AAIB Bulletin: 2/2009

the bulb filaments had failed. Further investigation took place which showed that the correct bulbs had been fitted and no cause of overheating could be found. The troubleshooting documentation and the unit were then sent to the manufacturer for examination; to date no apparent reason for the smoke has been determined. 3 © Crown copyright 2009 AAIB Bulletin: 2/2009 N850KF EW/G2008/05/19 ACCIDENT Aircraft Type and Registration: Cessna 310Q, N850KF No & Type of Engines: 2 Continental IO-470-VO piston engines Year of Manufacture: 1968 Date & Time (UTC): 10 May 2008 at 1030 hrs Location: Jersey Airport Type of Flight: Private Persons on Board: Crew - 1 Passengers - None Injuries: Crew - None Passengers - N/A Nature of Damage: Right wingtip, flap, aileron and propeller damaged Commander’s Licence: Commercial Pilot’s Licence Commander’s Age: 48 years Commander’s Flying Experience: 1,820 hours (of which 120 were on type) Last 90 days - 64 hours Last 28 days - 38 hours Information Source: Aircraft Accident Report Form submitted by the pilot and subsequent enquiries by the AAIB Synopsis The right main landing gear failed to lock down after the landing gear was selected down on the approach to land. This resulted in the failure of a component in the gear retraction/extension mechanism after landing, causing the right gear to partially retract. The right wingtip, flap, aileron and propeller contacted the runway and were damaged. The right gear was subsequently found to be stiff in operation; this was attributed to inadequate lubrication. History of the flight When the pilot selected the landing gear down approximately 3.5 miles out on an ILS/DME approach to Jersey Airport, the green gear down and locked light for the right main landing gear failed to illuminate. He initiated a go-around and when safe to do so, replaced the light bulb, but this made no difference. He then attempted to lower the landing gear using the normal and the emergency extension systems and tried sideslipping the aircraft in an attempt to lock the gear down, but these efforts were also unsuccessful. The green light for the right main gear remained off, leading the pilot to conclude that the gear had not locked down. The pilot then made a visual approach to Runway 27, using full flap and reducing the speed to the blue-line figure. On rounding out, he held off until the stall warning occurred, to achieve a gentle touchdown. He 4 © Crown copyright 2009 AAIB Bulletin: 2/2009 N850KF EW/G2008/05/19 held the aircraft straight, allowing it to slow down without applying the brakes, in the hope that the right main gear, even if not locked down, would continue to support the aircraft. As he moved the mixture controls towards the fully closed position, the right wing started to drop. He then turned off the fuel, electrics and magnetos and once the aircraft had come to a halt, he exited via the main door. Subsequent examination revealed that a link securing the right leg in the extended position had suffered an overload failure. The damage was consistent with the effect of the component, designed to carry only retraction and extension loads, being subjected to the ground loads due to the failure of the landing gear to reach the locked position. During retraction and extension tests performed following subsequent repairs, considerable stiffness of operation of the right leg was noted. This was attributed to poor lubrication of the pivot bearings. It was concluded that the resulting stiffness in operation had prevented the leg from reaching the fully extended position. 5 © Crown copyright 2009 AAIB Bulletin: 2/2009 TC-MBG EW/C2008/02/01 ACCIDENT Aircraft Type and Registration: Fokker F27 Mk 500 Friendship, TC-MBG No & Type of Engines: 2 Rolls-Royce Dart 532-7 turboprop engines Year of Manufacture: 1971 Date & Time (UTC): 1 February 2008 at 2115 hrs Location: Stand 201, Edinburgh Airport Type of Flight: Commercial Air Transport (Cargo) Persons on Board: Crew - 3 Passengers - None Injuries: Crew - None Passengers - N/A Nature of Damage: Propeller, engine and ground power unit severely damaged Commander’s Licence: Airline Transport Pilot’s Licence Commander’s Age: 56 years Commander’s Flying Experience: 4,080 hours (of which 2,745 were on type) Last 90 days - 74 hours Last 28 days - 26 hours Information Source: AAIB Field Investigation Synopsis The aircraft was scheduled to operate a night cargo flight from Edinburgh to Coventry. The weather conditions at Edinburgh Airport were wintry with snowfall, which required the aircraft to be de-iced. Shortly after both engines had been started, the commander signalled to the marshaller to remove the Ground Power Unit (GPU) from the aircraft, which was facing nose out from its stand, down a slight slope. As the marshaller went to assist his colleague to remove the GPU to a safe distance prior to the aircraft taxiing off the stand, the aircraft started to move forward slowly, forcing them to run to safety. The flight crew, who were looking into the cockpit, were unaware that the aircraft was moving. It continued to move forward until its right propeller struck the GPU, causing substantial damage to the GPU, the propeller and the engine. The ground crew were uninjured. No cause as to why the aircraft moved could be positively identified. History of the flight TC-MBG was operating from Stand 201 on the North Cargo Apron, at Edinburgh Airport. Its operator had been subcontracted by another operator which regularly uses the airport. The crew had flown the aircraft together the night before from Coventry to Edinburgh without incident. On that sector and the planned sector back to Coventry, the commander was line training the co-pilot. Due to forecast high winds at Edinburgh, the aircraft was repositioned 6 © Crown copyright 2009 AAIB Bulletin: 2/2009 TC-MBG EW/C2008/02/01 by the handling agent to face into wind, after the crew had gone off duty. This placed the aircraft pointing nose out of the stand, facing down a slight slope. Prior to the accident, the crew, which included a travelling company engineer, reported for duty at 1900 hrs for a scheduled departure time of 2050 hrs. A GPU was connected to the right side of the aircraft when they boarded, to provide electrical power prior to engine start. There were two ground handlers in attendance to oversee the departure, a marshaller, who supervised the start up and an assistant. It was dark at the time. The aircraft was loaded and prepared for departure without event. De-icing was necessary due to falling snow and this caused a delay. When it was completed, the company engineer went outside to inspect the aircraft and collect the de-icing certificate. At approximately 2113 hrs, with the ‘Pre-Flight’ checklist completed, the co-pilot requested and received start clearance from ATC. The co-pilot then commenced the ‘Before Start’ checklist. As he called “Parking brake”, expecting to hear the commander reply “Set” to confirm the parking brake was on, they were interrupted by the return of the company engineer, who verbally confirmed to the commander that the nosewheel was chocked. The co- pilot’s parking brake call-out was not subsequently responded to by the commander. Using hand signals, the commander then requested and received clearance from the marshaller to start the aircraft’s engines. After start, with the engines stabilised, the commander noticed that the main and brake pneumatic system pressures had fallen to 1,600 psi. He advanced the engine power levers in a bid to restore the pressure to 1,800 psi. The commander then signalled to the marshaller to disconnect the GPU, by indicating a ‘T’ with his hands which he then pulled apart. The co-pilot then started to read out the ‘After Start’ checklist to the commander. Upon receiving the signal, the marshaller went to assist his colleague remove the GPU to a safe distance. As the marshaller reached the GPU, the aircraft started to move forward slowly. Noticing this, he shouted to his colleague, who was between the GPU and its tug. They both ran clear of the aircraft as it continued to move forward. The flight crew were still progressing through the ‘After Start’ checklist when they heard a loud ‘bang’ from the right side of the aircraft. The commander checked the engine instruments and noticed that the right engine had failed. He shut down the left engine and secured the aircraft by pulling both engine shutoff handles, before vacating the aircraft with the company engineer, followed shortly by the co-pilot. Once outside, the commander noticed that there were no chocks in the vicinity of the nosewheel. The Airport Fire and Rescue Services (AFRS) were on scene within two minutes. Upon arrival they chocked the nosewheel, as no chocks were present and laid a blanket of foam beneath the right engine to cover the leaking fuel. Edinburgh Airport Managing Director’s Directive 04/07 Managing Director’s Directive (MDD) 04/07, ‘Aircraft Pushback and Powerback Procedures’, was issued by Edinburgh Airport on 28 March 2007. It stated the following: ‘Straight pushbacks are forbidden from any North Cargo Apron stand. If any aircraft on these stands has been previously repositioned to face out (e.g. because of prevailing wind conditions), such aircraft must be pulled off stand and lined up on the taxiway centreline before starting engines.’ 7 © Crown copyright 2009 AAIB Bulletin: 2/2009 TC-MBG EW/C2008/02/01 The airport operator commented that MDD’s are sent out electronically to the general managers and station managers. They added that it is the responsibility of the handling agents to ensure that all applicable Airport Notices are brought to the attention of any new operator or airline company. Commander’s comments The commander stated that when he had operated from Edinburgh on the two days before the accident, the aircraft was positioned in the same manner as when the accident happened, ie on the same stand and facing outwards, to avoid high tailwind conditions. On these previous occasions ATC had given clearance for the aircraft to taxi off the stand under its own power. He added that he had not received a copy of MDD 04/07 from his company until 25 February 2008, 24 days after the accident. The commander stated that the brake pressure gauge was reading approximately 1,800 psi when he checked it during the ‘Pre-Flight’ checks and that he had checked that the parking brake was set during the ‘Before Start’ checklist. He added that throughout the ‘After Start’ checklist they were not aware of the aircraft moving prior to the impact with the GPU. Marshaller’s comments The marshaller stated that he had supervised TC-MBG when it had operated from the North Cargo apron, facing nose out with the same operating crew, over the preceding two days. He added that on the night of the accident, the aircraft’s nosewheel was chocked when he went to assist his colleague to remove the GPU. He stated that he had felt slightly under pressure to expedite the departure as there had been a delay due to the aircraft requiring de-icing. ATC controller’s comments The ATC controller stated that at the time of the accident, he was working both Tower and Ground frequencies. When TC-MBG called for start clearance, he was initially unsure of its callsign due to the poor quality of the co-pilot’s transmission. He was not aware that the aircraft was facing out of the stand and could not see it from his position in the control tower, due to the darkness and the distance involved. Recorded data The event was captured on the 30-minute Cockpit Voice Recorder (CVR). The recording indicated that the commander was providing instruction to the co-pilot. Although the flight was delayed awaiting de-icing services, the checklists were completed in an unhurried fashion. The number two engine was started, followed by the number one engine. Sound spectrum analysis of the recording showed that both engines stabilised at around 8,000 rpm. The commander then said “A LITTLE BIT MORE POWER TO CHARGE THE SYSTEM A LITTLE BIT”, after which the engine speeds increased to around 8,600 rpm. Approximately 20 seconds later, whilst progressing through the ‘After Start’ check list, the sounds of the propeller striking the GPU were heard. This started with 0.7 second of propeller strike noise followed by a one second gap, a further one second period of propeller strike noise and then a louder mechanical sound, possibly associated with the engine breaking free of its mounting. The aircraft is of an age when only a very limited number of parameters were required to be recorded by Flight Data Recorder (FDR). None of these would have assisted with this investigation. No data on the accident were recorded in any case, as the FDR start/stop logic had not yet triggered it to start recording. 