Total Delay and Total Uncertainty in UTC Time Link Calibration - A BIPM Pilot Study

Z. Jiang

Abstract:	Coordinated Universal Time (UTC) is distributed in the form of the difference [UTC-UTC(k)] between UTC and its local realization at the laboratory k. The algorithm of calculation of UTC is based on clock differences, obtained from the time link values of UTC(l)-UTC(k) in a procedure that uses a unique reference (called “pivot”), at present UTC(PTB). The dominant part in the uncertainty of the realizations of UTC comes therefore from the time transfer. The largest contribution to the uncertainty of [UTC-UTC(k)] is that of the calibration, which is 1 ns for TWSTFT and 5-7 ns for GNSS. In [7] and [8] we introduced the time link calibration based on the concept of total delay aiming at unifying the calibrations of all the UTC links. To do this, a calibration scheme denominated METODE (MEasurement of TOtal Delay) was designed at the BIPM. The underlying concept is to make possible the cross-calibrations. Practically, we use a GPS PPP time link to calibrate the other links, such as the TWSTFT, the C or P codes of GPS and GLONASS. This is impossible by differential receiver calibration which is limited between the same type of observations. The total delay of a GNSS receiver system is the total electronic delay between the phase center of the antenna and the UTC point UTCP(k). Correspondingly, its uncertainty is the total uncertainty. The classic differential receiver calibration determines the so called ‘internal delay’, a part of the total delay, requires measuring all sub-delays. Consequently, its uncertainty is the combined uncertainties of all the sub-delays. While METODE measures directly the total delay and hence the ‘sub-uncertainty’ are not accumulated. The largest uncertainty source of the METODE is the short-term instabilities of the GNSS receivers, saying within a calibration tour of typically 3 months. However it is not affected by the long-term stability is not mandatory. Classic receiver differential calibration requires both the short and the long-term stabilities of the reference receiver which is absolutely calibrated. Long-term stability is very difficult to attain even for a stationary receiver working in the laboratory condition. In fact, for the purpose in the UTC time link calibration, the long-term stability is not necessary. In earlier studies, the uncertainty estimation is based on modelling the uncertainty sources, such as the uncertainties in the ephemerids, the multipath, the atmosphere delays etc. To evaluate the total uncertainty, we use directly the measurement data statistics without any hypotheses. In fact, we compare directly the METODE results to the independent but more accurate measurements: the time interval counter (±0.3 ns), a moving Cs standard (0.5~1 ns) and the optical fiber link (±0.2 ns). A BIPM pilot study on the METODE link calibration was launched in 2011 and following laboratories participated in: BIPM and OP in France, PTB in Germany, AOS and PL in Poland and TL in Taiwan. Experiments have been carried out in 2013 at and between these laboratories, including a comparison to the 420 km optical fiber link PL-AOS. Based on the above experiments, we conclude that a total uncertainty of 1~2 ns is attainable suing the METODE calibrator under a delicate scheme. They are now ready for the application in the UTC time link calibrations. After validation of the experimental phase, it is planned to include the time link calibration in the BIPM calibration guidelines.
Published in:	Proceedings of the 45th Annual Precise Time and Time Interval Systems and Applications Meeting December 2 - 5, 2013 Hyatt Regency Bellevue Bellevue, Washington
Pages:	112 - 125
Cite this article:	Jiang, Z., "Total Delay and Total Uncertainty in UTC Time Link Calibration - A BIPM Pilot Study," Proceedings of the 45th Annual Precise Time and Time Interval Systems and Applications Meeting, Bellevue, Washington, December 2013, pp. 112-125.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In