Loss of control

Sequence of events

On the morning of 8 April, the pilot flew the aircraft from Archerfield to Gatton to practise aerobatics. That afternoon he carried out two flights during which he performed an aerobatic sequence known as a "free program", which is a sequence of manoeuvres designed by the pilot to meet certain criteria. The following morning, the pilot carried out a check flight with an instructor to obtain approval in accordance with Civil Aviation Regulation 155(3), to conduct aerobatics down to a minimum height of 1,500 ft above ground level. The minimum height for aerobatics is 3,000 ft above ground level unless this approval is obtained.

Early in the afternoon, the pilot commenced a flight to practise his "free program" again. The flight was observed from the ground by other pilots, one of whom was in radio communication with the pilot. On this occasion it was planned to critique the flight via the radio. The pilot on the ground, who was an experienced aerobatic pilot, stated that about halfway through the sequence he could see there was a problem. The aircraft had been inverted for 5 or 6 seconds after a half loop, before performing a half loop downwards and returning to upright flight. The observing pilot said that after reaching a horizontal attitude, the nose of the aircraft pitched up 40-60 degrees and the aircraft rolled to the right. He immediately queried the pilot as to what he was doing but received no response. The aircraft continued rolling slowly to the right and several more transmissions were made, but there was still no response. The aircraft then appeared to stabilise, and the pilot responded saying that he thought he might have blacked out, but that he was all right now. He then said that he thought he would land and have a rest for a while.

The landing appeared to be normal, and shortly afterwards the events of the flight were discussed with the pilot. The pilot said that he could remember the nose of the aircraft coming up to the horizon, but he did not remember anything after that until the aircraft was near the downwind leg of the circuit. A discussion was held on "G-LOC" (g-induced loss of consciousness) and its possible effects including disorientation. After further discussion the pilot decided that he would make another flight later that afternoon, as he did not appear to have suffered any aftereffects. A technique was then discussed which would reduce the g force during the manoeuvre and help prevent the situation from recurring during the sequence.

At about 1730 EST, the pilot took off to practice the "free program" sequence again. He was in radio communication with his brother, a highly experienced pilot and flying instructor. The pilot had said that he would increase the base height for the sequence as an added safety margin.  The sequence was commenced and proceeded normally and the manoeuvre which had previously caused the problem was performed with what appeared to be the suggested modifications. However, shortly after recovery to the straight and level attitude from the downward half-loop, the nose of the aircraft pitched rapidly to 10 degrees nose-up and the aircraft began to diverge slowly to the right with the nose continuing to rise slowly. The aircraft then commenced a continuous roll to the left. During this time, the ground observer made several unsuccessful attempts to contact the pilot by radio. The nose of the aircraft dropped, and the aircraft dived almost vertically into the ground from a height estimated by witnesses as 2,000-2,500 ft.

Wreckage examination

Examination of the wreckage did not reveal any defects or failures which may have been present before impact, and which could have contributed to the development of the accident.

The canopy had been previously broken when a radio detached from its mounting in the rear cockpit during aerobatics. Two parallel cracks extended laterally across the top of the canopy over the rear cockpit. A repair had been carried out by stitching the cracks with locking wire. The impact site and wreckage did not initially yield a representative quantity of Perspex (only about 20 per cent was found) and the site was then excavated.

Approximately 60 per cent by weight of the canopy Perspex was recovered from the impact crater, including the lockwire stitching with some Perspex still attached. The remainder of the canopy Perspex probably broke into fragments smaller than those recovered. A thorough ground and air search was carried out for Perspex under the flight path, but none was found. This evidence indicates that the canopy is unlikely to have failed in flight.

Impact information

Microscopic examination of both airspeed indicator faces revealed a pointer impact indicating 195 kts on one instrument. The aircraft impacted the ground vertically making a crater approximately 1.3 m deep. The impact forces were not survivable.

Medical

The pilot was assessed as medically fit for issue of a Class 2 medical certificate in December 1994 and appeared to be in good health on the day of the accident. Pathological examination showed evidence of coronary artery stenosis which could have affected the pilots g tolerance. However, medical opinion was that this could not be categorically stated to be the cause of a low g tolerance, or a contributing factor to the accident.

Research and analysis

From the available evidence it appears that the fatal flight was a repeat of the previous flight, except that on this occasion the pilot was unable to regain consciousness before the aircraft struck the ground. The aerobatic sequence up until the time of control loss was, in the opinion of witnesses, flown with more precision and was of a higher standard than on the previous flight. The manoeuvres were described as being tighter, and the pilot was probably generating more g than on the previous flight. He had been advised to reduce power during the manoeuvre that had previously caused the symptoms consistent with G-LOC, to enable the downward portion of the manoeuvre to be conducted at a slower speed and therefore reduced g. Witnesses said that a power reduction was heard during the inverted part of the manoeuvre, but the entry speed is not known.

A BASI Air Safety Research Report produced in February 1988, "The possibility of G-Induced loss of consciousness (G-LOC) during aerobatics in a light aircraft", stated that with sustained moderate rates of onset of g force, G-LOC is preceded by visual symptoms such as grey-out and then black-out (loss of vision). G-LOC is due to the reduced flow of blood to the brain when the magnitude of g in the vertical axis (Gz) passes beyond a particular value, the G-LOC threshold. Beyond this value, neurones fail to function in the absence of the oxygen replenishment provided by the normal blood flow and G-LOC occurs.

The report stated that with a rapid g onset (1g per second for more than 3 to 5 seconds) sudden loss of consciousness may occur without the prior visual warning symptoms. High performance aerobatic aircraft such as the Pitts S2 are quite capable of achieving rapid g applications, and rapid g onset is frequently experienced by competitive aerobatic pilots.

Centrifuge research has shown that there is a period of functional incapacitation following G-LOC lasting an average of 15 seconds. Full recovery may take 30 seconds or longer from the initiation of the manoeuvre that induces G-LOC. An aircraft descending vertically at an average speed of 120 kts will take approximately 15 seconds to reach the ground from 3,000 feet.

