Two-way Satellite Frequency Transfer
with Highest Ever Precision, Developed by NICT
- Demonstration with three orders of magnitude better precision over 9,000-km baseline -


In order to determine Coordinated Universal Time (UTC) and Japan Standard Time, research institutes around the world main- tain and manage atomic clocks that generate time and frequen- cies, and these must be compared with each other with high pre- cision. Generally, methods for international transfer of time and frequency use artificial satellites, but unfortunately the precision of methods currently in use (Figure 1, solid blue and black lines) is several orders of magnitude lower than that of the atomic clocks that are the basis for standard time (Figure 1, dotted gray line).

As an example, measurements must be continued and averaged over a full day to reach the precision of a Cesium Atomic Fountain Pri- mary Frequency Standard, which is used to define the second. To obtain the equivalent precision to an optical clock (Figure 1, green line), which is the most promising future candidate for redefining the second, measurements over more than 100 days would be re- quired.

Figure 1 Instability of frequency comparison between NICT and PTB

Figure 1 Instability of frequency comparison between NICT and PTB
Red: TWSTFT carrier-phase, Blue: TWSTFT code, Black: GPS carrier phase. Gray dashed line shows a typical frequency instability of Hydrogen maser. Each frequency comparison is performed using a reference signal supplied from a local standard time generated from Hydrogen maser. The long-term instability is limited by the local time.

To improve on this situation, NICT has developed an opti- cal carrier transmission system using optical fibers, incorporating an efficient fiber-noise cancelling technology, and has conducted the first-ever direct comparisons of optical-lattice clocks between NICT and the University of Tokyo. These confirmed that the fre- quencies of optical lattice clocks at the two facilities coincided to a high precision of 16 decimal places. However, using optical fiber is expensive and international comparisons would require submarine cable, so the method is still not practical for intercontinental trans- fer.
Because of this, NICT is working to increase the precision of satellite comparison and transmission technologies, and compari- son technologies using VLBI techniques that receive noise signals from radio stars outside the galaxy.

Satellite antenna group for comparing time and frequency

Satellite antenna group for comparing time and frequency
TWSTFT: Two-Way Satellite Time and Frequency Transfer QZSS: Quasi-Zenith Satellite System

Two-way frequency comparison by geostationary satellite
using carrier phase

Methods for comparing time and frequency by satellite are shown in Figure 2. Comparison using GPS satellites is done by receiving signals sent from navigation satellites. The precision of these methods, using codes or carrier phase, are approximately 5 ns and 50 ps, respectively (Figure 1, black line). Conversely, with two-way comparison using geostationary satellites, in which two earth stations send and receive signals simultaneously for compar- ison, the signal path is identical in both directions, so atmospheric delay, orbital effects, and other factors almost completely cancel out and higher precision can be achieved. Conventionally, approxi- mately 500 ps precision (Figure 1, blue line) can be obtained using a 2.5 Mbps code phase. Even higher precision could be obtained using a higher-speed modulated signal, but use of a geostationary satellite is expensive, so in the past 10 years, there have not been significant improvements to this precision.

To resolve this, we began research and development to intro- duce the use of carrier phase, as is done with GPS, to two-way sat- ellite comparison. We developed a method that cancels the phase noise that the signal receives when the frequency is converted at the satellite and frequency shifts caused by Doppler effects due to small oscillations in the geostationary satellite orbit. Then, we did experiments using the ETS-VIII satellite, followed by more using a real, commercial geostationary satellite, and verified im- provements in precision using this method, from approximately 500 ps conventionally, to 0.2 ps (Figure 1, red line). Thus, NICT has achieved for the first time, the highest precision in the world for frequency transfer using satellites. A 0.2 Mbps modulated sig- nal was used in the measurements to assist signal capture, and measurements were completed using a satellite transponder having 1/10 the bandwidth of those used with conventional methods.

Figure 2 Time and frequency transfers using satellite

Figure 2 Time and frequency transfers
using satellite

Intercontinental frequency transfer with Germany

The Physikalisch-Technische Bundesanstalt (PTB) in Germany plays a central role in the time-comparison network that defines UTC, and is a top-class laboratory for frequency standards development. We first checked the measurement precision of two-way comparison using carrier phase with experiments over the short distance between Kashima and Okinawa. Then, we con- ducted a long-distance, international comparison experiment over 9,000 km with PTB, confirming the same precision as in the short-distance experi- ments (Figure 1). However, because of the long distance and local time difference, there were ionospheric delay effects, appearing as a diurnal variation of approximately 150 ps.

This error ap- peared due to the frequency difference between uplink and downlink frequencies (14 GHz uplink, 11 GHz downlink), and was not observable with earlier two-way comparison measurements using code phase due to the lower precision. We con- firmed that the reduction of this error was possible by correcting for the effects of ionospheric delay using total electron content (TEC) maps created by NICT and the Royal Observatory of Belgium for the skies over Japan and Europe (Figure 3).

Figure 3 Ionosphere effect in NICT-PTB TWSTFT link

Figure 3 Ionosphere effect in NICT-PTB TWSTFT link
15-min mean of difference between TWSTFT and GPS results. GPS result is ionosphere free. Green: Ionospheric delay, Blue: without ionosphere correction, Red: with ionosphere correction. The difference shows the disagreement of two results. When the ionospheric delay becomes larger (area covered by ellipse), we can see that the correction works.



Future prospects

By improving measurement precision by three orders of magnitude over conventional methods, new factors contributing to error have come to light. The effects of the ionosphere are one such effect, but there are others, such as degradation of mid and long-term measurement precision due to phase fluctuations caused by temperature fluc- tuations in field instruments. These were not an is- sue at conventional levels of precision. NICT will continue to collaborate with international research institutes to develop systems that can control such error factors and achieve still higher precision, and to conduct tests to verify these systems.


Senior Researcher Fujieda which looks down on a parabolic antenna.

Senior Researcher Fujieda which looks down on a parabolic antenna.