The calibration of reciprocal radar systems has been studied previously, where it was shown: (1) that full polarimetric calibration of radar systems can remove cross-polarization errors from the measurements, and (2) that for reciprocal radars Fourier analysis of polarimetric data obtained using a rotating dihedral can reduce the noise and background errors in the calibration. To calibrate nonreciprocol radar systems, we must obtain full polarimetric data from two objects whose scattering matrices have independent eigenvectors. Thus, in addition to a rotating dihedral, a sphere or a flat plate is needed to solve for the transmitting and receiving characteristics of the system. We find the current methods of solution of the calibration equations unattractive for the following reasons: (1) noise and clutter rejection is not built into the analysis, and (2) the mathematics seems unnecessarily complicated. In addition to mathematical simplification, several ways to verify data integrity are presented, so that system problems can be detected early in the calibration phase.