1. The pilots of CYC were conducting in-flight familiarization training using the emergency power lever. That procedure was not contained in the aircraft manufacturer's pilot operating handbook.
2. The engine manufacturer's documentation contained information on the use of the emergency power lever, which did not preclude the use of the emergency power lever for in-flight familiarization training.
3. The engine sustained a flameout at an altitude above mean sea level from which reignition of the engine was not successfully completed.
4. Erosion of the first-stage compressor blades would have reduced the aerodynamic efficiency of the compressor blades.

The pilots of CYC were conducting in-flight simulated engine failure training, which involved activation of the emergency power lever (EPL). The engine ignition switch was not in the ON position during the initial operation of the EPL during this training. The pilot's operating handbook (POH) contained a requirement to place the engine ignition switch in the ON position during an actual malfunction of the fuel control unit (FCU). However, because the aircraft manufacturer only included requirements for an actual FCU malfunction, the POH did not address the engine control settings for training of this type.

The POH contained a caution which stated that the use of the EPL was for emergency purposed only, and did not mention the use of the EPL for in-flight or ground familiarization training. The engine manufacturer's Service Information Letter (SIL) noted the use of the EPL for familiarization training, while suggesting that this training be completed on the ground. The discrepancy between these two documents may have led to the flight crew's belief that the use of the EPL for familiarization training in-flight was acceptable.

Disassembly and inspection of the engine and fuel control unit did not identify any anomaly of the engine that could have resulted in the in-flight shutdown (flameout). However, the erosion of the first-stage compressor blades to the engine manufacturer's maximum service limits would have reduced the aerodynamic efficiency of the compressor blades. This may have affected the optimum compression ratio of the compressor section and resulted in decreased airflow through the engine.

According to the pilots, the engine power setting was low (reported at 55% Ng) at the time fuel was manually introduced using the EPL. If the airflow through the engine had been affected by the first-stage compressor erosion, the engine would have had a further reduced airflow rate for that given power setting and excessive fuel may have been introduced.

Analysis of the Engine Condition Trend Monitoring (ECTM) information, which indicated higher parameters for engine fuel flow, T5, and engine gas generator speed (Ng), should have alerted maintenance personnel to the erosion of the compressor. The inspection of the compressor completed 42.4 airframe hours before the incident should have noted and recorded the erosion of the first-stage compressor blades. Additionally, the ECTM information indicated increases in fuel flow and Ng, with marginal decreases in T5 temperatures. The decrease in T5 temperatures below the baseline values was believed to be related to the insulation anomaly of the two thermocouple probes.

A Cessna C208 aircraft, registered VH-CYC (CYC), with two pilots on board, was being operated for pilot type endorsement training. Air Traffic Control (ATC) had cleared the pilots to conduct upper level air work between 4,000 and 5,000 ft above mean sea level (AMSL) within a 5 NM radius of Green Island, Queensland. Following the upper level air work, the crew requested, and were granted a clearance for, a simulated engine failure and descent to 2,000 ft.

The pilot in command (PIC) reported that while completing the simulated engine failure training, he had retarded the power lever to the FLIGHT IDLE stop and the fuel condition lever to the LOW IDLE range, setting a value of 55% engine gas generator speed (Ng). The pilot under training then set the glide attitude at the best glide speed (for the operating weight) of about 79 knots indicated airspeed (KIAS). The PIC then instructed the pilot under training to place the propeller into the feathered position, and maintain best glide speed. The PIC reported that he instructed the pilot under training to advance the emergency power lever (EPL) to simulate manual introduction of fuel to the engine.

According to the PIC, he then noticed that there was no engine torque increase, with the engine inter-turbine temperature (ITT or T5) and Ng rapidly decreasing, and a strong smell of fuel in the cockpit. While the pilot under training flew the aircraft, the PIC placed the ignition switch to the ON position and also selected START on the engine starter switch. He then reportedly placed the EPL to the CLOSED position, the propeller to the UNFEATHERED position and the fuel condition lever to the IDLE CUTOFF position to clear the excess fuel from the engine. The PIC reported that they then increased the aircraft airspeed to 120 KIAS, at which point he reintroduced fuel into the engine by advancing the fuel condition lever. He reported that following these actions, the strong fuel smell persisted.

