Research Of Auto-Coupling Control System And Clock Recorery Algorithms In Optical Sampling Systems

With the rapid development of modern optical networks, traditional electronic instruments would be limited to measure and monitor the ultra-high speed signals. Optical Sampling is a technology of sampling ultra-high speed signals optically instead of electronically, which can overcome the bottleneck limit of electronics, thus extend the bandwidth of sampling technologies. Therefore, this dissertation mainly focuses on the technologies of the auto-coupling control system which is applied to sampling gate in optical sampling system and the clock recovery algorithms to measure the eye diagrams. The main research efforts are summarized as follow.1) Waveguide-to-fiber coupling theory is introduced first to give the fundamental concept of designing auto-coupling control system. Software synchronization is introduced for clock recovery in sampling as well. GPIB technology and serial ports technology are presented, so as MATLAB scheme.2) The design, realization and testing of auto-coupling control system is discussed. The hardware design and the software design in the auto-coupling control system are given. Main work focuses on the alignment algorithms. A binary-tree data-truncation (BTDT) algorithm and return-traversal search (RTS) algorithm based on a threshold were designed for single axis searching. The experimental results show that the RTS algorithm provides a higher accuracy which also achieves multiple kinds of accuracy; while the BTDT is faster to search the maximum point than the RTS. Finally the auto-coupling control system is applied to waveguide testing. The experiments show that insertion loss of tapered fiber can drop to8-9dB including the transmission lossof waveguide in15minutes while aligning manually usually takes more than0.5hour and system can be kept stable in15-20minutes. Besides, with auto testing system sweep of wavelength and power can be completed in8minutes for2200points. Auto-coupling system has good effect, stability and practicability.3)Parameters like sampling rate, transforming function are defined by simulation and precision specification is proposed for clock recovery. Then, the ZOOM-FFT algorithm is realized by simulation and an improved ZOOM-FFT spectral fitting algorithm is given. Firstly, the secondary spectral fitting method is proved by analyzing interpolation theory. Secondly, higher order interpolation methods are simulated. It is shown that with samples generated by kinds of sampling rate or signal rate, there is little difference between the Fpresize values got from different interpolation order methods; however, as the orders become higher, the amount of fitting calculation increases greatly. As a result, considering time and precision factors, it is obvious that the secondary spectral fitting method is more effective. Finally, the simulation results for the secondary spectral fitting are analyzed and the value of Fpresize has improved by0.0043%compared with ZOOM-FFT algorithm while the time complexity remains the same.