Improvements in the Measurement of Very Low Cross Polarization Using the Three Antenna Polarization Technique

The Three-antenna polarization measurement technique is used to determine the axial ratio, tilt angle and sense of polarization of three antennas from measurements on each of three antenna pairs. The three antennas are generally nominally linearly polarized and the measurement data consists of the change in amplitude from the initial antenna orientation where they are co-polarized to the orientation where one of the antennas is rotated about its axis to the null amplitude position. The sign of the phase change is also noted and the phase change at the null position is known from theoretical calculations to be either plus or minus 90 degrees. The correct sign is determined from the sign of the phase change. For antennas with axial ratios in the range of 50 to 80 dB that will be used as near-field probes or as feeds for reflector antennas, it is imperative to measure the polarization parameters as accurately as possible. The primary source of uncertainty in the measurement is due to scattered signals in the measurement range that arise from multiple reflections between the two antennas and from the absorber on the chamber walls. For antennas with very large axial ratios, the scattered signals can be larger than the true measurement signal. These scattered signals can change the sign of the phase and produce large errors in the amplitude at the null. If the separation distance between the antennas is adjusted after rotating to the null to produce a maximum amplitude, the scattered signal is in phase with the true measurement signal. If the distance is adjusted for the minimum at the null, the scattered and true signals are out of phase. Measurements at these two positions will produce the best measurement of the phase sign and the true amplitude. But if measurements are being performed at a number of frequencies, the maximum and minimum amplitude positions will be different for each frequency, and this will complicate automated multifrequency measurements. New improvements have been developed in the details of the measurements that greatly improve the determination of the phase sign and the amplitude at the null for multiple frequency measurements and these will be described and illustrated in the following paper. With these improvements, the estimated uncertainty of a 60 dB axial ratio is on the order of 1.8 dB. A new technique has also been developed to improve the source correction of the pattern data for probes with large axial ratios that guarantees that the on-axis polarization of the pattern data will be identical to the results of the Three-antenna measurement. The probe correction processing will then produce the highest accuracy results for the polarization of the AUT.