Actuator fault diagnosis is often studied under strong assumptions on available sensors. Typically, it is assumed that the sensors are either fault free or sufficiently redundant. The purpose of this paper is to present a new method for actuator fault diagnosis which is robust to sensor distortion. It does not require sensor redundancy to compensate sensor distortion. The essential assumption is that sensor distortions are strictly monotonous. Despite the nonlinear and unknown nature of distortions, such sensors still provide useful information for fault diagnosis. The robustness of the presented diagnosis method is analyzed, as well as its ability to detect actuator faults. A numerical example is provided to illustrate its efficiency.