An external calibration system for DBF receiver arrays at Ka-band

The approach of phased arrays with digital beam forming (DBF) is interesting for satellite communication systems. However, a complex system of large-size arrays with more than 1000 channels is necessary. Thus, one of the most critical tasks is integrating and controlling a huge number of active and passive components in a compact system. The topic of the present dissertation is the development and the analysis of an online calibration system with probes surrounding the DBF receiver array. In comparison to other concepts in literature, this so-called external concept is chosen for large-size arrays because of its low impact on the integration density and its moderate hardware complexity. This thesis discusses its applicability to the calibration of channel deviations and its extension to further include their mutual coupling in a joint procedure. Three aspects of the calibration procedure for channel deviations are improved, namely the probe design, the model of the probe-receiver coupling, and the algorithmic part. - The probes are realized as coaxial monopole antennas with an approximately omni-directional radiation pattern to improve their robustness against fabrication tolerances. - An analytical model of the probe-receiver coupling is developed to enhance the accuracy with acceptable computational complexity. - A combination of the arithmetical and the geometrical mean is proposed to improve the stability of the calibration procedure. In addition, the concept is extended towards a joint calibration of the channel deviations and their mutual coupling. A simplified model of the coupling matrix of a receiver system is shown to be necessary to limit the hardware cost. Additionally, three decoupling algorithms are considered. The common Least-Square (LS) method is the first approach in case of a large number of probes and calibration signals with a high signal-to-noise-ratio (SNR). Furthermore, the necessity of at least one separate reference channel is demonstrated. In the second approach, the LS algorithm is complemented by a Principle Component Analysis (PCA). The accuracy is shown to be improved for a limited number of probes. However, the approach is suitable only for the calibration of small-size receiver arrays because of the required high signal-to-noise ratio (SNR) and the huge computational cost. Finally, an extension of LS with multiple reference receivers (MRR) is suggested to diversify the modeling errors and to achieve the best possible calibration accuracy for a limited SNR. Although the hardware cost is higher, this approach is relevant for the calibration of a large-size array. The usefulness of the proposed calibration procedure is demonstrated in laboratory experiments with linear and planar receiver arrays. The correction of the main beam direction as well as the improvement of the side-lobe level of the receiver array are reported. Finally, a tolerance analysis is performed to determine the technical requirements of the calibration system, such as the number of probes and reference channels or the required SNR of the calibration signals.