The aircraft was not fitted with a flight data recorder or a cockpit voice recorder. However, the crash was recorded by multiple cameras, several of which also recorded audio. When combined with data retrieved from the memory in a GPS receiver located in the wreckage, a fairly accurate reconstruction of the aircraft's final moments could be created. The takeoff roll proceeded normally; however, after leaving the ground, the aircraft continued to pitch upwards, reaching a nose high attitude of 34 degrees within four seconds of rotation. At this point, the aircraft rolled to the right, consistent with a power-on stall (due to the direction of rotation of the propellers, the right wing will stall before the left one at high engine power). The aircraft reached a maximum right bank angle of 152 degrees, at which point the direction of the roll reversed. The aircraft struck the ground at a 70 degree right bank and 38 degree nose-down pitch, just ten seconds after lifting off. Analysis of the audio from several security cameras found no evidence of engine problems, nor could any engine issue explain the aircraft's performance. The aircraft had undergone substantial modifications to its flight control system in 1985; several cables had been re-routed to accommodate the special surveillance equipment installed. The aircraft's stick pusher system, known on the Merlin as the stall avoidance system (SAS), had also been altered for this reason. An examination of the bulb filaments in the cockpit
annunciator panel revealed that at some point prior to impact, the SAS FAULT annunciator had illuminated. In addition, recent maintenance had been performed to the aircraft's flight control system, but maintenance instructions for the modified version of the system were not available when this was done. Damage to the wreckage prevented a conclusive analysis of the primary flight controls or the SAS. Based on the performance of the aircraft, the evidence recovered from the wreckage and a study of previous incidents involving aircraft in the same family, the BEA narrowed the cause of the crash to three possibilities.
Possible causes The most plausible of the three would have been a failure or jamming in the nose down cable of the elevator control system. This scenario would explain why the aircraft continued to pitch up beyond the recommended climb attitude after rotation. In this case, the pilot would not have been able to counteract the upwards moment through pushing forward on the control column. Depending on the nature of the failure, the control column itself would not necessarily have jammed, which would have made the problem more difficult to recognize immediately. The only way to overcome an elevator control failure would be through the pitch trim system, but the aircraft reached a stall situation too quickly for this to be effective. As the aircraft approached a stall, the stick pusher would have activated. As the pusher servo is connected to the same control cabling as the control column, it would have been unable to function correctly, explaining the illumination of the SAS FAULT light. A less likely but still possible hypothesis was an inadvertent activation of the stick pusher system. This was considered possible due to the relocation of the SAS components to an area near a water evacuation drain on the aircraft. If the pusher servo had activated at the point of rotation, the pilot may have reflexively pulled against it with enough force to disable it, causing an abrupt pitch-up. Inadvertent SAS activation had happened on other SA-227 aircraft. However, the aircraft's flight path did not fully support this scenario. If the stick pusher had activated erroneously and been over-corrected by the pilot, the aircraft would likely have pitched down slightly before rapidly pitching up. Instead, the upwards pitch was more or less constant until the moment the aircraft stalled. Nevertheless, without more accurate flight data, the theory could not be ruled out completely. The least likely of the three involved the antenna for the HF radio (consisting of a cable extending from the upper forward fuselage to the tip of the vertical stabiliser) snapping during rotation. Afterwards, the cable may have become wrapped around the right-hand elevator, jamming it upwards. The HF antenna is not standard equipment on the SA-227, and the type installed had been known to fail occasionally. In 1998, a cable had jammed the elevator of a SOCATA MS.893 into a nose-down position, causing a fatal crash. Regardless, no components from the cable were found on the runway, and the physics involved to cause the cable to jam the aircraft's elevators were considered extremely improbable. The final report concluded that whichever malfunction occurred "probably originated in the specific modifications of the aircraft and in the application of an inappropriate maintenance to these modifications." ==Notes==