| Résumé | A new method of uncertainty analysis for non-white noise is presented and used to evaluate traceable holdover capabilities of a new NRC system for time and frequency dissemination to remote rubidium-cell (Rb) clocks. As in many similar systems, these remote clocks are disciplined to track UTC using the Global Positioning System and the common view time transfer method. When disciplining data is unavailable, we extend traceability into this holdover period with evaluations of uncertainty from the measurements of the Rb clock's non-white noise: modelled, matched, and simply presented as a time dependent dispersion in Monte Carlo simulations. This leads to improved simplicity, reliability and economy for the new method, with a post processed standard uncertainty down to 6 ns when disciplined, with a holdover period of t seconds introducing an uncertainty component of 1.82 × 10⁻³t nanoseconds (combined standard uncertainty of 160 ns for a 24 h holdover). |
|---|
