Sequence of events

At 1940 eastern standard time (EST), on 4 October 2004, the Boeing Company 737-300 aircraft, registered ZK-FDM, became bogged (Figure 1) at Melbourne Airport. The aircraft was being taxied to runway 27 for a scheduled international passenger service to Hamilton, New Zealand.

Figure 1: Left main landing gear tyres after penetrating the blast-protection surface adjacent to taxiway Papa.

The flight crew commenced taxiing the aircraft to runway 27 from bay 'Delta 4' at the international terminal. Taxiing to Papa¹ via Tango was not available because of the stage A2a works in progress on the northern apron (depicted at Attachment 1), and the surface movement controller (SMC) instructed the crew to taxi the aircraft via Sierra and Echo to the runway 27 holding point on Papa. As the aircraft approached Sierra the crew requested confirmation of the taxiing instructions, and the SMC repeated those previously given. The SMC then told the crew to contact the aerodrome controller (ADC) on frequency 120.5 MHz.

The flight crew subsequently reported that while taxiing east along Echo, their attention had been drawn to the lighting associated with the apron works near the intersection of Tango and Papa. There was a holding bay located north of Echo, and the first lead-in light to the holding bay was about 40 m west of the first lead-in light to Papa. As the aircraft approached the lead-in lights to the holding bay, the pilot in command stopped the aircraft. The copilot asked the ADC "is it hard left Papa here?"

The ADC was located in the aerodrome control tower, which was about 1,450 m to the west-southwest of the aircraft's position. From the ADC's vantage point, and in the night conditions, the aircraft would have appeared to have been at the lead-in to Papa, and the ADC confirmed to the crew that it was Papa.

The pilot in command, in response to the advice from the ADC, turned the aircraft left and began to enter the holding bay, thinking it was taxiway Papa. It then became obvious to the crew that they were not on Papa, and the pilot in command turned the aircraft to the right to regain taxiway Echo and the lead-in to Papa. He subsequently reported that another aircraft taxiing on Quebec for runway 27 had its lights 'very brightly shining', and that he misidentified the double lines on the taxiway shoulder as being the taxiway centreline. As a result, the pilot in command inadvertently steered the aircraft onto the blast-protection surface adjacent to the taxiway shoulder pavement. The left main landing gear tyres of the aircraft penetrated the blast-protection surface, and the aircraft became bogged.

The aircraft could not be moved from its bogged position, and the 104 aircraft occupants, which comprised two technical crew, three cabin crew and 99 passengers, were disembarked and taken back to the international terminal.

The aerodrome operator reported that all airside signs, markings and lights complied with ICAO standards and CASA regulations, and that they were serviceable and operating normally at the time of the occurrence. The circumstances of the occurrence did not suggest that there were any deficiencies in the aerodrome airside signs, markings and lighting systems.

The aircraft weight and balance data for the planned flight revealed that its taxi weight was 54,434 kg, and the centre of gravity was at 20.4 percent mean aerodynamic chord. Under those conditions, about 90 percent (49,500 kg) of the aircraft's total landing gear pavement load was exerted by both the main landing gears, that is, each main landing gear pavement load was about 24,750 kg. The left main landing gear pavement load was of sufficient magnitude to result in the left main landing gear tyres penetrating the blast protection surface.

The aircraft was recovered from its bogged position on the following day and subjected to a heavy landing engineering inspection. The inspection revealed no damage resulted from the incident, and the aircraft was returned to service.

ANALYSIS

The circumstances of the accident were consistent with a loss of tail rotor thrust following the failure of the tail rotor drive shaft as the helicopter entered the hover.

The time at which the damage to the drive shaft occurred was not able to be determined. However, given the absence of rotational damage to the tail rotor blades, it is unlikely that it occurred during the accident flight.

The action of the pilot in increasing engine power when faced with the loss of tail rotor thrust was inappropriate and exacerbated the situation.

FACTUAL INFORMATION

At 1215 Eastern Standard Time¹ on 27 September 2004, a Kawasaki Heavy Industries, 47G3B-KH4² helicopter, registered VH-MTF, was being operated on a tourist flight with one adult and a young boy as passengers. The flight included landing on a 1 m high wooden platform in the Teepookana Forest in north-west Tasmania.

