The foreign registered Boeing 747SP was on a flight from Melbourne to Sydney and was under the control of Sydney Approach (APP). Whilst being radar vectored for an extended downwind leg to runway 16R, the aircraft was issued with a clearance to descend to 5,000 ft, which was read back by the flight radio operator. The controller then instructed the crew to change radio frequency, and re-confirmed the cleared altitude as 5,000 ft. The weather at Sydney was CAVOK.

The aircraft was subsequently observed on radar to descend to 2,600 ft, while on the downwind leg. Several attempts by the controller to contact the aircraft were unsuccessful. At about this time, the altitude readout of the aircraft's transponder ceased to be visible on the radar screen, with only primary returns being received.

About 2.5 minutes later, the Approach controller re-established radio contact and instructed the aircraft to climb to 3,000 ft, at which point the altitude readout re-appeared on the radar screen.  The aircraft was then processed for a landing without further incident. The Approach controller subsequently reported that he had re-confirmed the instruction to descend to 5,000 ft because he was unsure the crew had understood the clearance.

For international operations the airline concerned delegates the responsibility for all English language radio transmissions to a flight radio operator. The radio operator on this flight had accumulated some 25 years’ experience, including several flights into Sydney over the last seven years. When the flight radio operator read back "five thousand" on the second occasion, the controller was satisfied that the clearance had been understood. Recorded radio communications confirmed that the controller had used correct radio phraseology and that the flight radio operator had correctly read back the descent clearance.

It was later found that the flight radio operator believed the instruction had been to descend to "two five thousand", which he initially interpreted as 25,000 ft, although the aircraft had already descended to 6,000 ft. He chose not to question the clearance as he rationalised the controller had meant 2,500 ft.  He called out "two thousand five hundred” to the non-handling pilot, who entered these numbers in the altitude select window on the mode control panel.  When the controller issued the frequency change, and confirmed the descent clearance, the flight radio operator believed that 5,000 ft was the altitude at or below which the frequency change was to be made.  All crew members on the flight deck were listening on headphones at the time of the occurrence and said they were somewhat confused by the clearance instruction but chose not to query it.

SAFETY ACTION

As a result of the investigation, the Bureau issued the following Interim Recommendation to Airservices Australia on 17 July 1996:

"IR960050

The Bureau of Air Safety Investigation recommends that Airservices Australia amend radiotelephony phraseology in the Australian Manual of Air Traffic Services and Aeronautical Information Publication to eliminate the possibilities for misunderstanding in the use of the word "to" in climb/descent instructions." On 27 August 1996, Airservices Australia responded as follows:

"Reference is made to BASI Air Safety Interim Recommendation No. IR960050 relating to the use of the word "to" in climb/descent instructions.

This matter has been widely debated in a number of workshops held with the aviation industry, Defence and CASA during the review of phraseologies used in Australia. In the interests of international harmonisation and in concert with our agreed position with Australian aviation interests, Airservices will continue to align with ICAO and retain the use of the word "to" with level change instructions. Use of the word "to" in this context is consistent with ICAO Doc 4444, PANS RAC, Part IX   Phraseologies Section 3, paragraph 3.1 .2.

Given the need for clarity in communications and to reinforce the point made by ICAO at Part IX, paragraph

2.1 " ATS personnel and other ground personnel will be expected to use appropriate subsidiary phraseologies which should be as clear and concise as possible and designed to avoid possible confusion by those persons using a language other than one of their national languages", the Bureau's Interim Recommendation has been forwarded to our ATS Training Section for further action.

With regard to the publication of instrument approaches on charts to assist VFR pilots, as discussed in IR 950206, I understand that CASA is in the process of reviewing this recommendation. In addition, publication of instrument approaches on charts is currently under discussion in the Airspace 2000 proposals on pages 36 and 68 of the document titled "Airspace 2000 - A Plan for the Future Management of Australian Airspace"."

The response has been classified as OPEN.

Approaching home base, towards the end of a charter flight, the pilot became concerned about a low fuel quantity indication on the fuel gauge, so he advised his company by radio that he was landing in a paddock about 7 nm south-south-east of Moorabbin. He also requested that fuel be brought to him.

After landing and shutting down the engine, the pilot realised that the nearest access to the aircraft by road was blocked by a canal and a fence. He therefore decided to reposition the aircraft to facilitate refuelling. After starting the engine and having hover-taxied a short distance with an estimated 20 kt tailwind and at about 10 ft above the ground, the helicopter's engine flamed out due to fuel exhaustion. The aircraft touched down, heels of the landing skids first, on uneven terrain. The main rotor severed the tail boom, and the helicopter rolled over.

