The use of GNSS positioning in highly automated systems, like packet delivering using drones or self-driving cars, demands a reliable and accurate position estimation. At the same time the risk of radio-frequency interference (RFI), intentional or unintentional, increases. A promising countermeasure is to suppress the interference in the spatial domain using an antenna array. To apply this technique, a weighting vector is calculated and regularly updated during operation. This weighting vector is applied to the input signals resulting in a single output signal, which can drive a common GNSS receiver. However, these weighting vectors vary in time to adapt to the direction of arrival (DoA) of the satellites and the RFI. Therefore, this technique may induce a phase error to the carrier phase measurements. The carrier phase measurements are obligatory for high precision approaches like Precise Point Positioning (PPP) or Real Time Kinematics (RTK). In this work two new approaches reducing the phase error induced by spatial signal processing are introduced. The first approach aims to steer the phase of an updated weighting vector based on the previous weighting vector and the second approach aims to steer the phase of an updated weighting vector based on a constant weighting vector for all satellites.