ERA21LA289 · Piper PA-23

During an instructional flight, while performing a practice approach-to-landing stall with the flaps and landing gear extended, the pilot receiving instruction applied full power to both engines at the airframe buffet. While recovering, the left engine lost total power and the left propeller feathered uncommanded. At that time, the airspeed was about 57 mph and the airplane immediately rolled about 45° to the left, consistent with a Vmc roll. The flight instructor immediately took control of the airplane to recover. As he applied full right rudder, his seat slipped back, which released pressure on the right rudder pedal. The airplane entered a second Vmc roll, from which he was able to recover, but the airplane was at an altitude of about 1,100 ft agl and in a descent of about 800 ft per minute (fpm). The pilot receiving instruction was unable to restart the left engine, and the flaps and landing gear remained extended. When the airplane was flying below 700 ft agl and descending rapidly, both pilots realized that the airplane would not be able to reach the airport. The flight instructor subsequently landed the airplane with its landing gear and flaps still extended in a cornfield, during which the left wing and fuselage were substantially damaged. The airplane came to rest upright with both propellers feathered.

During a postaccident test run, the left engine produced full-rated power and manifold pressure with no hesitation noted during rapid full-throttle applications. No discrepancies were noted with either engine when a propeller oil control leak test procedure was performed on each. While the installed left propeller blades had a basic diameter that was 2 inches longer than the approved propeller, the propeller was not counterweighted as specified by the applicable supplemental type certificate, and the air charge in the propeller cylinder was about 2 psi less than the lower specified limit in the Propeller Owner’s Manual and Logbook, it is unlikely that those conditions would have resulted in to the uncommanded feathering of the left propeller. The propeller governor oil pump flow rate was measured to be slightly less than the minimum specified, the reason for the lower flow rate could not be determined because the test bench was not capable of measuring the internal leakage. The investigation also could not determine if the lower flow rate was sufficient to maintain the accident propeller at a low pitch while in a low power and airspeed flight condition. Based on this information, the reason for the total loss of engine power of the left engine and the reported uncommanded feathering of the left propeller could not be determined.

During the postaccident examination of the right front seat, a deformation in the tube-seat catch release was noted. That condition, which likely existed at takeoff, prevented one of the two seat catch plungers from extending into the seat track and likely allowed the seat to move back when the flight instructor applied right rudder. Given the described sequence of events provided by the flight instructor, the seat slipping directly contributed to the second Vmc roll and loss of altitude before he was able to recover. With an inoperative left engine, extension or retraction of the flaps and landing gear would have to be accomplished manually using a hand pump in the cockpit. If the seat slippage (and the subsequent loss of control and altitude) had not occurred, it is possible that either pilot would have had the time to manually retract the landing gear and flaps, and that the flight instructor could have flown with a single engine to a diversion airport.

Verbatim NTSB analysis and probable cause from the NTSB dataset