The pilot reported that as he brought the helicopter to a 1 m hover above the platform, the helicopter began to rotate slowly to the right. He unsuccessfully attempted to counter the rotation by applying left tail rotor control input. The pilot then increased engine power in an attempt to regain tail rotor control and to move the helicopter clear of the landing platform. That action had the effect of rapidly increasing the rotation of the helicopter to the right and it began to ascend, reaching about 5 m above ground level. The pilot then lowered the collective control and the helicopter impacted the ground heavily on its right side, several metres from the landing platform. The pilot and adult passenger released their seatbelts and then both assisted the young boy to exit the wreckage. The pilot and passengers received minor injuries.

The pilot described the wind conditions at the time of the accident as a headwind with an approximate strength of 8 kts. That assessment was consistent with the wind data for the Strahan area provided by the Bureau of Meteorology³. The pilot also reported that the main rotor RPM indications were normal and that the helicopter had sufficient power to complete the approach⁴. At the time of the accident, the weight and balance of the helicopter were within prescribed limits. There was no evidence that the helicopter had collided with anything during the approach.

The pilot was appropriately qualified and endorsed to operate the helicopter type and held a valid medical certificate. He was a very experienced agricultural aeroplane pilot and had obtained a commercial pilot (helicopter) licence 14 months before the accident. He had accrued at total of 292 hours in helicopters since that time; 286.4 hours of which had been in the Bell 47 helicopter type. The pilot was experienced with operations into and out of the Teepookana Forest landing platform.

The landing platform was located within a dense forest in an area that was cleared of trees but covered by 1 m high scrub. The trees closest to the clearing had been trimmed to a height of about 5 m to allow a 'fly-in, fly-out' approach. There was no requirement to conduct a vertical approach to the platform.

The helicopter's fuselage structure was deformed by the impact and the tail boom was bent in a downward direction at approximately station 100⁵. There was corresponding bending damage to the tail rotor drive shaft assembly long shaft at the same point. The operator reported that examination of the damaged tail rotor pitch control system revealed that the controls were intact and would have been capable of normal operation. All parts of the helicopter were accounted for by the operator at the accident site.

The two-blade tail rotor assembly, mounted on the right side of the tail boom, was intact and correctly attached to the helicopter. There was no evidence of rotational damage to the leading edges or tips of either blade (Figure 1). During the ground impact one blade had been bent outward at the tip and the other was bent in toward the tail rotor gearbox.

Figure 1:     Tail rotor blade damage

The helicopter's tail rotor drive shaft assembly consisted of a series of two short shafts and one long shaft that were situated on the top of the tail boom assembly. The long shaft was supported in eight hanger bearing assemblies and was secured at its front and rear by drive coupling assemblies. The operator inspected the tail rotor drive system and found that the long shaft assembly tubing was fractured and the pin situated through the front drive coupling assembly was sheared. There was also significant distortion of the corresponding pin in the shaft's rear coupling.

Inspection of the tail rotor drive system, including the drive shaft bearings, tail rotor extension housing and tail rotor gearbox, with the exception of the long drive shaft, revealed nothing that would have prevented normal operation.

ATSB specialist examination of the failed components (Appendix A) attributed the tail rotor drive shaft failure to a significant torsional overload event, leading to a loss of coupling security and the subsequent slippage, frictional heating and shear fracture of the shaft. That examination was unable to determine when the torsional overload occurred or what specific events may have contributed to it.

At the time of the accident, the helicopter had logged 72 flight hours since the issue of the current maintenance release. The last recorded maintenance carried out on the helicopter was a spark plug change on 21 September 2004, 1.0 flight hour prior to the accident. On 31 August 2004, 5.7 flight hours prior to the accident, one tail rotor blade was replaced because of delamination of the leading edge wear strip.

Information received from the operator and from the maintenance organisation indicated that there had been no known tail rotor strike or sudden rotor stoppage since the helicopter was placed on the Australian aircraft register in 1992. The helicopter's prior history was not examined.

The company operations manual contained the published normal and emergency procedures affecting aircraft operations. An appendix to the manual contained the flight check systems and operating procedures specific to each aircraft type operated by the company, with the exception of the Kawasaki-Bell 47G3B-KH4 helicopter. The company did however, make available to pilots a copy of the Civil Aviation Safety Authority approved Kawasaki-Bell 47G3B-KH4 helicopter flight manual.

With reference to tail rotor failures, that flight manual stipulated:

1. Immediately execute an autorotative descent and maintain an airspeed of 34 KIAS at least.
2. Execute a normal autorotative descent and landing.

The flight manual did not contain any specific advice for pilots in response to a tail rotor drive failure when hovering.

