R. T. Kenneth Jaldehag

Geodesy using the Swedish Permanent GPS Network∷ Effects of snow accumulation on estimates of site positions

We have observed variations at the several centimeter level in estimates of the vertical coordinate of site position. The estimates are obtained from our analysis of data acquired from the Swedish permanent Global Positioning System (GPS) network. The observed variations are strongly correlated with changes in the indirectly inferred accumulation of snow, which we assume collects on the radomes and pillars; the GPS sites could not be observed directly due to their remoteness. Numerical simulations which assume a simple geometry for the snow cover are used to study the effects of snow accumulation on GPS phase observables and hence on estimates of the vertical coordinate of site position obtained from these observables. Our results indicate that the variations in the vertical coordinate of site position can be fully explained by reasonable accumulations of snow which retard the GPS signals and enhance signal scattering effects.

Time and frequency transfer in an asynchronous TCP/IP over SDH-network utilizing passive listening

A technique for time and frequency transfer over an asynchronous TCP/IP network is being developed by SP, Swedish National Testing and Research Institute together with STUPI. When implemented, users will be able to compare their clocks by connecting to the system. The technique is based on passive listening to existing data traffic in the network. Since the network is asynchronous, intermediate clocks are located and compared at each router. We use the frame alignment bytes of the SONET/SDH protocol as references in order to compare these clocks. As a test bed for the experiment, we will use the Swedish University Computer Network (SUNET). A preliminary assessment of the technique in a lab environment will be performed late 2005

A fiber based frequency distribution system with enchanced output phase stability

Experimental results on the stability of the output phase of a frequency distribution system from several days of measurement is presented, in addition to a discussion regarding the influence of control loop parameters. The setup handles the issue that the output phase stability of a system depends on perturbations along the transmission length. This is especially critical if the signal is transmitted through optical fiber, at lengths of a few 100 m. An experimental evaluation using a laser based transmitter at a wavelength of 850 nm, and 625 m of multimode fiber where 575 m where placed outdoor, a temperature dependence of 100 ps/°C was detected. To compensate for these slow variations in real time, a setup using two-way transmission, in conjunction with an adjustable optical delay, was constructed. This device is adjusted to induce a delay variation of equal magnitude but opposite direction, in comparison to the delay change of the fiber. Calculating the modified Allan deviation of the transmitted signal, it is apparent that without active compensation, the deviation at τ below 1000 s is comparable to the values from the measurement system without transmission. At longer integration times, however, the slow variations in the fiber transmission will deteriorate the modified ADEV substantially. When activating the dynamic adjustment of pre-delay in the system, the deviation at shorter times will increase with a few dB, however, the modified ADEV decreases continuously with τ, eventually below the values for the uncompensated system. In conclusion, activating a dynamically controlled pre-delay in a fiber based frequency transmission system will induce a small penalty on fast variations of the output phase, however giving a remarkable improvement on slower variations. The usefulness of this added functionality must therefore be determined by the application of the signal.

VLBI frequency transfer using CONT11

Geodetic VLBI is an independent technique, which does not rely on third parties. This makes it a viable future alternative for time- and frequency transfer over long baselines. Frequency link instabilities in the order of 1.5e-15 for time periods of one day are comparable to those achievable with methods using GNSS carrier-phase observations. Data of the continuous VLBI campaign CONT11 were analyzed and compared to results from GPS PPP analysis on collocated/common clock stations.

Active detection of propagation delay variations in single way time transfer utilizing dual wavelengths in an optical fiber network

Several communication systems of today rely on the real time accessibility of accurate time and frequency measures and there is an increasing demand for the development of new and redundant methods for the distribution of these measures. The classical two-way method is able to compensate for the inevitable variations in the time and frequency propagation delay. The two-way method is used for time transfer in free space, electrical or optical domain, but has the disadvantage of often using two different paths for transmitting back and forward. The paths may be of equal length and have equal propagation delay, but nevertheless there is often a remaining asymmetry in the propagation paths. The inevitable asymmetry between the paths in the time transfer delay must be detected and compensated for, if an accuracy better than µs-level is needed for transmission distance exceeding a few km. Furthermore, if the number of users is high, there will be a complex and large network of two-way time signal transmissions. Therefore, a solution using one-way broadcasting would be more desirable, and would be possible if the variations in transmission time could be estimated from the received data at the far (user) end. The one-way method uses only one path of transmission and is possible to implement in existing Wavelength Division Multiplexing-networks. Proof of concept and results of this one-way time transfer technique based on transmission of a repetitive signal, modulated on two lasers at different wavelengths 8 nm apart and transmitted through an optical fiber, has been presented previously. These data showed a strong correlation between a change in transfer time at one wavelength, and the transfer time difference for the signals at the two wavelengths. In this paper, the setup and the measurement results have been improved and new data is collected which shows improvement in the reliability and quality of this technique. The stability is improved through component analysis and minimizing error sources. The distance is improved from 38km to 160km.

