The effects of clock errors on timescale stability

Abstract

The weighting scheme for the cesium clocks and hydrogen masers constituting the USNO timing ensemble is reexamined from an empirical standpoint of maximizing both frequency accuracy and timescale uniformity. The utility of a sliding-weight relation between the masers and the cesiums is reaffirmed, but improvement is found if one incorporates inverse Allan variances for sampling times of 12 and 6 hours for the cesiums and masers, respectively, with some dependence on clock model.