As the aircraft approached 1,500 ft, the PIC broadcast a MAYDAY, informing ATC that they had a 'flameout' of the engine and that they were going to complete a forced landing water ditching near Green Island. While the pilot under training flew the aircraft, the PIC placed the propeller into the feathered position, closed the fuel condition lever to the IDLE CUTOFF position and turned off the starter and ignition switches. They then completed a successful landing in a depth of about 2 m of water near Green Island. The pilots evacuated the aircraft without injury.

The aircraft, which sustained minor damage during the ditching, but subsequent substantial damage due to saltwater immersion, was recovered to the mainland. Following examination of all connections and control linkages, the engine was removed for examination under the supervision of the Australian Transport Safety Bureau (ATSB) at the engine manufacturer's overhaul facility. The engine trend monitoring (ETM) data logger was also removed from the aircraft for examination.

Engine information

The general condition of the engine was good except for exfoliation corrosion of the magnesium and aluminium alloy components as a result of salt-water immersion. The first-stage axial compressor blades displayed significant erosion of the blade leading edges at the blade root portion of the airfoil. The erosion of the blades measured a maximum of .250 inch (.635 cm). According to the engine manufacturer's maintenance manual, the erosion limits of the compressor blade at the root was .250 inch without repair. The eight engine thermocouple probes were also examined. Testing of the probes indicated that they all passed the heat response test, but two probes did not pass the insulation test. The engine igniters operated satisfactorily when tested.

The engine fuel control unit (FCU) was removed to another facility for disassembly and examination under ATSB supervision. The examination of the FCU revealed no evidence of any internal component failure or anomaly, which would have prevented normal operation prior to salt-water immersion.

Engine temperature indicating system

The engine temperature indicating system consisted of a bus-bar assembly, eight individual thermocouple probes connected in parallel, a wiring harness incorporating a terminal block and an adjustable trim harness incorporating a T1 thermocouple probe.¹ The T1 thermocouple probe was connected in parallel with the T5 wiring harness to bias the T5 signal and give the system a reference point. The engine manufacturer's maintenance manual included a note in the Engine Condition Trend Monitoring (ECTM) Shift Fault Isolation Chart stating that the T5 indication usually decreased when the thermocouple probes were unserviceable. The chart also noted that:

If several probes are broken or damaged, the loop resistance would not necessarily fall outside the allowable tolerance. However, erroneous temperature indications could occur due to reduced sampling.

An anomaly with the insulation of the thermocouple probes would typically manifest itself in abnormally low temperature readings.

Engine trend monitoring data logger

The ETM data logger recorded Ng, bus voltage, engine torque, T5, engine shaft horsepower (SHP), fuel consumption, airframe hours, engine total cycles, starts and duration and system exceedances. This information was electronically stored on a removable data key. Information stored on the data key could then be downloaded into a computer. The pilot reported that the data key was installed into the ETM display unit at the commencement of flight. However, after the aircraft was recovered, the data key was not found.

After preservation, the ETM data logger was shipped to the manufacturer for downloading. The manufacturer successfully recovered nineteen exceedances covering a period from 1 January 2003 to 3 February 2004. The majority of those logged were exceedances of propeller RPM and engine SHP.

Engine emergency power lever

The EPL, which was connected through linkages to the manual over-ride lever on the FCU, governed the fuel supply to the engine should a pneumatic section malfunction occur in the FCU. The EPL permitted the pilot to restore engine power by activating the lever to manually introduce fuel to the engine.

The aircraft manufacturer's Service Kit, SK208-142, provided for the installation of mechanical devices allowing for the installation of copper witness wire to the EPL. If the EPL was been moved from the NORMAL position, the copper witness wire would fracture and provide a physical indication that it had been activated. The installation of SK208-142 was not mandatory for Australian registered aircraft. However, the requirements of SK208-142 had been complied with on the aircraft.

Cessna Alert Bulletin, CAB01-15, included a requirement to ensure the fitment of the copper witness wire to the EPL of all aircraft that had SK208-142 installed. Compliance with the requirements of CAB01-15 was not mandated for Australian or United States (US) registered aircraft. According to the flight crew, no copper witness wire was installed on the aircraft at the time of the occurrence.