Information in the company operations manual regarding pilot response to a tail rotor drive failure in another piston-engine helicopter (Robinson R44) included:

LOSS OF TAIL ROTOR THRUST DURING HOVER

1. Failure is usually indicated by right yaw which cannot be stopped by applying left pedal.
2. Immediately roll throttle off into detent spring and allow aircraft to settle.
3. Raise collective just before touchdown to cushion landing

The generally accepted procedure for pilot actions in the event of a tail rotor failure is to quickly roll off the throttle or snap close the throttle and perform a hovering autorotation⁶,⁷,⁸,⁹ For example:

The likely worst place for loss of tail rotor thrust to happen is in the hover, and the reaction is quite simple - get rid of the engine power and land the helicopter from a hovering engine failure condition. Easy to do on those machines that have throttle(s) on the collective¹⁰.

1. The 24-hour clock is used in this report to describe the local time of day, Eastern Standard Time (EST), as particular events occurred. Eastern Standard Time was Coordinated Universal Time (UTC) + 10 hours.
2. The Kawasaki Heavy Industries, 47G3B-KH4 helicopter is a single pilot/single flight control helicopter manufactured under licence from Bell Helicopters. It is commonly known as the KH4 helicopter and is a derivative of the Bell 47.
3. Given that the pilot positioned the helicopter into wind during the approach and landing, the risk of loss of tail rotor effectiveness (LTE) was negligible.
4. There were no external conditions that would have placed the pilot at risk of overpitching or drooping the main rotor.
5. Positioned 100 inches aft of the datum. The datum was located 2 inches forward of the rotor mast centre-line.
6. Coyle, S. (2003). Cyclic & collective - More art and science of flying helicopters. Mojave, CA: Helobooks, pages 341and 342.
7. Federal Aviation Administration. (2000). Rotorcraft flying handbook (FAA-H-8083-21).  Washington, DC: FAA.
8. Newman, R. (1999). Helicopters will take you anywhere: A manual for helicopter pilots. Mentone, Vic: The Helicopter Book Company.
9. Becker, M. (1997). Mike Becker's helicopter handbook. Noosaville, QLD: Becker Helicopters Australia.
10. The Kawasaki-Bell 47G3B-KH4 helicopter had a throttle of this design.

At 1215 Eastern Standard Time on 27 September 2004, the pilot of a Kawasaki Heavy Industries, 47G3B-KH4 helicopter, registered VH-MTF, was being operated on a tourist flight with two passengers in north-west Tasmania. The pilot reported that as he brought the helicopter to a 1 m hover above the raised landing platform, the helicopter began to rotate slowly to the right. The pilot unsuccessfully attempted to counter the rotation by applying left tail rotor control input. The pilot then increased engine power, however, that action had the effect of rapidly increasing the rotation of the helicopter to the right and the helicopter climbed to about 5 m above the ground. After the pilot lowered the collective control, the helicopter impacted the ground heavily on its right side. The pilot and passengers received minor injuries.

The helicopter's tail rotor drive shaft had failed during the occurrence. ATSB specialist examination of the failed drive shaft, attributed the failure to damage from a significant torsional overload event, leading to the shear fracture of the shaft. The examination was unable to determine when the torsional overload occurred, however, examination of the wreckage indicated that it was likely that it had occurred prior to this accident

Information received from the operator and from the maintenance organisation indicated that there had been no known tail rotor strike or sudden rotor stoppage since the helicopter was placed on the Australian aircraft register in 1992. The helicopter's history prior to that time was not examined.

The action of the pilot in increasing engine power when faced with the loss of tail rotor thrust was also examined.

ANALYSIS

The investigation found that there was no evidence of a pre-existing defect in the helicopter that may have contributed to the occurrence, nor was there any evidence of a medical condition that could have affected the pilot's ability to control the helicopter. Consequently, the investigation concluded that in the prevailing environmental conditions, the accident was consistent with pilot spatial disorientation. This analysis examines the development of the occurrence and highlights a significant risk associated with night VFR operations.

The pilot departed for Eurella homestead 6 minutes after civil twilight in moonless, overcast, and probably showery conditions that were likely to restrict visibility to less than the required 5 km. Except for the homestead lights, the ground lighting or celestial illumination required by the Aircraft Flight Manual was not available. Although the pilot had flown at night on two recent occasions (23 and 26 August 2004), those flights did not fully satisfy the night VFR recency requirements and were probably over a well lit area. Given the pilot's limited recent and overall helicopter night flying experience, and the forecast weather conditions, it is unlikely that the pilot planned to conduct the return flight at night. The pilot had probably used the helicopter to move cattle and that task may have taken longer than expected. The proximity of the homestead, the local knowledge of his passenger, the night VFR capability of the aircraft and access to GPS information may have influenced the pilot to attempt the return flight.

