The pilot reported that just after liftoff, at about 60 knots, the engine lost power without warning. He described the power loss as a wind-down of RPM over four to five seconds and not a sudden cut. There was no abnormal engine indication before the failure. The aircraft landed heavily on the runway causing the right canard to fail and damaging the rear fuselage. Inspection of the aircraft revealed that the carburettor float bowl had been fitted with four perspex windows which had been glued to the outside of the bowl. Two of the windows were found on the floor of the trailer which had been used to recover the aircraft while the remaining two were still in place. The pilot, who built the aircraft, fitted the windows to assist in setting the carburettor float level. The windows were not a standard fit, or an approved modification. The adhesive used was 24 hour Araldite. This substance is known to deteriorate when in contact with aviation gasoline. Had the windows fallen from the float bowl during the takeoff, fuel would have escaped and possibly resulted in the symptoms described by the pilot. However, the investigation was unable to determine whether the windows became detached before, or as a result of, the accident. No weather observations are recorded at Proserpine. However, at Mackay (93 km to the south-east) at 0800 hours on the day of the accident, the wet and dry bulb temperatures were 6`C and 7`C respectively. Under these conditions, there is the likelihood of serious carburettor icing occurring at any engine power. It is possible, therefore, that the engine lost power due to carburettor icing.

The pilot had flown the aircraft to Merimbula three days previously. On the return journey the aircraft was about 50 kilometres from the destination when both engines failed within three minutes of each other. The pilot selected alternative fuel tanks but the engines did not respond. The pilot prepared for a forced landing and commenced operating the hand pump to lower the landing gear. This procedure had to be abandoned close to the ground, and the aircraft touched down in the selected paddock with the gear only partially extended. Investigation revealed that the engines had failed when the fuel in the selected tanks was exhausted. Fuel sufficient for about 11 minutes of flight time was drained from the auxiliary tanks at the accident site. No mechanical fault or defect was discovered which might have led to the accident. The pilot advised that when the engines failed, the appropriate fuel gauges were indicating that substantial fuel remained. Examination at the accident site revealed that the gauges were reading virtually empty. However, one of the gauges was known to have given inaccurate readings in flight and the discrepancy in gauge indications was not resolved. The pilot was aware that fuel had vented in flight from this particular aircraft. The company provided tape with which to seal the fuel filler cover flaps, in an effort to prevent further fuel loss by venting. The pilot had applied this tape as suggested, and although there was no evidence of fuel venting in flight, this possibility remains. The pilot had no means of accurately determining the fuel quantity on board prior to DEPARTURE for the return flight from Merimbula. He had very limited experience on the type, and did not know what the fuel consumption figures would be, using his fuel management techniques. He was also aware that one of the fuel gauges was inaccurate. Although fuel was readily available at the aerodrome, the pilot had elected not to refuel the aircraft.

Circumstances:

Before commencing spraying, the pilot conferred with the farmer who pointed out three sets of power lines. He checked the location of each of these lines in relation to an adjacent highway and a tree line in the area to be sprayed. He also completed an aerial survey. The pilot successfully sprayed the first three quarters of the paddock and then commenced on what he considered to be the more difficult section. As the aircraft descended over the line of trees to spray height, the pilot heard and felt a thump. He flew the aircraft back to the strip, where he discovered damage to the leading edge of both wings. The aircraft had struck a fourth set of wires. This two strand set was connected to a cross tree on an 11,000-volt pole and strung for about one kilometre through trees to a pole beyond the highway. The farmer had neglected to mention this pair of wires and the pilot did not see it on his aerial survey.

Significant Factors:

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

1. The pilot failed to detect the fourth set of wires.

2. The aircraft struck the wires while commencing to spray the final section of paddock.

3. The aircraft sustained damage to both wing leading edges.

Circumstances:

The floatplane was taking off towards the south into a wind reported as south-east to south gusting to 30 kt. Water conditions were described as choppy. After a run of about 300 m, just prior to the aircraft becoming airborne, the pilot reported hearing a loud bang. The aircraft immediately commenced a sharp left turn, followed by several rocking oscillations. The pilot shut down the engine after which the aircraft slowed and stopped in a left wing low attitude with the wingtip in the water. The front landing wire attachment bracket on the left float and the aft landing wire attachment bracket at the top of the rear right strut had both failed at their respective bolt holes. Metallurgical examination found that both brackets were extensively corroded, with the bracket cross sections being reduced by about 30 percent in both cases. It was considered that the reduced cross-sectional area was insufficient to sustain normal applied loads. It appeared that the front left float bracket failed first, followed by the rear right strut bracket, resulting in pivotal movement of the left float.

