Factual Information

The pilot departed from his property in his ultra-light aircraft at around 0830 EST for a local flight. He flew over several of his friends' properties and over the local township. The weather as reported by witnesses was a sunny day, wind light and variable, scattered high level cloud with a temperature of 18 degrees C.

Witnesses reported that they had observed the aircraft flying in a south westerly direction at an altitude of approximately 150 ft. The aircraft was observed to be flying at a constant speed with the engine noise remaining constant. A father and son said that they waved as the aircraft flew over where they were standing and the pilot acknowledged by waving back to them. The aircraft then banked steeply to the right, with the bank angle estimated to be 90 degrees. At approximately 180 degrees into the turn, the aircraft was observed to stall. At this point the nose dropped, and the aircraft proceeded to rotate. The aircraft was in a steep nose down wings level attitude, when the aircraft crashed through the safety rails of a bridge and impacted the ground, fatally injuring the pilot. The point of impact was approximately 75m from where the witnesses were standing.

Examination of the wreckage did not find any pre-existing defects that would preclude normal operation of the aircraft. There was no evidence that the pilot had any pre-existing medical conditions that would affect his ability to fly the aircraft. The pilot had accumulated some 188 hours flying time in ultra light aircraft, the majority of which was in this aircraft. He was currently undergoing additional flying training to gain his Private Pilot's Licence.

Analysis

The final flight path observed by the witnesses during the turn was consistent with the aircraft stalling and entering an autorotation. In attempting the manoeuvre, the pilot allowed the aircraft performance to degrade by not maintaining sufficient airspeed, resulting in the aircraft departing controlled flight at an altitude insufficient to recover.

FACTUAL INFORMATION

History of flight

The pilot and his passengers arrived at Hotham Heights at about 1735 EST on 21 September 1993. At the time the temperature was minus three degrees Celsius, it was snowing lightly, and the visibility was poor due to low cloud. The pilot parked the helicopter facing in a northerly direction in the day car park overnight. Snow deflector baffles were not fitted to the airframe and covers were not placed in the engine air intakes or on the exhaust stacks. Overnight it snowed enough to cover tracks from the previous day. The average windspeed, recorded from midnight, varied from 4 knots to 13 knots, varying in direction from the north through to the west.

Shortly after 0900 on 22 September 1993 the pilot threw a few cupfuls of water from a bucket into both sides of the particle separator which is the air filtering device for the engine air intake. He also threw water onto the tail rotor assembly and the windshield. Throwing water into the particle separator is not standard practice on a daily inspection prior to flight. About ten minutes later the pilot returned to the helicopter, loaded the two passengers, untied the main rotor, and started the engine. When the main rotor started to spin, snow was flung off the main rotor blades. The pilot ran the engine for five to ten minutes. Then the helicopter lifted into a hover briefly before it departed to the south west and descended into a valley.

The helicopter wreckage was not located until 1245 on 23 September after a prolonged search. The accident site was on a steep, snow covered, timbered slope in mountainous, alpine terrain about 1,300 ft lower than Hotham Heights. The pilot and the two passengers did not survive.

Flight notification

The pilot had telephoned his wife from Hotham Heights at about 0920 on 22 September to say that he would be departing shortly and that he would arrive at Moorabbin by 1200 local time. By about 1315 the pilot's wife became so concerned that her husband had not arrived that she telephoned a friend who contacted Moorabbin Airport to check on the whereabouts of VH-FUX. It was then realised that the helicopter had not arrived, and a search was initiated.

The pilot had not submitted a flight plan to the Civil Aviation Authority (CAA). By telephoning his wife, he had opted for the approved alternative of leaving a Flight Note with a responsible person. The investigation could not determine whether the pilot had acquired a weather forecast before departing Hotham Heights. A post-accident estimate of the fuel on board indicated that it was sufficient for the conduct the flight.

Weather

When the helicopter departed Hotham Heights, the temperature was minus 2.2 degrees Celsius, the relative humidity was 100% and the wind was 256 degrees at 8 kts. It was not snowing. The mountain peaks and tops of the ridges were covered by seven eighths of cloud which was showing signs of dissipating with the sun shining through in patches. The cloud base was slightly below the level of the Hotham Heights day carpark. Cloud conditions on the ridgeline immediately west of the accident site were probably similar. The valley into which the helicopter descended was reported to be clear of cloud with visibility being about 2 km.

Survival

The pilot survived the crash and died from injuries and/or hypothermia before the helicopter wreckage was located.

The helicopter was equipped with a Narco ELT10 survival beacon and a basic survival kit. The survival beacon was ejected from the helicopter at impact and was damaged. Once damaged, the beacon was incapable of transmitting a distress signal to assist in locating the helicopter.