Time transfer using an asynchronous computer network: Results from three weeks of measurements

We have performed a time transfer experiment between two atomic clocks, over a distance of approximately 75 km using an 10 Gbit/s asynchronous fiber-optic computer network. The time transfer was accomplished through passive listening on existing data traffic and a pilot sequence in the SDH bit stream. In order to assess the fiber-link clock comparison, we simultaneously compared the clocks using a GPS carrier phase link. The standard deviation of the difference between the two time transfer links over the three-week time period was 243 ps.

VLBI time-transfer using CONT08 data

One important prerequisite for geodetic Very Long Baseline Interferometry (VLBI) is the use of frequency standards with excellent short term stability, i.e. hydrogen masers. This makes VLBI stations, which are often co-located with Global Navigation Satellite System (GNSS) receiving stations, interesting for studies of time- and frequency-transfer techniques. In this paper we present an assessment of VLBI time-transfer based on the data of the two week long consecutive IVS Cont08 VLBI-campaign by using GPS Carrier Phase (GPSCP). Cont08 was a 15 days long campaign in August 2008 that involved eleven VLBI stations on five continents. For Cont08 we estimated the worst case VLBI time link stability between the station clocks of ONSALA and WETTZELL to about 1.5e-15 at one day. Comparisons with clock differences estimated with GPSCP confirm the VLBI results. The paper also indicates time-transfer related problems of the VLBI technique as used today.

Two-Color One-Way Frequency Transfer in a Metropolitan Optical Fiber Data Network

Abstract: Two-color one-way frequency transfer through one strand of optical fiber is an alternative method to two-way frequency transfer, and is useful if unknown asymmetries exist in the link due to different paths for each direction. The term “two-color” refers to the ability to send signals at two different wavelengths utilizing the same fiber in one direction. The method is suitable for implementation in existing urban Single Mode Fiber networks, for instance in networks that are utilized for data and television communication. It is therefore able to coexist with data channels in wavelength-division multiplexing (WDM) systems. It performs as a dynamical control of transit time and simultaneously enables a real-time phase stabilized output signal. This paper presents results from a comparison of two cesium beam frequency standards separated by about 3 km over an optical fiber network located in a metropolitan area in Sweden. The cesium standards were simultaneously compared to each other with a Global Positioning System (GPS) satellite link and over optical fibers, so that the optical fiber technique could be evaluated with respect to the GPS technique. The difference in frequency stability between the two methods is shown to be about 3 × 10−15 over an averaging interval of 10 000 s.

Time transfer using frame detection in fiber-optical communication networks: New hardware

In this paper, a new, recently developed hardware for time transfer using passive listening and detection of SDH frame headers in fiber-optical networks is described. The method has been presented earlier, and results, using prototype equipment and an experimental fiber-link, have shown that time transfer with a precision of a few nanoseconds is possible over links with network distances exceeding 500 km. In order to further develop the method and make it available to regular users of time keeping equipment, it has been essential to minimize the space requirements of needed hardware and to make the implementation and installation more easily and straight forward.

A GPS carrier-phase aided clock transport for the calibration of a regional distributed time scale

Clock transportation is a historically proven time transfer method for the calibration of time links and time scales. With the establishment of satellite-based time transfer methods, however, clock transportation has become less attractive especially on long baselines. In order to match for instance the GPS common view time transfer method with calibration uncertainties of a few nanoseconds, it is necessary to transport high quality, expensive clocks such as caesium beam frequency standards. The stability of the clock during transportation and the duration of the transport set the limit of the prediction uncertainty. Being able to measure the clock during transportation instead of predicting it would yield some major advantages: (a) the use of less expensive and small clocks such as rubidium or quartz oscillators for transportation, (b) no need for environmental conditioning of the transported clock, and (c) the duration of the transport is not critical as long as the clock can continuously be measured. One solution to the clock measurement problem during transport is the use of GPS carrier-phase observations as described and evaluated in this paper. It is shown that a calibration uncertainty of less than one nanosecond is potentially achievable.