Gz tolerance is reduced by various factors such as fatigue, hypotension (low blood pressure), hypoglycaemia (low blood sugar), dehydration, and illness. The pilot was, by all accounts, in good health at the time of the accident, but it is not known what factors may have affected his g tolerance.

Conclusion

Findings

1. The pilot was correctly licenced and endorsed to carry out the flight.
2. The pilot probably experienced G-LOC on the flight previous to the accident flight and was subsequently advised on a means of avoiding a recurrence.
3. The accident flight was commenced some hours later for the purpose of the pilot practising the same manoeuvres as on the previous flight.
4. The aircraft was seen to go out of control at the same part of the sequence as on the previous flight.
5. The pilot did not answer repeated radio transmissions after the aircraft had gone out of control.
6. The pilot had a medical condition which may have reduced his tolerance to g.
7. There were no mechanical defects found which may have contributed to the development of the accident.

Significant factors

1. For reasons which could not be positively determined, but probably related to G-LOC, the pilot lost control of the aircraft.
2. The pilot did not regain control of the aircraft before it struck the ground.

FACTUAL INFORMATION

History of the flight

Pylon racing was conducted at Valley Field, south of Launceston, from 14-19 February 1995. Following completion of the races five T-28 aircraft were to be flown back to the mainland from Launceston. The plan was for the T-28s and a PA-31 to return to the mainland in loose formation.   The role of the PA 31 was to provide search-and-rescue support in the event of a T-28 ditching.  On the morning of the accident the formation leader briefed the pilots involved.  Take-off was to be in formation pairs, spaced a short time interval apart.  The first two pairs were to be T-28 aircraft.  The pilot of VH-LAO asked to be in the last pair which meant flying in formation with the accompanying PA-31 aircraft.  The flight was briefed as a normal departure, no mention was made of the performance of any aerobatic manoeuvres, nor did the pilot of VH-LAO discuss this possibility with any of the other formation members.

The formation take-off on runway 32 proceeded as planned with the pairs taking off at 15-20 second intervals.  Due to the performance of the PA-31 being lower than the T-28, the climb speed of the last pair was limited to about 120 knots.  Observers reported that the third pair made a normal take-off and initial climb, with the T-28 to the left of the PA-31.  At a height of about 500-600 feet the nose of VH-LAO was seen to rise and the aircraft started rolling to the left.  The left roll continued with altitude remaining about the same until the aircraft had rolled inverted. At about that stage of the roll, the nose started to drop.  After about 270 degrees of roll, the aircraft's attitude was very nose-low, and the rate of roll had decreased.  Observers indicated the pilot appeared to be trying to recover from the dive, but there was insufficient altitude to regain level flight before the aircraft, still 10-15 degrees nose-down, struck the ground.  Some witnesses thought the engine continued to operate at about the same power as at the start of the manoeuvre, but one person indicated the power was cut while on the final descent.

Damage to aircraft

The aircraft was severely damaged by impact forces and a post-impact fire.

Wreckage trail/wreckage examination

The aircraft struck the ground approximately 100 metres to the left of the extended runway centreline.  The wreckage trail, which extended for about 140 metres, was towards the north-east (about 90 degrees to the right of the take-off path).  The landing gear and flaps were in the retracted position. Examination of the wreckage did not identify any evidence of a pre-impact defect that could have affected the safe operation of the aircraft.  Also, nothing was found that could have caused an uncommanded roll to the left.

Weight and balance

At the time of take-off, the aircraft was loaded to approximately the maximum permissible take-off weight.  Balance was within limits.

Passenger background

The passenger did not have any pilot qualifications.

Pilot history

During 1986 the pilot had completed part of an RAAF pilot course.  He was suspended from this course following unapproved low flying manoeuvres.  Subsequently he took up civilian flying.  At the time of the crash, he held a Grade One fixed wing instructor rating.  In October 1989 he had received an endorsement to conduct basic aerobatic manoeuvres.  At about the same time he also received an endorsement authorising him to teach basic aerobatic manoeuvres, including loops, aileron rolls, slow rolls and barrel rolls.

After obtaining his aerobatic approvals, his logbook showed he had done only a moderate amount of basic aerobatic flying and aerobatic instruction.  Most of this was in simple general aviation type aircraft such as the Cessna 150. These types are small and light and have a very moderate performance compared to the much heavier, higher powered and faster T-28. The pilot did not hold an approval to conduct aerobatic manoeuvres below 3,000 feet.

Reports indicated that during his time as an instructor he had occasionally been involved in abnormal flight activities.  These included continuing flight into poor weather conditions and making multiple very low passes over a country airstrip.  On two separate occasions with student pilots, he had taken the controls and conducted a 360-degree roll manoeuvre.  The aircraft type on which these manoeuvres were flown was not approved to perform aerobatic manoeuvres.  For approximately the last two years the pilot had been employed as a co-pilot on regular public transport operations, but he still maintained an active interest in private flying.

The pilot obtained a T-28 endorsement on 2 April 1994.  During the endorsement he carried out two or three horizontal roll manoeuvres. These were executed from an entry speed of 180 knots, the normal speed for entering such rolls.  A few days later a flight included horizontal roll manoeuvres, using entry speeds of about 190 knots. The pilot did not fly the T-28 again until 12 February 1995, when he did a short flight with the owner of another T-28.  On 15 February he was permitted to fly this T-28 to Launceston with the owner-pilot on board.

At Valley Field the pilot received a low-level check required by the organisers for participants in pylon races.  This check was only for pylon racing and did not include any aerobatic manoeuvres. While at Valley Field, he was trained to fly the pace plane used to start aircraft in the T-28 pylon races. The pilot did not fly in pylon race events but did fly the pace plane on four or five occasions.  The pace plane leads the competitors to the start line and then pulls up out of the way.

The pylon races were conducted at a minimum height of 200 feet, but competitors were permitted to descend to 100 feet on the straight leg.  The races ended on the straight, at which stage the aircraft were at relatively high speeds. At the end of the race a pull up/climb manoeuvre was required, to a height of 2,000 feet.  While doing this pull up/climb, some competitors executed a 360-degree roll manoeuvre.

