Validating and comparing GNSS antenna calibrations

Abstract

GNSS antennas have no fixed electrical reference point. The variation of the phase centre is modelled and tabulated in antenna calibration tables, which include the offset vector (PCO) and phase centre variation (PCV) for each frequency according to the elevations and azimuths of the incoming signal. Used together, PCV and PCO reduce the phase observations to the antenna reference point. The remaining biases, called the residual offsets, can be revealed by circulating and rotating the antennas on pillars. The residual offsets are estimated as additional parameters when combining the daily GNSS network solutions with full covariance matrix. We present a procedure for validating the antenna calibration tables. The dedicated test field, called Revolver, was constructed at Metsähovi. We used the procedure to validate the calibration tables of 17 antennas. Tables from the IGS and three different calibration institutions were used. The tests show that we were able to separate the residual offsets at the millimetre level. We also investigated the influence of the calibration tables from the different institutions on site coordinates by performing kinematic double-difference baseline processing of the data from one site with different antenna tables. We found small but significant differences between the tables.