were terminated. The separation technique being used by tower controllers was one of visually monitoring the outbound aircraft on a track that was plotted on a map to be clear of the procedural tolerance, (in this case 5.2 degrees off the VOR radial, plus an additional 1 NM lateral separation buffer), of the inbound aircraft. No documentary evidence could be found to permit the application of this standard.

The pilot in command (PIC) of VH-NJY was a check captain and the copilot (CP), a first officer, was undergoing command training. The PIC had requested an NDB approach as a training exercise and had selected the VOR selectors off to simulate their failure. He was aware of VH-EWS departing on the 137 radial and tracking to Gladstone. At 10 NM he noted that the trainee had allowed the aircraft to diverge to the right of the 155 radial. The PIC did not bring this to the attention of the trainee, in order to see if he was going to fly outside tracking tolerances. The PIC said that at no time did the ADF needles indicate the aircraft was any further east than the 150 radial. He was satisfied that the aircraft was well clear of the Hay Point helicopter pad which is located near the 137 radial. He also had VH-EWS in sight passing to the right and was satisfied he was well clear. In summary, the PIC of VH-NJY was operating outside tracking tolerances and outside the terms of his airways clearance. This was due to deselection of the most accurate navigation aid (the Mackay VOR) to simulate its failure for training purposes. A non documented separation standard was used by the tower controllers. An instruction has been issued to all Northern District Towers that until a separation standard is developed and approved, outstation staff are to apply only those separation standards that are included in the Manual of Air Traffic Services (MATS).

The radar sector providing enroute control to VH-NJY did not advise the pilot that radar services were terminated when he was handed off to Mackay tower. The District Safety and Quality Management Branch has recommended that radar controllers advise pilots proceeding into procedural tower environments within radar coverage, when radar services are terminated. Weather conditions were fine, with Mackay under the influence of a high-pressure system, QNH 1023, nil cloud, visibility in excess of 10 km, and a light south-easterly wind.

The pilot reported that he lowered the gear on the downwind leg of the circuit and the green light illuminated, indicating the gear was fully down and locked. The remainder of the landing checks were completed before turning base. A gentle touchdown on the main gear was made on the threshold of runway 15. As the nosewheel was lowered onto the runway, the operating lever came out of the lock and the gear retracted.

A foreign registered Boeing 747 was proceeding from Brisbane Qld to Cairns Qld, via airway H106. The flight was being conducted as a check flight. As the aircraft was being processed for arrival, the crew were cleared to descend to 6,500 ft, which the crew confirmed in their first contact with Cairns Approach. The aircraft's progress was being monitored by Air Traffic Control (ATC) on radar.

ATC instructed the aircraft to expect a left circuit for runway 15 with further descent in 14 track miles. The crew read back "...14 track miles". Shortly after, the controller observed the aircraft's Mode C SSR readout indicating 5,300 ft at 19 NM from Cairns. The radar lowest safe altitude in this sector was 4,400 ft. The crew were instructed to maintain 5,500 ft. Subsequently, the crew reported that the aircraft was clear of cloud and that they had sight of the surrounding terrain.

ATC processed the aircraft via a left circuit for runway 15 at Cairns without further incident.

During the investigation the crew reported that they understood the ATC instruction "further descent in 14 track miles" to mean that they were cleared to descend to 4,500 ft at 14 DME Cairns on a DME arrival. The crew also stated that ATC "was a little bit difficult to understand". Regardless of this perceived difficulty, the crew did not seek confirmation of the ATC instruction.

The recording of the communication between ATC and the aircraft confirmed that prior to this incident no further descent instructions were passed to the aircraft.

SIGNIFICANT FACTORS

1. The crew misunderstood the ATC instruction "further descent in 14 track miles".
2. The crew did not question the ATC instruction before descending below the last assigned level.

The pilot reported that when he landed, the nosewheel was cocked sideways, and collapsed when it contacted the runway. The pilot said the nosewheel steering link was probably broken before the aircraft departed Innisfail. He believes the nosewheel steering may have been inadvertently forced past the steering limit, while being towed.

The failed nosewheel fork was examined by a CASA airworthiness inspector, and in his opinion the nosewheel strut failure was caused by overload after the tyre and wheel failed. He said the possible causes of the tyre and wheel failure could be a heavy landing in crosswind, or the nosewheel impacting with something on the runway.

The strut failed rearwards and substantially damaged the fuselage structure back as far as the wing strut attachments. The cause of the tyre and wheel failure could not be positively determined.

