The Institut für Erdmessung (IfE) at the Leibniz University Hannover calibrates GNSS antennas determined by the well known robot based concept in the field. Our group has recently improved the calibration procedure to support absolute receiver antenna calibrations for all GNSS frequencies (GPS L1/L2/L5, GLONASS L1/L2 and Galileo E1/E5). To fully use the advantage of multi GNSS processing capabilities for absolute and relative precise positioning, accurate and consistent receiver antenna calibration patterns of ground stations are required. Currently, in IGS and EPN they are available by chamber calibration method for some antennas. However, systematic and sometimes significant differences exist between both approaches (field robot and chamber) that have to be studied in detail to identify the causes and to fix these issues.In this contribution, we present first robot based multi GNSS patterns from IfE. Furthermore, a study is performed, to show the impact from the observation domain (differences of antenna phase centre patterns) to the parameter domain (position, troposphere, ambiguities). In this study, we used reference stations from the EPN network which provide individual antenna patterns obtained from both approaches (field robot and chamber). Baselines of lengths ranging from 150-600km and one of 1670km are analysed. We found that on the one hand, differences between both pattern sets do not met the proposed "1mm-rule-of-thumb" in the most of studied cases. On the other hand, we show that in cases of consistent antenna pattern sets no significant differences in the parameter domain are obtained. However, mixing patterns from different approaches implies deviations of up to 1cm (in one case up to 2cm) mostly in the topocentric up component. These differences are directly aligned to the differences in the antenna pattern.