The aircraft maintenance manual stated that if the EPL witness wire was broken or missing, a determination was to be made as required by the engine maintenance manual, to assess if the engine limitations had been exceeded.

The aircraft manufacturer's information on the operation of the EPL stated that the use of the EPL was for emergency purposes only, and contained cautions about the use of the EPL for any other purposes. Further information about the aircraft manufacturer's use of the EPL is contained in appendix A.

The engine manufacturer's Service Information Letter (SIL) Number PT6A-053R2 addressed the use of the EPL. Although it also stated that the EPL was for emergency purposes only, it mentioned the use of the EPL for training purposes under supervision to maintain emergency practices proficiency. It included a note which suggested that familiarization training using the EPL be conducted on the ground. Further information on the operation of the EPL is contained in appendix B.

The pilot reported that, based on the reference to familiarization training in the SIL, he considered that the use of the EPL for in-flight familiarization training was acceptable.

Civil Aviation Regulation 1988, Part 50E addressed inconsistent requirements relating to aircraft operation and maintenance. Part 50E noted that by order of priority, the aircraft manufacturer's requirements superseded the requirements of an aircraft component manufacturer such as an engine manufacturer.

Recent engine maintenance

There were no engine logbook entries concerning engine compressor erosion. Engine compressor washes had been completed on a periodic basis as required. Civil Aviation Safety Authority requirements for the aircraft are contained in appendix C of this report. Engine trend data for the aircraft's engine are contained in appendix D.

The circumstances of the accident are consistent with a loss of control due to insufficient main rotor RPM being maintained, and incompatible control inputs from the instructor and the student following the initiation of the simulated engine failure by the instructor. The reported actions by the instructor indicate that he was attempting to recover the situation and allowing the student to follow him through on the controls. The student also recalled attempting to manipulate the helicopter's controls during the descent. It was unlikely that the instructor could have maintained effective control of the helicopter with both pilots manipulating the controls. Procedures for clarifying who is in control at all times, should be established and followed.

The helicopter manufacturer warned that to recover lost main rotor RPM, the pilot must immediately roll on throttle and lower the collective simultaneously. Both pilots reported that they could not lower the collective to the full down position. The activation of the low rotor RPM warning horn during most of the descent confirms that the collective was seldom in the full down position. The instructor reported attempting to increase the throttle position, but it felt like the student had frozen on the throttle. There were no defects found in the examination of the helicopter that would have explained why the collective was not able to be lowered to the full down position or the throttle increased. The manufacturer cautions that once the main rotor RPM decreases below 80%, pilots may not be able to recover control even if the flight controls are correctly positioned. Both pilots recalled seeing the rotor RPM needle in the vicinity of 80% during the descent. The student's recollection suggested that the rotor RPM may have reduced to below 80%.

The investigation was unable to resolve the differences between the statements by the instructor and the student with reference to the way in which the throttle was reduced.

The seat structures are designed to deform during a high G vertical impact, reducing the load transmitted to the seat occupant and increasing survivability. However, in deforming, the seat structure loses significant strength. In this case, the seat structure lost sufficient strength to allow the left anchor point of the left seat lap belt to tear free, increasing the risk of injury to the seat occupant.

An instructor and student were conducting a training flight from Caloundra aerodrome in a Robinson Helicopter Company Model R22 (R22) helicopter, registered VH-HBI. The weather was fine with a light north-west wind.

The instructor reported that shortly after the helicopter reached the intended initial cruising altitude of 1000 ft, and over a suitable area that was clear of other traffic, he reduced the throttle setting to idle to simulate an engine failure. The instructor said the purpose of the exercise was to test the student's alertness and ability to enter and maintain a stabilised autorotation. He intended to terminate the practice engine failure by introducing engine power at about 500 ft. During interview, the instructor said that he rolled the throttle off quickly but gradually. In a subsequent letter, he advised that he slowly reduced the throttle setting to idle. The instructor advised that he did not announce the simulated failure to the student, and that he had previously initiated 'unannounced' engine failure exercises to the student as part of the student's training.

