Gentlemen, in view of the report limitations, the future identification of the cause of the failure of the duplex bearing will be very important. Without specific experience of this type, I would venture the following thoughts. After the duplex bearing became effectively locked and the shaft rotated at T/R speed, I would guess that (with an input axial load) the castleated nut would friction weld to the pin-carrier in a short time (probably measured in seconds) then, followed almost immediately by the unscrewing of the nut and T/R runaway. The question must be, how to mitigate this process in the event of a duplex bearing failure.
OAP