A detailed comparison of two continuous GPS carrier-phase time transfer techniques

Global positioning system (GPS) carrier-phase time transfer, as a widely accepted high-precision time transfer method, frequently shows a data-batch boundary discontinuity of up to 1 ns, because of the inconsistency of the phase ambiguities between two consecutive data batches. To eliminate the data-batch boundary discontinuity, several techniques have been proposed in recent years. The question is how much the solutions of these techniques differ from each other and how well the solutions are faithful to clocks. To answer these questions, this paper chooses two techniques to study: revised RINEX-shift (RRS) technique [1, 2], and phase integer common-view (Phase-CV) technique [3]. This paper shows that the time deviation of the difference between the two techniques is below 100 ps, for an averaging time of less than 10 d. Especially, for an averaging time of less than 1 d, the time deviation is less than 30 ps. We also find that both RRS and Phase-CV match TWSTFT (two-way satellite time and frequency transfer) and TWOTFT (two-way optical-fiber time and frequency transfer) quite well. Especially, the difference between RRS/Phase-CV and TWOTFT is less than +/- 0.25 ns for more than 20 d, for a baseline of 268 km. These results show that both RRS and Phase-CV agree very well, and they are both faithful to clocks. However, this is based on the assumption that there is no obvious change in the GPS receiver reference time. When there is a sudden change in the reference time, Phase-CV cannot follow the time change. In contrast, RRS still follows the time change and represents the clock well.