The instructor reported that the student correctly applied right anti-torque pedal and pulled the cyclic control rearwards, but did not lower the collective lever. He also reported that the low rotor RPM warning horn came on within a few seconds and the airspeed reduced rapidly to between 30 and 40 kts. The instructor said that the helicopter then started shuddering and he lowered the collective as far as it would go and pushed the cyclic forward. He said that, as they approached 500 ft he attempted to increase engine power, but it felt like the student had 'frozen' on the throttle and he was not able to rotate the twist grip. The instructor said that the low rotor RPM horn was on for most of the descent. He said that the rotor RPM was just above the horizontal, around 83%, and the engine RPM was at idle on the engine and rotor RPM indicator (see Figure 1). He said that he did not advise that he was taking control of the helicopter, but provided 'spoken instruction, supported by directive pressure on the controls', because he wanted the student to feel the control movements. Although the instructor recalled applying up collective to cushion the helicopter for landing, the pilots were not able to terminate the helicopter's descent.

The helicopter impacted tidal mudflats near the northern end of Bribie Island in a nearly level attitude with a high rate of descent and low main rotor RPM. The floor of the helicopter sustained significant deformation and physically trapped the student's feet. Both occupants of the helicopter sustained serious injuries during the accident.

An examination of the wreckage did not identify any defect that would have prevented normal operation of the helicopter prior to the accident. The collective was able to be moved to the full down position and there was no indication of any restriction.

The student was unable to recall much of the sequence of events during the occurrence. He said that the instructor did not mention the possibility of unannounced engine failures during the flight. In an interview with the ATSB, he said that the instructor had rolled off the throttle. Later, in a letter he said that the instructor snapped the throttle off very rapidly. The student described the correct response to a simulated engine failure, but said that on this occasion, the collective did not feel like it went all the way down. The student recalled seeing the engine RPM and rotor RPM needles below the horizontal position on the engine and rotor RPM indicator (see Figure 1), when the helicopter was descending through about 700 ft. The student said that during the descent the collective was about halfway up and although he tried to push it down, it didn't feel like it moved.

The helicopter manufacturer advised that deformation of the seats provided additional absorption of vertical energy beyond that required for certification. Both crew seat structures significantly deformed during the impact with the right seat being more affected. The right seat had a plastic first aid kit under it, which had been crushed. The left anchor point of the left seat lap belt had torn free of the seat pan. The rivets attaching the anchor point had stretched, but not separated, however the sheet metal had failed, allowing the anchor point to come free.

The helicopter was operating with a valid maintenance release, and had accrued 387 hrs total time in service. The instructor had accumulated approximately 925 hrs total helicopter experience, of which about 158 hrs was in the R22. He had completed his instructor rating with the same operator, and had a total of about 105 hrs instructional time, all in R22 helicopters. The student pilot held the equivalent of a helicopter student pilot license. He had accrued about 94 hrs in the R22, of which about 16 hrs was in command.

The R22 Pilot's Operating Handbook stated that during an autorotation (prior to the flare), the collective should be adjusted to maintain the rotor RPM in the green arc between 97% and 104%, or approximately 90% to obtain maximum glide distance.

Section 4, page 10 of the R22 Pilot's Operating Handbook stated:

CAUTION - During simulated engine failures, a rapid decrease in rotor RPM will occur, requiring immediate lowering of the collective control to avoid dangerously low rotor RPM. Catastrophic rotor stall could occur if the rotor RPM ever drops below 80% plus 1% per 1000 ft of altitude.

The engine RPM and rotor RPM needles would have been in the horizontal position when the respective RPM values were 80% (see Figure 1). The low rotor RPM warning horn was designed to activate when the rotor RPM was less than 97%. The warning horn did not activate if the collective was in the full down position.

The R22 Pilot's Operating Handbook also included three Safety Notices pertinent to this accident.

The first, titled 'Surprise throttle chops can be deadly', stated:

The student may freeze on the controls, push the wrong pedal, raise instead of lower the collective, or just do nothing. The instructor must be prepared to handle any unexpected student reaction.

The second Safety Notice, titled 'Fatal Accidents caused by Low RPM Rotor Stall', stated:

No matter what causes the low rotor RPM, the pilot must first roll on throttle and lower the collective simultaneously to recover RPM before investigating the problem. It must be a conditioned reflex.

