The aim of the flight was to check the front seat pilot's capabilities in assessing and flying slope-lift generated by mountain slopes. The area near the mountain became unsuitable for continued operations due to mist and the presence of other aircraft. As a result, the crew decide to return to their base. The instructor reported that as the glider passed through a saddle, with sufficient height and speed to reach the airfield, it was caught in a series of downdrafts. A significant amount of altitude was lost in the turbulence and the glider continued to lose altitude rapidly once it reached clear air. The air brakes were found in the fully deployed position. They had not been selected by either pilot. The glider was 300 ft above ground level and over trees when the air brakes were selected in. The glider had insufficient performance available to reach the airfield.

The pilots decided to complete an outlanding on a road. The space between the trees on either side of the road was less than the glider's wingspan and the right wing and canopy collided with trees during the landing. The glider came to a stop in the trees at the side of the road.

It is probable the airbrakes were deployed by the loadings encountered during the turbulence.

FACTUAL INFORMATION

History of the flight

The aircraft was scheduled to complete a flight test following major repairs to the left wing and fuselage. These were carried out in Broome to repair damage resulting from a birdstrike which occurred in May 1996 (BASI report 9601590). The crew for the flight comprised two experienced pilots, one of whom was a qualified test pilot, and a flight test engineer. There were no passengers. When the landing gear was selected up, after take-off, the nose landing gear door unsafe and landing gear handle lights remained on. The gear was selected down, and safe indications were obtained. The gear was selected up a second time and it retracted normally. The test flight was continued, with sequences being completed up to the landing gear tests. Landing gear down selection resulted in a safe indication on the main gear (two green lights) and unsafe indications on the nose gear (a red gear light, no amber door light and a light in the gear handle). Actions were completed in accordance with the abnormal and emergency checklist however, the nose gear continued to indicate unsafe.

An inspection by ground engineers, during a fly-past, determined that the nose gear was still retracted but the gear doors were partially open. Positive-gravity manoeuvres and a touch-and-go landing failed to shake the nose gear free. The crew decided to land at Broome once all other traffic was clear, and emergency services were in place. Advice from the manufacturer's representative in Broome indicated that the propellers would remain clear of the runway during the landing. A landing was completed with the aircraft sliding to a stop on its main wheels and nose, 460 m after touchdown.

Nose gear locking mechanism

To release the nose gear, using either the normal or alternate systems, the over-centre up-lock initially moves further away from the neutral position before moving back through neutral to the gear extension position.

Movement of the up-lock causes a corresponding movement in the upper drag strut. Inspection determined that the nose landing gear had been jammed in the up position as a result of a lack of clearance between bolt heads securing a cover plate on the nose leg and the upper drag strut. The lack of clearance prevented the initial unlocking movement. Witness marks were found on the upper drag strut where the bolt heads had been making contact with it.

Nose leg history

The nose leg had been removed from another aircraft following a ground incident some months prior to this accident. As it was damaged, it was sent to an overseas workshop for repair. When the nose leg was returned to the operator a cover plate, which is designed to protect microswitches associated with the nosewheel steering system from dust and stone damage, was missing. The overhaul shop advised the operator to obtain a replacement. The operator believed it already had a suitable replacement in a box in its store in Adelaide and no further action was taken. The original nose leg from VH-JSI was removed, during the wing repair, and fitted to the ground incident aircraft to help restore it to flying status as soon as possible. The overhauled nose leg was fitted to JSI once the wing repair was almost complete.

Nose leg cover plate

When the box containing the replacement cover plate was delivered from Adelaide, and opened in Broome on 2 August 1996, it was found that some parts were missing. The manufacturer advised that a 36-week lead time was required to replace the missing parts. Maintenance personnel in Broome then decided to manufacture substitute parts. The operator's engineering staff, outside Broome, were not consulted. The operator's maintenance control manual required all non-approved parts to undergo a design and approval process prior to being fitted to its aircraft. A Brisbane-based contract engineer would have been responsible for this design and approval had it been requested. No action was taken to meet the design and approval requirements.

