ROB :: TIME

Introduction

Among the studies developed at the Royal Observatory of Belgium, the exact time realization has a secular tradition. When the Observatory was born, the objective was indeed the realization of star catalogues from meridian observations; such a program imposed the installation of the best clocks available at that time and determinations of time as precise as possible. At the beginning of the twentieth century, new pendulums were installed in the basement specially designed for that; these rooms were maintained at stable temperature (with a stability of some tenths of degrees) and the clocks were attached to stone pillars knocked in the ground, and independent of the building. The clocks, named “fundamentals” were 4 Rieffler pendulums, of which two maintained sidereal time and the two others maintained the mean solar time. The time was determined from meridian observations of stars, and the time was disseminated using telephone, in public locations as the Anvers port for example. Later, and until 1980, the time of the Observatory was used for synchronization of speaking clock as well as time signals emitted by the radio.

Near 1950, the first quartz clocks appeared, and other important technological developments followed in the frame of clocks as well as in the frame of remote clock comparisons. The ROB time laboratory evolved accordingly, installing the first quartz clocks in 1955, and the first atomic clock in 1968, together with the modern techniques for international clock comparisons (Loran-C since 1969, and GPS since 1984).

The atomic time

Until the year 1960, the time standard (the second) was defined after the Earth rotation. The second was defined as a part (1/86400) of the solar day (time interval between two passages of the sun in the meridian of an observation site); since 1967, the second is defined upon the quantum properties of the atom cesium, improving largely the precision of the time unit.

It was already known by the astronomical community that the "Earth rotation clock" was not very stable as the observation of the stars showed irregularities. As the length of the day is variable, the second defined as a fraction of the day had also variable duration, which is inadequate for a measurement unit. However, no clock was able to generate a time scale more stable than the Earth rotation; the most precise determination of the time was obtained from the astronomical observations. The situation changed when the atomic clocks entered into the game. The scientific community has then been provided with a time standard whose stability is incomparably better than the Earth rotation. The second was then redefined by the 13^th Conference Générale des Poids et Mesures, as 9 162 631 770 periods of radiation of the transition between the two hyperfine energy levels of the fundamental state of atom Cs 133. The principle of the atomic clock is the following. The atoms can be in different energy levels, corresponding to quantum states. To make a transition from one quantum state to another, an atom needs to receive a radiation whose the frequency is exactly the frequency of its transition. In the atomic clock, the frequency of the excitation is adjusted in order to get a maximum of atoms in the excited state, and the so-obtained frequency is used (by count of 9 192 631 770 cycles for the cesium clock) to generate the seconds (one pulse per second). The BIPM (Bureau International des Poids et Mesures, Paris) has the responsibility to compute the TAI (International Atomic Time). The TAI corresponds to a weighted mean of about 250 atomic clocks situated in about 40 laboratories in the world. One of these laboratories is at the Royal Observatory of Belgium.

As the Earth rotation is much less stable than the atomic clock, the solar time will wander away from the TAI. For practical purpose, it is necessary to keep a precise time definition that stay very close to the solar time. For this reason, the UTC (Universal Time Coordinated) was introduced in 1971, it is obtained by adding the so-called "leap second" if necessary, i.e. if the disagreement between the Earth rotation and UTC becomes too important, in order to keep the difference between UTC and the Earth rotation time scale lower than 0.9 second. The local time is obtained by adding to UTC the time zone correction of the location, and an eventual summer time correction.

UTC realization

Each time laboratory maintains a physical realization of UTC, i.e. a clock (or a set of clocks) synchronized to UTC and of which the frequency is adjusted on the UTC frequency. This realization is called UTC(k), where ‘k’ is the acronym of the laboratory. For the Royal Observatory of Belgium, it is UTC(ROB).

The clocks participating to TAI must be compared each other. In a same laboratory, the clocks are regularly compared to UTC(k) using a time interval counter which measures the time delays between the pulses (one pulse per second) of the clocks and the pulses of UTC(k). For clocks located in different laboratories, two techniques are used: the measurement of a two-way signal going from one laboratory to the other through a geostationary satellite, and the common view of GPS satellites. Only this latter is used at the ROB; it consists in:

In a first time, all the laboratories observe all the visible GPS satellites at given epochs, provided by the BIPM, and for each epoch the GPS receiver determines the clock synchronization error between UTC(k) and the GPS time. Each laboratory sends once per month to the BIPM the data for the previous month concerning the difference between UTC(k) and GPS time, and the differences between the other atomic clocks and UTC(k).
In a second step the BIPM combines the stations two by two, and as the GPS time is a common reference, deduces for each observation epoch the synchronization error between the UTC(k)’s. The clocks synchronization errors between the laboratories are then easily obtained.

Finally, the BIPM computes, from all local clock data compared by either two-way or common view, a « Free Atomic Scale » (or in french, Echelle Atomique Libre, EAL), based on an ensemble algorithm, in which the weight attributed to each clock is determined from its long term stability. This weighted average is then combined with an ensemble of primary frequency standards(*) developed in some laboratories, in order to obtain the TAI second as close as possible from the official definition of the second

The TAI is therefore determined a posteriori: the clocks needs first to work in order to provide data that will be sent to the BIPM which will make the computation of TAI a posteriori. The BIPM informs the timing community, each month for the previous month, about the status and evolution of each UTC(k) with respect to UTC. Consequently, each timing center knows how well it predicted UTC and what was the difference between its UTC(k) and the official time UTC, and from that, readjusts the frequency of UTC(k) if necessary, i.e. if UTC(k) goes away from UTC.

______________________________________________________________________________________________________________
(*) Primary frequency standards are very precise atomic clocks, manufactured in some laboratories, and which directly realize the definition of the second.