Wreckage examination

Examination of the wreckage did not reveal any pre-existing defects which may have contributed to the accident. Damage sustained by the helicopter during the accident, as well as the damage to the trees, indicated that at initial impact the helicopter had a moderate rate of descent, with low rotor RPM and very little forward airspeed. The helicopter impacted the ground on its left side after which it slid about 20 m down a steep slope before coming to rest against tree trunks. It was within its approved centre of gravity and gross weight limits at the time of the accident.

No evidence of significant torsional twisting was found on any of the drive shafts. The main rotor mast was not torsionally twisted. The engine compressor case-half liners suffered minor internal damage consistent with engine RPM being very low at impact. This evidence is consistent with the engine having flamed out before impact.

The engine was successfully test run after the accident in an approved engine test cell.

Radio transmissions

At 0937, Melbourne Flight Service received a radio mayday call. A call sign was not received, and the Flight Service officer had no idea who transmitted the words 'mayday, mayday, mayday' but did correctly identify which frequency, from the group of frequencies he was monitoring, on which the call was made. He immediately declared the distress phase and notified the search and rescue mission coordinator (SARMC). As no further radio calls were made to indicate that an aircraft was in distress, the search and rescue phase was cancelled.

The recording of the mayday call was subsequently analysed by BASI and identified as having originated from VH-FUX. The mayday call lasted 2.1 seconds. After voice modulation ceased, there was approximately 0.42 seconds where noise was recorded. A second transmission was made approximately one second later and was approximately 0.23 seconds in duration. The results were compared with the helicopter manufacturer's data and with inflight recordings obtained during inflight trials conducted in VH-FUX on 16 September 1993 when sound and instrument readings were recorded to assist in another accident investigation. A tone which was considered to have been related to aircraft operation was detected and compared with prior recordings of VH-FUX in flight. The tone indicated that the aircraft may have been in an autorotation with the main rotor RPM reducing from 85% to 82% over a 3.3 second period. The normal range for a power off autorotation is 90% to 107%.

Flight in icing conditions

The approved Flight Manual for the Bell 206B states that the helicopter is certified for operations under non-icing conditions. The Flight Manual also states that engine anti-icing shall be selected on for flight in visible moisture in temperatures below plus 4.4 degrees Celsius. At the accident site the anti-ice valve on the engine was found on and the anti-ice switch in the cockpit was found in the on position. It is not known when the pilot selected anti-icing on.

In the Antarctic, the pilot had flown helicopters which were fitted with the same type of engine. He would have been familiar with the potential problems associated with ice. Compressor surge/stall has previously occurred in a Bell 206B in similar weather conditions in the Australian alps. In that case, the surge/stall occurred before the helicopter lifted off. Compressor surge/stall is audible. Witnesses who saw and photographed VH-FUX at Hotham Heights prior to and during its departure, reported no unusual engine noise.

Advice was sought from the Allison engine manufacturer. The engine manufacturer raised the possibility of a slug of ice/slush/snow being sucked into the intake causing a flameout. Such flameouts have been known to occur with the Allison 250 C18 and C20 engines fitted to Hughes 369 helicopters (also known as Hughes 500s) during flights in the Antarctic. In the Antarctic occurrences, the engine flameouts occurred in flight after about 15 to 20 minutes, but the outside air temperatures were colder than in the Australian alps. The Allison 250 B17, which has the same compressor, has been known to flameout in a Nomad floatplane due to water spray during taxiing. A twin engine Bolkow 105 helicopter also suffered a flameout on both of its Allison 250 C20 engines due to snow/sleet ingestion inflight despite the fact that the anti-ice was operating on both engines. In the case of a slug of ice/slush/snow or water causing an engine flameout, usually no damage is subsequently found in the engine.

Tests And Research

An experiment involving throwing cups of water into a particle separator, which was not fitted to an aircraft, proved that a significant amount of water will pass through the swirl vanes of the particle separator into the plenum chamber. There were no drain holes in the plenum chamber. This area is well sealed off to allow only filtered air to enter the engine.

ANALYSIS

Pre-flight actions

Because the helicopter was parked overnight in falling snow and moderate winds, without compressor intake covers installed, it was quite possible that snow made its way into the particle separator. The pilot probably saw frost/snow in the particle separator prior to starting the engine, which would account for his throwing water into it.

However, some of the water thrown into the particle separator by the pilot, plus some of the overnight snow, probably accumulated in the plenum chamber. It is unlikely that the pilot would have been able to see into the plenum chamber, if he tried to check for the presence of snow or water, because the Perspex viewing ports would have been covered in frost. Also, it is normally difficult to see clearly into the chamber except in very bright sunlight or with the aid of a bright torch. The pilot was not seen using a torch and ambient light at the Hotham Heights carpark was diminished by the foggy conditions.

Pilot decisions

The pilot probably elected to depart from Mount Hotham because the cloud on the high ground was just beginning to dissipate, and the valleys appeared clear. As there was no evidence of compressor surging/stalling, it is likely that the pilot turned on the engine anti-ice soon after starting the engine. He then ran the engine for some time, probably to warm it and to determine that there was no intake ice forming. With the rotors turning for several minutes before take-off, the pilot should have been able to feel whether or not the helicopter was vibrating as a result of ice/frost accumulation on the rotor blades.

