ERA14LA329 · Flight Design Gmbh CT-SW 2006

The private pilot was taking off in the light sport airplane for a local flight. Witnesses reported seeing the pilot begin the takeoff. During the rotation, the airplane appeared to pitch up higher than normal, followed by up-and-down pitch oscillations and left bank oscillations; it climbed no higher than about 75 ft. The airplane began a slow left bank, which was not consistent with a normal takeoff procedure, before impacting trees south of the runway in a left-wing-low attitude. The pilot indicated to his son before he died that the airplane experienced flight control issues related to the autopilot.

Postaccident examination of the airplane revealed that the flaps were symmetrically extended 15 degrees, and there was no evidence of preimpact failure or malfunction of the flight controls for roll, pitch, or yaw. According to an airplane performance study, the pilot was operating the airplane about 4 knots above its stall speed during the left turn; however, the location of the main wreckage with respect to the airplane's last GPS data point indicated that the bank angle likely increased and exceeded the airplane's critical angle-of-attack, which resulted in an aerodynamic stall.

Data from the engine's recording device indicated that, during the beginning portion of the takeoff sequence while the airplane was over the runway, the engine rpm decreased about 50 rpm and then increased nearly 900 rpm about the point and time when the airplane banked left and hit the trees. Postaccident operational testing of the engine revealed that it produced full-rated power with no evidence of preimpact failure or malfunction. The reason for the reduced power setting at takeoff could not be determined.

Although the pilot reported that the pitch-and-roll oscillations during takeoff were related to the autopilot, it could not be determined during examination of the autopilot whether the autopilot was engaged during the accident flight. Postaccident testing of the autopilot controller and roll servo revealed no evidence of preimpact failure or malfunction; the override torque value of the roll servo was within limits.

The autopilot controller minimum airspeed was found set to a value of 0, which would have disabled the minimum airspeed alert if the autopilot were engaged. The pitch servo was found inoperative due to a failed voltage regulator; however, this condition would not have caused any increased torque or servo runaway. Following replacement of the failed component, the pitch servo tested satisfactorily, and the override value was within limits. Therefore, if the autopilot had been engaged and an autopilot malfunction had occurred, the pilot would have been able to override any pitch and yaw servo commands. Further, if the autopilot had been engaged and the controller minimum airspeed had been set to an appropriate value, it is likely that a stall alert would have occurred that provided the pilot with adequate time to respond to and avert an aerodynamic stall.

Verbatim NTSB analysis and probable cause from the NTSB dataset