On one occasion at Valley Field the pilot, while flying a T-28, was observed to carry out a barrel roll aerobatic manoeuvre at a height of about 1,500 feet.  On the day prior to the accident, he flew as a passenger-observer in a T-28 flown by a well-known American air-show pilot, who conducted an impressive low-level aerobatic display.

Medical information

The pilot passed his last pilot licence medical examination on 18 April 1994.  Following the accident a post-mortem examination was performed on both the pilot and the passenger.  No evidence was found to indicate that either the pilot or the passenger had any medical condition that might have contributed to the accident.

Meteorological information

Information from the Bureau of Meteorology indicated that at about the time of the accident the weather was fine. The temperature was about 19 degrees, the surface wind was from 270 degrees at three knots and there were five octas of cumulus cloud with a base of 3,500 feet.  There was no significant turbulence.

Wake turbulence

The only possibility of wake turbulence would have been from the preceding T-28 aircraft, but this could not have caused the manoeuvre flown.

Survival aspects

The accident was not survivable.

Tests

The normal entry speed for roll manoeuvres in the T-28 was about 180 knots.  The speed at the time of the roll that preceded the accident was probably about 120 knots.  Information was sought on the roll capability at the slower speed, and limited testing was done by an experienced competition aerobatic pilot who also owned a T-28.  The aircraft was loaded to a similar weight.  Tests showed that an aileron roll could be completed without losing significant altitude.

Barrel rolls to the left were attempted but in every case the recovery occurred with a 60-70 degree deviation to the right.  The testing pilot used different methods of entry for these manoeuvres.   Significant altitude losses could not be avoided and were approximately 600-800 feet.  On one of the rolls the throttle was closed about two-thirds of the way through the roll. This resulted in an additional altitude loss of about 200 feet.

ANALYSIS

The weather conditions were good.  The only wake turbulence that could possibly have been encountered was from preceding aircraft, which were of the same type.  This could not have caused an upset of this magnitude.  No evidence was found to suggest any abnormality existed with the aircraft.  There was no evidence to suggest that anyone, apart from the pilot, had any idea that an aerobatic manoeuvre might be conducted after take-off.

Evidence indicated that on occasions the pilot had carried out unapproved flying activities. The information indicated a strong probability that on the accident flight the pilot had attempted to carry out a low altitude roll manoeuvre.  He was not approved to do low-level aerobatics, and his background training did not prepare him for such manoeuvres.  He had little experience on the T-28, which is a large, heavy, high powered, single engine aircraft type.  From the evidence, it appears that the pilot was attempting a roll manoeuvre and mishandled the aircraft.  The high degree of hazard associated with attempting such a manoeuvre at such a low height was significantly increased by the relatively low airspeed at which it was started.

CONCLUSIONS

Findings

1. The pilot was correctly licensed and endorsed to fly the T-28.
2. The pilot held an aerobatic endorsement but did not hold an approval to conduct aerobatic manouevres below3,000 feet.
3. The pilot had a moderate level of basic aerobatic experience and had little aerobatic experience on the T-28.
4. The weather conditions were good and did not contribute to the accident.
5. There was no evidence of any defect in the aircraft.
6. There was no evidence to show that anyone apart from the pilot had any idea that an aerobatic manoeuvre would be performed after take-off.
7. The evidence indicated that the manoeuvre carried out was intentionally initiated by the pilot.
8. The relatively low airspeed at the start of the manoeuvre was well below that recommended for the conduct of horizontal rolls.

Significant Factors

1. The pilot had limited experience on the aircraft type and had never been trained for, or approved to do, low-level aerobatic flight on any type.
2. The pilot probably attempted to carry out an aerobatic manoeuvre at an unsafe height.

The aircraft was operating from an agricultural airstrip 600 ft above mean sea level, spreading superphosphate over moderately steep undulating terrain. The duration of each flight was 6-7 minutes. The accident flight was the seventh and probably intended to be the last for the day.

A witness, who was situated under the flight path, reported that the aircraft was tracking east-north-east in what appeared to be normal flight. Her attention was distracted for a few moments and when she next saw the aircraft it was in a near vertical dive with the upper surface of the wings facing her. The aircraft then struck the hillside and burst into flames.

Examination of the wreckage did not reveal any pre-existing defect which may have contributed to the accident. Impact marks on the propeller indicated that the engine was operating at impact.

The superphosphate load remained in the hopper and the emergency dump system actuating lever was in the closed position. Inspection indicated that the dump system was serviceable prior to impact.

Calculations indicated that at the time of the accident the aircraft, although heavily loaded, was operating within the flight manual maximum weight limitation.

A light north-easterly wind was observed at the airstrip. However, at the accident site, which was about 250 ft higher, the wind was a moderate west-north-westerly. Sky conditions were clear with a visibility of 30 km.

The aircraft probably experienced windshear and turbulence as it encountered a quartering tailwind approaching the ridgeline. The result would have been a reduction in climb performance, and it is likely that the pilot attempted to turn the aircraft away from the rising terrain. During the turn it appears that the aircraft stalled and that the pilot was unable to regain control before it struck the ground.

The reason the pilot did not dump the load when the climb performance was reduced could not be determined.

Significant factors

The following factors were determined to have contributed to the accident.

1. Shifting wind conditions conducive to windshear and turbulence were present in the area.

2. The aircraft was climbing at near to maximum allowable weight.

3. Control of the aircraft was lost with insufficient height available to effect a recovery.

FACTUAL INFORMATION

History of the Flight

The pilot had carried out a local flight with his wife, and a number of circuits and landings were made at the property airstrip. His wife then alighted from the aircraft and went inside the residence which was adjacent to the runway. The pilot had said he intended to do a few more circuits and landings. The aircraft was heard taking off a few minutes later, and witnesses located to the south of the strip saw it making what appeared to be a "flat" left turn onto a northerly heading. The witnesses said the aircraft was about 100 m above ground level and that the engine noise level was low. The aircraft then went out of their field of view. A short time later, the crashed aircraft was discovered by a neighbour who was driving towards the property. The weather was fine and there was a light southerly wind at the time.