According to the pilot, the fuel gauge registered 45 US gallons when he commenced the charter. He believed this was sufficient fuel to complete the flight using a fuel burn off rate of 25 US gallons per hour. However, since leaving base, the engine had operated for an estimated one hour and 50 minutes, including the time spent at ground idle. Unusable fuel is 10 pounds (1.7 US gallons). Only a very small amount of fuel was found in the tank after the accident.

No fault with the engine or the airframe has been reported to have contributed to the accident. Both of the electric fuel boost pumps were serviceable and operating at the time. The aircraft was not equipped with the optional, low fuel warning light system.

Significant Factors

The following factors are considered relevant to the development of the accident:

1. The pilot did not exercise sound fuel management procedures.

Approaching Tintinara on a cross-country flight, lift conditions deteriorated, necessitating an outlanding. The pilot chose what appeared to be a suitable paddock and made an approach and landing. During the landing roll, as the glider came up over a rise in the terrain, a fence appeared across the landing path. The glider could not be manoeuvred to avoid the fence and a collision ensued. The pilot had not seen the fence during the visual inspection of the intended landing area from the air.

The pilot reported that he had refuelled the aircraft to 80 litres per tank prior to departure. Pre-flight and engine checks were normal. Soon after take-off, as the flaps were retracted at about 120 ft, the fuel pressure dropped to zero and the engine stopped abruptly. The pilot immediately landed on the remaining runway but was unable to stop the aircraft overrunning the runway into rough terrain, collapsing the nose landing gear. He reported that during the shutdown checks the fuel selector was turned from the left tank to the off position.

No defects were found when the engine was stripped after the accident. It was reported that the engine ran normally during a test run after reassembly, using all of the original accessories.

The reason for the sudden engine stoppage was not determined.

The pilot took off to conduct a low-level aerobatic display. Shortly after liftoff, at a very low height, he rolled the aircraft inverted. He applied forward elevator to hold the nose up to avoid a descent.

At that point a blanking nut came off the top of the canopy and the pilot instinctively reached for it. As he did so, he inadvertently allowed the aircraft to descend inverted onto the runway surface. While still inverted the aircraft slid to a stop on the runway.

Significant Factors

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

1. The presence of the loose blanking nut.

2. Low height of the aircraft.

3. The pilot allowed himself to be distracted from his primary task of flying the aircraft, at a crucial stage of the flight.

The aircraft was being flown on a freight charter between Canberra and Goulburn, with a company approved check pilot occupying the right control seat. He had earlier conducted a route check on the left seat (handling) pilot who had been acting as pilot in command under supervision (ICUS).

At about 10 NM from Goulburn, the crew heard a radio transmission from an aircraft at Goulburn reporting a flock of birds on the approach to runway 04. The check pilot stated that as the aircraft joined the circuit, his attention was primarily focussed on searching for the reported birds.  As the aircraft entered the downwind leg of the circuit the check pilot advised the handling pilot to tighten the approach and look for the birds.

The handling pilot did not lower the landing gear prior to landing, although both pilots reported having seen the gear down indicators illuminated. They later stated that the indicator lights may have been illuminated by the late afternoon sun. The aircraft subsequently landed with the landing gear retracted, scraping the lower fuselage and damaging both propellors.

Whilst the check pilot had intended to spend the day route checking the handling pilot, he noted that the handling pilot was already qualified on the Canberra to Goulburn route.  The check pilot asked the handling pilot if he was happy to fly the sector to Goulburn, to which the handling pilot replied that he was.  The check pilot assumed that he had passed the pilot in command (PIC) responsibility to the handling pilot and therefore had no formal role in the operation of the aircraft, other than that of cursory supervision.  The PIC name on the flight plan was not amended, and with no formal handover of the PIC responsibility, the handling pilot continued to consider himself ICUS.

Company operations in the aircraft were normally single pilot; however, the conduct of a route check involved two pilots, a PIC and a pilot acting ICUS.  The cockpit during the route check was in effect a multi crew environment, although neither pilot had received any specific training in multi crew operations or crew resource management. The check pilot had not received training in conducting route checks.