8 © Crown copyright 2009 AAIB Bulletin: 2/2009 TC-MBG EW/C2008/02/01 The standards for more modern aircraft require more parameters to be recorded. Retrospectively increasing the number of parameters recorded by an FDR on older aircraft may be prohibitively expensive due to interfacing issues. However, current imaging technology potentially provides a cheaper alternative means of capturing a wide array of additional parameters via cockpit image recording. Minimum standards for such equipment have been specified in EUROCAE document ED-112. Work is currently underway to incorporate ED-112 into ICAO requirements and introduce cockpit image recorders. Once such recorders have become available and their associated costs are better understood, consideration should be given to reviewing the cost/safety benefit case for retrofitting them to aircraft with limited FDR parameter sets. Examination of the accident site A photograph of the accident scene, taken on the morning after the event, is shown at Figure 1. Chocks had been placed at all the wheels at this stage. According to the AFRS, no chocks were found in the vicinity of any of the wheels on their arrival and they chocked the wheels as a precaution, prior to applying foam. Two chocks were found in the wreckage of the GPU, with one of them being visible in Figure 1. The power lead was found trailing on the ground between the GPU and the aircraft, although the ground power receptacle door on the aircraft had been closed. Following disconnection, the lead would normally be folded into one of the recessed trays that run the length of each side of the GPU’s chassis and which are also Figure 1 View of the accident site. Position of chock is indicated. The other is located within GPU debris in foreground Chock 9 © Crown copyright 2009 AAIB Bulletin: 2/2009 TC-MBG EW/C2008/02/01 used to store chocks. The tractor unit was not attached to the GPU; the ground handler was in the process of attaching it at the time of the occurrence. (Note: the handling agent had modified its GPUs in order to reduce the possibility of towing a unit away whilst connected, via its power cable, to the aircraft. As a result the GPU must first be unhitched from the tractor and the towbar raised to a near vertical position. The latter action applies the wheel brakes and operates a mechanical interlock which, by means of an associated relay, allows electrical power to be supplied to the aircraft.) The GPU had sustained substantial damage as a result of being struck by the right propeller. The sliding portion of the cover, which was made from steel, had been torn from the chassis and thrown some 7-8 metres, landing in front of the outboard section of the aircraft’s right wing. The control panel had also been removed from its mountings but had remained attached to the GPU by electrical cables. The roof of the tractor had suffered a glancing blow from a propeller blade and the rear window had received a number of impacts from flying debris. As can be seen in the photograph, the aircraft’s right engine nacelle had almost separated from the wing, remaining attached only by the exhaust duct and some conduits. The instability of the nacelle during the accident sequence and the close proximity of the fuselage had resulted in the propeller striking and breaking an adjacent window transparency. Description of the aircraft pneumatic system The F27 aircraft is equipped with a pneumatic system in which pressurised air is used to operate the brakes, nosewheel steering and landing gear. A schematic diagram is presented at Figure 2a and b. It consists of two separate systems, the main and alternate/ emergency, each with an air storage bottle and an additional bottle for the main braking system. The entire system is charged by means of compressors, one driven from the accessory gearbox of each engine, or from a compressed air supply via a charging valve in the rear of each engine nacelle. The nominal working pressure is 3,000 psi. The aircraft manufacturer stated that, with both engines running at 10,000 rpm, the charging rate is around 20 minutes per 1,000 psi increase in pressure. An isolating valve is incorporated within the system, which, as can be seen in Figure 2a, actually consists of two valves, both operated by a single control rod. The purpose of the valve is to preserve stored pressure in the event of a leak elsewhere in the system. The Maintenance Manual noted that: ‘…operational requirements allow a leakage of 100 psi per hour based on pneumatic system capability.’ Fokker F-27 expanded checklist The operator of TC-MBG stated that they use the aircraft manufacturer’s checklists, as published in the Airplane Flight Manual. The ‘Pre-Flight’ checklist pneumatic system check reads as follows: ‘Pneumatic pressures…………………..Check Min for take-off ● MAIN system 1500 psi BRAKE system 1500 psi ALTN system 2500 psi’ 10 © Crown copyright 2009 AAIB Bulletin: 2/2009 TC-MBG EW/C2008/02/01 Tests on the braking system When the aircraft was first examined by the AAIB the brake pressure gauge was indicating close to zero. A charging trolley was obtained and the system was charged to approximately 2,000 psi, with the intention of attaching a tug to the aircraft in order to test the efficacy of the parking brake. However, a tug did not become available for approximately one and a half hours, during which time it was observed that the brake pressure had decayed to around 1,100 psi. Following the arrival of the tug, the system was recharged and, with the tug gently pulling and pushing the aircraft, satisfactory operation of the parking brake was demonstrated. At a later date, it was decided to conduct a more accurate assessment of the leakage rate of the system. Accordingly, the pneumatic system was charged to Figure 2a Pneumatic system schematic diagram See Fig 2b 3,000 psi with the isolating valve open. The parking brake was set and the isolating valve closed. One hour later the pressure readings were observed as follows: main 2,800 psi, brakes 2,550 psi, alternate/emergency 3,000 psi. The isolating valve was then opened, thus connecting together the main and brake systems, with the pressure equalising at 2,700 psi; the alternate/ emergency system remained at 3,000 psi. Forces acting on the aircraft The accident occurred shortly after the engines were accelerated beyond 8,000 rpm. Information from the propeller manufacturer indicated that the propeller blades would remain at the zero angle pitch stop until around 13,000 engine rpm, with the result that the total thrust from both propellers in the prevailing conditions of 0°C and 6 kt headwind was only of the order of 55 kg force. 11 © Crown copyright 2009 AAIB Bulletin: 2/2009 TC-MBG EW/C2008/02/01 Analysis There are three unresolved issues surrounding this accident. The first concerns whether the aircraft was, in fact, chocked when it began to move. The commander stated that he signalled to the ground crew to remove the GPU and not the chocks and the marshaller stated he did not remove them at this time. According to the AFRS, no chocks were seen in the vicinity of the nosewheel when they arrived at the scene, although two were found in the wreckage of the GPU. The degree of interference with the debris following the accident (See Fig 2a) The slope of the hard standing where the aircraft was parked was approximately 1.5%. The aircraft was facing down the slope, with the result that the component of the aircraft weight of around 17.5 tonnes acting down the slope was some 260 kg force. The only other force contributing to the forward movement of the aircraft was the exhaust efflux from the engines, but the aircraft manufacturer indicated that this force would have been “negligible”. Figure 2b Brake system schematic diagram 12 © Crown copyright 2009 AAIB Bulletin: 2/2009 TC-MBG EW/C2008/02/01 cannot be established with certainty, so the reported absence of chocks does not necessarily mean that they were absent at the time of the event. The second issue is whether the parking brake was set. The crew could be heard going through the ‘Before Start’ checklist on the CVR, but, as a result of an interruption from the travelling engineer, there was no verbal response to the challenge “Parking brake set”. Nevertheless, the commander stated that he had set the brake correctly and no comment was heard on the CVR to indicate that the parking brake was not set, or reset, after the impact. The third issue concerns the amount of leakage in the aircraft pneumatic system and the likely brake pressure available immediately prior to the accident. Subsequent tests showed that the leak rates for the brake and main systems were respectively 450 and 200 psi/hr. These values clearly exceeded the Maintenance Manual limit of 100 psi/hr, although it is possible that the pneumatic pipes within the right engine nacelle were subjected to a series of shocks and vibrations during the accident, resulting in the exacerbation of existing leaks or the generation of new ones. Pneumatic system leaks can be expected on a 37-year- old aircraft such as this, but even with minimal leakage, the charging system (which is capable of generating 1,000 psi increase every 20 minutes at 10,000 rpm) could struggle to maintain adequate pressure in the event of a long taxi with frequent brake and steering applications. The alternate/emergency system is available for occasions when brake pressure falls below the minimum value, although it would not normally be used when starting the aircraft. If it is assumed that the aircraft arrived on the stand earlier in the day with the system fully topped up at 3,000 psi, it is likely that this pressure would have almost entirely dissipated, with the as-found leak rate, during the 15 hours or so the aircraft was parked. Furthermore, the aircraft was moved once during the day, using a tug, in order to position it facing in the direction of forecast high winds. This would have involved at least one parking brake release/set cycle, which would have further reduced the stored pressure. The crew reported observing a brake pressure of 1,800 psi when they boarded the aircraft prior to the accident, which would suggest that the leak rate may have been considerably less than the subsequent tests indicated. However, if