The third Safety Notice, titled 'Practice autorotations cause many training accidents', stated:

As the aircraft descends through 100 feet AGL, make an immediate power recovery unless all of the following conditions exist:

1. Rotor RPM in middle of green arc,
2. Airspeed stabilized between 60 and 70 KIAS [knots indicated airspeed],
3. A normal rate of descent, usually less than 1500 ft/min,
4. Turns (if any) completed.

This Safety Notice also stated:

Practice autorotations continue to be a primary cause of accidents in the R22 and R44. Each year many helicopters are destroyed practicing for the engine failure that very rarely occurs.

A review of the ATSB database identified 18 accidents between 1985 and 2003 that occurred during autorotation training in Robinson R22 helicopters.

Most flight instructor manuals emphasise the importance of establishing and using procedures that at all times identify which pilot has control of the aircraft. For example, the Flight Instructor Guide - Helicopter (1995) issued by Transport Canada stated:

CONTROL OF AIRCRAFT

2. There should never be any doubt as to who has control of the aircraft. ...:

(d) when the student has control, you must not "ride" the controls. Your student may feel that you are taking control and this could lead to a dangerous situation. Additionally, you may rob your student of the feeling of accomplishing the manoeuvre independently. This is particularly difficult during critical manoeuvres, such as full-on autorotations, when there is little time available to the instructor to correct errors. This procedure must be adhered to at all times.

The pilot was forced to land on a road when the useable fuel on board the aircraft had been consumed. The aircraft fuel quantity had been verified with an incorrect fuel dipstick that indicated unusable fuel as being useable fuel. The aircraft fuel gauges were capable of indicating fuel quantities that were consistent with the last fuel calibration card in the cockpit.

The incorrect markings on the dipstick accounted for a fuel discrepancy of 23 litres. The investigation could not determine the source of the remaining 20 litre fuel discrepancy.

The pilot had flown the Cessna 172 aircraft, registered VH-DBG, from Derby to Fitzroy Crossing earlier in the day. The pilot reported that, prior to departure from Derby, he had dipped the tanks using a dipstick, which displayed that there was 100 litres of fuel on board. After arrival at Fitzroy Crossing, the pilot added 44 litres of fuel to the aircraft. The pilot reported that he again used the dipstick to check the amount of fuel. He said that the dipstick displayed that there was 100 litres of fuel on board the aircraft.

The aircraft departed from Fitzroy Crossing and the pilot conducted a scenic flight on the return leg to Derby. As the aircraft was approaching Derby, cruising at 3000 ft, the engine began to splutter and then lost power. The pilot turned the aircraft towards a sealed road and, after transmitting a PAN call, conducted a forced landing onto a road. There were no reported injuries. The pilot reported that the flight time for the return flight was 1.8 hours. He also reported that he had leaned the mixture during the cruise portions of the flight in accordance with the engine manufacturer's operating manual. The ATSB did not attend the site, however the investigation was conducted with reference to information provided by the pilot in command, the operator and several other parties.

The pilot arranged for a licensed aircraft maintenance engineer to attend the aircraft on the road. The engineer reported that when he arrived at the aircraft and checked the fuel tanks, there was no useable fuel in either of the tanks. The pilot also reported that there was no useable fuel remaining in the fuel tanks. The engineer then checked the aircraft engine and fuel system and, after adding fuel, the aircraft was flown back to Derby with no reported problems.

Once the aircraft was back at Derby, further checks of the aircraft and its systems were conducted, with no reported defects found. The engineer reported that he added approximately 60 litres of fuel to each tank in 30 litre increments and found that the cockpit fuel gauges were showing quantities that were consistent with the fuel calibration card that was present in the cockpit.

A post occurrence flight plan of the proposed flight, in accordance with the company operations manual, revealed that the aircraft was required to carry 104 litres of fuel. With this amount of fuel onboard the aircraft, and a normal inflight fuel burn, there would have been 43 litres of fuel remaining in the aircraft at the point where the engine stopped.

The operator conducted an investigation into the circumstances of the occurrence. The operator's investigation found that the dipstick used by the pilot in command to check the fuel quantity incorrectly indicated unusable fuel as useable fuel. The amount of unusable fuel totalled 23 litres.