The track information recovered from the hand-held GPS showed manoeuvring after 1830 that suggests that the pilot, probably using GPS information, made several attempts to track to the homestead, but was unable to do so. It is likely that during the manoeuvring the pilot was at a low altitude, attempting to maintain visual contact with surface features, possibly with the assistance of the landing lights. Such visual contact would have enabled control of the helicopter and clearance from terrain. In the absence of a consistently discernable horizon, any visual contact with the homestead lights would not have enabled the pilot to determine the helicopter's attitude. Prior to the impact, the pilot may have lost visual contact with the surface due to cloud and/or rain and become spatially disorientated.

The pilot may have attempted to control the helicopter by reference to the flight instruments. However, he had not logged any instrument flight time in a helicopter and had not been exposed to significant night-flight away from metropolitan areas. The relative instability of the helicopter and the different operating environment meant that the pilot's considerable aeroplane night and instrument flight experience was not directly transferable to night VFR helicopter operations. Consequently, spatial disorientation could have developed rapidly.

Flying the helicopter at a low altitude at night with cloud and/or showers in an area with little lighting was a very demanding task with little margin for error. However, once the helicopter became airborne after civil twilight, there were few options available to the pilot. The pilot's lack of helicopter instrument flight experience would probably have precluded consideration of climbing to the lowest safe altitude and tracking to an aerodrome with an instrument approach. Given that the adverse weather was widespread, diversion to another location while maintaining external visual reference was also an unlikely option.

A landing at a location other than the homestead was an option. It is possible that the accident occurred when the pilot became spatially disorientated in the adverse conditions while attempting to land the helicopter. However, it is also possible that, unable to communicate with the homestead, the pilot avoided an out-landing due to the consequent difficulty in reaching the homestead without transport.

Illumination of the clutch light as indicated by the stretched filament may have resulted from clutch operation during flight or from disruption during the impact. If the clutch light had illuminated during flight it may have distracted the pilot and contributed to spatial disorientation.

As a result of the separated ELT antenna cable, the search and rescue effort was significantly affected. However, in this case, the nature of the impact and the extent of injury to the occupants indicated that the search and rescue effort would not have influenced their survivability.

The circumstances of this occurrence highlight the risk of spatial disorientation during night VFR operations and reinforce the significance of the cautions included in the helicopter manufacturer's safety notices SN-18 and SN-26.

SIGNIFICANT FACTOR

The pilot departed after civil twilight in conditions where a natural horizon was probably not discernible and consistent visual reference to surface features was not likely.

FACTUAL INFORMATION

History of the flight

On 8 September 2004, the owner/pilot of a Robinson Helicopter Company R44 Raven II helicopter, registered VH-JWX, conducted a private flight under the visual flight rules (VFR) from Coffs Harbour, NSW to Eurella Station, Qld. The flight included a landing at Roma, Qld where the pilot refuelled the helicopter with 180 L of Avgas from the bulk underground fuel storage supply.¹ The pilot then continued to Eurella Station, located approximately 54 km west of Roma, arriving at 1705 Eastern Standard Time. The pilot shut down the engine and the property owner boarded the helicopter for a pre-arranged local flight. The pilot made several attempts to start the engine, during which it backfired a few times. Once started, the engine seemed to function normally.

The helicopter departed the homestead at 1725 in a northerly direction. A person on an adjoining property about 7 km north of Eurella homestead saw the helicopter operating to the east late in the afternoon. He reported that the helicopter conducted a number of take-offs and landings in what appeared to be the same general area over a period of about 30 minutes. He saw the helicopter depart in a southerly direction at about 1830.

The next reported sighting was by a person at Eurella homestead who, in poor light conditions, saw what appeared to be the helicopter's landing light to the north of the homestead. The light moved toward the west of the homestead. Soon after, that person again saw the light to the west and expected the helicopter to land at the homestead within a few minutes. However, she became concerned when the helicopter did not arrive and telephoned an employee of the property owner to report her concern. The employee contacted the Australian Search and Rescue organisation (AusSAR) and search action was initiated. The helicopter was located the following morning in open, rolling country, 3 km west of Eurella homestead. The two occupants were fatally injured, and the helicopter was destroyed.

