This paper presents a method for stand-still identification of parameters in a permanent magnet synchronous motor (PMSM) fed from an inverter equipped with an three-phase LCtype output filter. Using a special random modulation strategy, the method uses the inverter for broad-band excitation of the PMSM fed through an LC-filter. Based on the measured current response, model parameters for both the filter (L, R, C) and the PMSM (L and R) are estimated: First, the frequency response of the system is estimated using Welch Modified Periodogram method and then an optimization algorithm is used to find the parameters in an analytical reference model that minimize the model error. To demonstrate the practical feasibility of the method, a fully functional drive including an embedded real-time controller has been built. In addition to modulation, data acquisition and control the whole parameter identification method is also implemented on the real-time controller. Based on laboratory experiments on a 22 kW drive, it it concluded that the embedded identification method can estimate the five parameters in less than ten seconds.