Impact Information

The aircraft impacted flat ground whilst upright, descending, and slightly left-wing low. It then bounced beneath powerlines which were approximately 11 m high, and across a road in a direction of 315 degrees. The canopy, ailerons, and various small pieces of aircraft separated before the aircraft came to rest upright, 52 m from the initial impact point. There was no fire. The point of impact was about 300 m to the north of the airstrip which is aligned northeast-southwest.

Wreckage Examination

Detailed technical examination of the aircraft was carried out. Specialist examination of a portion of the exhaust pipe confirmed that the engine had been operating at the time of the accident, albeit at low power, as evidenced by the nature of damage to the propeller and the spinner. There were no mechanical defects found that would have prevented the engine from operating normally.

The structure of the aircraft was examined, and no pre-existing structural defects were found. However, a fibreglass fairing from the base of the tailfin had become detached in such a way that it could have restricted the movement of the elevator control surface. Two small screws had pulled out of the lower rear edge of the fairing. A white paint witness mark on the fairing matched the top inboard edge of the left elevator and showed that the two surfaces had at some stage been in contact. When the fairing was placed against the elevator with the witness marks aligned, it was evident that the detached section of fairing could restrict the upward movement of the elevator. In addition, the elevator trim tab was found in the fully down position.

ANALYSIS

The available evidence does not support pilot incapacitation being a factor in this accident.

Examination of the aircraft indicated that it was capable of operating normally at the time of impact. However, there is a possibility that a fibreglass fairing had detached in flight and restricted up movement of the elevators. The elevator trim tab appears to have been fully down at the time of impact, as it had punctured the rudder surface in that position. This trim-tab position is that which normally gives full nose-up trim. However, if the elevator were restricted in upwards movement and only the tab was movable, then downwards movement of the tab would result in a nose-down effect, the reverse of that expected by the pilot.

A control working in the opposite sense to that expected would be extremely confusing to the pilot. A pilot placed in this situation might reduce the power, because the effect of controls is influenced by power and airspeed. The witnesses who saw the aircraft said it seemed to be making very little noise and was flying slow and at a low altitude. The effect of reducing power, however, would be for the aircraft to adopt a nose-down attitude, compounding the difficulty faced by the pilot.

The pilot was possibly attempting to fly the aircraft back onto the airstrip. From the position of the aircraft before ground impact, and in relation to the airstrip, the pilot would have had to make a left descending turn to align the aircraft with the strip. At this point it would have also been necessary to make a power reduction. This would have further compounded the nose-down tendency of the aircraft which the pilot would have instinctively tried to overcome with the use of nose up trim, further exacerbating the nose-down tendency.

While the above hypothesis offers an explanation for the circumstances of the accident, it relies on an assumption that the fibreglass fairing became detached before impact. However, it has not been possible to determine conclusively whether the fairing became detached before or after impact.

CONCLUSIONS

Findings

1. There was no evidence to support pilot incapacitation being a factor in this accident.
2. The aircraft was operating normally on the flight preceding the accident flight.
3. The weather was fine.
4. The aircraft was seen operating at a low altitude with low power setting immediately before the time of the accident.
5. The engine was capable of normal operation but was delivering low power at the time of impact.
6. The nature of the ground impact indicates that the pilot lost control of the aircraft.
7. The elevator trim tab was in the fully down position at the time of impact.
8. The tailfin fairing was detached and showed evidence of having been in contact with the left elevator control.
9. It could not be determined if the fairing became detached before or as a result of the impact.

Significant Factors

1. Control of the aircraft was lost at low altitude for reasons which could not be determined.
2. The pilot was unable to prevent ground impact.

The pilot was scheduled to spray 50 acres of rice with Londax. The paddock was level and open with no significant obstructions.

The aircraft departed Wakool with a fuel quantity of 200 L (144 kg) of AVGAS and approximately 510 kg of a mixture of Londax and water in the spray hopper.

About five minutes before arriving overhead the treatment area, the pilot established radio contact with the farmer, who was to act as the swath marker. After completing the first spray run in a northerly direction, a second run was completed in a southerly direction. During this run the marker observed that the spray equipment was operating. The aircraft departed the area, heading approximately towards Wakool, however, the pilot did not advise the marker of any reason for departing. When unable to re-establish radio contact, the marker arranged for the operator at Wakool to be advised.

The operator's chief pilot commenced an airborne search from Wakool to the paddock where the aircraft had been spraying.  Finding nothing, he returned to Wakool and arranged for the police to be advised that the aircraft was missing, before re-commencing the air search.  He found the wreckage near the base of trees about 2 km south-east of the spray site, close to a private airstrip.  This strip, aligned 080/260 degrees, was about 850 m long and suitable for a Pawnee Brave landing.

The accident site was in trees about 100 m east and about 50 m south of the unmarked 260 degree strip threshold.

At the time of the accident, the temperature was about 20 degrees Celsius.  The weather was fine with a north-north-easterly wind of 5 to 10 kts.  There were no significant gusts, no significant cloud and visibility was excellent.

The aircraft was within its approved centre of gravity and gross weight limits at the time of the accident and the fuel on board was adequate for the flight.

The pilot was appropriately endorsed on the Pawnee Brave, however, he was relatively inexperienced in agricultural operations.  His Grade 2 Agricultural Rating was issued on 5 July 1994, and he completed 10 hours under direct supervision on 28 October 1994, and then a further 8 hours unsupervised prior to accident.

There were no known witnesses to the accident.  Examination of the impact marks on the trees, fence and the ground indicated that the aircraft had struck the trees while tracking 230 degrees, displaced to the south of the extended strip centreline, with an angle of bank of about 110 degrees to the right and an attitude of 30 degrees nose down.  The short distance of about 20 m from the initial impact point to the final resting position of the wreckage was indicative of a relatively slow horizontal velocity.  Damage to the propeller was consistent with the engine producing power at impact. Flaps were ascertained to have been at 15 degrees which was the recommended landing setting for a heavily loaded Pawnee Brave. There was no fire.

