A measurement-based error-vector-magnitude model to assess non linearity at the system level

K measurement-based Error-Vector-Magnitude (EVM) extraction and modeling is proposed to obtain a least squares estimate of the EVM for a class of modulated excitation signals sharing a common probability density function (pdf) and Power Spectral Density (PSD). The method splits the influence of the linear dynamic and the nonlinear distortion on the EVM. The dependence of the EVM on the input signal power is hereby obtained. The influence of the measurement noise on the measured EVM is compensated, wrestling in a clear improvement of the measured quantity. The results are validated on measurements obtained by a Vector Signal Generator and Vector Signal Analyzer (VSG-VSA) measurement setup.