the pressure had been observed to be low, the crew had the option of summoning a charging trolley, and, moreover, would have had the time to do so whilst awaiting the de-icing vehicle. Despite the delay, the CVR indicated that the checklists were being worked through in an unhurried manner, with no evidence to suggest an intention to make up for lost time with a rushed departure. Thus, lack of system pressure is perhaps the least likely of the possible scenarios. Of the forces acting on the aircraft causing it to move forward unexpectedly, gravity would have been the most significant, with a small contribution from the propeller thrust (assuming the propellers were at their ground fine settings) and a smaller contribution from the jet efflux from the engines. The contracted operator and the subcontracted/aircraft operator had not received a copy of MDD 04/07 prior to the accident and the flight crew of TC-MBG had been given approval to self-manoeuvre the aircraft off the North Cargo Apron on several occasions prior to the accident, contrary to the instructions in the MDD. Had these instructions been followed, the accident is unlikely to have occurred. This is due to the fact that once the aircraft had been towed onto the taxiway centreline, the 13 © Crown copyright 2009 AAIB Bulletin: 2/2009 TC-MBG EW/C2008/02/01 fall of the taxiway would have been laterally across the aircraft, so that the component of gravitational force would have acted sideways, instead of forwards. Even if the aircraft were not restrained with the brakes or chocks, it is unlikely to have moved. Handling agent’s actions A representative for the handling agent stated that they have several procedures in place to audit the performance of their ground crew both internally and externally, covertly as well as overtly. These processes are formally recorded and actions are taken to address any deficiencies found. The handling agent had assumed that the airport operator sent MDDs to operators directly, but this accident showed that this was clearly not the case. As a result, the handling agent is introducing a formal procedure to ensure that in future all MDDs are sent to aircraft operators. Conclusion The aircraft moved forward inadvertently after engine start, causing its right propeller to strike a GPU. Possible explanations include that the parking brake was not set, the chocks had slipped from the nosewheel, or the chocks were removed prematurely. There was insufficient evidence to determine which of these scenarios was the most likely. Contributory factors were: the aircraft was facing down a slight downslope, the ramp was slippery due to the weather conditions and the flight crew increased engine speed to top up the pneumatic system pressure. The airport operator’s instructions contained in MDD 04/07 required aircraft facing nose-out on North Cargo Apron stands to be towed onto the taxiway centreline, prior to starting engines. Had these instructions been complied with, the accident would probably have been avoided. 14 © Crown copyright 2009 AAIB Bulletin: 2/2009 VP-AAG EW/B2008/02/01 ACCIDENT Aircraft Type and Registration: Pilatus Britten Norman BN2A 26 Islander, VP-AAG No & Type of Engines: 2 x Lycoming IO-540-E4C5 piston engines Year of Manufacture: 1969 Date & Time (UTC): 2 February 2008 at 1420 hrs Location: Wallblake International Airport, Anguilla Type of Flight: Commercial Air Transport Persons on Board: Crew - 1 Passengers - 2 Injuries: Crew - 1 (Minor) Passengers - 2 (Minor) Nature of Damage: Substantial Commander’s Licence: Air Transport Pilot’s Licence issued by the FAA Commander’s Age: 41 years Commander’s Flying Experience: 4,217 hours (of which 693 were on type) Last 90 days - not reported Last 28 days - 17 hours Information Source: Aircraft Accident Report produced by the British Virgin Islands office of Air Safety Support International Synopsis The aileron gust lock was not removed prior to flight, resulting in loss of control after takeoff. Distracted by efforts to accommodate a non-revenue passenger on this cargo flight, the pilot did not complete a pre-flight check or check the full and free movement of the flight controls before takeoff. History of the flight The commander intended to fly the aircraft from Anguilla Wallblake International Airport (AXA) to the neighbouring island of St Maarten (SXM) to await cargo inbound on another flight. The cabin of the aircraft was configured for cargo operations with no passenger seats fitted, as the only other planned occupant was the operator’s Chief Engineer, who would be sitting beside the commander in the right hand seat. However, the commander asked the operator if he could take a family member with him to SXM. The operator agreed and an extra seat was fitted. Witnesses stated that the commander appeared “rushed” prior to departure. The commander stated that he partially carried out the normal pre-flight inspection. He then started the engines. Before taxiing he realised that the nose landing gear chocks were still in place so he shut down the left hand engine, removed and stowed the chocks and then restarted the left engine. 15 © Crown copyright 2009 AAIB Bulletin: 2/2009 VP-AAG EW/B2008/02/01 The aircraft took off from Runway 10 at 1415 hrs. At between 100 ft and 150 ft the commander initiated a left turn but after some initial movement the ailerons jammed. When he discovered that he was unable to straighten the ailerons he attempted to return to land on Runway 10. The other flight controls did not appear to be restricted. With the ailerons jammed, the aircraft continued to turn to the left, losing altitude as it flew over a settlement to the north of the aerodrome, until pointed directly at the Air Traffic Control tower, causing the Air Traffic Control Officer (ATCO) to abandon the tower. The commander judged that the aircraft was too fast and high to attempt a landing and therefore initiated a go-around, applying full power. He continued the left turn, losing height and speed to position the aircraft for another approach but, as the aircraft descended over the northern edge of the runway, its left wing struck the perimeter fence. On impact the aircraft spun about its vertical axis with its wings level and continued sliding sideways on its right side for approximately 80 ft before coming to rest facing north-west. The commander made a radio call to inform ATC that everyone on board was safe. The aircraft was substantially damaged but there was no fire or obvious fuel leakage and no serious injuries to the three occupants. On vacating the aircraft the commander noticed that the left aileron gust lock was still in place between the inboard section of the aileron and the fixed trailing edge of the wing. Aircraft examination Examination of the left aileron and trailing edge revealed damage that was inconsistent with the impact sequence, indicating that the aileron and trailing edge had been deformed by an external object. There were no indications of an internal defect that would have contributed to jamming of the aileron. A review of the technical logbook and previous aircraft scheduled maintenance work pack covering the previous 12 months did not reveal any irregularities. Further inspection of the aileron gust lock revealed that it was not the type supplied by the aircraft manufacturer, which comprises two plates that clamp the aileron from above and below to prevent movement. The gust lock in use was a triangular metal cap positioned over the trailing edges of the wing and aileron and secured with a bungee cord. A hood on the opposite end of the bungee cord formed the pitot/static head cover. As the aileron gust lock had remained in place, so had the pitot/ static cover, rendering the altitude, airspeed and rate of climb instruments unreliable. The aircraft manufacturer supplies a separate pitot/static cover. Conclusions The commander was probably distracted from his normal duties whilst arranging additional seating to accommodate the second passenger. He did not complete the requisite pre-flight check or the subsequent check of full and free movement of the flight controls, either of which would have revealed an obstruction to proper operation of the ailerons. Overseas Territories Report Please note that a more comprehensive report, produced for the Governor of Anguilla, by Air Safety Support International, is available from: Governor’s Office, Old Ta, PO Box 60, The Valley, Anguilla AI-2640, West Indies All accidents and serious incidents that occur in the United Kingdom Overseas Territories are now investigated directly by the AAIB and a full report published by the AAIB. 16 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 INCIDENT Aircraft Type and Registration: AS332L2 Super Puma, G-CHCF No & Type of Engines: 2 Turbomeca MAKILA 1A2 turboshaft engines Year of Manufacture: 2001 Date & Time (UTC): 20 November 2007 at 2057 hrs Location: Aberdeen Airport, Scotland Type of Flight: Training Persons on Board: Crew - 3 Passengers - None Injuries: Crew - None Passengers - N/A Nature of Damage: None Commander’s Licence: Airline Transport Pilot’s Licence Commander’s Age: 50 Commander’s Flying Experience: 13,199 hours (of which 2,040 were on type) Last 90 days - 118 hours Last 28 days - 37 hours Information Source: AAIB Field Investigation Synopsis A Training Captain was conducting an Operational Proficiency Check (OPC); the pilot under training was required to demonstrate a clear area rejected takeoff. The helicopter was equipped with a Training Idle System (TIS) which was in use to simulate a failure of the left engine. The helicopter took off along Runway 16 at Aberdeen; at about 28 kt the commander simulated a failure of the left engine and the takeoff was rejected. The pilot flared the helicopter to reduce speed and descended towards the runway. As the collective control lever was raised to reduce the rate of descent, the overspeed protection system shut down the right engine. Main rotor rpm (Nr) decayed rapidly and the helicopter touched down firmly before rrpm could be restored. The right engine freewheel unit had failed causing that engine to overspeed; this was contained by the overspeed protection system shutting down the engine. Four Safety Recommendations are made. History of the flight The purpose of the flight was to conduct standardisation training and OPCs on two pilots who had recently completed their type training and Licence Skills Test. The weather was good with a surface wind from 140° at 5 kt, visibility was in excess of 10 km with a few clouds at 3,700 ft, the temperature was +7°C and the QNH 1010 hPa. The intention was to commence the training by carrying out a maximum performance 17 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 rejected takeoff. The flight had been fully briefed and the performance, weight and balance calculations had been completed prior to departure. The helicopter performance calculations were based on the training weights permitted in the Flight Manual Supplement relating to the engine TIS. The TIS allows the commander to simulate an engine failure on either engine by reducing its power to a training idle condition. The engine not selected to training idle powers the rotor system and is referred to in this report as the operating engine. Should the operating engine fail, the engine at training idle automatically accelerates to power the rotors. Following a normal start on both engines a freewheel check was carried out and both freewheel units operated normally. The TIS was tested in