The wreckage was examined by engineers at the accident site and the engine and propeller assemblies were later subjected to more detailed inspection.  No pre-existing faults were found with the aircraft which may have contributed to the accident.

The aircraft was not equipped with a survival beacon.

It is possible that the pilot was intending to land into the west on the airstrip near which he crashed.  The approach path for a landing to the west was over open, flat, dry pasture clear of obstacles.  However, on the southern side of the approach path a tree line converged towards the airfield's southern boundary.

Given the impact position in relation to the airstrip, it is possible that the pilot conducted a low-level right turn onto final for a downwind landing into the west and that the aircraft stalled at a height from which recovery was not possible before ground impact.

Significant Factors

The following factor was considered relevant to the development of the accident:

1.   The pilot probably underestimated the effect of the tailwind component during the turn onto final. 

The flight was planned as a regular public transport flight from Williamtown to Lord Howe Island. After departure, the pilot reported to Sydney Flight Service that he intended climbing to flight level 230. However, as the aircraft passed flight level 200 the pilot advised flight service that he was now climbing to flight level 210.

The pilot did not report at the first scheduled position code named 'Shark' on time but subsequently advised that he was descending to flight level 130. Shortly afterwards, the pilot reported having crossed 'Shark' and provided an estimate for the next scheduled position, 'Shrimp'. He also stated at this time that the aircraft was maintaining flight level 160. No further communications were recorded by Air Traffic Services from the aircraft. However, during the above period the pilot was in radio communication with two other company aircraft, both bound for Lord Howe Island. One aircraft was crewed by the company managing director and the company chief pilot.

When Sydney Flight Service did not receive the 'Shrimp' position report communications checks were commenced, and following the failure of these checks to establish contact with the aircraft, a search-and-rescue uncertainty phase was declared.

After the managing director arrived at Lord Howe Island, he contacted Melbourne Rescue Co-ordination Centre in response to a request from that centre and inquired about the aircraft. The crews of both other company aircraft subsequently reported hearing a radio transmission from the pilot of VH-SVQ stating that he had 'lost it'.

An extensive air and sea search failed to locate the aircraft or its occupants. Only a small number of pieces of the aircraft were found floating on the sea surface.

The investigation determined that the flight was not a regular public transport flight as the company did not hold the required approval from the New South Wales Air Transport Council to operate such flights over the Williamtown to Lord Howe Island route.

The factors that directly related to the loss of the aircraft could not be determined. However, a number of factors relating to the operation this flight, the operation of the company and the oversight of that operation by the regulator were identified.

The report concludes with a number of safety recommendations.

1. FACTUAL INFORMATION

History of the Flight

The aircraft had returned to Walgett late on the afternoon of the day before the accident, having completed a five-day charter to the Gulf of Carpentaria.

On the day of the accident the Walgett Aero Club held a barbecue and flying competition. The pilot of the Cessna 337 indicated that he did not intend to take part in this competition. Later in the day, he advised the flying instructor who was supervising the flying competition that he wished to carry out a low pass over the aerodrome. The instructor had no objection to this request.

At approximately 1550 hours the pilot took off from runway 18 with three passengers. After what appeared to be a normal circuit and approach, the aircraft made a high-speed pass, with the landing gear retracted, parallel to runway 18 at approximately 20-30 ft above ground level (AGL).

At 100-150 m from the runway intersection, witnesses observed the aircraft enter a steep climb. Witness estimates of the attitude adopted by the aircraft ranged from 40 to 70 degrees nose-up. The aircraft remained in this high nose attitude for 6-10 seconds until an altitude of approximately 700-1,000 ft AGL was reached.

At this point the aircraft's left wing dropped, the nose lowered steeply, and witnesses noted that the engine noise reduced significantly. The instructor supervising the competition stated that after the aircraft appeared to stall, he saw the rudder surface on the tailplane fully deflect in a direction opposite to the observed rotation. The aircraft rotated slowly to the left in an extreme low-nose attitude. Another witness commented that when the aircraft had descended to approximately 200-300 ft AGL, it appeared to adopt a slightly higher nose attitude. This change of attitude was transitory. The nose attitude lowered again quickly, and the aircraft impacted the ground in a very steep nose-down attitude.

Wreckage Examination

The wreckage was located on the Walgett aerodrome, 42 m to the south of the runway strip markers of runway 36.

The aircraft had impacted the ground in a steep nose-down attitude, wings level, with negligible rotation after impact. It did not slide along the ground after impact. The structural deformation was related to the onset of impact loads.

All aircraft extremities, including doors and all control surfaces, were present in the wreckage.

Technical examination of the engines and propellers showed them to be capable of normal operation prior to the impact. No indication was found of any aircraft system malfunction which may have contributed to the accident. However, destruction of the cockpit and the instruments precluded the individual systems' pre-impact status being determined.

The accident was not survivable.

Pilot Information

The pilot was the holder of a commercial pilot licence (aeroplanes). He held a valid medical certificate with a requirement to wear glasses. He was also the aircraft owner. He held an air operator's certificate, re-issued by the Civil Aviation Authority on 30 June 1994, which allowed him to carry out charter operations in VH-DRI.

The last entry in the pilot's logbook was made on 23 August 1994. At this time, he had accumulated approximately 3,200 hours total flight time (3,100 hours in single-engine aircraft and 3,050 hours as pilot in command).

The pilot had completed his endorsement training on the C337A on 11 June 1994. This was his first multi-engine endorsement. At the time of the last logbook entry, he had accumulated 102 hours in the aircraft type, most of which was in VH-DRI.

On 27 June 1994 the pilot undertook a flight check with a Civil Aviation Authority flying operations inspector in order to complete his chief pilot requirements and to include the C337A on his air operator's certificate.  On this occasion it was considered that the pilot met requirements but was to complete further training in the handling of emergency procedures. The CAA pilot file notes that an approved air test officer completed this training and advised that the pilot's handling of emergency procedures was satisfactory. The variation on the air operator's certificate was issued on 30 June 1994.