The Ted Smith Aerostar 601 aircraft, registered VH-WRF, departed Coolangatta at 1301 ESuT with a flight instructor and a commercial pilot on board. The aircraft was being operated on a dual training flight in the Byron Bay area, approximately 55 km south-south-east of Coolangatta. The aircraft was operating outside controlled airspace and was not being monitored by air traffic control. The weather in the area was fine with a south-easterly wind at 10 - 12 kts, with scattered cloud in the area with a base of between 2,000 and 2,500 ft.

The purpose of the flight was to introduce the commercial pilot, who was undertaking initial multi-engine training, to asymmetric flight. At approximately 1445, the operator advised Australian Search and Rescue that the aircraft had not returned to Coolangatta, and that it was overdue. Recorded radar information by Airservices Australia revealed that the aircraft had disappeared from radar coverage at 1335. Its position at that time was approximately 18 km east-south-east of Cape Byron. Search vessels later recovered items that were identified as being from the aircraft in the vicinity of the last recorded position of the aircraft. Those items included aircraft checklist pages, a blanket, a seat cushion from the cabin, as well as a number of small pieces of cabin insulation material. No item showed any evidence of heat or fire damage. No further trace of the aircraft was found.

Air traffic control received normal radio transmissions from the aircraft during the departure from Coolangatta and the transit to the operating area. No other transmissions from the aircraft were received. In particular, there were no radio transmissions from the aircraft around 1335.

The instructor was the chief flying instructor of the organisation that owned the aircraft. He held a Grade 1 flight instructor rating. He had extensive civilian and military experience as a pilot and flight instructor, most of which was on single-engine fixed and rotary wing aircraft. At the time of the accident, his total flying experience was 7127.2 hours, with 447.7 hours on multi-engine aircraft. The instructor had accrued 294 hours on Aerostar aircraft of which 194 were as a flying instructor.

The student pilot held a commercial pilot's license and had about 283 hours flying experience at the time of the accident. He commenced flying training on 23 April 2001 on a Mooney M20J aircraft, and first flew solo on 23 July 2001, after 38.3 hours dual instruction. On 21 December 2002, he gained an unrestricted private pilot's license, having flown a total of 181.1 hours. On 19 September 2003, with a total flying experience of 257 hours, the student pilot failed his first commercial pilots license test. On 17 October 2003, he passed the second attempt with a total flight time of 276.9 hours. Excluding the accident flight, the commercial pilot's flight time on Aerostar aircraft was 3.5 hours.

The accident flight was the commercial pilot's fourth in the Aerostar aircraft and was the third flying exercise sequence in the operator's multi-engine training syllabus. The objectives of the exercise included controlling the aircraft after the failure of an engine, recovering from a stall in the take-off configuration, and entering and recovering from a minimum control speed (Vmca) situation. (Vmca is the minimum control speed in flight with one engine inoperative.)

Recorded radar data for the flight showed that the aircraft proceeded from Coolangatta to the east of Byron Bay where it conducted a series of manoeuvres in an area approximately 18 km square, at between 2,500 and 3,000 ft, with occasional brief excursions below 2,500 ft. Between 1313:00 and 1328:00 ESuT, there were three instances, about 4 minutes apart, where the recorded groundspeed of the aircraft decreased rapidly from approximately 150 kts to between 100 and 110 kts before increasing again. Each speed reduction was accompanied by an altitude loss of 200 - 400 ft. Between 1328:40 and 1329:05, the groundspeed decreased from 140 to 118 kts. It then fluctuated between 121 and 112 kts for the next 1 minute and 25 seconds while the recorded altitude reduced from 2,600 to about 1,900 ft. The recorded altitude and ground speed then steadily increased for the next 3 minutes to a maximum of 2,800 ft and 123 kts respectively. During the next 60 seconds, the recorded altitude reduced to 2,500 ft while the groundspeed decreased to 110 kts at 1335:00. At 1335:29, the recorded altitude was 2,600 ft and the groundspeed 108 kts. The last valid radar data was at 1335:37 when the recorded altitude was 2,100 ft, and the groundspeed was 100 kts. The aircraft was tracking in a south-easterly direction from about 1331 until radar contact was lost. There were no sudden or significant changes in the recorded track during that period.

The aircraft was being operated on a valid maintenance release and there were no maintenance items outstanding at the time of the accident. The aircraft was flown from Coolangatta to Sydney and return on the night before the accident and was reported to have operated normally during those flights. The aircraft was not fitted with a stall warning system.