Phased array observations with infield phasing

We present results from pulsar observations using the Giant Metrewave Radio Telescope (GMRT) as a phased array with infield phasing. The antennas were kept in phase throughout the observation by applying antenna based phase corrections derived from visibilities that were obtained in parallel with the phased array beam data, and which were flagged and calibrated in real time using a model for the continuum emission in the target field. We find that, as expected, the signal to noise ratio (SNR) does not degrade with time. In contrast observations in which the phasing is done only at the start of the observation show a clear degradation of the SNR with time. We find that this degradation is well fit by a function of the form SNR(τ)=α+βe−(τ/τ0)5/3$(\tau ) = \alpha + \beta e^{-(\tau /\tau _{0})^{5/3}}$, which corresponds to the case where the phase drifts are caused by Kolmogorov type turbulence in the ionosphere. We also present general formulae (i.e. including the effects of correlated sky noise, imperfect phasing and self noise) for the SNR and synthesized beam size for phased arrays (as well as corresponding formulae for incoherent arrays). These would be useful in planning observations with large array telescopes.