Post-mortem examination of the pilot revealed the presence of marked atherosclerosis of the coronary artery but there was no evidence of a coronary occlusion. Some alcohol was detected in liver and muscle fluid which was used for testing as sampling of blood or vitreous humour was not possible.

Weather

The weather on the day of the accident was mild with a temperature of 23 degrees C. A high-pressure system was located over south-east Australia.  The sky was clear and there was a gentle breeze from the south-west of up to 5 kts. The visibility was good.

2. ANALYSIS

Aircraft Handling Characteristics

Cessna aircraft are generally docile in most areas of handling. A number of pilots who had extensive experience on the C337, including flight instructors, agreed that placing the aircraft in the attitude that was witnessed on the day of the accident would have resulted in a much more aggravated stall than would be experienced as a result of a stall from straight and level flight. Witness statements agree that the aircraft was being operated at or near full power during the manoeuvre. It could not be ascertained whether the pilot reduced power before or after the point of the stall.

As the aircraft was seen to adopt and maintain a very high nose attitude, the stall that resulted would have occurred quickly due to the rapid loss of airspeed. The height required to recover from such a stall would have been significant and probably greater than that which was available.

The observed full deflection of the rudder surfaces was consistent with the actions of a pilot who may have been attempting to counter an incipient spin.

The C337A was certified under the United States Civil Aviation Regulations Part 3 which preceded the Federal Aviation Regulations Part 23.  As this aircraft was considered a multi-engine aircraft, it was not required to undergo spin testing as part of its type certification. Consequently, no data is available to indicate the typical height loss expected as the result of a spin.

VH-DRI was certified for operations in the normal category. The flight manual stated that operation shall be limited to normal flying manoeuvres but may include straight and steady stalls and turns in which the angle of bank to the horizontal is 60 degrees or less. Other acrobatic manoeuvres shall not be performed.

The Australian Civil Aviation Regulations define aerobatics as "manoeuvres intentionally performed by an aircraft involving an abrupt change in its attitude, an abnormal attitude, or an abnormal variation in speed". A glossary of aeronautical terms used for accident investigation by the US Department of Transportation Safety Institute, Oklahoma, defines an aerobatic manoeuvre as "a pre-planned flight manoeuvre in which the aircraft exceeds either 60 degrees of bank or 30 degrees of pitch".

The observed manoeuvre is consistent with the pilot's probable intention to attempt a wingover or possibly a stall turn. By any of the above definitions, wingovers and stall turns are aerobatic manoeuvres and are outside the normal flight envelope for this aircraft type.

Fuel System

VH-DRI was equipped with a main fuel tank of 174 litres usable fuel capacity in each outboard wing panel and a sump tank of 2.7 litres fuel capacity in the lower portion of each tail boom. Fuel flows to the sump tanks via two outlets in each main tank, one at the bottom forward edge and one at the bottom rear edge of each tank. Fuel then flows from the sump tanks through a bypass in each electric auxiliary fuel pump (when the pump is not operating) to selector valves located at the wing roots.

The inclusion of sump tanks reduces the risk of interruption of the fuel flow when the aircraft is placed in a range of flight attitudes including those that were witnessed during the accident. Had fuel been unable to drain from the main cells to the sump tanks due to the observed manoeuvre, it would have taken approximately two minutes and thirty seconds to unport the fuel lines in the sump tanks with the aircraft operating at full power, assuming that the sump tanks were full at the time the aircraft entered the nose-high attitude. As witnesses reported that the elapsed time between the aircraft entering the pull-up and the nose lowering at the top of the manoeuvre did not exceed 10 seconds, it is unlikely that engine failure would have occurred as a result of interruption to the fuel flow due to unporting of the fuel outlets from the main fuel cells.

VH-DRI was also equipped with an optional auxiliary fuel tank (68 litres usable capacity) in each wing between the cabin and the tail boom. The auxiliary tanks feed directly through the engine-driven fuel pump to the engine. The fuel from each auxiliary tank drains via a single outlet near the bottom of the tank approximately halfway between its forward and aft edges.

If the auxiliary tanks were selected it might be possible, at very low fuel levels, to unport the outlet to the tanks if the aircraft were in a very nose-high attitude for a considerable length of time.

The C337A engine is fuel injected. The fuel injection system delivers fuel, under pressure, to the inlet manifolds of the engine and is unlikely to be affected by the placement of the aircraft in unusual attitudes.

It was not possible, due to the severity of the damage, to accurately determine the position of the fuel selectors in the cockpit prior to impact. The flight manual stated that main fuel tanks should be selected for take-off, landing and the first 60 minutes of flight. It is therefore most likely that the main tanks were selected for this flight. Calculations have determined that there was approximately 150 litres of fuel on board the aircraft prior to its last flight. It was also not possible to accurately determine the quantities of fuel in each tank as every fuel cell was ruptured on impact with little or no fuel being observed in each cell. Nonetheless, it is unlikely that either engine failed due to fuel starvation.

Weight and Balance

Calculations of the weight and balance of the aircraft were based on the following:

1. The pilot was in the front left pilot seat while the three passengers occupied the right front seat and the two centre seats directly behind.
2. According to fuel agent records, the pilot purchased fuel at Birdsville but did not purchase fuel after return to Walgett. The aircraft flew direct from Birdsville to Walgett, a flight time of four hours. Assuming full tanks at Birdsville and a fuel consumption of 85 litres per hour (based on the pilot's operating handbook), fuel remaining at Walgett was calculated at 154 litres. After return to Walgett the pilot told the passengers that the aircraft had consumed 90 litres per hour. The pilot also informed the instructor supervising the flying competition that the aircraft had approximately 150 litres of fuel on board.
3. A bag of flour was found in the cabin area of the aircraft.  It was estimated to have weighed 20 kg.

Using these figures, it was determined that the aircraft was within weight limits but that the centre of gravity (c.g.) was outside limits, marginally forward of the c.g. envelope.

