Time Domain Statistics for Evaluating Residual Noise, Including Instabilities in Time/Frequency Transfer Systems

The instabilities in time and frequency transfer systems, a form of residual noise, can contribute significantly to the total uncertainty in time or frequency comparisons. Understanding the characteristics of transfer instabilities is increasingly important with the new high-stability optical frequency standards being developed. First-difference statistics such as the rms Time Interval Error (TIErms), the Frequency Transfer Uncertainty (FTU), and ADEVS (a novel use of the Allan deviation equation) provide a more direct and accurate measure of residual noise than second-difference statistics such as the Allan Deviation (ADEV), the Modified Allan Deviation (MDEV), and the Time Deviation (TDEV). A unifying discussion on the use of existing first-difference statistics with residual noise, introduced individually in two previous publications, is presented here. Simulated noise data is then analyzed to illustrate the differences in the various statistics. Their strengths and weaknesses are discussed. The impact of pre-averaging phase (time) data is also shown.