The cover plate was non-load bearing, and it was reported that it did not appear to be a significant part of the nose leg structure. The original plate was shaped to provide clearance between it and the microswitches. Because of the limited manufacturing capability in Broome, the replacement cover was manufactured from flat plate and spacers were used to provide similar clearance.

The original plate was secured using counter-sunk screws. This aspect went unnoticed by the maintenance personnel although the screws were depicted in the Dash 8 parts manual. The holes in the replacement plate were not drilled to accept counter-sunk screws and it was secured using hexagonal-head bolts and washers. The use of spacers and hexagonal-head bolts and washers meant that the tops of the bolt heads protruded 8.5 mm beyond the normal position of the counter-sunk screw heads. The measured space available between the tops of the counter-sunk screws and the upper drag strut, on a serviceable Dash 8 with its nose leg retracted, was 5 mm. The difference of 3.5 mm was sufficient to cause the landing gear to jam although hydraulic pressure was able to overcome the lack of clearance on at least one occasion after take-off.

The maintenance engineer who fitted the replacement plate to the nose leg reported that, had the holes in the plate been counter-sunk he probably would have recognised that counter-sunk screws were required. Had counter-sunk screws been used instead of hexagonal-head bolts and washers, the plate, with its spacers, would still have protruded 6.6 mm instead of 8.5 mm. That is 1.6 mm beyond the space available.

While the replacement plate was being manufactured the nose leg was refitted to the aircraft and landing gear retraction tests successfully completed. No retraction tests were attempted after the cover plate was fitted.

Organisational issues

Three significant organisational issues were identified during the investigation. These involved project management, pressure and fatigue and loyalty and motivation.

Project management

The repair of JSI was completed by two groups of engineers. The wing and fuselage repair was supervised by an engineer employed by the aircraft manufacturer and completed by an overseas contractor with assistance from the operator's staff. The reassembly and certification was completed by a licensed aircraft maintenance engineer, with Dash 8 approvals, using a mixture of the operator's staff from Broome, Perth and Darwin.

The operator's technical manager, who had ultimate responsibility for the work on JSI, was based in Adelaide and was only able to visit Broome on four occasions, the most resent being between the 12-16 May. It was reported that he departed on 16 May as the reassembly was essentially complete. He was not consulted about the plan to manufacture a replacement cover plate. A Broome-based licensed aircraft maintenance engineer was originally appointed to coordinate the repair, but he was not given any formal project management authority. This engineer did not have Dash 8 approvals, but he did have extensive experience in the maintenance and repair of large airline aircraft and was familiar with the requirement to have the correct approval for all parts fitted to transport category aircraft. The licensed aircraft maintenance engineer, with Dash 8 approvals, who assumed responsibility for reassembly and certification, was sent to Broome on temporary assignment, to provide licence cover, and arrived after the wing repair had started and the coordinating maintenance engineer had commenced his task. It was reported that this led to some conflict amongst maintenance staff. The result was that neither of the licensed aircraft maintenance engineers, or anyone else, exercised any overall responsibility for the project. At no stage was a project manager appointed to supervise the whole job. Although the maintenance personnel were aware of the maintenance control manual's requirements for non-approved parts, action was not taken to ensure that it was complied with.

Pressure and fatigue

A deadline had been set by the operator's management team for JSI to depart Broome on         18 August 1996. Reassembly and certification of JSI could not be completed until after the wing repair had been finished. Because of the complexity of the wing repair the contractor declined to give the operator a definite completion date. The wing and fuselage repair was not completed until the week commencing 12 August 1996. As a result, time to complete the whole process was limited. The nose and left main landing gear and left engine and propeller had to be reinstalled. Electrical systems, damaged during the birdstrike, had to be repaired and tested. A test flight was needed, and the aircraft had to be repainted. The local engineers believed that the time required to get the Brisbane-based engineer to complete design and approval of the replacement cover plate would probably have exceeded the limited time available.