Had the weather conditions caused the accumulation of significant airframe icing before departure, particularly on the rotors, the pilot and the passengers would probably have felt significant airframe vibrations soon after the engine was started. Such vibrations would have prompted the pilot to shut down the engine and postpone the departure. Also, if the aircraft had suffered serious airframe icing after departure, it would more likely have crashed under power, which was not the case.

A likely scenario is that the pilot, while flying in the valleys, saw a possibility of tracking north west and attempted to track in that direction in the hope of clearing the alpine area sooner than if he persisted in the valleys. In so doing, the pilot is likely to have flown at reduced airspeed as he approached the ridgeline which was slightly north west of the impact site. The evidence indicates that there probably was low cloud on or near the ridgeline as well as associated reduced inflight visibility.

Engine flameout

Given the weather conditions which prevailed in the Mount Hotham area at the time of the accident, there are two likely reasons for a flameout of the helicopter's Allison 250 C20 engine - compressor intake icing or a slug of ice/slush/snow being ingested into the compressor intake.

If the pilot had left the anti-ice switch in the off position for some time, ice could have built up on the engine compressor intake and caused the engine to surge/stall or even flameout. However, the pilot's actions after start, and the lack of any evidence of surge/stall at that time, make this less likely.

The helicopter probably encountered a slightly higher outside air temperature as it flew in the valleys. Also, the particle separator may have warmed up slightly as a result of engine heat soak, particularly if the helicopter was flying at low forward airspeed. An increase in temperature probably caused a large enough slug of ice/slush/snow to dislodge from the plenum area and the particle separator and enter the engine compressor intake causing an instant flameout. The helicopter was not fitted with the optional auto reignition system which relights the engine very quickly in the event of a flameout.

Because of the low height above the ground the pilot would not have had time to restart the engine. The combination of a flameout, low height, low airspeed, mountainous terrain, and tall trees resulted in a very heavy landing. The low height and lack of time would also account for the pilot not managing to transmit his call sign during the mayday call.

Survival

Had the Emergency Locator Transmitter (ELT) beacon been capable of transmitting a distress signal, it is possible that the signal would have been intercepted by the monitoring satellite or by an overflying aircraft. The time taken to locate the helicopter could have been significantly shorter in duration, which may have increased the pilot's chances of survival.

CONCLUSIONS

Findings

1. The pilot parked the helicopter outdoors and did not fit it with engine intake covers.
2. The helicopter was exposed to blowing snow and freezing conditions overnight.
3. The engine anti-ice switch was found in the on position.
4. The helicopter was not fitted with the optional engine auto-reignition system.
5. The ELT fitted to the helicopter was damaged during the impact and could not transmit a distress signal.
6. The pilot survived the initial crash impact but died from injuries and/or hypothermia before the helicopter was located.

Significant Factors

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

1. The helicopter was parked overnight, without engine intake covers being fitted, in conditions of blowing snow.
2. The pilot may not have inspected the plenum chamber for snow or water accumulation before take-off.
3. An engine flameout occurred at a low height over terrain unsuitable for a forced landing.

SAFETY ACTION

The issues relating to the ELT survival beacon raised in this report are being considered in the light of this and several other occurrences. They will be the subject of a future Bureau of Air Safety Investigation report.

After completing a TB-20 aircraft type endorsement at Nowra the pilot hired the aircraft for a return flight to Forbes via Bankstown.  During the afternoon of Friday 1 October, he completed the Nowra to Bankstown sector but had to postpone the flight to Forbes due to en route weather conditions which precluded flight under visual flight rules (VFR).

On Sunday 3 October the pilot obtained area weather forecasts, and an aerodrome forecast for Forbes which indicated that the weather at Forbes was fine but the en route weather was not suitable for VFR flight. The pilot was also advised that two helicopters bound for Bathurst had turned back to Bankstown due to poor visibility.  Despite the forecasts and reports of adverse en route weather the pilot decided to attempt the flight to verify the accuracy of the forecasts which had predicted an improvement after 0900 hours.  The pilot did not submit details of his proposed flight to the CAA, nor did he request a SAR watch or leave a flight note.

VH-JTI subsequently departed Bankstown at 0830 EST and after the pilot changed from the tower frequency there was no further radio contact with the aircraft.  At about 0900 bushwalkers in the Kowmung River valley saw a single engine light aircraft flying in a southerly direction in the valley. The aircraft, which was not identified, was observed turning west to enter the Christies Creek Gorge.  The unbroken cloud base obscured the tops of the ridges and the aircraft, which appeared to be operating normally, was flying about midway between the valley floor and the cloud base.