Search and rescue

AusSAR reported that it was notified at 1947 that the helicopter was overdue. Weather conditions were unsuitable for an air search, but a surface search was initiated. AusSAR advised that no ELT signal was received on 8 September by satellite or by aircraft at high altitude passing within 130 km of Eurella Station. An ELT signal was detected on two satellite passes early on the morning of 9 September. The signals were identified as originating from separate locations; one approximately 22 km to the south-west, and the other approximately 22 km to the south-east, of Eurella Station. However, those signals were not merged by the satellites as coming from the same source, so they were of little assistance in the search. Local aircraft were tasked to begin a search at daylight on 9 September and the wreckage was located at 0708 by the crew of a search aircraft. Accident site information

The accident site elevation was about 30 m below the ground elevation at the homestead. The homestead was not visible from the accident site.

Figure 1: Aerial view of the accident site

GPS track information

The helicopter was fitted with a fixed global positioning system (GPS) receiver, and also a handheld GPS receiver mounted in a cradle on the instrument panel. The fixed receiver did not contain a non-volatile memory card, but the handheld unit did. Track and ground-speed data for the occurrence flight was retrieved from the non-volatile memory card. Altitude information was not retained in the memory card.

Figure 2 displays the GPS recorded track of the helicopter overlaid in blue on a 1 in 250,000 scale topographical map of the area. The local times that the helicopter was at various locations are depicted.

Figure 2: GPS track overlay, with the landing sites A, B, C and D

The GPS data showed that the helicopter landed five times during the flight. Those positions are depicted on the map and are described as follows:

The data indicated that, after the helicopter departed position B at 1827, it initially tracked almost directly toward the homestead, but that the track then veered south-west. That track was clear of the high ground indicated by the 400 m contour near Mt Muttaby, as depicted on the chart at Figure 2. There are distinct features in the helicopter's track after 1830, indicating that the pilot turned toward the homestead on four separate occasions between 1830 and 1840, only to turn away each time. The accident occurred on the fifth occasion that the helicopter's recorded track turned in the approximate direction of the homestead.

Subsequent to the occurrence, an employee from Eurella Station found that cattle had been moved from the paddock that included positions A, C, and D, to an adjoining paddock. Those paddocks were linked by a gate adjacent to position B. The employee recalled that the property owner had intended to move the cattle to the adjoining paddock and that the gate adjacent to position B was the gate through which he would have expected the cattle to be moved.

Pilot information

The pilot held an air transport (aeroplane) pilot licence and a command multi-engine instrument rating. He had extensive aeroplane flying experience, including regular public transport turbo-jet aircraft and corporate turbo-jet aircraft operations in Australia and overseas. His aeroplane flying experience exceeded 10,000 hours and included 1,418 hours of night flight and 711 hours of instrument flight.

The pilot obtained a private pilot (helicopter) licence on 23 September 1998 and had about 582 hours helicopter experience. He obtained a night VFR (helicopter) rating on 12 September 2000 and since that date had recorded about 11 hours helicopter night flight. Almost all of the logged flights were in the Sydney metropolitan area. Helicopter night flying recorded by the pilot in the two years prior to the occurrence was 0.4 hours on 23 August 2004 and 0.6 hours on 26 August 2004. That night flying most likely occurred during the latter stages of flights to the Sydney metropolitan area.

There was no record of the pilot having received any specific training in operating helicopters in remote areas or dark night conditions where there was little or no ambient lighting. No helicopter instrument flight time was logged.

The pilot held a valid medical certificate. Post-mortem and toxicology examinations did not reveal any pre-existing condition that might have affected the pilot's ability to safely conduct the flight.

Helicopter information

The pilot purchased the helicopter new in early August 2004. At the time of the occurrence the helicopter had operated for 34.1 hours. The maintenance release was valid and the documentation indicated that all applicable maintenance and regulatory requirements had been met.

The helicopter was equipped and certified for night VFR operations. Instrumentation included an airspeed indicator, artificial horizon, sensitive pressure altimeter, turn coordinator, horizontal situation indicator, global positioning system indicator, and vertical speed indicator.

The helicopter was equipped with twin landing lights in the lower nose section. The lights were fitted with 100 watt spot globes and, according to the Aircraft Flight Manual, were 'set at different angles to increase the pilot's field of vision'. Both lights were activated by the one switch which was mounted on the cyclic control centre post.

