Shorts 360 EI-BPD departed Dublin at 18:14 for a flight to Sheffield. The aircraft was cleared for the ILS/DME procedure for runway 28 and the crew requested the QFE which was 980 hPa. The decision height for the approach was 400 feet. At 1918:11 the crew reported that they were established on the localizer. When the aircraft intercepted the glidepath, the flaps were set to 15ø correctly configuring the aircraft for the approach. The handling pilot recalled that initially the rate of descent was slightly higher than the expected 650 ft/min leading him to suspect the presence of a tailwind, however, the rate of descent returned to a more normal value when approximately 4 nm from the runway. The propellers were set to the maximum rpm at 1,200 feet agl. When the crew reported that they were inside 4 nm they were cleared to land and passed the surface wind, which was variable at 2 kt; they were also warned that the runway surface was wet. Both pilots saw the runway lights when approaching 400 feet agl; the flaps were selected to 30ø and confirmed at that position. Both pilots believed that the airspeed was satisfactory but, as the commander checked back on the control column for the landing, the rate of descent increased noticeably and the aircraft landed firmly. Both pilots believed that the power levers were in the flight idle position and neither was aware of any unusual control inputs during the landing flare. The aircraft bounced before hitting the ground again, this time with the nose wheel first, before bouncing once more. The aircraft was then seen to travel about half way along the runway before slewing to the left and running onto the grass. When the aircraft stopped the left wing tip appeared to be touching the grass. CONCLUSION: “Evidence from the CVR indicated that the flight was conducted in a thoroughly professional manner in accordance the operator’s normal procedures until the final stages of the approach. The recorded data indicate that three seconds prior to touchdown the propeller blade angle changed from the flight range to the ground range. Coincident with this change the CVR recorded sounds consistent with the propellers ‘disking’ and the FDR indicated that the aircraft then decelerated longitudinally and accelerated downwards. The engineering investigation revealed that the propeller control rigging and the operation of the flight idle baulk were correct. Selection of ground fine requires the pilot to firstly release the flight idle baulk and then lift and pull the propeller levers further back, this combined action rapidly becomes a programmed motor skill in the routine of daily operations. It is therefore possible that the handling pilot unintentionally selected the propellers into the ground fine position whilst still in the air.”