JTI was reported missing during the evening of Sunday 3 October after it failed to arrive at Forbes.  An intensive air search was commenced the following day.  The bushwalkers did not report sighting an aircraft in the valley because a Sydney daily newspaper they had read incorrectly identified the date on which the aircraft was reported missing.

The wreckage was located on the eastern side of the Boyd Range on Sunday 24 October by a bushwalker who was about 2 kms from the site on the opposite side of the valley. The accident site was at an elevation of 2,500 ft on steeply sloping terrain which was moderately timbered with trees 18 to 25 m in height. The occupants survived the impact forces but suffered burns from a fuel fed fire which started when the aircraft began to break up during the collision with trees. Several days later they perished from the combined effects of their burns and exposure after they had walked about 2 km down to Wheengee Whungee Creek.

Examination of the wreckage found the landing gear and the wing flaps retracted.  Propeller inspection indicated that the engine was operating at a low power setting at ground impact.  No defects which may have influenced the circumstances of the accident were identified.

A removable emergency locater transmitter was attached to the floor of the luggage compartment at the rear of the cabin. The transmitter assembly had a self-contained aerial and was therefore capable of operating remotely from the aircraft. The transmitter was destroyed by intense fire which gutted the cabin area and there were no reports of any transmissions being received from the transmitter.

CONCLUSION

Significant Factors

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

1. Forecast and observed weather conditions over the Blue Mountains and Great Dividing Range in the general area of the proposed route were unsuitable for flight in accordance with visual flight rules.
2. The pilot continued flight over mountainous terrain in weather conditions which were unsuitable for flight under visual flight rules.
3. The pilot was inexperienced, both generally and on the accident aircraft type.
4. The aircraft was not configured for the optimum manoeuvring speed in the prevailing conditions.

The pilot was returning from a local sheep mustering flight on his property. When the gyroplane failed to return to the airstrip, and the pilot could not be contacted by CB radio, a search was commenced. The wreckage of the gyroplane was found about 230 metres west of the airstrip, having struck the ground in a steep descent with little forward speed, fatally injuring the pilot. The weather conditions at the time of the accident were reported to have been affected by strong gusting north- westerly winds, associated with blowing dust and reduced visibility.

An inspection of the wreckage revealed no evidence of any pre-existing mechanical fault or defect. However, whilst the gyroplane was in flight, the rotors had struck the tailplane and fractured the tail boom to the rear of the engine mounting. This damage was consistent with excessive flapping of the rotor blades associated with air flow reversal through the rotor disc. Air flow through the rotor disc normally enters from beneath the plane of the disc and exits above. Other effects of reversal of airflow are the irreversible deceleration of the rotor blades, and a nose down pitch change. In this accident the gyroplane appeared to have tumbled end over end, as well as rotating about the vertical axis.

The reason for the onset of airflow reversal on this occasion could not be positively established. However, in conditions of strong gusty winds, large changes in the vertical wind component could lead to airflow reversal. The likelihood of such an occurrence would have been increased if the gyroplane was being flown at relatively high speed.

The pilot suffered from epilepsy and was taking prescribed medication at the time of the accident. It could not be determined if the pilot's medical condition was a factor in this accident.

Following the completion of 27 jumps over a 9-month period using a student rig, the parachutist commenced conversion training to his own rig. This training consisted of three dual training jumps and two solo jumps. After a further five jumps using his own rig, the student took a two and a half months holiday during which he made no jumps.  Two weeks prior to the accident he made three jumps using his own rig.

The accident jump was part of the student's training towards a "B" licence. The training consisted of a "B relative" jump requiring him to perform certain manoeuvres relative to a tutor during the free fall period of the jump. He used his own rig for this jump.  On completion of the free fall component of the jump the two parachutists separated for the deployment of the parachutes. When his main parachute had deployed the tutor looked down and noted that the student's main parachute had not deployed normally and was being discarded. The tutor did not see the reserve parachute deploy before the student hit the ground.

No defect was found in the equipment which could have caused either the main parachute problem or the non-deployment of the reserve parachute. Damage to the Spandex pocket in which the main pilot parachute is kept suggested that the main pilot parachute throw-away procedure was mishandled resulting in the need for the main parachute to be discarded.

The main parachute was released at sufficient altitude for the reserve parachute to deploy. If correctly followed, the procedure taught for the rig in use should have ensured that the reserve deployed immediately after the main parachute was discarded.

It is possible that under the stress of the abnormal event the parachutist reverted to an earlier training awareness by waiting for some time for the reserve to automatically deploy following the release of the main parachute, as it does on the student rigs. Student rigs have a reserve static line (RSL) or an automatic activation device which automatically deploys the reserve parachute should the main parachute be discarded.

On-scene evidence indicated that the parachutist remembered too late that he had to manually deploy the reserve. If an RSL or an automatic activation device had been attached the accident may not have occurred.