A row of eight amber warning lights located at the top of the flight instrument panel included a clutch warning light. A further six warning lights were positioned at the top of the centre pedestal.

The helicopter's engine was coupled to the rotor drive system via four double-stranded vee-belts. After engine start, an electric actuator would tension the belts when the pilot engaged the clutch switch. The actuator sensed belt tension and was automatically energised when the tension was less than required. The clutch warning light would illuminate whenever the clutch actuator circuit was activated. The Aircraft Flight Manual included a note regarding the clutch in Section 3, Emergency Procedures. The note stated that stretching of the belts often resulted in illumination of the clutch warning light for brief periods as the drive actuator readjusted belt tension. The note also included actions that the pilot should take after 7 or 8 seconds of illumination of the clutch light. One of those actions was to pull the clutch circuit breaker.

The helicopter was fitted with a Pointer (TSO-C91A) Model 3000-10 emergency locator transmitter (ELT). The unit was located on the left side of the rear fuselage.

The total flight time from Roma until the time of the occurrence was about 1 hour 35 minutes. Assuming a fuel usage rate of 60 L per hour, approximately 95 L would have been consumed during that time. On that basis, approximately 95 L should have remained at the time of the occurrence.

Wreckage information

opposite to the helicopter's direction of travel at impact. The impact severely crushed most of the cabin area and deformed the fuselage and tail boom structures.
Two distinct main rotor blade impact marks on the ground forward and to the right of the initial nose impact position, and the damage to the main rotor blades, indicated that the rotor blades were being driven by the engine at impact. The tail-rotor system was intact and there was no evidence that the fuselage was yawing at impact. There was no indication that the helicopter had struck any of the trees in the vicinity of the impact site.

The left fuel tank ruptured during the impact sequence and was empty. The right fuel tank was also empty. With the helicopter lying on its left side, the right fuel tank vent line was at the lowest part of the tank and would have allowed fuel to drain out. There was a strong smell of Avgas in the vicinity of the wreckage on the day after the accident.

The hydraulic system switch was found in the ON position.

Instrument panel light globe and instrument examination confirmed that electrical power was available to the instruments. There was no evidence of malfunction of any of the instruments.

The six warning lights at the top of the centre pedestal were destroyed by impact forces, preventing an assessment being made of their status at impact. The eight warning lights at the top of the flight instrument panel were intact. Examination of those light globes revealed stretching of the clutch warning light filament. Stretching indicates that the filament was hot and that electrical power was applied to the globe when it was subject to forces during the impact sequence. It was not possible to determine the length of time that the globe had been illuminated. Filament stretch was not evident in any of the other seven warning light globes from the top of the instrument panel.

Damage to the landing light globes prevented any assessment being made regarding their status at the time of the occurrence. The damage to the landing light switch indicated that it was in the ON position at impact.

The circuit breaker panel was destroyed by impact forces. The clutch actuator fuse was serviceable. The wreckage examination did not reveal any fault in the clutch system. Although the circuit breaker panel was destroyed, the evidence of electrical power to the clutch warning light indicates that the circuit breaker was engaged, and therefore the system was powered at the time.

The coaxial cable from the ELT unit to the external antenna had separated at the connector to the antenna base on the inside of the antenna mounting panel. The separation of the coaxial cable trapped the transmitted signal within the fuselage compartment. That rendered the ELT unit ineffective and prevented satellite detection of the signal. The separation of the cable appeared to have been as a result of impact forces. As a result, the search and rescue effort was significantly affected.

Specialist examination of the ELT revealed that it had activated upon impact and, when connected to a suitable antenna, was capable of transmitting a normal signal.

The engine was test run after removal from the wreckage and operated normally. The hydraulic pump and three hydraulic servos that formed part of the main rotor flight control system were removed from the wreckage for functional testing. The tests were conducted at the helicopter manufacturer's facility in the USA and supervised on behalf of the ATSB by a representative from the US National Transportation Safety Board. The tests confirmed that the hydraulic system components met the specifications for normal operation.

Meteorological information

Documents found in the helicopter included an Area 41 weather forecast valid from 0900 to 2100 on the day of the occurrence and the Roma terminal area forecast (TAF) valid from 1200 to 2400 on the day of the occurrence.

The area forecast indicated that the weather in the vicinity of Eurella Station would include areas of rain with locally moderate falls, scattered showers and isolated thunderstorms. The Roma TAF indicated that between 1500 and 2400 there would be 60 minute periods in which the visibility would be 2 km in heavy rain, with broken cloud² at 700 ft.

