A real-time performance improvement method for composite time scale

The composite time scale (CTS) provides a stable, accurate, and reliable time scale for modern society. The improvement of CTS’s real-time performance will improve its stability, which strengths related applications’ performance. Aiming at this goal, a method achieved by determining the optimal calculation interval and accelerating adjustment stage is proposed in this paper. The determinants of the CTS’s calculation interval (characteristics of the clock ensemble, the measurement noise, the time and frequency synchronization system’s noise and the auxiliary output generator noise floor) are studied and the optimal calculation interval is obtained. We also investigate the effect of ensemble algorithm’s initial parameters on the CTS’s adjustment stage. A strategy to get the reasonable initial parameters of ensemble algorithm is designed. The results show that the adjustment stage can be finished rapidly or even can be shorten to zero with reasonable initial parameters. On this basis, we experimentally generate a distributed CTS with a calculation interval of 500 s and its stability outperforms those of the member clocks when the averaging time is longer than ∼1700 s. The experimental result proves that the CTS’s real-time performance is significantly improved.