SAFETY ACTION

Paragraph 5.1.16 of the Safety Requirements section of the Australian Parachute Federation Operational Regulations has been amended to read:

'All descents made by parachutists who do not hold a Certificate "E" must be made with equipment fitted with a functional reserve static line or an automatic activation device. The DZSO may permit exemptions to this rule for specific descents.

Note:  This regulation became effective on 1 February 1994 where the descent is being made with a harness/container manufactured after 30 July 1993, or on 1 February 1995 where the descent is being made on a harness/container manufactured before 1 August 1993.'

The main purpose for investigating air safety occurrences is to prevent aircraft accidents by establishing what, how and why the occurrence took place, and determining what the occurrence reveals about the safety health of the aviation system.

Such information is used to make recommendations aimed at reducing or eliminating the probability of a repetition of the same type of occurrence, and where appropriate, to increase the safety of the overall system.

To produce effective recommendations, the information collected, and the conclusions reached must be analysed in a way that reveals the relationships between the individuals involved in the occurrence, and the design and characteristics of the systems within which those individuals operate.

The student hang glider pilot was performing his first soaring flight. His glider and another were observed to be flying about 200 feet above the same hill. The gliders collided, and the glider flown by the student spiralled into the ground. The student received fatal injuries, and the other pilot was uninjured and able land his glider.

The surviving pilot advised that he did not see the other glider, possibly because he was flying into the setting sun at the time of impact, but felt something strike the left wing of his glider.

Information concerning this accident was provided to the Bureau by the Safety Director for the Victorian Hang Gliding Association.

History of the flight

The helicopter carrying one pilot and three passengers departed Gladstone aerodrome at 1035 hours EST on a charter flight to Heron Island.  It was equipped with emergency flotation and overwater survival equipment consisting of life vests and a raft.

En route three miles out to sea near Facing Island and when cruising at an estimated 1500 ft, the pilot transmitted a mayday call on the mandatory traffic advisory frequency. He advised that he was returning to Gladstone because the helicopter had experienced a severe jolt in flight.  Soon after the mayday call, the pilot advised that he was ditching.

The only survivor, a passenger with no aviation qualifications, was seated in the right rear seat.  She recalled that the helicopter airframe gave a kick in flight immediately before the pilot transmitted a mayday.  During the descent, she heard the engine making an unusual shuddering noise and saw the nose of the helicopter twitching left to right several times.

The emergency popout floats comprising six float bags were deployed before the helicopter impacted the water at 1043 EST.  The severity of the structural deformation and the break-up was consistent with the helicopter impacting the water with high vertical deceleration forces and some forward speed.  Only two float bags remained inflated after impact and the helicopter sank in 24 metres of water with the pilot and the front passenger still strapped in their seats. The passengers in the left and right rear seats were thrown from the cabin on impact.

The helicopters buoyant emergency locator beacon was lost at sea and the non-buoyant beacon, which was attached to the airframe, sank with the helicopter.  No signal was heard from either beacon.  The life raft was found afloat but still packed in its valise.  The pilot and the passengers were wearing serviceable life vests which were not inflated. Pathological examination of the three fatalities revealed that the cause of death was drowning.

The helicopter was recovered at latitude 230 16 S longitude 1510 24 E at a position 20 kms NE of Gladstone and about 5 kms from Facing Island, the nearest landmass.

Wreckage Examination

An examination of the wreckage revealed no evidence of airframe or system malfunction which may have contributed to the accident.  The gross weight and centre of gravity were within prescribed limits.

A detailed examination of the engine and its accessories revealed that most of the stator vanes of stages three and four of the compressor were missing.  Five of the fourth stage stator vanes and all but one of the third stage vanes were fractured in the upper case-half and the majority of the third stage vanes were fractured in the lower case-half. All fracture surfaces were destroyed by secondary damage.  Available evidence indicated that the reason for the vane failures was a fracture of one or more of the third stage stator vanes probably resulting from erosion in the upper case-half.

Erosion to the lining of the compressor halves was found to be present around a number of adjacent vanes in the first, second, fifth and sixth stages.  The engine operations and maintenance manual requires the removal of the vane assemblies in these conditions.  Measurement of the first stage vane showed that it did not meet the chordal and thickness erosion limits specified in the manual and a second stage vane did not meet thickness requirements. Particles embedded in the eroded region of the first stage vanes were high in silicon content.  None of the compressor rotor blades were fractured.

The case-halves of the engine compressor were last inspected by a licensed aircraft maintenance engineer at

Gladstone on 18 March 1993.  The engineer advised that the inspection was performed in accordance with the Allison engine maintenance manual and the compressor was deemed to be serviceable. Between 18 March 1993 and the accident, the compressor had operated for 135 hours.

The engine manufacturer requires compressor inspections at intervals not exceeding 300 hours time in service with more frequent inspections required when operating in harsh conditions such as salt water and sandy or dusty environments, where abrasive particles may enter the compressor.   No other defect was found with the engine which could have contributed to the engine malfunction.