Significant Factors:

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

1. Inadequate inspection.

2. Extensive corrosion of landing wire attachment brackets.

3. The weakened brackets failed under the operating conditions prevailing.

The pilot reported making an approach to land in light winds. The balloon skipped once and then made a slightly heavier than normal landing after brushing through a small tree. A female passenger complained of injury while still in the basket after landing. The balloon was deflated and the injured passenger was left in the basket for the trip to the hospital. A later report revealed that the passenger had suffered tibial plateau fractures to both legs. Witnesses and other passengers made no adverse comments about the severity of the landing. The injured passenger was wearing suitable footwear, but may have assumed an incorrect position when bracing for the landing.

The King Island aerodrome weather was forecast to include temporary periods of moderate to heavy rain showers, six eighths of cloud cover base 800 feet and visibility reduced to 3000 metres. The forecast surface wind was 340 degrees at 20-30 knots. The pilot's qualifications required a minimum visibility of 5000 metres for operation at night in Visual Meteorological Conditions, (NGT VMC). It was planned that another pilot, qualified for Instrument Flight Rules (IFR) operations, would act as pilot in command, but he became unavailable. The flight plan submitted by the pilot indicated he would be operating under the IFR category and when queried on taxiing he confirmed that this was so. Shortly after DEPARTURE, the pilot requested a weather report from an IFR pilot who had just landed at King Island. This gave a cloud base of 2000 feet, heavy rain showers and visibility of 2000 metres. Flares had been laid to allow the other pilot to use Runway 35 because of the strong northerly wind. The pilot of VH-CAY activated the electric lighting for Runway 28. He reported that the weather was satisfactory enroute and he could see lights ahead on the island. On crossing the coast flight conditions became rough in moderate to severe turbulence. The pilot advised he arrived over the aerodrome at 1500 feet above mean sea level and observed the lighted wind sock was horizontal, with the direction fluctuating rapidly between north and west. He turned to the south and broadcast his intention to land on Runway 28. Neither the pilot nor the passenger had any further recall of the events leading to the accident. VH-CAY was heard passing over the aerodrome and the engine sound was very loud, suggesting to the listener that the aircraft was low. It had been raining continuously for more than an hour, sometimes very heavily, and it was still raining at the time. The aircraft was subsequently seen flying at a very low height some six kilometres south of the aerodrome, tracking approximately north. It was raining very heavily in that area and the wind was very strong. Soon afterwards there was a sound of impact and a flash of light. The aircraft had struck the tops of trees 30 feet high, then descended to the ground. After the aircraft came to rest it was destroyed by a fire. Examination of the wreckage was severely hampered by the extreme fire damage sustained, but no evidence was found of any defects that might have contributed to the accident. The aircraft had evidently been under control at the time it collided with the trees. A post analysis of the conditions by the Bureau of Meteorology indicated the possible presence of strong up and down drafts, horizontal wind shear, turbulence, and estimated visibility as 2-3000 metres in rain. The evidence suggested that the pilot may have been lower than he believed as the aircraft overflew the aerodrome. Having passed overhead, there there would have been few external visual references under the existing conditions to alert him that the aircraft was inadvertently being descended into the ground.

During a daily inspection three days prior to the accident, the pilot noted a fuel leak in the area where the fuel filter bowl fits on top of the fuel pump. A new gasket was made, from a thinner material than the original, and a silicon sealer was therefore applied as an added precaution. The aircraft subsequently operated without fuel problems, until the engine failed completely following a particular takeoff. The pilot attempted to land straight ahead, but the airspeed was less than the optimum and the aircraft touched down heavily, collapsing the right hand gear leg. The investigation determined that the fuel filter outlet was blocked by the silicon sealer that had been used on the gasket. The pilot believed that the silicon that he used was fuel resistant. It is not certain whether the silicon in fact did react with the fuel and degrade or whether an excessive amount was used resulting in the excess breaking away inside the filter bowl and blocking the outlet. The failed landing gear component was inspected and it was determined that it was manufactured from a material as specified in the plans and had failed due to overload.