accordance with the operator’s Standard Operating Procedures (SOPs) and found to be fully serviceable. The helicopter was ground taxied to Runway 16, which is 1,829 m long, 46 m wide and has an asphalt surface. Following two demonstrations of an engine failure in the hover, using the TIS, the commander then demonstrated the rejected takeoff profile using a Takeoff Decision Point (TDP) of 60 kt. Following this demonstration, the helicopter stopped on the runway approximately half way along its length and was then ground taxied back to the threshold of Runway 16. A rejected takeoff was then flown by one of the pilots under training using the same TDP and, as before, approximately half the length of the runway was used. Since there was sufficient runway length remaining, the pilot repeated the exercise. The helicopter was initially established in a 10 ft hover and then accelerated along the runway. As the airspeed passed through 28 kt at a height of 39 ft, the commander simulated a failure of the left engine using the TIS. The pilot lowered the collective control lever and pitched the nose up to 20° in order to reduce speed. As the speed decayed, the nose was lowered and the helicopter descended normally. The collective control lever was raised to cushion the landing, but at about 10 ft the crew heard the sound of an engine running up, accompanied by a loud bang and the sound of the low Nr warning. The commander took control of the helicopter, adopted the landing attitude and raised the collective control lever to its maximum limit. The helicopter continued to descend and touched down firmly with the left engine accelerating. The Nr which had decayed to 68% just prior to the touchdown began to increase and eventually stabilised at 90%. The crew noted what appeared to be smoke or vaporised fuel on the right side of the helicopter and requested the attendance of the Airport Rescue and Fire Fighting Service (ARFFS). The crew identified from the cockpit indications that the right engine had suffered an overspeed condition and carried out the engine shutdown drill in accordance with the emergency checklist. Following confirmation from the ARFFS that there were no signs of fire, and noting that all other helicopter systems were normal, the commander taxied back to the operator’s parking area. After a discussion with engineering control, the helicopter was shut down. The helicopter was examined in accordance with the requirements of the engine overspeed inspection, after the removal of the right engine, at the operator’s maintenance facility. The main rotor gearbox right freewheel shaft was found to rotate freely in both a clockwise and an anti-clockwise direction. The gearbox was removed for investigation. Weight and balance The maximum permitted takeoff training weight for the ambient conditions was 8,880 kg, with a maximum permitted landing weight of 8,450 kg. The actual takeoff 18 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 weight at the commencement of the takeoff was 8,261 kg. The CG limits for the AS332L2 helicopter at this weight are 4.5 m to 4.95 m aft of the datum; the CG position of G-CHCF at the time of the incident was 4.67 m. Main rotor gearbox The AS332L2 is fitted with Makila 1A2 engines, which provide an additional 132 shp for single engine operation compared to the Makila 1A1 engine fitted to the AS332L1. Power from each of the engines is transmitted to the main rotor gearbox, which is common to all AS332 variants, through two input drive gearboxes fitted to the forward face of the main rotor gearbox. In the event that an engine fails or is shut down, a freewheel unit within the input drive gearbox prevents the engine being back driven through the gearbox by the remaining operating engine. The freewheel is a ‘ramp and roller’ unit, see Figure 1. The rollers are positioned on the engine driven ‘ramped’ freewheel shaft by a cage. As torque is applied to the freewheel shaft, the rollers are forced up the ramps locking the freewheel shaft to the gearbox input shaft allowing the engine to ‘drive’ the gearbox. In the event that engine torque is lost the rollers move down the ramps due to the relative rotation of the freewheel shaft and the gearbox drive shaft, disengaging the engine from the gearbox. The roller cage is fitted with a spring which holds the rollers towards the upper end of the ramps to minimise roller slip during engagement. Engineering investigation Main rotor gearbox The main rotor gearbox was disassembled at the operator’s overhaul facility under AAIB supervision and in the presence of the manufacturer’s representative. No mechanical defects were found within the gearbox with the exception of the right engine freewheel unit which had been severely damaged. Examination of the freewheel unit showed that the roller cage had rotated to a point where the rollers had overridden the freewheel ramps, and moved into the adjacent ‘trough’ in the freewheel shaft disengaging the engine output shaft from the gearbox. Figure 1: ‘Ramp and roller’ freewheel unit Engineering investigation Main rotor gearbox Figure 1 ‘Ramp and roller’ freewheel unit 19 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 The freewheel anti-rotation stops had failed as a result of the roller cage rotating into them with significant force and all the rollers showed signs of deformation and mechanical damage. The shaft exhibited signs of significant mechanical wear on the ramps together with some burring of the ramp lips, produced when they had been ‘over-ridden’ by the rollers. Metallurgical examination confirmed that there were no material abnormalities within the freewheel shaft, the rollers or the roller cage. No evidence of a failure was identified during the inspection of the right engine which may have contributed to the failure of the right freewheel unit. Manufacturer’s experience and actions Whilst there have been previous failures of the freewheel during engagement, the incident experienced by G-CHCF was the first failure of a fully engaged AS332 freewheel. Prior to February 2007, the allowable wear limits for the freewheel shaft ramps had been the same for both the AS332L1 and L2 gear boxes. During the overhaul of gearboxes, the manufacturer had identified that the wear rate of the AS332L2 freewheel assemblies was higher than that seen on the AS332L/L1 fleet and that the right freewheel shaft was subject to significantly higher wear rates than the left freewheel shaft. Whilst the reason for this higher rate of wear was not fully understood, it was believed to be due to variations in the torsional loading and rigidity within the rotor drive system. The increased wear rate of the L2 freewheel shafts had led to a significant increase in the number of shafts being scrapped due to excessive wear, with most right freewheel shafts being scrapped at the first exposure (3,000 hrs). In order to prevent an ‘in service’ freewheel failure due to excessive ramp wear, in February 2007 Eurocopter issued Repair Letter (RL) 214, which introduced tighter wear limits for the L2 freewheel shaft ramps. There was no requirement to re-inspect units overhauled prior to the release of RL 214. As a result of this incident a review of the overhaul records of all AS332L2 main rotor gearboxes was completed by the manufacturer which identified those which may have been exposed to the potential of freewheel failure in operation. These units fell into two groups: those in which both freewheel units may have operated for at least 3,000 hours in the right input gearbox, and those where only one of the freewheel units had operated in that position for more than 3,000 hours. Those units within the first group were exposed to the potential of a double freewheel failure whilst those in the second were exposed to a single failure. On 20 December 2007 the manufacturer issued Alert Service Bulletin (ASB) 01.00.74 to require the removal and inspection of the first group of gearboxes within 40 hours or before 31 December 2007 (whichever was the earlier) and the second group within 100 hours and before 31 January 2008. This action was mandated by the publication of EASA Airworthiness Directive 2007-0312-E on 21 December 2007. Maintenance records The published overhaul life of the AS332L2 main rotor gearbox is 3,000 hours. The operator’s records confirmed that the gearbox fitted to G-CHCF had operated for 2,644 hours since its last overhaul in November 2005. It was confirmed that the right freewheel shaft had been installed for a total of 8,942 flying hours and the left shaft for 5,652 hours at the time of the incident. A review of the records showed that the failed freewheel unit had been installed in the right input gearbox for 5,652 hours. They also confirmed that both the left and right freewheel shafts had been inspected and found within the published limitations applicable at the time of the last overhaul. Tooling calibration records confirmed that all the tooling used had been correctly calibrated at the time of this inspection. 20 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 Recorded information Flight Recorders The helicopter was equipped with both a Combined Voice and Flight Data Recorder1 (CVFDR) and a Health and Usage Monitoring System (HUMS). The CVFDR was capable of recording just over eight hours of data and 90 minutes of audio respectively. Parameters pertinent to this investigation were the main rotor speed, engine free power turbine speed (Nf), gas generator speed (Ng) and engine torque. A time history of the relevant parameters recorded during the incident is shown in Figure 2. The HUMS, and its associated data, is discussed later. CVFDR Recorded Information The CVFDR was removed from the helicopter and replayed at the AAIB. Data indicates that the right engine (No 2 engine) was started at 1909 hrs and the left engine (No 1 engine) five minutes later. Both the engine start and subsequent pre-flight checks were normal and the commander confirmed the TIS was operating correctly. At 1952 hrs the helicopter entered Runway 16 with the pilot in the first officer position (FO) as the handling pilot. The FO flew the helicopter into a 10 ft hover, before the commander simulated the failure of the left engine, using the TIS, and the helicopter slowly descended to the ground. This exercise was repeated from the same height, before the commander demonstrated a rejected takeoff, with the TIS being used to simulate the failure of the left engine as the helicopter approached 60 kt at a height of 40 ft. About three minutes later, the helicopter was cleared to re-enter Runway 16, the FO flew the helicopter Footnote 1 Penny and Giles manufactured CVFDR, part number 900/51508, serial number 1030/10/93. into the hover and the commander briefed him as to when he would activate the TIS. The helicopter then transitioned into the climb and at about 50 kt and 50 ft, the commander activated the TIS. The helicopter landed without incident and came to a stop on the runway. The commander confirmed that they would perform a further rejected takeoff. During the transition, the commander activated the TIS to simulate the failure of the left engine when the helicopter was at 39 ft, at which time the airspeed was about 28 kt. Initially everything appeared normal, but at a height of about 10 ft, the right engine Nf speed rapidly increased to 115%, before the engine was automatically shutdown and the main rotor speed started to decay. (See points A and B, Figure 2). Almost immediately, the low rotor speed aural warning