The operator's company policy required pilots based at Gladstone to record relevant engine power trend parameters on each day a helicopter flew.   From 18 June 1993 until 26 July 1993, pilots recorded thirty-one sets of engine power trend figures for VH-PCR and this trend monitoring showed no degradation of power in the engine and gave no warning of an imminent compressor failure.

Pilot Information

The pilot held a Commercial Pilot Licence (Helicopters) with a valid class 1 medical certificate.  He had been rostered off duty for the three days immediately prior to the accident and he had adequately rested prior to the flight which was conducted within his normal duty period.  He had no known medical problems at the time of the accident.

The pilot received endorsement training on Bell 206 helicopters during February and March 1993. This training involved five hours of dual instruction, followed by 4.6 hours of operations in-command-under-supervision, flying a Bell 206 between Gladstone and Heron Island.  His most recent dual check was on 31 May 1993 at Gladstone in a Bell 206. Since 1989 he had accrued over 1000 hours as co-pilot of twin engine helicopters engaged in offshore operations.  His total experience as pilot-in-command was 254 hours of which 95.4 hours had been flown in Bell 206 helicopters.

The pilot did not hold a helicopter utility float endorsement nor was there a legal requirement in Australia for the pilot to hold such an endorsement for over-water operations in a helicopter equipped with emergency popout floats. Popout floats are designed for emergency use only and do not inflate until activated by the pilot.  Practice autorotative descents in helicopters equipped with emergency popout floats are not normally carried through to termination onto the water and popout floats are not normally inflated during such practice descents.  In contrast, utility floats are expected to be permanently inflated when fitted to a helicopter.  They are more robust and are designed for regular landings onto water, including termination onto water following a practise autorotative descent.

No evidence was found to indicate that the pilot had flown a helicopter with inflated floats or that he had received any training in simulated engine-off touchdowns onto water. He had successfully undergone helicopter underwater escape training on 31 July 1991.

Weather

Forecast weather was for south-south-easterly winds at 10 to 15 kts turning more easterly in the afternoon with a sea swell up to one metre.

The crew of a company helicopter which arrived overhead the survivor and the floating wreckage about 20 minutes after the accident reported that the sea surface at the accident site had about half a metre swell with occasional white cap waves. Visibility was in excess of 40 km and the wind was blowing from the east-south-east at about 10 kts. The crew considered the sea surface conditions to be suitable for a safe ditching.

Rotor RPM

The low rotor RPM warning system fitted to a Bell 206B helicopter consists of a warning light on the instrument panel and an audible tone from a cockpit speaker. It is activated when the rotor RPM is less than 90% and the collective pitch lever is raised above the fully down position.

Seven radio transmissions made by the pilot were recorded on the AVDATA logging recorder at Gladstone.  The recorded information was examined, and the results were compared with the helicopter manufacturers data and with inflight recordings obtained during subsequent flight tests in a Bell 206B.

During the first three transmissions, when operations appeared normal for taxiing and departure from Gladstone, rotor RPM remained constant at approximately 99.5% (98-100% being normal).  Throughout the remaining transmissions including the mayday and declaration of ditching, the main rotor RPM appeared to range from

81.6 % to a maximum of 96.6% and then decrease to 75.9% over a period of approximately 28.4 seconds.  The last transmission occurred 19.6 seconds later and was approximately one second in duration and it was not possible to determine the rotor RPM during this transmission.  An audible tone, which may have indicated the low rotor RPM aural alert, was also heard during periods of the fourth and fifth transmissions and continuously during the sixth and seventh transmissions.

Inflight yaw

Difficulties associated with maintaining balanced flight would be increased by fluctuations in torque.  It is also more difficult to maintain balanced flight with inflated floats than with standard high or low skid landing gear, or with emergency floats stowed. The survivor's description of the nose of the helicopter moving left to right during the descent was consistent with torque fluctuations.

Descent over water

The helicopter was equipped with a radar altimeter and after the accident, the warning light on the radar altimeter was found set to activate at 200 feet above terrain or water. The check captain who gave the pilot the Bell 206B endorsement advised that he had verbally explained the operation of the radar altimeter to the pilot.  However, during the endorsement and subsequent training, they did not practice power terminations over water or land by reference to the radar altimeter.  The engine malfunction occurred when the helicopter was over water with no other visual cues (such as land mass or trees) to assist the pilot with depth perception for flight termination onto the water. It could not be determined if the pilot used the radar altimeter to assist with judgement of ditching on the day of the accident.

Warning and caution lights

The filaments of the globes in the warning and caution panel were inspected.  The engine out warning filaments showed some signs of stretching including a broken filament.  The low rotor RPM filaments showed definite evidence of stretching including a broken filament.

This indicates that electrical power may have been present in the filaments during the impact sequence.  It is probable that the engine-out and low-rotor RPM warning systems activated during the descent.

