Currently, imaging methods are used to diagnose loosening of endoprosthetic implants, but fail to achieve 100% accuracy. In this study, a passive sensor array which is based on the interaction between magnetic oscillators inside the implant and an excitation coil outside the patient was investigated. The excited oscillators produce sound in the audible range, which varies according to the extent of loosening. By performing several experimental tests, the sensor array was optimized to guarantee reproducible and selective excitation of the sound emission. Variation in the distance between the oscillators demonstrated a definite influence on the quality of the generated sound signal. Furthermore, a numerical design analysis using the boundary element method was generated for consideration of the magnetic field and the selectivity of the oscillators during excitation. The numerical simulation of the coil showed the higher selectivity of a coil with a C-shape compared to a cylindrical coil. Based on these investigations, the passive sensor system reveals the potential for detection of implant loosening. Future aims include the further miniaturization of the oscillators and measurements to determine the sensitivity of the proposed sensor system. Full article