The pilots were of similar age and experience, and the check pilot had become confident in the handling pilot's ability during the previous route sectors. This had the effect of him relaxing his supervision of the handling pilot. As the aircraft approached Goulburn the check pilot's attention became focussed on activities outside the cockpit, rather than supervising the performance of the handling pilot prior to landing. The handling pilot normally conducted the prelanding checks on early downwind.  The check pilot's advice to tighten the circuit and look for birds was made early on downwind, probably interrupting the handling pilot's habit pattern, resulting in landing gear extension being omitted from the pre-landing actions.

SAFETY ACTIONS

The Bureau is considering the safety issues revealed in the investigation. Any safety output will be published in the Quarterly Safety Deficiency Report.

ACTUAL INFORMATION

The B747 flight was enroute from Sydney to Kansai, via overhead Cairns then A216. The aircraft had been processed by Sydney and Brisbane ATC sectors and transferred to Cairns area. Cairns area radar position was manned by a trainee controller under supervision. This was the first occasion the trainee had manned the console.

The Cairns planner, who prepares flight strips for use by area radar controllers, did not include coordination annotations on the strip. Prior to handing the flight strip to the radar controller the planner should have annotated the strip with a 'B'. The radar controller places ticks alongside the 'B to indicate completed coordination actions. Later, as the aircraft passed overhead Cairns, the training officer and trainee were so engrossed in training aspects that they failed to update the estimate to the next control agency (Brisbane Sector 10).

Overhead Cairns the aircraft was approximately ten minutes earlier than estimated. Consequently, the aircraft transferred to Brisbane sector earlier than expected. Brisbane sector was unable to provide Port Moresby air traffic service with the 30 minutes notice required by international agreement.

There was no breakdown in separation.

ANALYSIS

Brisbane Sector 10 uses procedural control methods and does not have access to radar information. Consequently, if estimates for aircraft transiting the sector are not notified or amended by adjacent sectors, there is no other means for air traffic controllers to maintain the disposition of air traffic.

Aircraft overflying Cairns northbound have approximately 35 minutes flying time before they enter the Papua New Guinea flight information region. The need to provide 30 minutes notice to the PNG air traffic service allows only a short period for coordination to be implemented. In turn, this limits opportunities to alleviate errors should coordination be delayed or fail to be implemented.

The Cairns planner controller was required to annotate each flight strip with a 'B' to indicate a requirement to coordinate with Brisbane Control. Normally, the 'B' would have reminded the radar controller (and trainee) to notify the estimate to Brisbane sector. However, on this occasion the lack of a visual cue and the controller’s pre-occupation with training reduced the possibility of successful point to point coordination to such a degree that the system failed.

Point to point coordination is a logical process for transfer of flight information through an air traffic system. However, the lack of redundancy in the process means that the system fails if voice coordination is not implemented. Overall, the lack of a safety net for point-to-point voice coordination leaves little room for error.

FINDINGS

1. The B747 crew planned and operated the flight as scheduled.
2. The aircraft was accepted by and handed off to Cairns area control.
3. The Cairns planner did not place coordination annotations on the flight strip.
4. Cairns area control failed to update the pilots estimate to Brisbane Sector 10.
5. The aircraft was transferred to Brisbane Sector 10 approximately ten minutes earlier than expected.
6. Brisbane Sector 10 was unable to provide 30 minutes notice to Port Moresby air traffic service.

SIGNIFICANT FACTOR

1. There was a breakdown in point-to-point coordination.

SAFETY ACTION

As a result of the investigation Airservices Australia Northern District has implemented a local instruction for sector controllers. Brisbane controllers are to check with the Cairns planner if an updated estimate for northbound aircraft is not received by an aircraft's Cairns estimate.

FACTUAL INFORMATION

History of the flight

The aircraft was engaged on a charter flight for the BHP Cannington mine, and departed from Townsville at 0630 EST. The flight was uneventful, and the aircraft landed at Cannington at 0840. The aircraft was refuelled and at about 0945 seven passengers and baggage were loaded. The aircraft was started and taxied for runway 36. During this time, the necessary checks were completed. The take-off was commenced, and after the aircraft became airborne and was accelerating with a positive rate of climb, the landing gear was selected up.