ANALYSIS

Introduction

The two main factors considered relevant to the development of the accident were the failure of the engine compressor in flight and the ensuing termination of the autorotative descent.  The pilots report of a severe jolt, as well as the kick and engine shuddering described by the survivor, were consistent with symptoms of an engine compressor failure.

Low rotor RPM during the descent

The analysis of the AVDATA tape recording revealed a rotor RPM lower than prescribed operating limits.  During the sixth transmission, rotor RPM reached 75.9% and at such a low RPM level the helicopters rate of descent would have been considerable.  It is possible that the pilot was trying to ascertain the extent of available power and any attempt to increase power by raising the collective pitch lever, would probably have increased the likelihood of compressor surging or stalling.  This would cause corresponding torque fluctuations, and the damaged engine would not have been able to produce sufficient power to maintain rotor RPM within limits.

Sustained low rotor RPM

At the operating weight at the time of the occurrence, the normal rate of descent during an autorotative descent would be approximately 1,500 ft/min.  If the pilot had maintained the collective lever in a slightly raised position for several seconds, including the period of the descent through the last 100 ft, the main rotor RPM would have decreased to  a less than desirable level with a descent rate considerably greater than 1,500 ft/min.  With main rotor RPM reaching a low of 75.9%, the inertia retained in the rotor would be considerably reduced.  If the pilot then attempted to reduce the rate of descent near the sea surface by increasing collective pitch, the rotor would not have been capable of producing sufficient lift to arrest the abnormal rate of descent.  Left unchecked the remaining rate of descent would have ensured a hard landing on the water.

High collective pitch pull increase

The pilot may have achieved a stable autorotative descent, but before he had descended to the height where autorotation would normally be terminated, he may have experienced problems with depth perception.  If collective pitch was increased with the helicopter too high above the water, the rate of descent would reduce initially, but then increase considerably as rotor RPM decreased. This would result in a hard impact with the water.

CONCLUSIONS

Findings

1. The pilot was suitably licensed and qualified to undertake the flight, and there was no evidence that he was suffering from any illness or incapacity during the flight.
2. Prior to the accident, the pilot had been briefed on the operation of the emergency popout floats of the Bell 206B. There is no evidence that he had flown any helicopter with inflated floats before the accident flight.
3. There was no requirement for pilots operating emergency flotation equipped helicopters to satisfy the water endorsement criteria of the flying training syllabus.
4. The helicopter was equipped for overwater operations and the pilot and passengers wore approved life vests. The life raft and vests were found to be not inflated after the accident.
5. The helicopter was fitted with emergency flotation equipment which was inflated by the pilot before the helicopter impacted the water.
6. The fuselage was severely broken up and deformed in a manner consistent with high vertical deceleration forces.
7. The weather and sea conditions should not have precluded a safe ditching.
8. The helicopter's gross weight and centre of gravity were within prescribed limits.
9. The engine compressor failed, and the engine then lost power in flight to a degree that the pilot was forced to ditch the helicopter.
10. The main rotor RPM decreased to 75.9% during the descent.
11. The cause of the compressor vane failures was a fracture of one or more of the third stage stator vanes resulting from erosion in the upper case-half.
12. Regular engine power trend monitoring records revealed no degradation of engine power and gave no warning of an imminent compressor failure.
13. The compressor had operated for 135 hours time in service since the last compressor case-half inspection.
14. No airframe defect was found which may have contributed to the accident.

Significant factors

1. The engine compressor failed due to the fracture of one or more of the third stage stator vanes. Erosion in the upper case-half of the compressor may have contributed to the failure.
2. For reasons which could not be determined a safe autorotative landing onto the water was not made.

SAFETY ACTIONS

Safety action implemented by helicopter operator

Since the accident, the operator has trained its Gladstone based helicopter pilots in operations with utility floats fitted, including complete touchdown autorotations onto water.

Recommendation

As a result of the investigation into this occurrence, the Bureau has made the following recommendation R940236:

The Bureau of Air Safety Investigation recommends that the Civil Aviation Safety Authority review the requirements for pilots operating helicopters fitted with emergency flotation equipment.  The review should consider the applicability of a requirement to satisfy the water endorsement criteria of the flying training syllabus.

The pilot who had assembled the aircraft was also distributing the aircraft in Queensland. He had planned to fly the aircraft from Caboolture to an airshow at Hervey Bay.

The flight from Caboolture to Gympie took about two hours because of headwinds encountered en route. At Gympie the pilot refilled the left fuel tank of the aircraft from a 20-litre container he carried in the aircraft. He was then given a ride to a nearby service station where the container was refilled. Upon returning to the aircraft the refilled container was strapped to the seat beside the pilot. The pilot boarded the aircraft and after starting the engine the aircraft was taxied for take-off. Witnesses reported that after the aircraft became airborne it climbed overhead the airfield before setting course to the north. The witnesses also stated that the aircraft appeared to be operating normally and that engine operation was also normal.