An analysis by the Bureau of Meteorology indicated that during the late afternoon on the day of the occurrence, a surface trough was located from Camooweal to St George, with cold south-west winds to its west and northerlies to its east. The surface trough combined with an upper level trough over the southwest of the state to bring a large cloud band with widespread rain and isolated thunderstorms to the interior. The analysis of satellite imagery and synoptic reports, concluded that there was a high probability of rain in the Eurella Station area around the time of the occurrence, and most likely greater than 5 oktas of cloud cover. However, because the nearest weather radar station was about 200 km distant at Charleville, the amount of cloud cover in the area of the occurrence could not be confirmed.

Persons at and near Eurella Station variously reported that the weather conditions during the day of the accident were windy, with heavy cloud and showers.

Astronomical information

According to information published on the Geoscience Australia website, sunset and twilight times at Eurella Station on the day of the occurrence were:

Sunset 1757³
Civil Twilight 1820⁴

Other information on the website indicated that the moon set at 1409 and was 79 degrees 31 seconds below the horizon at 1830 that evening.

Helicopter night VFR

The pilot's night VFR (helicopter) rating authorised him to act as pilot in command of private or aerial work flights at night under the VFR. Once issued, a night VFR rating remained permanently valid. To exercise the privileges of the rating, a pilot needed to complete a 1-hour night flight during the previous 12 months and one take-off and landing at night during the previous 6 months. There was no requirement for the holder of a night VFR rating to have any recent instrument flight time prior to conducting a flight at night.

A pilot operating under the VFR at night was required to operate in visual meteorological conditions that included a minimum of 5 km visibility. The Aircraft Flight Manual, Section 2, Limitations, included the following statements:

VFR operation at night is permitted when landing, instrument, and anti-collision lights are operational. Orientation during night flight must be maintained by visual reference to ground objects illuminated solely by lights on the ground or adequate celestial illumination.

At the time of the occurrence there was a 1,000 watt flood light on each of the northern and western walls of Eurella homestead, as well as lights in other buildings. However, there were many trees in the vicinity of the homestead, some of which were higher than the homestead roof. Depending on the altitude and position of the helicopter, the trees could have prevented those lights being seen from the helicopter (Figure 1). There was no other lighting in the general area, including at the airstrip adjacent to the homestead. The homestead lights, in effect, formed a 'point' source of light.

Spatial Disorientation

Spatial disorientation refers to a situation in flight in which the pilot fails to sense correctly the position, motion or attitude of the aircraft. When the condition is fully developed, the pilot is unable to tell which way is 'up'.

The risks of non-instrument rated pilots flying in conditions in which they are not able to orientate the aircraft by visual reference have been well known for over 50 years. During testing conducted on a group of non-instrument rated pilots, the average time before loss of control of the aeroplane, after visual reference was lost, was 178 seconds.⁵

US FAA Advisory Circular 60-4A, Pilot's Spatial Disorientation, was published in 1983 and was intended to inform pilots of the hazards associated with disorientation caused by loss of visual reference with the external environment. It included the following information:

Tests conducted with qualified instrument pilots indicate that it can take as much as 35 seconds to establish full control by instruments after the loss of visual reference with the surface.

The helicopter manufacturer issued a safety alert and safety notices (SN) as a result of various occurrences and incidents, and included those notices in the Aircraft Flight Manual Section 10, Safety Tips. Two of the notices related to night flight - SN-18 Loss of Visibility Can Be Fatal, and SN-26 Night Flight Plus Bad Weather Can Be Deadly (see Appendix A). Safety notice SN-18 stated in part:

Helicopters have less inherent stability and much faster roll and pitch rates than airplanes. Loss of the pilot's outside visual references, even for a moment, can result in disorientation, wrong control inputs, and an uncontrolled crash.

Appendix A

On 8 September 2004, the owner/pilot of a Robinson Helicopter Company R44 Raven II helicopter, registered VH-JWX, conducted a private flight under the visual flight rules (VFR) from Coffs Harbour, NSW to Eurella Station, Qld. The flight included a landing at Roma, Qld where the pilot refuelled the helicopter with 180 L of Avgas from the bulk underground fuel storage supply. The pilot then continued to Eurella Station, located approximately 54 km west of Roma, arriving at 1705 Eastern Standard Time. The pilot shut down the engine and the property owner boarded the helicopter for a pre-arranged local flight. The pilot made several attempts to start the engine, during which it backfired a few times. Once started, the engine seemed to function normally.