While the gear was still in transit, there was a sudden power loss from the left engine. The pilot described a simultaneous height loss, roll, pitch, and yaw accompanied by a sound similar to a buzz saw from the left engine. The pilot immediately recognised that the left engine had failed and attempted to maintain speed and directional control. Power was reduced on the right engine to maintain directional control, and it was the pilot's intention to land the aircraft with gear retracted beyond the end of the runway. However, the left wingtip struck a steel fence post, and this spun the aircraft to the left. The aircraft struck a low earth bank while travelling sideways and rearwards. When the aircraft came to rest, all windows were obscured, and the pilot believed the aircraft was on fire. The pilot tried unsuccessfully to open the main cabin door, and the passengers were then evacuated through the emergency exit.

Wreckage examination

The fuselage was structurally intact, but both wings were distorted at the attachment points. Both engines were ripped from their mounts, and both main wing fuel tanks were ruptured. All passenger seats failed inside load, which was the basic cause of most injuries sustained by the occupants. On-site examination of the left engine revealed that both spark plugs in number four cylinder had been damaged. The damage was indicative of a foreign object being present in the cylinder during engine operation. Visual inspection of the piston and bore also revealed extensive damage. No foreign object was recovered from the cylinder during this inspection.

Strip examination of the left engine revealed that:

1. The propeller was physically disconnected from the engine. Examination of the propeller reduction gearbox revealed that all four attachment bolts of the stationary gear assembly had failed.
2. The stationary gear assembly appeared to have excessive backlash between the plate and gear. Disassembly of this item revealed excessive wear on the mating spline teeth of the two components.
3. There was significant fretting present on the mating surfaces of the stationary plate and reduction gearbox housing.
4. Number 4 cylinder was extensively damaged. A foreign object had entered the cylinder and severely peened the piston. The spark plug points were peened over, effectively closing the gap to zero.
5. The supercharger impeller was severely damaged, with the intake section of the vanes being machined off by an unknown object. Additionally, a segment of an impeller vane had been broken off. No foreign objects or remains of the supercharger impeller were found in the intake system.

Meteorological information

The weather was fine. The wind was 090 degrees at 15 kts, and the temperature was 30 degrees Celsius.

Weight and balance

The aircraft was loaded within prescribed weight and balance limitations.

Take-off weight (TOW) was 3908 kg. Maximum allowable TOW was 3992 kg.

Aircraft performance

The pilot said that the aircraft had accelerated to 85 kts when he raised the nose to the take-off attitude, and the aircraft became airborne at approximately 88 kts. The take-off weight chart showed that take off safety speed was 86 kts. Take-off safety speed is the speed which ensures that adequate control can be maintained under all conditions, including turbulence and sudden and complete engine failure during the climb after take-off. When the failure occurred, the pilot said he applied full right rudder, right aileron, and lowered the nose attitude. To maintain directional control, he reduced power on the right engine. There was then no alternative but to land straight ahead.

ANALYSIS

There were three distinct failures on the left engine. The stationary gear assembly was disconnected from the propeller reduction gearbox housing. The number 4 cylinder was not producing power as an unknown foreign object had damaged the spark plugs. The supercharger impeller had been severely damaged by the ingestion of an unknown foreign object.

Of these failures, only the failure of the stationary gear assembly would have resulted in the complete loss of power reported by the flight crew.

The disconnection of the stationary gear assembly was attributed to the fatigue failure of its four attaching bolts. In turn, the fatigue failure of these bolts can be attributed to a markedly reduced fatigue life resulting from increased cyclic loading. This loading resulted from excessive backlash between the mating spline between the stationary gear and its attachment plate.

The pilot stated that after the failure he was unable to control the yaw and roll that occurred and elected to land the aircraft wheels up straight ahead. The effect of the sudden and complete loss of power from the left engine above the take-off safety speed should not normally have rendered the aircraft uncontrollable. However, in this case, as distinct from an engine failure, the propeller had disconnected from the engine and was windmilling at high RPM and producing far more drag than it would have done after normal engine failure.

This had the effect of rolling and yawing the aircraft to the left. In addition, there was a 15-kt crosswind from the right, which presented an adverse environment for control inputs required to keep the aircraft straight. The pilot assessed that he may not be able to maintain control in the few seconds available, and he closed the throttles and landed the aircraft.

SIGNIFICANT FACTORS

1. The retaining bolts for the propeller gearbox stationary gear assembly failed when the aircraft had just become airborne.
2. The pilot was unable to maintain directional control and landed the aircraft with landing gear retracted.

SAFETY ACTION

As a result of the investigation into this occurrence, the Bureau of Air Safety Investigation issued Safety Advisory Notice SAN960154 to the Civil Aviation Safety Authority. The notice highlighted deficiencies in the left engine propeller gear box and provided copies of the detailed engineering reports for evaluation.

