A time scale is a procedure for combining the data from an ensemble of clocks or frequency standards. The input data to the ensemble algorithm are generally the time (or frequency) differences between each of the members and the reference device for the system. Therefore, the overall time and frequency of the ensemble are free parameters and must be continuously adjusted using external data and a steering algorithm. In addition, the outputs of the procedure are the average time and frequency of the group and the characteristics of each device (time, frequency, frequency aging, prediction error, etc.) with respect to this average, and a second job of the steering algorithm is to realize this computed average time and frequency in a physical device. I will discuss the considerations that govern the design of steering algorithms in general. I will illustrate these considerations using the algorithms that realize UTC(NIST) from the ensemble average of the time differences of the cesium standards and hydrogen masers that are located at the NIST laboratory in Boulder. I will also discuss the design of the steering of the backup for UTC(NIST), which is based on an ensemble of cesium standards located at the NIST radio station in Fort Collins. The backup time scale is intended to support the NIST services should the primary time scale become unavailable, so that it must track UTC(NIST) as closely as possible, which implies a tight coupling between the two scales. At the same time it must remain independent of UTC(NIST) so that it does not fail if its external reference becomes unavailable, which implies a loose coupling. I will discuss the actual design, which is a compromise between these two noncompatible requirements.
Proceedings Title: Proc. 2008 PTTI Conf.
Conference Dates: December 1-4, 2008
Conference Location: Reston, VA
Pub Type: Conferences
AT1, steered clocks, time scale, UTC(NIST)