The performance of GPS time and frequency transfer: comment on 'A detailed comparison of two continuous GPS carrier-phase time transfer techniques'

The paper 'A detailed comparison of two continuous GPS carrier-phase time transfer techniques' (Yao et al 2015 Metrologia 52 666) presents the revised RINEX-shift (RRS) method, a technique using 'classical precise point positioning (PPP)' solutions on sliding batches and aiming at providing continuous time links. The authors claim the superiority of the RRS technique with respect to 'classical PPP' in terms of frequency stability and solving for discontinuities due to data gaps. It is shown here that these conclusions do not rely on physical principles, and are erroneous as they are driven by misinterpreted or corrupted PPP solutions. Using state-of-the-art PPP computation on the same data sets used in Yao et al's paper (2015 Metrologia 52 666), we show that the stability of RRS is at best similar to that of ` classical PPP' (within statistical uncertainties). Furthermore, the RRS method of removing discontinuities in case of data gaps by interpolating the phase data should not be applied systematically as it can cause erroneous clock solutions when the data gaps are associated with a true phase discontinuity.