Abstract

The deployment duration of ocean bottom seismometers (OBS) is limited by non-optimum timing brought about by large and non-linear clock drifts that increase with deployment duration. Precise and accurate timing is a requirement not only for measuring earthquake locations and spatio-temporal migration rates, but also for measuring temporal changes in rock velocity indicative of stress changes that can be indicators of near-term rupture. Engineers at the Woods Hole Oceanographic Institution (WHOI) have developed the capability for complete data retrieval from autonomous OBS using a high-speed optical telemetry modem capable of sustained transfer rates of 10 Mbits/second from a variety of platforms, including fully autonomous underwater vehicles. In addition to retrieving data, the offset of the OBS clocks relative to GPS-referenced time can be accurately measured and logged via the optical link. We've expanded the concept of the Optical Modem as an appliance concept by letting the user feed a Pulse-per-Second (PPS) signal into the Master optical modem that will be recovered on the Slave modem and delivered to the remote equipment with minimal additional jitter (40ns RMS). The tradeoff between jitter and remote timing lock is field-configurable. A Network Time Protocol (NTP) server is used to capture and transfer the vehicle time-of-day. Using this new advancement in WHOI Optical Modem technology, the OBS clock is able to compare its own clock drift to the freshly disciplined vehicle clock without perturbing the data offload capacity, at a slight cost of available bandwidth. To achieve this, the Optical Modem Physical Layer (PHY) has been modified to interleave measurement and configuration data frames, referred to as Out-of-Band (OOB) frames, alongside the legacy payload frames. Those new frames provide a direct deterministic, bottom-level communication channel for optical modems to exchange link information. The PPS recovery mechanism estimates the transmission (time-of-flight) and modem processing delay by measuring the time, in Field Programmable Gate Array (FPGA) clock cycles, between a determined point in the frame transmission process and the correlation peak of the frame detector on both sides of the optical link, exchanges that information over the Out-of-Band channel and feeds it to a local tracking loop. This local loop, independent from the reference optical modem can accommodate a fairly big frequency error between master and slave FPGA clocks. Because the only variable delay is the time light takes to travel between the two optical modems, a side benefit of the work done to make the timing transfer possible, is the ability to optically measure range between optical modems. While the raw resolution is quite poor expressed in FPGA clock increments, the measurement distribution is observed to be Gaussian, and the sampling frequency is 100Hz, thus the estimated range can be easily improved with averaging.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call