An Improvement of TA(TL): Using the Combination of a Hydrogen Maser And a Cesium-Clock Ensemble

We developed a model to predict and modify the long term behavior of a hydrogen maser. We found the 1st order phase difference of a hydrogen maser, relative to TA(TL), was roughly homogeneously increased by time in whole analyzed period. Base on it, we modeled the frequency offset of a hydrogen maser as a linear function of time, remove the frequency drift and get modified phase combined the TA(TL) with a hydrogen maser. We examined the modified phase data by IGS GPSCP time transfer final data. The results showed the modification did not affect the short-term and mid-term stability; its Allan deviation is almost the same as the original hydrogen maser when the average time is under 3 days. We also found the longer period we used the better long-term performance we got, the periods of 30, 50, and 100 days showed comparable stability to TA(TL) when average time is more than 10 days.