The pilot had purchased the aircraft earlier in the day and was flying it home to Kingaroy. During the pre-flight inspection he noted that the fuel quantity indication in the sight gauge was 35 litres. This was assessed as adequate for the 20-minute flight, plus reserves.

About 8 km before the destination the pilot again checked the fuel quantity, which was reported to be 10 litres. This was assessed as adequate for the remainder of the flight. Shortly after, the engine failed, and a forced landing was conducted. The paddock chosen appeared to be covered in smooth grass. During the landing roll a wheel struck a clump of the long grass, and the aircraft nosed over, coming to rest inverted.

Examination of the aircraft found that less than three litres of fuel were in the aircraft when the engine failed. The pilot was not aware that the fuel sight gauge was accurate only in flight. There was no mention of this in the aircraft manual.

Factors

1. The pilot had little experience on the aircraft type.

2. The engine failed due to fuel exhaustion.

3. A clump of grass caused the aircraft to nose over.

A video camera was mounted in the cockpit to record the pilot’s actions and comments. This also recorded some of the data presented on the instrument panel and some of the view out of the front windscreen. Most of the record from the video survived the impact and was able to provide a comprehensive record of the flight. Further information was gained from a flight data acquisition unit fitted to the aircraft and from a ground-based video camera which recorded the spin sequence. Later in the investigation the second prototype aircraft was flown and provided additional information.

The aircraft had been spun approximately 60 times prior to this flight. The spins had started with the aircraft set up with low weight and optimal C of G and had progressed to this flight which was conducted at the most critical weight and C of G configuration required for certification.

The spinning flight immediately prior to the accident flight was conducted at a slightly lower weight and at a not so critical C of G. The pilot was not able to recover from the spin on this flight until he had dumped the ballast and deployed the anti-spin parachute. The manufacturers investigation determined that the pilot had been slow to apply the correct control inputs, and the elevator control rigging did not allow full down movement of the elevator. The control cables were reset to ensure that the elevators were able to operate to the stops. After the accident there was not sufficient data available to show whether these corrective actions were effective in restoring full elevator control during a spin, although ground checks had shown full and adequate movement was available.

The investigation into the accident determined that inadequacies in the design of the fin and rudder, and in the rudder control system, had combined with the airflow blanking effect of the horizontal stabiliser, the elevators, and the slab sided fuselage, to preclude adequate spin recovery.

The fin and rudder were assessed as having insufficient area outside of the blanking flow when the aircraft was spinning. The manufacturer has since extended the chord and height of the fin to increase the area outside of that which is blanked during a spin.

The rudder has been extended in chord and lowered in position relative to the tailplane. The fin has been increased in height thereby increasing its aspect ratio. A ventral fin has been fitted to the underside of the fuselage. These measures should increase the effectiveness of the empennage.

The second prototype aircraft was initially flown with essentially the same empennage as the accident aircraft. After some flights on the second aircraft the rudder hinge moment was found to be inadequate. This factor had been masked in tests on the accident aircraft by the use of bungies in the nose wheel steering system. Installation of springs in the nose wheel steering system of the second aircraft showed up the hinge moment inadequacy which the manufacturer has corrected in the redesign of the rudder system.

Analysis of the accident data, and of subsequent test flights, has raised some doubts as to the effectiveness of the application of the elevator control during critical spin recovery. Tests are continuing in an endeavour to assess whether or not there is a problem in this area.

After the accident a review of literature concerning the effectiveness of anti-spin parachute installations was undertaken. This disclosed that the use of a parachute on a long lanyard to pitch the aircraft out of a spin has been rejected in favour of stopping the rotation through use of a larger parachute with no lanyard and short risers.

Analysis

This was a prototype aircraft and some deficiencies and/or problems during testing are to be expected.

With this particular aircraft the fact that the inadequate rudder hinge moment was masked throughout flight testing meant that the inadequate rudder performance during critical spin recovery was not clearly detected until it combined with other factors to become critical. These other factors included an ineffective anti-spin parachute, extensive blanking of the fin and rudder, and flight at the extremes of the weight and C of G envelope.

It is not known what, if any, effect the previous rerigging of the elevator controls had on this flight.

Significant factors

1. The rudder and fin effectiveness was inadequate for the spin test being undertaken.
2. The anti-spin protection systems were ineffective.

The aircraft was not able to be recovered from an intentional spin.