activated and the commander took control of the helicopter, just before it landed firmly. As the right engine had started to run down, the left engine responded with an increasing Ng speed. By the time the left engine had reached its normal operating speed, the helicopter had already landed. From the point at which the right engine had started to rundown, the left engine had taken about three seconds to increase from the TIS setting of 69% to 91% Ng. The main rotor speed had decayed from about 97% to about 73% in two seconds. When the right engine had started to rundown, the engine torque indication had rapidly decreased and increased twice (See point C, Figure 2). The indications were later attributed to a misalignment of the torque sensing components, which had become misaligned as a result of the freewheel unit failure. Analyses of the CVFDR data did not identify any abnormalities in either the operation of the helicopter or the characteristics of the freewheel unit. However, it did identify one defect in the recording of data from 21 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 Figure 2 - Salient FDR Parameters Figure 2 Salient FDR Parameters 22 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 the FDR systems mandatory tri-axial accelerometer2. The defect did not affect the operation of the HUMS or any other helicopter system. It was found that rather than the normal acceleration parameter quiescent value indicating +1g, it indicated -1g. A check by the operator on 27 November 2007 found that the tri-axial accelerometer had been incorrectly installed. Maintenance records indicated that the accelerometer had last been removed and replaced during July 2007. A fleet-wide check was carried out and no further helicopters were affected. The incorrect installation of the sensor had the effect of inverting the sense of the normal acceleration parameter and also reversing the operation of the lateral acceleration parameter. The operating range of the lateral acceleration parameter was not affected; however, that of the normal acceleration parameter was no longer compliant with the legislative requirement. Instead of having an operating range of between +9g to -3g, this was reversed to +3g to -9g. Although the Aircraft Maintenance Manual (AMM) installation procedure provided a diagram of the correct orientation of the sensor, there was no requirement to carry out a post-installation test. On this occasion, the incorrect installation of the tri-axial accelerometer did not result in a loss of information, although under different circumstances, information from a mandatory parameter could have been compromised. The helicopter manufacturer has responded to the installation error by confirming that it will be carrying out a review of the AMM procedure. Footnote 2 The FDR system is equipped with a dedicated tri-axial accelerometer, which provided normal, lateral and longitudinal acceleration data. Health and Usage Monitoring System Overview and system description In accordance with legislation3, the helicopter was equipped with a Health and Usage Monitoring System (HUMS). The system is designed to record vibration data from sensors that are strategically placed around the helicopter. Data can then be analysed to detect incipient defects in the major components of the helicopter, before they can become a hazard to flight. The system may also be used to improve the reliability of the airframe and its components by identifying sources of abnormal or increasing vibration. HUMS data trending is predicated on the comparison of data that has been obtained during as stable and as consistent a period of flight as practicable. For this reason, data is most typically recorded when either on the ground or in the cruise. First generation systems, such as North Sea HUMS and Integrated HUMS (IHUMS), were developed in the late 1980s and early 1990s during North Sea helicopter operations. These early systems were installed, developed and supported by the helicopter operators and HUMS equipment manufacturers, with approval from the CAA. G-CHCF utilised a later generation system known as EUROHUMS. This differed from the first generation systems by being developed and supported by the helicopter manufacturer. In common with the design philosophy of other rotary wing health monitoring systems, EUROHUMS does not record data continuously in flight. At pre-defined Footnote 3 The Civil Aviation Authority (CAA) issued AAD 001-05-99, which became effective on 7 June 1999. The AAD made the installation and use of health monitoring systems mandatory for United Kingdom registered helicopters issued with a Certificate of Airworthiness in the Transport Category (passenger), which had a maximum approved seating configuration of more than nine passengers. 23 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 time intervals, during either the ground or cruise phase, snapshots of vibration information specific to particular engine or drive train components are recorded. Components with high rotational speeds, such as the engine input shafts, would be recorded more frequently than lower speed components, such as those within the main gearbox. At the time of the incident, data pertaining to the engine input shafts were recorded once every 20 minutes when the helicopter was in the cruise phase. The engine input shafts were not recorded during any other phase of flight. Recording intervals for components had been refined over a number of years and had been demonstrated as providing suitable levels of detection by the helicopter manufacturer. In accordance with UK legislative requirements, HUMS data is downloaded and analysed once per day. Data is downloaded into a ground-based analysis tool which detects any imbalance; misalignment; damaged, eccentric or cracked gears, or bearing wear within the main rotor gearbox and engines. An indication of the health of important components is thus provided and caution or warning indications can be provided if predefined limits are exceeded. If a caution or warning for a component is generated, or an adverse trend is identified, the ground-based system is able to provide the operator with details of corrective maintenance action. After maintenance, ongoing HUMS monitoring and in-service inspections are used to ensure that any corrective action was successful. Although HUMS has demonstrated that it provides an effective means of monitoring, there are certain components, such as freewheel units, which may not exhibit any detectable levels of vibration during normal operation and, as such, cannot be monitored effectively by a vibration monitoring system. HUMS data Prior to the incident, the operator had been monitoring a progressive increase in the right engine input shaft vibration level. On the 18 November 2007, two days before the incident, the operator replaced the right engine. In the following two days, a small number of data points were recorded by the HUMS. These points indicated a reduction in vibration for the right engine input shaft. However, results from the subsequent stripdown did not identify any keys areas of damage and the operator and helicopter manufacturer discussed whether the increasing trend may have been due to wear of the freewheel unit. The helicopter manufacturer reviewed the HUMS data and confirmed that a progressive increase in vibration levels relating to the right engine input shaft had been detected and that, following engine replacement, the vibration levels had reduced. They confirmed that vibration levels recorded prior to the engine replacement had been at a low level and that the subsequent stripdown findings were not unusual considering these low levels. Both the helicopter manufacturer and operator concluded that the increasing vibration trend had been as a result of normal wear within the engine or its coupling to the main rotor gearbox and was not related to vibration of the freewheel unit. Operational aspects Takeoff and landing profiles The AS332L2 has takeoff and landing profiles for both clear area and helipad operations. These profiles ensure that the helicopter complies with Performance Class 1 requirements when operated at a weight appropriate for the ambient conditions. In order to ensure that the helicopter can either land or 24 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 fly away in the event of an engine failure the profiles have Takeoff Decision Points (TDPs). During a clear area takeoff the TDP is calculated on the distance available for landing should the helicopter have to abandon or reject the takeoff and/or the maximum weight, whichever is the more limiting. The TDP is based on Indicated Airspeed (IAS) and an associated height. The TDP may be varied between 20 kt and 60 kt at 10 kt increments and associated heights of 20 ft to 40 ft in 5 ft increments. The heavier the helicopter, the higher will be the IAS and height that define the TDP. With a limiting reject distance, the TDP IAS and height will be lower. If an engine fails prior to TDP the takeoff is rejected and the helicopter should stop within the pre-determined distance. Should an engine fail after the TDP, the helicopter can be flown away providing a target IAS is maintained and the correct power is set on the operating engine. The Landing Decision Point (LDP) for a clear area landing profile is a fixed height of 100 ft and requires an IAS of 35 kt at that height with a rate of decent less than 400 fpm at the LDP. In the event of an engine failure before LDP the helicopter may continue to land or the pilot may go around and climb away. After LDP the helicopter must be landed and should stop within the promulgated landing distance. When operating at a helipad the profile requires that the helipad must have a minimum diameter of 24 metres. The TDP is then a fixed point, 130 ft above the pad and with a horizontal back up distance from the pad of 125 m. LDP is the same as for a clear area but the approach is steeper. The takeoff, landing and rejected takeoff profiles are set out below. Clear area profiles Takeoff with Single-Engine Failure recognised at or before the TDP Abort takeoff as soon as engine failure occurs. Simultaneously ● reduce the collective pitch while maintaining a rotor speed of at least 250 rpm (94%), and adopt a nose-up attitude of 10° to 20°, allowing the aircraft to climb slightly. As aircraft begins to sink, ● control attitude and cushion touchdown. On the ground, reduce collective pitch to ● minimum and use wheel brakes to stop the aircraft. Takeoff with Single-Engine Failure at or after the TDP Continue the takeoff procedure. VI (IAS) = V.TOSS (IAS) - 10 kt [where V.TOSS is the Takeoff Safety Speed] will be lower. If an engine fails prior to TDP the takeoff is rejected and the helicopter should stop within the pre-determined distance. Should an engine fail after the TDP, the helicopter can be flown away providing a target IAS is maintained and the correct power is set on the operating engine. The Landing Decision Point (LDP) for a clear area landing profile is a fixed height of 100 ft and requires an IAS of 35 kt at that height with a rate of decent less than 400 fpm at the LDP. In the event of an engine failure before LDP the helicopter may continue to land or the pilot may go around and climb away. After LDP the helicopter must be landed and should stop within the promulgated landing distance. When operating at a helipad the profile requires that the helipad must have a minimum diameter of 24 metres. The TDP is then a fixed point, 130 ft above the pad and with a horizontal back up distance from the pad of 125 m. LDP is the same as for a clear area but the approach is steeper. The takeoff, landing and rejected takeoff profiles are set out below. Clear area profiles Takeoff with Single-Engine Failure recognised at or before the TDP Abort takeoff as soon as engine failure occurs. • Simultaneously reduce the collective pitch while maintaining a rotor speed of at least 250 rpm (94%), and adopt a nose-up attitude of 10° to 20°, allowing the aircraft to climb slightly. • As aircraft begins to sink, control attitude and cushion touchdown. • On the ground, reduce collective pitch to minimum and use wheel brakes to stop the aircraft. Takeoff with Single-Engine Failure at or after the TDP Continue the takeoff procedure. VI (IAS) = V.TOSS (IAS) - 10 kt [where V.TOSS is the Takeoff Safety Speed] • Control NR • Accelerate to or maintain V.TOSS. • At V.TOSS reduce collective pitch to 2-minute One Engine Inoperative (OEI) rating and simultaneously shift Ng stop to 2-minute OEI rating position. • At a height of 200 ft, level off and accelerate from V.TOSS to Vy (best rate of climb airspeed). 