Having more or less fuel on board would not have significantly affected this result, with the c.g. close to or just forward of the c.g. envelope.  The only factor that would have made a significant difference to the position of the

c.g. would have been the seating position of the passengers. Had they been seated in the rear row of seats or some combination of centre and rear seats, the position of the c.g. would have moved to within the envelope.

Despite this finding, discussions with C337 pilots indicate that it is unlikely that the forward position of the c.g. had a significant effect on the handling characteristics of the aircraft.

Seat Mechanism

There have been documented occasions when control seats have dislodged from their previously locked position and moved backwards when loads have been imposed by rotation for take-off, or by g-loads in aerobatic manoeuvres or in turbulence. This has been attributed to excessive wear in the seat adjustment mechanism.

A sudden rearward movement of the seat could make it difficult for the pilot to reach the control column. There is also the possibility that the pilot could instinctively grab at the control column in an attempt to counter the seat movement, causing an abrupt nose-up change in attitude.

The Civil Aviation Authority issued an Airworthiness Directive (AD) in relation to the seat adjustment mechanism in September 1988 (AD/CESSNA 337/27 Amt. 1 Seat Adjustment Mechanism). This AD was required as a periodic inspection every 100 hours or 12 months, whichever occurred earlier.

The aircraft logbook indicated that the AD had been complied with. The most recent inspection was carried out on 5 July 1994, approximately three months prior to the accident.

Due to the severity of the damage to the cockpit area, it was not possible to determine if the seat had dislodged, in flight, from the desired position.

Pilot Performance

Estimates of any degree of pilot performance degradation due to the presence of alcohol in liver and muscle tissue should be treated with caution. It was therefore not possible to accurately determine the blood alcohol level.

There was no evidence to suggest that the pilot had consumed alcohol on the day of the accident.

There was no evidence that pilot incapacitation was a factor in the accident.

Weather

The weather was not considered to be a contributing factor to this accident.

Summary

As the aircraft was observed to adopt a very nose-high attitude and to sustain it, airspeed would have reduced significantly. As the pilot attempted to turn out of the nose-high attitude, the observed subsequent flight path of the aircraft was consistent with a stall and spin. The altitude of the highest point in the flight path was insufficient to permit the pilot to effect a recovery from a spin.

Loss of engine power at or near the highest point in the flight path would have reduced the power available during the manoeuvre intended to bring the aircraft out of the nose-high attitude. This could have increased the rate of airspeed loss and may have slightly advanced the time at which control was lost. It is doubtful, however, that an engine failure would have precipitated the loss of control.

3. CONCLUSIONS

Findings

1. The pilot held a valid licence and was endorsed on the aircraft type.
2. The pilot carried out a manoeuvre for which the aircraft was not certified.
3. The aircraft appeared to stall at the highest point in its flight path.
4. The aircraft descended in a steep nose-low attitude and impacted the ground shortly after.
5. The aircraft was within maximum weight limits.
6. The centre of gravity of the aircraft was marginally forward of the forward limit of the c.g. envelope.
7. The engines and propellers were capable of delivering power prior to impact.
8. No other pre-existing airframe or system malfunction that could have directly affected the flight was found.
9. There was no evidence of a medical condition that could have affected the pilot's ability to control the aircraft.

Significant factors

The pilot lost control of the aircraft at an altitude which was insufficient to permit a recovery before the aircraft impacted the ground.

The aircraft was operating from a 400 m grass airstrip orientated approximately 080 degrees M. The strip was on a property where the aircraft had been in storage for some six months prior to the accident. However, the pilot flew the aircraft for about 1.5 hours a few days before the accident. This activity included a flight to Mareeba Airport where the aircraft landed and the pilot borrowed some tools, including a soldering iron, to work on the aircraft. No information about the work done was available. However, the aircraft later departed Mareeba and apparently functioned normally during the return flight to the property.

On the day of the accident, the pilot secured two jerry cans of fuel in the right seat of the aircraft and placed two carry bags on the shelf behind the seat. The pilot indicated to witnesses that he intended to fly one circuit and land, bid farewell to those at the strip, and then depart for Innisfail, his first intended landing point.

A witness reported that the pilot started the engine and allowed it to idle for between 3 and 5 minutes before commencing the take-off. As the aircraft accelerated for take-off, it drifted towards the left side of the strip, but the pilot corrected this, and the aircraft became airborne. The witness, who had observed the aircraft take off from the strip a number of times, considered that the length of strip the aircraft used to become airborne was greater than he had observed on previous occasions. He believed, also, that the aircraft did not climb as well as usual and was making little headway against the wind. At an estimated height of 20 m above ground level, the aircraft entered a left turn at about 30 degrees angle of bank. When heading approximately north-west, it suddenly rolled further left and spiralled to the ground.

The Bureau of Meteorology advised that, based on observations taken at Mareeba Airport on the day of the accident, the estimated weather conditions around the time of the accident were: wind, south-easterly at 5-10 kt, temperature 19-20 degrees C, and relative humidity 80 percent.

Reports from Mareeba Airport indicated that the wind was gusty at times during the day, making the conditions unsuitable for circuit flying. Witnesses at the accident site reported the weather conditions as a light south-easterly breeze with occasional drizzle, although the sun was shining when the aircraft took off. This information appears to be supported by a photograph taken of the aircraft shortly before take-off. However, other photographs taken shortly after the aircraft took off, and looking towards the south-east, show a darkened sky and low cloud.

Examination of the aircraft wreckage did not reveal any abnormality which might have contributed to the accident. Metallurgical examination of a section of the exhaust pipe confirmed that the engine was developing power at impact.

The aircraft engine was fitted with a carburettor heat control. This was found in the off position during the wreckage examination. The atmospheric conditions which existed at the time of the accident were such that serious carburettor icing was likely at idle engine power. It is possible, therefore, that there was a buildup of carburettor ice in the period the engine was idling before take-off. However, any such buildup would cause a decrease in engine performance by lowering the available maximum RPM, an effect which would be evident to the pilot via the cockpit engine instruments. The photographs referred to above show that the aircraft became airborne well before the end of the strip. The presence of carburettor icing cannot, therefore, be confirmed. However, any reduction in engine power would have reduced aircraft performance, particularly during a climbing turn.