No defects were found with the cabin door mechanical locking system. Prior to take-off, the locking mechanism of the door was checked by the safety pilot and no problems were reported. When the separated door was located, the external handle was not in the locked position and the latch bolts were not fully extended.

It was possible to move the door locking handle to a high resistance position, where the green line indicators on the latch bolts appeared to line up with the stationary arrow heads in the inspection windows, but the door locking mechanism was not fully engaged. At that position, if the top latch hook sense switch was short circuited due to the proximity of the sense switch terminal and the mounting nut, the door warning light would be extinguished. This condition would provide the crew with no indication that the door was not correctly locked.

Factual Information

The PIC reported that, while passing Flight Level 175 on descent, he heard a loud muffled thud and then air noise, and saw that the cabin door had opened and separated from the aircraft. There had been no door warning indication prior to the door opening.

Following the door separation, the cabin door warning light illuminated and the cabin oxygen masks deployed. The PIC completed emergency checks before diverting the aircraft to Thangool, Queensland.

On 12 September 2004, after a search by the aircraft operator, the cabin door was found approximately 3 NM north of Glenbar, Queensland (see Figure 1). When the door was found, the external handle was not in the locked position and the latch bolts were not fully extended.

Figure 1: Location of recovered door

Aircraft cabin door

The aircraft's cabin door was an outward-opening airstair type, hinged at the bottom, with a diaphragm and mechanical over-centre locking mechanism. It was also fitted with a pressurised seal.

The cabin door mechanical locking system consisted of four latch bolts. Two were located on the rear side of the door and two on the forward side of the door. In addition, two latch hooks located on the top of the door engaged onto stationary locking pins located within the fuselage (see Figure 2).

Figure 2: B300 cabin door locking mechanism

Locking and unlocking of the cabin door was accomplished by means of either the external or internal door handles. Once the aircraft was pressurised, the door's pressure seal inflated. As the pressure differential between the interior and exterior of the aircraft increased, the cabin door's internal diaphragm expanded and locked the interior handle to prevent inadvertent opening in-flight. The cabin door interior trim had three positions marked for the door handle, OPEN, LATCH and LOCK (see Figure 3).

Figure 3: Cabin door interior handle in lock position

When the locking mechanism was in the LOCK position, the door handle could not be moved to the LATCH or OPEN position until a release button on the internal diaphragm was depressed.

Each forward and rear latch bolt was marked with green line indicators, which aligned with stationary arrow heads that indicated that the cabin door was locked. The green line indicators were viewed through inspection windows at four locations on the inside of the door.

The recovered door was sent to the Australian Transport Safety Bureau for further examination. The investigation found that the mechanical latching mechanism was intact, however the door's rear latch hook sense switch terminal screw and the sense switch adjustment nut were located in very close proximity and intermittently contacted (see Figure 4).

Figure 4: Rear latch hook sense switch

The rear latch hook sense switch mounting plate showed no significant distortion and there was no evidence of in-service movement of the rear latch hook sense switch adjustment mechanism.

A circuit test of the three sense switches located within the door (two switches for the two latch hooks and one for the door handle position), determined that it was possible for the rear latch hook sense switch terminal to contact the switch mounting plate retaining nut, providing a short circuit to earth. This had the effect of rendering the door handle centre lock sense switch inoperative. In this condition, the door handle centre lock sense switch would not provide an indication to the crew of an unlocked door.

The examination also found that the door handle reached a position of maximum resistance between the LATCH and LOCK position just prior to the mechanism reaching the over-centre and locked position. It was possible to move the locking handle to this high resistance position, where the green line indicators on the latch bolts appeared to align with the stationary arrow heads in the inspection windows, but the latching hook had not yet activated the centre locking mechanism handle sense switch.

On 7 September 2004, at about 0710 Eastern Standard Time, a Raytheon B300 King Air aircraft, registered VH-KJD, was being operated on a private Instrument Flight Rules flight, from Brisbane to Truganinnie, Queensland, with the pilot in command (PIC) and, what the operator termed, a 'safety pilot'. The safety pilot was qualified on the aircraft type. Prior to departure from Brisbane, the aircraft's cabin door had been closed by the safety pilot, who stated that all of the pre-flight cabin door checks had been conducted in accordance with the pilot operating handbook.