Impact of the time error sampling period on telecommunications clock stability measures

Five stability quantities have been considered in the new telecommunications international standards for clock specification: the Allan Deviation, the Modified Allan Deviation, the Time Deviation, the root mean square of the Time Interval Error and the Maximum Time interval Error. A key issue to consider carefully - but sometimes overlooked - when performing stability measurements, in order to draw meaningful conclusions from the measurement data, is the impact of the TE sampling period on the behaviour of the stability quantities. Such a dependence, for some of the quantities above, has been already pointed out in different reference papers. which studied in depth the behaviour of the stability quantities analytically and under various aspects. In order to provide further, sound information on the impact that the TE sampling period may have in practice on the measured stability quantities, this paper reports a broad set of simulation results, under the presence of all the power-law clock noises and for all the stability quantities, together with some experimental results. The results obtained confirm, in excellent agreement with the theoretical analyses, that the behaviour of the Modified Allan Deviation and of the Time Deviation is substantially dependent on the chosen measurement sampling period in the observation intervals where white and flicker phase noises dominate.