With the increasing move to more electric systems in aircraft, ships and automobiles, there is a need to ensure that electromechanical actuators are designed to satisfy the conflicting specifications of low cost, low volume/weight, high performance and requiring little maintenance. The conclusion of the more for less design philosophy is that power electronic motor drives will be work harder, in harsher environments, for longer periods of time. Scheduled maintenance periods will be longer, and therefore it is imperative that drives, especially those used for safety critical applications will employ prognosis and diagnosis algorithms as part of their basic control structure, to predict and prevent in-service failure. The work proposed here will investigate the production of new signatures for indicating the condition of a motor drive and its load, and also determine how these signatures can be used to determine the type and severity of a fault. The aim is to embed the condition monitoring into the normal operation of an electromechanical actuator, in order to detect and distinguish between faults in the electrical machine, the power converter and the mechanical system.