The pilot stated that the purpose of the flight was to visit an engineering workshop in Charleville, and to demonstrate the aircraft which was for sale. Because he wanted to return from Charleville on the same day, and the flight time each way was approximately 4 hours 30 minutes, he had planned for an early departure. When he arrived at the airport, the aircraft was covered in a thick frost, and it took him over an hour to defrost the aircraft and complete a daily inspection. He departed Orange at 0620, and climbed to 6,500 ft. The weather was fine and sunny, but very cold so he pulled the heater control to full on. After passing Bollon, 120 NM southeast of Charleville, the pilot tuned to the Charleville NDB and noted that the needle rotated to indicate straight ahead. He recalled passing a strip about 25-30 miles from Charleville and made a mental note that he could land there if the weather deteriorated. He said that he then started thinking about the descent and recalled thinking about turning the heater down. It appears that the pilot then lost consciousness as the next thing he remembered was that the aircraft was descending through 1,000 ft at about 1,000 ft a minute.

The pilot said he pulled back on the control yoke and saw the horizon come into view. He banked hard left, and the engine coughed. The pilot then noticed what looked like a very long airstrip ahead and landed. After landing the pilot noticed the time was 1230 and realised that his ETA for Charleville had been about 1100. The pilot said his head was aching and he felt cold and nauseous.

The pilot eventually fixed his position roughly 150 NM northwest of Charleville, and after repairing the aircraft radio, he was able make contact with an overflying jet late the next day and was rescued.

Examination of the cabin heat system revealed a large amount of exhaust build-up in the scat hose leading to the cabin heat selector valve. The muffler was badly cracked around the outlet port and had a white soot stain around it. It was concluded that the pilot may have been effected by carbon monoxide which entered the cabin via the cracked outlet port of the muffler.

Approaching Adelaide, the aircraft made numerous diversions from track to avoid weather. On first contact with Adelaide approach, the pilot reported "....taking up a heading of 205 to intercept the localiser for runway 23, descending to 5000” The pilot was instructed to "...maintain 5000 make pilot intercept of the localiser, at this stage expect one holding pattern at Modbury".  The holding pattern was required for separation with other traffic.

The controller stated that the aircraft intercepted the runway 23 localiser beyond 20 NM from Adelaide.  When the aircraft was approximately 17 NM from Adelaide, the controller advised the pilot "...cancel the holding and expect ILS approach, descend to 3000". The pilot then requested confirmation they were cleared for an ILS approach.  The controller instructed the aircraft to "...descend to 2000 on the ILS" and a readback of that instruction was obtained.

A short time later the pilot reported maintaining 2000 ft. The pilot was instructed to maintain 2000 ft and was then asked if they were visual.  The pilot confirmed that they were visual, and the aircraft was then cleared to continue on a visual approach to runway 23.

Recorded radar data showed that the aircraft made a continuous descent from 5000 ft to 2000 ft and passed through

3000 ft at approximately 13 miles from Adelaide.  The aircraft reached 2000 ft at approximately 10 miles from Adelaide. The Adelaide control zone extends to 11 miles from Adelaide on the runway 23 localiser and from 11 miles to 20 miles the lower limit of controlled airspace is 2500 ft.  The descent profile depicted in the recorded radar data suggests that the aircraft was probably below the 2500 lower limit of controlled airspace for a few seconds prior to entering the control zone.

The captain advised that they were using Jeppesen charts and the runway 23 ILS chart has a note which says that aircraft arriving from the northwest may be radar vectored to intercept the localiser at 2000 ft.  The captain said that when they were cleared to descend to 2000 ft they were established on the localiser and both flight crew members thought they were cleared down to 2000 ft at that time.  They had obviously misinterpreted the intent of the instruction. He also pointed out that it is not normal to have a descent restriction imposed once cleared for an ILS approach.

The captain said that cockpit workload was high at the time in that the aircraft was being controlled by the flight management system for an automatic ILS approach.  Both glide slope and localiser had been captured and when the holding requirement was imposed, this created a potential extra workload in reprogramming the flight management system to get out of the approach mode and set up for a holding pattern. Fortunately, the holding requirement was later cancelled.

Factors

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

1. The flight crew misinterpreted an instruction from air traffic control.
2. There was a high workload in the cockpit at the time.
3. The air traffic controller did not notice the aircraft continue its descent below 3000 ft and in fact did not become aware the aircraft had reached 2000 ft until the pilot reported at 2000 ft.
4. The instruction given by the air traffic controller to "... descend to 2000 on the ILS", although correctly read back by the pilot, was open to misinterpretation.