25 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 Control N ● r Accelerate to or maintain V.TOSS. ● At V.TOSS reduce collective pitch to 2-minute ● One Engine Inoperative (OEI) rating and simultaneously shift Ng stop to 2-minute OEI rating position. At a height of 200 ft, level off and accelerate ● from V.TOSS to Vy (best rate of climb airspeed). At 200 ft but no later than when the OEI LO ● caption flashes, adjust collective pitch to maximum continuous OEI rating. Retract landing gear and continue climbing ● at Vy. NOTE: If landing gear is retracted below 60 kt, then red L/C caption will flash. Normal Landing After reaching LDP, proceed with a straight- ● in approach, reducing speed regularly to enter hover IGE at a height of 10 feet. Proceed with normal landing. ● Helipad profiles Takeoff with Single-Engine Failure at or before the TDP (before aircraft rotation) Abort takeoff as soon as engine failure occurs. Takeoff with Single-Engine Failure at or after the TDP (aircraft rotation started) Continue flight Training idle system The maximum OEI limits cannot be used unless an actual engine failure occurs. The TIS enables OEI flight training to be conducted using non-damaging power levels, provided that the helicopter weight has been reduced to the associated training limit. A guarded switch, associated with each engine, is provided on the overhead control unit. This switch is used to perform the following actions simultaneously: 1. Simulation of an engine failure by reducing the power of that engine to idle rating, with Nf being governed at a value slightly lower • At 200 ft but no later than when the OEI LO caption flashes, adjust collective pitch to maximum continuous OEI rating. • Retract landing gear and continue climbing at Vy. NOTE: If landing gear is retracted below 60 kt, then red L/C caption will flash. Normal Landing • After reaching LDP, proceed with a straight-in approach, reducing speed regularly to enter hover IGE at a height of 10 feet. • Proceed with normal landing. Helipad profiles Takeoff with Single-Engine Failure at or before the TDP (before aircraft rotation) Abort takeoff as soon as engine failure occurs. Takeoff with Single-Engine Failure at or after the TDP (aircraft rotation started) Continue flight Training idle system The maximum OEI limits cannot be used unless an actual engine failure occurs. The TIS enables OEI flight training to be conducted using non-damaging power levels, provided that the helicopter weight has been reduced to the associated training limit. A guarded switch, associated with each engine, is provided on the overhead control unit. This switch is used to perform the following actions simultaneously: 1. Simulation of an engine failure by reducing the power of that engine to idle rating, with NF being governed at a value slightly lower than 245 rpm, equivalent to 92.5% Main Rotor Rotation Speed (NR ). 2. De-rating the OEI stops of the other engine by approximately 1665 rpm (5%) in order to limit the Ng to a value not exceeding the takeoff power rating. The Flight Manual Supplement contains the following information regarding safety devices associated with the TIS: Function Safety Devices 26 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 than 245 rpm, equivalent to 92.5% Main Rotor Rotation Speed (Nr). 2. De-rating the OEI stops of the other engine by approximately 1665 rpm (5%) in order to limit the Ng to a value not exceeding the takeoff power rating. The Flight Manual Supplement contains the following information regarding safety devices associated with the TIS: ‘Function Safety Devices In addition to limiting the power to non-damaging ratings, the following safety devices are provided when using the TRAINING function: In the event of an incorrect training manoeuvre ● or an actual failure of the engine supplying the power, the required power can be obtained from the idling engine simply by pulling the collective pitch lever. The principle is as follows: as soon as the NR drops below 240 rpm (90.5%) the idling engine supplies the amount of power required until the actual 30-second OEI rating is reached (at NR = 220 rpm (83%)), using static droop effect. There is no risk of a false manoeuvre with ● the fuel flow control levers, since these levers remain in flight position; the OEI is simulated using specific controls. In the event of unintentional action on more ● than one TRAINING IDLE control, the function is inhibited and a minor governing fault is indicated. The function will be re-established after landing, shutting down both engines and re-starting the engines according to the standard procedure. Should the engine running in TRAINING IDLE ● mode fail, the procedure can be continued at the actual OEI rating (i.e. with a larger power margin) by setting the TRAINING IDLE control forward (switch guard down). Return to twin-engine flight is possible at any ● time by setting the TRAINING IDLE control forward (switch guard down).’ The Engine Monitoring Display (EMD) for the operating engine changes to the indications that the pilot would read in true engine failure condition. A letter ‘T’ appears in an inverted triangle to indicate to the pilot that the training mode is in operation. Helicopter certification The initial certification of the helicopter was carried out by the French, Direction Generale de l’Aviation Civile (DGAC). The UK CAA conducted a validation of that certification in 1991/92. No specific requirements were in place to establish the helicopter performance with the TIS selected and a subsequent failure of the operating engine. A requirement was in place which ensured that any ‘Option’ such as the TIS should not introduce an increased hazard. Prior to the introduction of the TIS, the only method of simulating a single engine failure was to retard a Fuel Flow Control Lever (FFCL). In the event of the operating engine failing that FFCL would then have to be advanced to reinstate the power available from that engine. This would have to be combined with a lowering of the collective control to prevent loss of Nr. If the FFCL were advanced too rapidly, the possibility of engine surge, flame-out or an overspeed shutdown of that engine was possible. 27 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 By introducing the TIS both FFCLs remained in the flight position and the decaying Nr, resulting from a failure of the operating engine, was used to trigger the acceleration of the engine at idle. The Digital Engine Control Unit (DECU) optimised the acceleration to restore safe flight. There was no record or data available of any tests carried out during development by the manufacturer or certification authorities to simulate the failure of an operating engine whilst the TIS was in use. The UK validation did not call for any testing of a failure of the operating engine but satisfactory flight tests of the TIS in operation were conducted. Operator’s safety actions General Following the incident involving G-CHCF, the operator ceased training and testing on the AS332L2 using the TIS. This was followed by a ban on using the TIS, which was imposed by the UK CAA for all operators. Tests and evaluation In order to understand the potential hazards which may arise when using the TIS the operator evaluated a number of test conditions in the AS332L2 flight simulator based in Marseille, France. The test points were identified within the takeoff and landing profiles for both clear area and helipad operations. The tests were based on the operating engine failing whilst the other engine was in the training idle mode. The test pilot then attempted to preserve Nr and safe flight whilst the other engine accelerated to the extent that the helicopter could either be landed or flown away. The Flight Simulator The flight simulator was a level D Synthetic Training Device (STD) with motion and visual display. The data on which the simulator was modelled was derived from the manufacturer’s AS332L2 flight test and certification programme. The certification programme confirmed the helicopter performance when using OEI 30-second power but a failure of the operating engine, with the other engine in the training idle mode, was not carried out. Although the modelling of the TIS was not derived from flight test data, it was considered that the simulator offered a reasonable indication of the likely outcome of the event being simulated. Test conditions The simulator was representative of the operator’s standard AS332L2 configuration. The conditions were set to sea level standard (1013 hPa, +15°C) with no wind. The runway at Hong Kong was used as it is at sea level and the model database gave good visual cues throughout the tests. Helicopter mass was set at 8,600 kg for the clear area work and 7,500 kg for the helipad exercises. These weights were representative training mass values (maximum training mass for the conditions would have been 8,900 kg and 7,600 kg respectively). Tests made Failures of the operating engine were investigated before and after TDP, and before LDP on a clear area, as well as before TDP on a helipad. Failures were initiated by selecting the left engine to training idle, at the target speed, using the TIS, followed one second later by injecting a failure into the operating engine. The operating engine was initially failed by introducing an Nf overspeed condition, but there was a marked delay in the left engine accelerating. By using the FFCL to stop 28 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 the right engine the time delay was reduced. The right engine failure was activated as the pilot flying (PF) was commencing the required recovery manoeuvre. The inherent delays in the simulation and the inertia of the helicopter meant that the actual failure of the operating engine during the takeoff exercises occurred at between 5 kt and 10 kt above the target speed. Each test point was recorded and the relevant data (Nr, Ng and radio height) could be noted. Findings The behaviour of the simulator was considered against the Flight Manual Supplement (FMS), Supp 3 (One Engine Inoperative (OEI) Flight Training Procedures), which contains the limitations, procedures and performance data for use of the TIS. It makes the following statement: ‘In the event of an incorrect training manoeuvre or an actual failure of the engine supplying the power, the required power can be obtained from the idling engine simply by pulling the collective pitch lever. The principal is as follows: as soon as the NR drops below 240 rpm (90.5%), the idling engine supplies the amount of power required until the actual 30-second OEI rating is reached (at NR =220 rpm (83%)), using the static droop effect.’ The tests revealed inconsistencies in the simulator modelling, not only in the response to low Nr and variation of collective movement, but also in the different responses depending on how the engine failure was initiated. The low Nr trigger at 220 rpm and removal of the training idle stop (release of real OEI 30-second power) were consistent, and appeared to be in accordance with FMS Supp 3 if the failure was introduced using the FFCL, but inconsistent if the failure was introduced via the Nf overspeed. This could be assessed in the helicopter without going to the actual OEI; it would be sufficient to see the release of the real OEI rating and an Ng increase through 90% in response to Nr decay. It was evident from the tests performed in the simulator that clear area rejected takeoffs, with failure of the operating engine during the reject (ie failures before TDP), could be recovered and the helicopter could be landed safely; this was the case relating to the incident with G-CHCF. It was also evident that failures just after TDP (with the intention to continue the takeoff) would result at best in a rejected takeoff, depending on the distance of suitable landing surface remaining. Any failure of the operating engine in the first segment (ie below 200 ft) would result in the helicopter either making a forced landing or descending to or below 35 ft in the flyaway. A similar conclusion could be drawn for failures before LDP, because the failure simulated at 150 ft resulted in a controlled forced landing with insufficient height to recover Nr or Ng to the extent that a safe go-around could be considered. It is probable that at least 200 ft would be required (possibly more because the helicopter is already descending at the moment of failure). Furthermore, real intervention times are likely to be greater and this can only have a detrimental effect on the potential outcome. Conclusions From the simulator tests, the Flight Manual Supplement, Supp 3, statement set out above does not accurately reflect the behaviour of the helicopter or the technique that the pilot should adopt. Simply pulling the collective lever did not restore Nr but caused it to decay into an over-pitching condition unless the collective lever was first lowered positively to prevent this. The helicopter also touched down beyond the rejected takeoff distance following a failure of the 29 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 operating engine after TDP and in the undershoot with the operating engine failing just before LDP. It was not possible to land or fly away safely during the helipad profile. Recommendations On the basis of the simulator test results, the operator considered that the TIS should not be used to simulate engine failures in the following cases: On clear areas during continued takeoffs in ● the first segment (ie below 200 ft) On clear areas during landing not below a ● height agreed with, and authorised by, the authorities On helipads at any time ● This restriction would remain in force until a risk assessment of engine failure training on the AS332L2 had been carried out. Analysis Engineering Excessive wear to the freewheel shaft ramps resulted in the freewheel rollers overriding the ramps, disengaging the engine from the main rotor gearbox. No metallurgical defects were identified within the freewheel unit. The gearbox had been overhauled in November 2005, prior to the release of Eurocopter RL 214, in accordance with the applicable limitations and procedures in force at that time. The possibility of an in-service freewheel failure has been significantly reduced by the introduction of RL 214. Eurocopter, in Alert Service Bulletin 01.00.74, identified all the AS332L2 main rotor gearboxes which were exposed to a potential freewheel failure with a defined timescale for removal and this was mandated by the publication of EASA Airworthiness Directive 2007-0312-E on 21 December 2007. All of the affected gear boxes were removed from service and are now compliant with EASA AD 2007-0312-E. Operations The crew were properly qualified to conduct the flight and the helicopter was being operated within the weight and the C of G envelope for the manoeuvre being flown. The training exercise had been fully briefed. The failure of the operating engine freewheel unit, as the collective control was being raised, occurred in an area of the takeoff profile where recovery was possible, and this was achieved through the prompt action of the helicopter commander in taking control and performing a safe landing. From the evidence provided by the FDR, the helicopter touched down before the left engine accelerated. Once on the ground the Nr was restored. There was no test data from the flight test or certification programme with which the TIS operation could be compared to establish whether it had operated correctly. The test points carried out in the simulator flying the clear area profile with a failure of the operating engine at or just before TDP showed that a rejected takeoff could be performed successfully. This relied on the prompt reaction of the pilot and demonstrated the need for the pilot to lower the collective control, if possible, to assist with restoring Nr. This action, combined with flaring the helicopter, would assist in reducing Nr decay and providing the accelerating engine with the best conditions for restoring Nr. It also showed that the action required in the Flight Manual Supplement, Supp 3, that ‘the required power can be obtained from simply by pulling the collective pitch lever’ is incorrect. 30 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 If an operating engine failure occurs at or after TDP, or just before LDP, attempting to fly away using the technique described in the Flight Manual Supplement, Supp 3, could result in significant over-pitching with the associated build-up in the rate of descent. The simulator modelling was not sufficiently reliable in the scenario being tested to identify an exact outcome. It is, however, probable that the helicopter would touch down or descend below the 35 ft minimum height required. The point of touchdown may be beyond the rejected takeoff distance available. Using the helipad profile, with an operating engine failure before or just after TDP or just before LDP the loss of Nr and high rate of descent may make the situation irrecoverable. The point of touchdown would be short of the pad in the early stages of the profile and beyond the pad if positive airspeed was achieved. It was considered highly unlikely a successful safe landing on the pad would be achieved. Conclusions The safe outcome of this incident was dependant upon a combination of the point at which the failure of the freewheel unit occurred and the prompt corrective action taken by the commander. The information presented in the Flight Manual Supplement, Supp 3, does not appear to accurately reflect the behaviour of the helicopter or the technique to be employed following a failure of the operating engine and may provide a false sense of security if using the TIS. The principle of having a system to accelerate an engine from a training idle position, following a failure of the operating engine, is a positive safety enhancement and avoids a rapid movement of the FFCL introducing an overspeed shutdown. However, the Flight Manual Supplement, Supp 3, should alert the pilot to the limitations of the system and in particular the technique to be used should the operating engine fail. Therefore: Safety Recommendation 2009-003 It is recommended that Eurocopter should review the operation of the Training Idle System on the AS332L2 helicopter in the event of the failure of the operating engine. Eurocopter should ensure that the behaviour of the helicopter in terms of Nr recovery and any height loss are included in the Flight Manual Supplement, Supp 3. The correct pilot technique for managing such an event should also be included. This information should be based on flight test data. Furthermore, the AS332L2 is one of a number of helicopters fitted with a Training Idle System, or similar system. As no certification requirements are stipulated for such systems, there may be other helicopters where the operation of the TIS is not accurately documented. Therefore: Safety Recommendation 2009-004 It is recommended that the European Aviation Safety Agency should review the accuracy of Flight Manual information covering Training Idle Systems fitted to all helicopter types or models. They should ensure that the information on the system, the behaviour of the helicopter and the correct pilot technique to be employed in the event of the operating engine failing are correctly documented. Moreover, there is no current requirement within the certification process for the Training Idle System to be evaluated with a failure of the operating engine. Data derived from such tests would ensure that the correct information was included in the Flight Manual and that accurate data was used for the modelling of flight simulators. Therefore, the following two Safety Recommendations are made: 31 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-CHCF EW/C2007/11/03 Safety Recommendation 2009-005 It is recommended that the European Aviation Safety Agency should require that when a helicopter is fitted with a Training Idle System, or similar system, the effects of a failure of the operating engine are determined during the flight test and certification process. Safety Recommendation 2009-006 It is recommended that the European Aviation Safety Agency should ensure that where a Training Idle System is fitted to a flight simulator the handling qualities and performance of the helicopter, following the failure of the operating engine, are accurately modelled. 32 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-AMVP EW/G2008/09/14 ACCIDENT Aircraft Type and Registration: Avions Fairey SA Tipsy Junior, G-AMVP No & Type of Engines: 1 Walter Mikron 2 piston engine Year of Manufacture: 1952 Date & Time (UTC): 15 September 2008 at 1250 hrs Location: Sandown Airport, Isle of Wight Type of Flight: Private Persons on Board: Crew - 1 Passengers - None Injuries: Crew - None Passengers - None Nature of Damage: Wooden propeller broken; engine cowling dented; port aileron underside scuffed Commander’s Licence: National Private Pilot’s Licence Commander’s Age: 82 years Commander’s Flying Experience: 5,500 hours (of which 273 were on type) Last 90 days - 11 hours Last 28 days - 2 hours Information Source: Aircraft Accident Report Form submitted by the pilot and occurrence reports submitted by airport operations staff Synopsis The aircraft was landing on Runway 05 at Sandown, in good weather conditions and with a surface wind from the north at 5 to 8 kt. The pilot reported that, as the aircraft was just above the runway prior to making a ‘three-point’ landing, the port wing stalled and struck the ground. Operations staff at the airfield reported that the wing drop occurred shortly after the aircraft had bounced on landing, following which the propeller struck the ground. The aircraft came to rest in a nose-low attitude and the uninjured pilot, who was wearing a full harness, vacated the open cockpit. 