The operator's aircraft maintenance personnel were required to work on JSI as well as complete their normal maintenance activities on other company aircraft. These processes resulted in some unusually long shifts (often 15 hours but up to 28.5 hours) and the onset of significant, observed fatigue.

Loyalty and motivation

Discussion with the operator's staff provided anecdotal evidence that there was strong company loyalty and motivation to get the job done. Staff excelled themselves in order to meet demanding deadlines. Whilst this approach is laudable, research and investigation has shown that it can lead to incorrect practices if the appropriate balance is not found.

Other issues

A number of other administrative and procedural issues were noted during the investigation. These were not directly related to this accident but provided supporting evidence that a culture aimed at "getting the job done" existed amongst the operator's staff.

ANALYSIS

It is probable that motivation, pressure and fatigue were the main factors in the decision not to complete the maintenance control manual requirements prior to manufacturing the replacement cover plate. It is likely that the maintenance engineers allowed their desire to get the job done to overcome any caution developed through experience. Limited time availability meant that ordering a replacement part or ensuring the completion of the required design and approval process, would prevent them meeting the deadline. Fatigue may have reduced the maintenance engineer's analytical ability and possibly contributed to the flawed decision.

The lack of a project manager to supervise and audit the complete engineering process meant there was no safety net to prevent incorrect decisions or practices, emanating from the workface, having a detrimental affect on safety. The lack of a project manager and the limited presence of the technical manager also meant there was no one of authority that the maintenance engineers could readily discuss the problem with before they made their final decision.

The fact that the cover plate was not a structural component and was only a protective cover probably misled the maintenance engineers into believing that the manufacture of a similar, but not exact, replacement part would not significantly affect safety.

The ultimate safety net, landing gear retraction tests, failed when the cover plate was not fitted prior to the tests. The plate was not fitted because it was not seen as a significant part of the landing gear system and time pressure probably led maintenance personnel to believe that the tests could not wait until assembly was complete.

SAFETY ACTION

The operator advised it has have taken action to address all of the management, administrative, procedural and maintenance issues noted during this investigation. Amended processes will be applied during the repair of damage sustained by JSI in this accident.

The Melbourne FSO notified Sydney FS that the Metroliner was taxiing at Dubbo and was provided with traffic information on another aircraft. The Metroliner departed Dubbo, and the crew reported their departure to Melbourne FS and broadcast a departure report on the Dubbo mandatory broadcast zone (MBZ) frequency. The Melbourne FSO notified the crew of the Metroliner the traffic information provided by Sydney FS and notified Sydney FS of the departure time of the Metroliner.

As the Metroliner climbed enroute, the crew observed another aircraft approaching overhead Dubbo from the northeast. The crew attempted to contact the pilot of the other aircraft via the MBZ and FS area frequencies. As the Metroliner passed 8,000 ft at approximately 5 NM from Dubbo the crew estimated they went through the level of the other aircraft.

The callsign of the other aircraft was established shortly after when the crew of the Metroliner overheard the pilot of the Mooney report overhead Dubbo at 8,000 ft.  The crew of the Metroliner changed to the Sydney FS frequency and confirmed the other aircraft's callsign and track details with the Sydney FSO.

ANALYSIS

The Melbourne FSO estimated the flight of the two aircraft would not be in conflict based on his experience and understanding of Metroliner aircraft performance. He considered the aircraft to be well separated despite the fact that the flight information on the flight progress strips indicated that the aircraft met the criteria for traffic in conflict detailed in the MATS.

The FSO had recently undergone a performance check and had received favourable comment in relation to his ability in assessing conflicts between aircraft. This may have led the officer to be somewhat over-confident in his abilities and possibly caused him to trust his own instincts in preference to standard procedures.

SIGNIFICANT FACTOR

The Melbourne FSO did not comply with the procedures in the Manual of Air Traffic Services for the provision of traffic information to IFR category aircraft.

