Flight Plan and Initial Phase
The commander filed a flight plan indicating the aircraft would depart Guernsey and proceed to the Berry Head VOR at flight level (FL) 100, then manoeuvre in the Plymouth area for a performance-related test flight. Start clearance was requested at 0902 hours, and after a short taxi, take-off clearance was granted at 0918 hours. Following departure, Guernsey ATC transferred the aircraft to the London Air Traffic Control Centre (LATCC) at 0930 hours.
Descent and Test Airspace
Upon reaching Berry Head at FL 100, the commander requested a descent to FL 60. The aircraft was cleared and handed over to Exeter ATC at 0943 hours. The controller confirmed the requirement for a block of airspace between FL 60 and FL 100, offering a radar advisory service. The aircraft was vectored onto a northerly heading to avoid departures from Plymouth Airport. The commander then requested further descent to FL 50, which was approved.
Loss of Contact
At 0947 hours, the Exeter controller noticed the transponder altitude readout indicated FL 47. The controller called the aircraft to confirm the local safe level of 3,500 feet but received no reply. From flight data recorder (FDR) timings, the crew would not have heard this call. Simultaneously, both transponder and primary radar returns disappeared. The controller made repeated calls without response and arranged for LATCC to inform the Distress and Diversion cell while notifying emergency services.
Witness Observations
Twenty-two witnesses were interviewed. All described a clear, bright sky with only some light cumulus clouds. No smoke or fire was seen emanating from the aircraft or its engines. Most witnesses observed the aircraft in a spin or spiral descent, predominantly to the left, though some described a falling-leaf motion. Four witnesses with clear views reported two or three turns of a left spin/spiral. Those who could hear the engines described a loud noise consistent with high power. Impact was followed immediately by a post-crash fire.
Wreckage Examination
The ground fire largely destroyed the wreckage, but structural integrity was established. The No 1 propeller was feathered, and the flaps were fully and symmetrically retracted. No mechanical malfunction was found. Both pilots had flown together previously; on this flight, the commander occupied the right seat and made all radio transmissions, setting engine power and recording data, while the first officer flew from the left seat.
Test Flight Configuration and Stall
The commander requested a block from FL 60 to FL 100, then a base of FL 50, consistent with a three-engine climb test. Normal practice would have involved configuring the aircraft during descent, but the flaps remained fully retracted instead of being selected to 25 degrees. In accordance with the three-engine climb configuration, the No 1 engine was shut down and its propeller feathered. With autopilot engaged and three engines at low power, the aircraft levelled at FL 50 and decelerated. The crew should have noticed the trim wheel rotating as the autopilot applied nose-up trim. The non-handling pilot might have misinterpreted this as manual trim input. Clear aural and tactile stall warnings via the stick shaker would have occurred. Both pilots were familiar with the test schedule, but the aircraft was incorrectly configured. The autopilot remained engaged until the stall, and no adequate response to the stick shaker was observed. The crew's failure to react correctly was deemed unusual; distraction could not be discounted.
Stall and Control Inputs
The aircraft stalled with the autopilot engaged. Power was increased on the three operating engines, and two seconds later the autopilot was disengaged. Asymmetric power caused a rapid left roll, countered by right rudder and right spoiler. The elevator was then moved to full nose-up. Except for decreasing right spoiler, these control inputs remained until just before impact, when engine power was reduced. Both pilots appeared involved in the control inputs, but some actions were unusual. The application of asymmetric power at the stall led to autorotation, which was inappropriate. The handling pilot applied opposite rudder normally, but the full aft control column after the stall was inexplicable, regardless of whether the pilot believed a spin or spiral dive had developed. Manufacturer's test data showed no elevator overbalance due to aerodynamic loads; it was considered most probable that the pilot placed the control column fully aft. The nose-up trim applied by the autopilot before disconnection would have created unexpected control forces, but that did not explain the subsequent full aft column. The rapid autorotation may have caused the handling pilot to become disoriented, and the high longitudinal control forces from the trim may have exacerbated the situation.
Probable Cause
Normal crew operation and co-ordination was lacking during this phase of flight.