33 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-ASJL EW/G2008/06/10 ACCIDENT Aircraft Type and Registration: Beech H35 (Modified) Bonanza, G-ASJL No & Type of Engines: 1 Continental Motors Corp IO-520-BA12B piston engine Year of Manufacture: 1957 Date & Time (UTC): 10 June 2008 at 1350 hrs Location: Compton Abbas Airfield, Dorset Type of Flight: Private Persons on Board: Crew - 1 Passengers - 1 Injuries: Crew - None Passengers - None Nature of Damage: Damage to nose and main landing gear, propeller and engine Commander’s Licence: Private Pilot’s Licence Commander’s Age: 41 years Commander’s Flying Experience: 195 hours (of which 37 were on type) Last 90 days - 13 hours Last 28 days - 6 hours Information Source: Aircraft Accident Report Form submitted by the pilot Synopsis Following a normal touchdown, the aircraft was approximately halfway along the runway during the roll-out when, without warning, the landing gear collapsed. The aircraft slewed to the left and came to a halt, blocking the runway. The pilot turned off the fuel and electrical services and both occupants vacated the aircraft without injury. The pilot reflected that it was possible he had inadvertently selected the landing gear switch to UP, having intended to retract the flaps. The aircraft was equipped with a weight-on-wheels microswitch that should have prevented landing gear retraction. However, it is possible that the slightly undulating nature of the runway may have caused the microswitch to open at some point during the roll out, allowing the gear to retract. 34 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-GSGZ EW/G2008/09/01 ACCIDENT Aircraft Type and Registration: CAP 232, G-GSGZ No & Type of Engines: 1 Lycoming AEIO-540-L1B5 piston engine Year of Manufacture: 1996 Date & Time (UTC): 2 September 2008 at 0940 hrs Location: Near Llay, 3 nm South of Hawarden Airport Type of Flight: Private Persons on Board: Crew - 1 Passengers - None Injuries: Crew - None Passengers - N/A Nature of Damage: Major damage to fuselage, tail and landing gear Commander’s Licence: Private Pilot’s Licence Commander’s Age: 56 years Commander’s Flying Experience: 957 hours (of which 689 were on type) Last 90 days - 34 hours Last 28 days - 10 hours Information Source: Aircraft Accident Report Form submitted by the pilot Synopsis The pilot reported that he misjudged a pre-flight visual inspection of the aircraft’s fuel quantity and the engine stopped due to fuel exhaustion during flight. The ensuing forced landing resulted in the aircraft pitching forward onto its back. This trapped the pilot who required external assistance to vacate the aircraft. History of the flight In preparation for a local aerobatic flight, the owner of the aircraft carried out the normal pre-flight checks, including a physical check of the fuel on board. The aircraft was fitted with a centre tank and a tank in each wing, with aerobatics only permitted with the wing tanks empty. The pilot confirmed the wing tanks were empty and on checking the centre tank, could see fuel and assessed the tank to be full. He did not use a dip stick. He estimated that there was sufficient fuel for at least half an hour’s flying depending on the amount of time that aerobatics were flown. The pilot strapped in and noted that the centre tank fuel gauge indicated 7/8ths full. This gauge is positioned in such a way that the pilot must move his leg to see it once strapped in. After a normal start and power checks, he took off from Hawarden Airfield and flew for approximately six minutes to reach the area in which he planned to perform the aerobatics. He completed his pre-aerobatic checks which included another check of the fuel and flew a 35 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-GSGZ EW/G2008/09/01 series of aerobatic manoeuvres lasting approximately seven minutes. The last manoeuvre was a stall turn during which the engine stopped for approximately 2-3 seconds before restarting. After recovering from the stall turn and with the aircraft level, another check of the fuel revealed the centre tank gauge indicating zero. The pilot started to return to the airfield, setting economical cruise and making a PAN call to inform ATC of his fuel state. Approximately 3 nm from the airfield the engine stopped and could not be restarted. As he was unable to glide to the airfield, the pilot conducted a forced landing in a freshly cultivated field. He reported that one reason for selecting this field was that there was a tractor working in it which might be of assistance after landing. The pilot described flying a satisfactorary approach but the roll out after touchdown was rough and bouncy. As the aircraft slowed, the nose pitched down and the aircraft flipped over onto its back, partly breaking the canopy. The pilot, who was wearing a full harness, had his head forced onto his chest by the earth which made breathing difficult. The driver of the tractor witnessed the accident and was able to lift the wing of the aircraft using the plough attached to the tractor. This allowed him sufficient room to dig away the soft earth and break the remaining canopy using a hammer. The pilot undid his harness and was able to escape from under the aircraft. Analysis The pilot considered the engine stopped approximately 15 minutes into the flight due to fuel starvation. He was not aware of any fuel leaks and despite his initial fuel checks indicating the centre tank was nearly full, he believes he misjudged the quantity and that it was more likely to have been only half full. The pilot reported that the fuel gauge did not move in a linear fashion and was therefore not an accurate method of assessing the quantity on board. In addition, due to the position of the gauge, it was not easy to continually monitor during flight. He commented that he would now fill a tank to the top as he considered this the only way to be sure of the fuel quantity it contained. The pilot did not routinely wear a parachute and had no option other than to carry out a forced landing. He reported that the accident had made him reconsider this and that he would now wear a parachute, only attempting a forced landing if there was a prepared surface on which to land. He believed this type of aircraft was likely to pitch forward onto its back on any other sort of surface, as on this occasion, and he considered himself lucky not to have suffocated or suffered a serious or fatal head injury. In this event, his situation would have been considerably more serious had it not been for the soft earth and the timely intervention of the tractor driver. 36 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-GAOM EW/G2008/09/30 SERIOUS INCIDENT Aircraft Type and Registration: CEA DR400/2+2, Dauphin, G-GAOM No & Type of Engines: 1 Lycoming O-235-H2C piston engine Year of Manufacture: 1977 Date & Time (UTC): 19 September 2008 at 0840 hrs Location: Runway 12, RNAS Culdrose, Cornwall Type of Flight: Private Persons on Board: Crew - 1 Passengers - 1 Injuries: Crew - None Passengers - None Nature of Damage: Main landing gear tyres burst Commander’s Licence: Private Pilot’s Licence Commander’s Age: 66 years Commander’s Flying Experience: 189 hours (of which 19 were on type) Last 90 days - 8 hours Last 28 days - 6 hours Information Source: Aircraft Accident Report Form submitted by the pilot, ATC recordings and further enquires by the AAIB Synopsis During a visual circuit at Royal Naval Air Station (RNAS) Culdrose, Cornwall, G-GAOM was cleared, by ATC, to “CONTINUE” as another aircraft had just landed. G-GAOM subsequently landed without clearance with the other aircraft still on the runway. As a result, G-GAOM braked very hard after landing, to avoid the other aircraft, and burst both tyres on the main landing gear. History of the flight G-GAOM was visiting RNAS Culdrose, Cornwall, from its base at Bodmin Airfield, with approximately 30 other light aircraft for a joint civil and military safety day. All visiting aircraft had been given a runway landing slot by ATC. When G-GAOM joined the circuit it contained three other aircraft and there were other aircraft joining behind. G-GAOM positioned right-hand downwind and the pilot was advised that he was number two in the landing sequence. As the aircraft turned onto right base, the pilot transmitted “TURNING RIGHT BASE BEHIND NUMBER ONE.” ATC replied “CONTINUE” because the aircraft ahead had just landed; the pilot did not acknowledge this clearance. G-GAOM subseqently landed on Runway 12 with the other light aircraft still on the runway. During 37 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-GAOM EW/G2008/09/30 the landing roll both main landing gear tyres burst before the aircraft stopped about 400 m behind the now vacating aircraft. G-GAOM was unable to vacate Runway 12 until additional assistance arrived several minutes later. Pilot’s comments The pilot commented that he did not hear the “CONTINUE” clearance transmitted by ATC. Although he saw the aircraft ahead on the runway, he considered that it would have vacated the runway before he landed. He added that he was concentrating hard on flying his aircraft, as it was high and fast on the approach, and he failed to ensure the runway was clear before he landed. He also felt compelled to land due to runway slot time pressures and the other aircraft joining. After landing he applied maximum braking to stop his aircraft before reaching the other aircraft. In hindsight he believes he should have flown a go-around. Safety actions As a result of this incident ATC at RNAS Culdrose is reportedly ensuring that all visiting civilian flying clubs are pre-briefed on the circuit procedures and are establishing several Visual Reporting Points to aid the sequencing of arriving VFR traffic. 38 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-WARP EW/G2008/10/08 ACCIDENT Aircraft Type and Registration: Cessna 182F Skylane, G-WARP No & Type of Engines: 1 Continental Motors Corp O-470-R piston engine Year of Manufacture: 1963 Date & Time (UTC): 15 October 2008 at 1100 hrs Location: Caernarfon Aerodrome, Gwynedd, Wales Type of Flight: Private Persons on Board: Crew - 1 Passengers - None Injuries: Crew - None Passengers - N/A Nature of Damage: Damage to propeller and engine shock-loaded Commander’s Licence: Private Pilot’s Licence Commander’s Age: 60 years Commander’s Flying Experience: 416 hours (of which 76 were on type) Last 90 days - 18 hours Last 28 days - 12 hours Information Source: Aircraft Accident Report Form submitted by the pilot Synopsis During the takeoff ground roll a gust of wind from the left lifted the left wing despite the pilot having held full into-wind aileron. The aircraft then pitched nose-down and the propeller struck the runway. The pilot taxied the aircraft off the runway and then shut down the engine. 39 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-AWUX EW/G2008/09/28 ACCIDENT Aircraft Type and Registration: Cessna F172H Skyhawk, G-AWUX No & Type of Engines: 1 Continental Motors Corp O-300-D piston engine Year of Manufacture: 1968 Date & Time (UTC): 27 September 2008 at 1550 hrs Location: Dowland Farm Strip, Devon Type of Flight: Private Persons on Board: Crew - 1 Passengers - 1 Injuries: Crew - None Passengers - None Nature of Damage: Left wing strut and cowling dented Commander’s Licence: Private Pilot’s Licence Commander’s Age: 53 years Commander’s Flying Experience: 305 hours (of which 242 were on type) Last 90 days - 5 hours Last 28 days - 4 hours Information Source: Aircraft Accident Report Form submitted b