Cocking of FPS is one of the means available to provide a reminder for further action by controllers. The supervising controller expected the controller being checked to cock the strip to act as a reminder to transfer the altitude to the appropriate box on the FPS. Controllers "cock" an FPS by placing it half out of the bay. In this way, the out of place FPS acts as a visual cue and requires a physical action for it to be restored to the correct location within a bay. The supervising controller did not notice that the controller had not completed the FPS correctly and that no reminder action had been taken.

The Melbourne controller instructed the crew of the A340 to contact Brisbane centre at PAVKO. The crew acknowledged the instruction but did not contact the Brisbane sector controller when the aircraft reached PAVKO.

The two Brisbane controllers did not notice that the aircraft had not called at PAVKO as co-ordinated. Approximately 15 minutes after the A340 was estimated to have reached PAVKO the controllers handed over responsibility for the sector to a new controller. Traffic was light at the time and the handover/takeover of the sector was completed without any specific mention of the A340.

After the handover/takeover, the on-coming controller expected further co-ordination on the A340 as the "altitude/flight" level box did not have a level. Also, he was only scanning the last two digits of the PAVKO estimate and did not notice that it indicated that the aircraft should have already passed the position. He believed the aircraft had a further 25 minutes to go before reaching PAVKO.

Twenty minutes later the new controller noticed that the level on the FPS for the A340 had been annotated in the "previous position" box but had not been transferred to the "altitude/flight level" box. He commenced checks to establish communications with the aircraft. The controller sought confirmation from Melbourne that the aircraft had passed Alice Springs. He was advised that the A340 had passed Alice Springs an hour ago. The controller then requested the crew of another aircraft in his sector to attempt to contact the crew of the A340. The crew of this other aircraft were able to contact the crew of the A340 by radio and, shortly after, normal air-ground-air communications were restored. There was no conflicting traffic while the aircraft was out of communication with ATC.

ANALYSIS

The supervising controller expected the controller being checked to perform to a high degree of efficiency because of the latter's long period of ATC experience. Consequently, he did not monitor the controller’s actions as closely as he would have for a less experienced controller.

The two Brisbane controllers were nearing the end of their shift, and this may have caused them to relax their vigilance in completing tasks. Consequently, neither controller completed their respective tasks in a satisfactory manner.

The handover/takeover would appear to have been conducted in a cursory manner due to the low traffic levels and the supervising controller’s expectations of the other controller’s performance. This aspect of expectation carried across to the on-coming controller who did not check the FPS. However, once the new controller recognised the error on the FPS, he was able to quickly recover the situation.

The reason for the crew of the A340 not contacting Brisbane centre at PAVKO as instructed could not be ascertained.

SIGNIFICANT FACTORS

1. The controller being checked did not annotate the FPS correctly.
2. The controller being checked did not ensure that all aspects relating to the sector were covered in the handover/takeover.
3. The supervising controller did not adequately monitor the actions of the controller being checked.
4. The crew of the A340 did not transfer to the Brisbane centre frequency at PAVKO as instructed.
5. The on-coming controller did not adequately check the flight progress strips after taking over responsibility for the sector.

Prior to the arrival of the aircraft to conduct spraying operations, the loader driver inspected the grass strip surface and informed the pilot that its condition was satisfactory for take-off and landing. The pilot landed the aircraft and subsequently inspected the condition of the strip himself, before taxiing along the centre of the strip to its western end in order to load the aircraft. He then taxied to the eastern end and commenced to take off towards the west.

The aircraft remained slightly left of the centreline of the strip during the ground roll and encountered an area of soft ground. The pilot lost directional control of the aircraft, which resulted in the left wingtip striking the ground, and the right main landing gear collapsing.

A treeline running north-south was located to the west of the strip. A gap in the treeline, slightly left of the strip centreline, was used by pilots during take-off to ensure obstacle clearance. The pilot reported that he may have been concentrating on the gap during the take-off ground roll and allowed the aircraft to veer to the left of the intended take-off path.