About five minutes later the aircraft was observed to the east of Gympie flying in a northerly direction. The engine was then reported to have misfired and stopped. The aircraft was turned to the west and overflew a golf course. The western side of the course was bounded by pine trees about 10 metres in height and the Bruce Highway. The aircraft cleared the pine trees but impacted the highway directly in front of a vehicle. The vehicle struck the fuselage of the aircraft which was dragged underneath the vehicle for a short distance.

The fuselage and inboard section of the wing were severely damaged by ground and vehicle impact. However, an inspection of the wreckage did not reveal any faults that may have contributed to the accident. The engine was removed from the aircraft and inspected. Apart from some minor accident damage it appeared in good condition. The engine was internally inspected and the only fault found was some minor scoring on the forward face of the number 1 piston which would indicate that the piston may have partially seized previously. The effect of this scoring on the operation of the engine prior to the accident could not be determined.

The pilot was experienced in the operation of ultralight aircraft and a senior ultralight flying instructor. He had operated this particular aircraft for all of the 25 hours it had flown since assembly. Following the reported engine problems, the aircraft was flown over the golf course which contained several fairways that would have been suitable for landing the aircraft in an emergency.

SIGNIFICANT FACTORS

1. The reason for the reported loss of engine power could not be positively determined.

2. The pilot overflew suitable landing areas without attempting a landing.

3. The pilot lost control of the aircraft at an altitude that was too low to effect recovery.

The aircraft, with only the pilot on board, was being flown from Archerfield to Caboolture via the light aircraft lane to the west of Brisbane in company with another aircraft. About five minutes after departing Archerfield, the pilot radioed that he was experiencing problems with both engines and that he was in an emergency situation. The pilot of the other aircraft advised him that there were suitable forced landing areas in and around a nearby golf course. However, the aircraft continued and slowly lost altitude before rolling inverted and diving steeply into the ground.

Ground witnesses reported hearing loud backfiring and fluctuating engine RPM from the aircraft. These sounds were accompanied by erratic rolling and yawing of the aircraft before it rolled to the left and inverted. The right wing was severed outboard of the engine as the aircraft impacted a large tree before crashing onto a road.

Wreckage examination revealed that the fuel selectors for both engines were set at the auxiliary tank positions, causing fuel for each engine to be drawn from the corresponding auxiliary tank in each wing. It was established that the aircraft had been refuelled to full main tanks prior to the flight. Further, the pilot had advised in a telephone conversation with an engineer before the flight that the contents of both auxiliary tanks was 60 litres or less. All fuel tanks except the left auxiliary tank were ruptured during the impact sequence. About one litre of fuel was recovered from this tank.

Examination of the aircraft engines indicated that the right engine was under power at impact while the left engine was not. The mechanical condition of the engines indicated that they were capable of normal operation.

The PA-31 pilot's operating handbook states that the main fuel tanks must be selected for take-off. However, the behaviour of the aircraft, the position of the fuel selectors, and the information concerning the contents of the auxiliary tanks suggest that the pilot probably commenced the flight with the auxiliary tanks selected. As the flight progressed and fuel was used, intermittent un-porting of the fuel outlet lines occurred. This caused temporary fuel starvation, resulting in engine surging. These interruptions to engine power would have caused the aircraft to lose altitude, as described by witnesses, and airspeed. The event in which the aircraft rolled to the left and inverted is consistent with the right engine suddenly surging to high power when the aircraft was flying at a low airspeed while the left engine was delivering little or no power.

The pilot gained a PA-31 type endorsement in July 1992. At the time of the accident, he had logged a total of 37.8 hours flying multi-engine aircraft, including 35 hours on this aircraft type.

It was established that the pilot did not use a written checklist. Had such a checklist been used, the incorrect fuel tank selection may have been detected. Notwithstanding this, fuel system management is a basic and essential aspect of aircraft operation. In particular, fuel tank selection is a standard check item in the event of engine malfunction during flight. The pilot's apparent failure to select the main fuel tanks may be explained, at least in part, by his relatively low level of aeronautical experience, both overall and on type. Additionally, the pilot's information processing capacity may have been affected by the stressful situation in which he found himself. There were indications from the radio transmissions made by the pilot that he was in a highly anxious state when he reported that he was experiencing difficulties.

There were areas beneath the aircraft's flight path upon which a forced landing could have been conducted, albeit with probable aircraft damage. The pilot's failure to conduct a forced landing is considered a factor in the severity of the accident.

Evidence obtained during the investigation and the circumstances surrounding this occurrence suggest that the pilot did not have an adequate understanding of the aircraft systems.

Significant Factors

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

1. The pilot did not use a written checklist.

2. The pilot operated the aircraft with the auxiliary tanks selected when the fuel contents of these tanks was low.

3. The pilot failed to conduct a forced landing