Safety action

The controller was counselled concerning the phraseology of the instruction he gave to the aircraft.  The captain was apprised of the situation from the air traffic controller's viewpoint and a cassette tape of recorded communications relevant to the incident was sent to him for review.

FACTUAL INFORMATION

A British Aerospace 146 (BA146) had departed Brisbane for Proserpine on climb to FL260. Four minutes later a Boeing 747-Special Performance (B747SP) departed Brisbane en route to Taipei on climb to FL350. The initial departure track was the same for both aircraft and was to the northwest of Brisbane. The Brisbane Sector 3B controller believed from past experience, that the rate of climb of the B747SP would far exceed that of the BA146. Consequently, he decided to monitor the horizontal separation on the radar display until vertical separation was achieved.

Approximately ten minutes after departure, the crew of the B747SP observed on the aircraft collision avoidance system (ACAS) display an aircraft ahead and at the same level. From the high rate of closure, the crew estimated that there would be a breakdown in separation very shortly unless action was taken to avoid the aircraft. They requested the Sector 3B controller advise them of the traffic. At the time, the horizontal separation between the aircraft was just over 5 NM and the groundspeed of the B747SP was 100 kts faster than that of the BA146.

The Sector 3B controller had been operating the position for the previous two and a half hours. He had been busy, but the number of aircraft had reduced in the last 15 minutes. During the last 30 minutes he had answered a number of radio calls from the pilot of a Visual Flight Rules (VFR) flight which was operating outside controlled airspace but in his area of responsibility for a radar advisory service (RAS). RAS had been introduced one month previously, and the sector had almost full radar coverage in non-controlled airspace. Consequently, the sector controllers often had pilots operating on the frequency using RAS procedures. The pilot of the VFR aircraft was incorrectly pre-fixing his radio transmissions with "Brisbane Centre" which required an acknowledgement from the Sector 3B controller. This had been annoying and was frustrating the controller. He had felt relieved when the pilot of the VFR aircraft transferred to an adjacent control position; Sector 3L.

When the crew of the B747SP transmitted to the controller requesting advice of the traffic ahead, the Sector 3B controller had been discussing aspects of the VFR flight with the Sector 3L controller. The Sector 3L controller was providing a RAS in his sector and had queried the Sector 3B controller with respect to the previous transmissions from the pilot of the VFR aircraft. The controllers had not utilised the intercom communications system to discuss the flight but had talked between the two consoles. Consequently, the Sector 3B controller's attention was diverted from his display as the horizontal separation between the B747SP and BA146 approached the minimum required.

After returning his attention to the display, the controller observed that the 5 NM radar separation standard was about to be infringed as the B747SP approached the BA146. The crew of the BA146 were unaware that the B747SP was approaching their aircraft from behind. The Sector 3B controller immediately instructed the crew of the B747SP to turn right in an endeavour to maintain separation between the two aircraft and then advised the crew that the traffic was a BA146. The crew of the B747SP complied with the Sector controller's instructions. The Sector 3B controller instructed the crew of the BA146 to maintain FL230 and then requested the crew of the B747SP to expedite climb until passing FL240. The horizontal separation reduced to 2.7 NM before vertical separation of 1,000 ft was established. There was a breakdown of separation.

ANALYSIS

The Sector 3B controller was relying on maintaining horizontal separation until vertical separation was achieved. The maintenance of adequate horizontal separation was conditional upon the regular scanning of the radar display by the controller. This was especially so because of the significant overtaking groundspeed of the B747SP.

Once the controller elected to monitor the situation instead of using a method of separation assurance, he needed to concentrate more on the radar display. However, he allowed himself to be distracted, firstly by the radio transmissions from the pilot of the VFR aircraft and secondly, by conversing with the adjacent sector controller. His concentration may also have lapsed to some degree because of the reduction in the number of aircraft on frequency in the previous 15 minutes.

Had the aircraft not been fitted with an ACAS, or had the crew not queried the controller, there was the possibility of a collision. Without an ACAS the crew may have observed the BA146 through the aircraft's windscreen; however, this would have been subject to the crew's cockpit workload. Their attention may have been inside the aircraft and consequently they would not have seen the BA146. The operation of the ACAS in the B747SP and the crew's request for traffic information to the controller were active defences in the situation.

SIGNIFICANT FACTORS

1. The Sector 3B controller did not utilise separation assurance techniques.
2. The Sector 3B controller's attention was diverted from monitoring the radar display during a critical phase.
3. The B747SP crew's use of the ACAS and subsequent request for traffic information alerted the controller to the possibility of a breakdown in separation.