The helicopter was being ground run after two of the three rotor blades had been refitted. The aircraft apparently experienced ground resonance and the main rotor struck the ground. The aircraft rolled over and came to rest against a tree. The sole male occupant was admitted to hospital.

Enquiries revealed the person operating the helicopter did not hold a pilots or engineers licence. The helicopter had apparently been purchased on behalf of another person with the intention of shipping it out of the country.

Since the person who was running the helicopter was released from hospital he has been pursued by the CASA for questioning but has disappeared.

Ground running of helicopters must be performed by a licenced helicopter pilot, and if the helicopter is tied down, may be performed by a licenced engineer.

During climb, passing 5,300 ft, a loud bang was heard. The pilot initially thought the noise to be a birdstrike, however it was noticed that the inner cowl on the left engine had fallen from the aircraft. The pilot elected to return to Brisbane. Post flight inspection revealed that the cowl had torn off along the hinge line and had struck the fuselage forward of the entry door causing denting and scoring of the fuselage skin.

The aircraft had undergone maintenance on the previous day, but no work had been done on the left engine, and there was no known reason for the left engine cowl to be unfastened. The copilot, who performed the daily inspection in the early hours of the morning (it was still dark) had not noticed the cowl was not properly fastened. The cause of the cowl being unfastened could not be determined.

The incident occurred on the first day of a two-day pageant to celebrate the 75th anniversary of Essendon airport. A flying display was programmed to commence at 1300 eastern standard time, the first event being a multiple parachute drop from VH-KTD. Air traffic services planned to give priority to programmed pageant events. Before and between events, several local operators took the opportunity to conduct joy flights. Joy flight traffic was heavier than anticipated by air traffic services.

Because the wind was a moderate south westerly, air traffic services established a left-hand racetrack pattern using runway 17 for most departures and runway 26 for arrivals.

Shortly before the incident, joy flight aircraft were being recovered to clear the airspace for the paradrop aircraft awaiting clearance for take-off from runway 17. The aerodrome controller decided to require the pilot of VH-NAY, a joy flight aircraft on approach to runway 26, to hold short of the runway 17 strip after landing so he could expedite the take-off of the paradrop aircraft. As NAY was already on final for runway 26 when the controller made this decision, he decided to wait until NAY had landed and slowed to taxying speed before issuing the hold short instruction. However, in the busy traffic situation the controller forgot to issue the hold short instruction to NAY and cleared the paradrop aircraft for take-off. NAY entered the runway 17 strip (gable marker line) as KTD became airborne north of the runway 17/26 intersection.

The incident would not have occurred if operations had been confined to a single runway. However, the use of runway 17 for departures and 26 for arrivals was a standard operating procedure in suitable weather conditions. The procedure increases traffic throughput and reduces aerodrome controller loads. Use of the "hold short" requirement, once the landing aircraft has reduced to taxying speed, was common.

The Essendon tower team leader was rostered on as an extra staff member in view of the expected heavy traffic. However, because he had attended the briefing for participating pilots, and was still relaying the details of the briefing to other tower staff, he had not taken up the position of assisting the aerodrome controller before the incident occurred.

Significant factors

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

1. Joy flight traffic was heavier than anticipated.

2. The extra tower controller was still briefing other tower staff and was not assisting the aerodrome controller when the incident occurred.

3. There was a degree of pressure on the aerodrome controller to ensure that joy flight aircraft were on the ground prior to the pageant commencing.

FACTUAL INFORMATION

An Australian registered B747 aircraft had departed Tokyo on a flight to Sydney. The crew had been radar-vectored by air traffic control around an active military restricted area (R116) and were commencing a left turn to intercept their planned route in accordance with the instructions received. As the aircraft climbed through flight level (FL) 220, the crew received a traffic alerting and collision avoidance system (TCAS) traffic advisory warning of unidentified aircraft crossing from right to left and climbing from a level beneath that of the B747.