The aircraft departed Gayndah at 0632 hours on a survey flight. This was the first flight of the day. The survey was terminated near a position about four kilometres north of Mt Goonaneman and the aircraft landed at Gayndah at 1138 hours. The crew assisted the oncoming crew to refuel and prepare the aircraft for the next flight. The aircraft was seen departing Gayndah shortly after 1200 hours. At about 1235 hours witnesses saw a column of smoke in the vicinity of Mt Goonaneman and notified Biggenden Police. The accident site was located by the police about 40 minutes later. The survey flights were being conducted at 80 metres above ground level and traversed the survey area on a flight path which was orientated in a north-easterly/south-westerly direction. The tracks being flown were about 40 kilometres long and 200 metres apart. The accident location indicates that the aircraft was on the first survey line adjacent to the position at which the earlier flight had finished, and was travelling in a north- easterly direction towards the mountain when the accident occurred. Witness observations were not conclusive in establishing the final flight path, however, some of the hearing reports were consistent in describing an unusual noise from the aircraft immediately before the accident. The engine noise was heard by one witness to cease and then recommence twice in quick succession. This could have been caused by a serious engine malfunction or changes made to engine controls by the pilot. Another witness reported seeing the aircraft make what he described as a steep climb and a sharp "U" turn near the mountain. This was not a normal manoeuvre for the aircraft to make and this witness may have observed the aircraft during the accident sequence. Unfortunately the time of this observation could not be established in relation to the time of the accident. The examination of the wreckage was hampered by the extensive destruction resulting from the post impact fire. Dust that was found in the fuel injectors from the left engine could have caused an engine malfunction but this could not be positively established. The right engine could not be fully examined due to fire damage. The propellers from both engines indicated that substantial power was being delivered at the time of impact. The nature of the ground impact was indicative of very low forward speed, high vertical speed, in a steep nose down, right wing low attitude. The pilot was very experienced, and it is considered to be most unlikely that he would allow such a situation to develop, or that he would attempt any unusual manoeuvre at low altitude near the mountain. There is no evidence to suggest that aircraft may have failed structurally, or that the pilot would not have been able to cope with any situation, other than that which had the potential to cause a sudden loss of consciousness. Pathology disclosed evidence of a heart condition in the pilot which could have caused sudden death or incapacity. The pilot's last medical examination, conducted in February 1988, had included an ECG. Had the pilot suffered incapacitation, the crewman, seated in the cabin behind the pilot would have had insufficient time, due to the low operating altitude, to correct the situation.

A witness, who lives alongside the airstrip, observed the aircraft flying in a north-south direction, about 300 feet above ground level and about 300 metres to the east of his position. The witness took particular notice of the actions of the aircraft and watched as it porpoised three times before diving steeply into the ground. Inspection of the wreckage revealed no defect which may have contributed to the accident. Weather at the time was fine, wind was light and variable, and the aircraft was being flown over flat pastureland. Investigation revealed that the pilot suffered from epilepsy, had heart disease, and that this was his first flight for over one year.

Circumstances:

The pilot reported that he joined downwind for runway 34 and completed downwind checks. He stated that normal landing gear extension occurred when he selected the undercarriage down, and a normal gear down indication was obtained (three green lights). The pilot stated that on late final approach the gear warning horn sounded when he reduced the throttle settings, and he then noticed there were no green lights. He said that he immediately went around and established the aircraft in the circuit at 1000 feet. The pilot stated that he extended and retracted the landing gear several times, and on each occasion the gear operated normally, and normal indications were obtained. A second approach was then commenced, and the pilot said that the lights were monitored all the way to touchdown. He reported that the landing was normal, but that shortly after touchdown the right wing began to sink. The pilot stated that he believed the right hand gear had failed, and that he attempted to recover the situation by applying full power and using aileron to level the wings with a view to going around, and proceeding to an aerodrome better equipped to handle an emergency landing. At this point, the left hand gear and the nose gear collapsed, and the aircraft slid to a stop on the gravel runway. Marks on the runway showed that the aircraft had landed with the landing gear extended and that the gear retracted soon after touchdown. Subsequent examination of the aircraft found no evidence of any pre-existing defect which may have contributed to the development of the accident. The landing gear switch was found in the down position, and the gear actuating mechanism was found to be in an intermediate position. Damage to the mechanism appeared to be consistent with the gear being driven down with the weight of the aircraft on it. This could occur if the gear was inadvertently selected up during the landing and then down again. The pilot is adamant, however, that this did not occur. The cause of the landing gear retracting during the landing roll could not be determined.

Significant Factors:

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

The landing gear retracted during the landing roll. The cause of the unscheduled gear retraction could not be positively identified.