The described behaviour of the aircraft in spiralling to the ground is typical of loss of aircraft control following wing stall. The height at which the event occurred would have precluded recovery to normal flight.

A further possibility is that the aircraft was affected by a change in wind conditions or turbulence and that the pilot lost control of the aircraft in these changing conditions.

Factors

The following factors are considered relevant to the development of this occurrence:

1. For reason(s) which could not be positively determined, the pilot lost control of the aircraft during a climbing turn after take-off.
2. The height at which the loss of control occurred precluded the pilot recovering the aircraft to normal flight.

Factual information

The glider was launched by aerotow for a local flight and at about 2,000 ft it was observed by the tug pilot to release and to fly straight and level.

Sometime later, the pilot a of glider flying in the Waikerie circuit area reported that there was a damaged glider in a field about 3 km south of the airfield. The tug pilot immediately commenced an aerial search and on locating the severely damaged glider, alerted emergency services. The pilot had not survived the impact.

On-site examination of the wreckage indicated that the glider had impacted the flat, grass covered terrain at high speed in a vertical, or near vertical dive. There was no evidence that the aircraft had been rotating immediately at impact and all aircraft components were located at the site except for the radio battery. Impact forces had destroyed the cockpit area, partially detached both wings, and broken the fuselage near the wing trailing edge.

The wreckage was recovered to a hangar at the airfield and inspected by engineers. Despite the extreme disruption of the airframe, all primary control system components were identified and examined. Many components had fractured during impact, however, it was determined that the control systems were all correctly connected prior to impact. The damage to componentry precluded a conclusive examination for jamming or obstruction of the control systems. The radio battery location was in an area of the aircraft well away from the control systems. The battery to electrical system connector wiring was found torn apart, consistent with separation of the battery from the aircraft at impact.

The glider had a valid maintenance release, and no maintenance deficiencies were identified. No evidence was found to suggest that a pre-existing failure of an aircraft component contributed to the accident.

The pilot was one of a group of foreign glider pilots visiting Waikerie for a period of intensive flying. He was correctly licensed and endorsed on the type. Prior to being approved to conduct solo flights at Waikerie, he was given flight instruction which included stall and spin recovery techniques.

A witness reported that on the day before the accident, a similar glider flown by the same pilot had pitched steeply nose-down and lost height immediately after release from the aero-tow.  The pilot was severely shaken, and he was subsequently debriefed by instructional staff before undertaking further flying.

The tow had appeared normal to the tug pilot who also briefly observed the glider in straight and level flight following the release. The descent was apparently not observed; however, in the light of the witness report of the previous day's incident, the possibility that the pilot lost control of the glider cannot be discounted.

The reason for the dive and the inability of the pilot to regain control of the glider could not be determined.

After an uneventful flight from Merimbula, the pilot overflew the airstrip to check the windsock. He noted that the windsock was swinging erratically around a mean direction of about 30 degrees left of the strip direction (approximately 140 degrees magnetic). The pilot then positioned the aircraft on the downwind leg to land towards the south-east.

Neither the pilot nor the surviving passengers reported any significant turbulence during the circuit. The pilot indicated that he was aiming to touch down about 100m in from the end of the strip. At what he thought was about 100 feet above the level of the strip, at an indicated airspeed of 85-90 kt, and with full flap selected, the aircraft rapidly lost altitude and landed very heavily a short distance in from the end of the strip. It then bounced, becoming airborne again.

The pilot reported that he was startled by the heavy landing. When the aircraft became airborne again, he momentarily applied power to go around but saw that the aircraft was headed towards trees, so he closed the throttle. His memory from this point was not complete but he did recall applying full right rudder, with no apparent effect, as the aircraft headed towards the trees. He also recalled the stall warning operating twice - once during the sink which led to the heavy landing, and again when the aircraft bounced.

Examination of the scene of the accident revealed that after the aircraft bounced, it contacted the ground again, left wing tip first, about 150m beyond the initial impact point and heading towards the left side of the strip. By this stage, the aircraft had developed a marked right skid. It continued in this manner across a dirt mound at the left edge of the strip and struck trees. The principal impact occurred when the aircraft fuselage just forward of the right-wing root struck a large tree, causing severe deformation to the right-side cockpit area.

The aircraft was equipped with an autopilot and a yaw damper. The pilot had conducted the cruise section of the flight with both these aids engaged. On arriving overhead, the strip he had disengaged the autopilot via the control wheel disengage switch. In normal operations, this switch also disengages the yaw damper. On this occasion, the pilot had checked that the autopilot had disengaged but did not confirm the yaw damper had disengaged. The pilot expressed the view that the difficulty he experienced in attempting to regain directional control when the aircraft was heading towards the left side of the strip could have been due to a problem with the yaw damper. However, a check of the yaw damper system after the accident did not reveal any faults.

The surface wind at the time of the accident was estimated by the Bureau of Meteorology to have been 110/15 knots.

The sequence of events described by the pilot indicates that the aircraft probably encountered windshear or a downdraft on late final approach. Given the prevailing wind conditions and the local topography, either of these phenomena could have been present. The pilot's recollection that he heard the stall warning sound during the bounce, plus the fact that the aircraft then contacted the ground left wing tip first indicate that the aircraft probably stalled during the bounce (possibly as a result of the pilot closing the throttle), causing the left wing to drop. The effect of the left wing tip dragging on the strip surface would have been to yaw the aircraft further (i.e. exacerbate the right skid). The decreasing speed of the aircraft as it headed towards the edge of the strip, along with the large skid angle, would have reduced rudder authority. This would then have prevented the pilot regaining directional control of the aircraft.

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

1. Windshear or downdraft conditions caused the pilot to lose control of the aircraft on late final approach.

2. The aircraft probably stalled during the bounce after a heavy landing, causing the left wing to contact the runway and the aircraft to yaw further left.

3. A severe right skid and decreasing speed reduced rudder effectiveness and prevented the pilot from regaining directional control.