Performance Evaluation Of A Clock Synchronization Over Fiber Data Links For Large Experiments

Yang Hu,Yonggang Wang,Jie Kuang
DOI: https://doi.org/10.1109/NSS/MIC42101.2019.9059891
2019-01-01
Abstract:Most large experiments for high energy particle physics are time measurement related, which requires a network supporting the broadcast communication in order to distribute the central clock to all endpoints. In addition to conventional solutions using a dedicated network just for time synchronization, several groups have proposed to embed the clock distribution network into the data stream network. Using only a single bidirectional fiber link for all network abilities, some special demands for these experiments, for example, limited space for hardware and long distance connection, can be satisfied. As the time resolution of detectors is constant increasing, the performance requirement for time synchronization based on this unified network is challenged. In this paper, we propose to use the state-of-art techniques on field programmable gate arrays (FPGA) and optical circulator based fiber connections to implement the unified network. By constructing a prototype system with three levels of hierarchical structure, the time synchronization performance is experimentally measured. Benefitting from the single wavelength fiber connection, high precision time-to-digital convertors (TDC), and phase interpolator (PI) components in FPGA SerDes transceivers, the clock offset of the distribution can be precisely measured and automatically compensated. Through multiple times of power-on and reset operation test, the clock synchronization accuracy over 5 km fiber connection is measured less than 40 ps. By employing the jitter cleaner devices in clock recovered endpoints, the recovered clock in different levels have independent jitter less than 4 ps. In addition, the achieved high performance is also proved to be insensitive to operation temperature and the connection distance. The test results show that the proposed implementation is a straight-forward and practical method with the high performance that can meet the requirements of most current large-scale experiments.
What problem does this paper attempt to address?