A pair of U.S. military FA-18 aircraft had departed from the Atsugi Naval Air Facility to operate in R116, which was located just east of the air route which the B747 was intercepting. Initially, the formation was proceeding under instrument flight rules but changed to visual flight rules on passing 9,000 ft, and tracked visually to the exercise area. Their rate of climb at this time was approximately 6,000 ft/min.

After receiving the TCAS traffic advisory, the crew of the B747 saw the military formation and, as the formation had changed course to be in conflict with the predicted track of the B747, they elected to increase the rate of climb and continue the left turn, to ensure separation from the FA-18s. During this manoeuvre, the rate of climb of the B747 reached 5,000 ft/min, higher than normal for this stage of flight. No traffic information was given by air traffic control in relation to the FA-18 aircraft.

While the crew of the B747 were carrying out their evasive action, the crews of the FA-18s had sighted the B747 and commenced a level-off manoeuvre to maintain FL 225. They calculated that their track to R116 would have conflicted with that of the B747, and assessed that as that aircraft had left FL 225, vertical separation would be achieved by the time horizontal separation was lost.

Radar analysis indicates that the aircraft passed with a minimum of 1,200 ft vertical difference while there was no horizontal separation. Because the appropriate standard is a minimum of 1,000 ft, no breakdown of separation occurred.

ANALYSIS

The FA-18 pilots were operating in accordance with the "see and avoid" principle of visual flight. They acted in accordance with the rules for aircraft transiting to or from the restricted area and maintained a level beneath that already vacated by the B747.

The crew of the B747 reacted to the visual sighting of military aircraft closing rapidly on their aircraft and on a track that was in conflict with their intended flight path. The TCAS equipment confirmed that the rate of closure and track change of the FA-18 formation would place the aircraft in imminent conflict. Even though there was no resolution advisory, the B747 crew elected to commence an evasive manoeuvre based on their observations. They had no information on the intent of the military crews and did not know of the decision of those pilots to level off below their aircraft.

The air traffic controllers were aware of the visual flight procedures used by the military formation and only had a requirement to pass traffic information on such flights if their other duties allowed. In this case, the controller was too busy to pass traffic to the Australian crew. This crossing point is a particularly busy section of airspace and it is normal for traffic information not to be given as military crews are required to remain clear of civil aircraft.

SAFETY ACTION

The Japanese Civil Aviation Bureau advised the military authorities to remind pilots of the airspace structure and of their requirement to remain well clear of civil traffic.

The Bureau requested that the operator fully brief crews on airspace requirements and military crossing points, and how TCAS advisories can be expected in this environment.

A group of pilots were undertaking endorsement training on the aircraft. They had been briefed on ground handling and starting procedures for the aircraft. The starting procedure involved turning the propeller through with full throttle and ignition off, in order to pass fuel mixture to the cylinders. After that process the throttle was set to be set to 1/4 inch open and the ignition switches turned on for hand starting.

On this occasion the instructor was outside the aircraft to hand swing the propeller and the pilot undergoing the endorsement was in the cockpit. The throttle had been set full open for the priming procedure and the instructor had asked for it to be set to 1/4 inch open. At this time the throttle was moved to a position which the trainee assumed was the appropriate setting. When the engine started it went quickly to a high rpm condition, the tail rose, and the propeller was destroyed by contact with the ground.

The trainee had not closed the throttle initially in order to obtain the correct datum from which to set the throttle for start. He later found that the throttle control movement was much greater than had been expected. In his previous experience, with nose-wheeled aircraft, a high rpm after start was of no concern from an aircraft attitude aspect, however with a tail-wheeled aircraft the tail can rise unexpectedly.

Factors

1. The trainee pilot was unfamiliar with the engine control movement;

2. The starting throttle position was not set from the closed throttle position;

3. The throttle was set too far open for start;

4. The flight controls were not set to help keep the tail low;

5. Timely action to reduce engine rpm was not taken;

6. The instructor did not check that the throttle and flight controls were properly set for start.