Monday, February 14, 2011

station communications update

This is a follow-up message to my last blog post in January, which described a weeklong maintenance operation on the Little Cayman CREWS station. During that week, we attempted to repair the station's radio link (which was entirely offline) and satellite transmitter (which was performing poorly). By the end of the week we had repaired the radio link but left the satellite communications entirely offline. At that point the near-real time data feeds were suspended. There have been several developments since January.

First, some background on CREWS communications. All CREWS stations are installed with both a TX312 satellite transmitter and an RF401 radio transceiver. Each station has an assigned GOES communications window of 20 seconds once per hour at 1200 baud, during which it transmits summary data for the hour. Some transmissions may experience interference due to weather/clouds or other sources; for a typical CREWS station, up to 5% of transmissions may be lost. These data are stored to a local flash memory card and are recoverable during the next (usually annual) visit, but they are not easily recoverable in the short term.

At most CREWS stations, the radio transceiver provides access to station programming and data only during our annual maintenance trips. This is a two-way 38400 baud connection that is typically used to check station operations from a laptop computer on the boat immediately after station power-up.

However, the Little Cayman Research Centre (LCRC) can make much more extensive use of its radio link by virtue of its near proximity to the offshore station, its dedicated computer resources, and its full-time research staff. During this January visit the centre's computer was reinstalled and reconfigured to automatically poll the CREWS station for new data once every five minutes. The land-based computer therefore holds a full repository of all the station's data -- not just hourly summaries, but details of every measurement recorded by the station, whether 6-minute (CTDs, CT), 1-minute (Vaisala weather transmitter), 30-second (BIC light sensors) or 5-second (anemometer, barometer, air temperature). It must be emphasized that this data link is self-correcting -- even if the land-based computer were to go down for an hour, a month, a year, it would automatically catch up on all missing data whenever it came back online.

The situation in January was that the new satellite transmitter failed during deployment, and weather conditions had worsened to the point where it was not immediately possible to switch back to the (poorly-performing) transmitter. Therefore, we went one step further with the radio connection and implemented some scheduled tasks on the research centre's computer to send the hourly data via FTP to a NOAA server. These data are in different format than the satellite transmission and required new programming to parse and interpret, but as of Friday, February 11th the Little Cayman CREWS station is once again online in much the same way as it had been via satellite communications.

The station's most recent twelve hours of data (updated hourly) can be found at http://www.coral.noaa.gov/static/data_lciy2_Web_12.html.

A longer 3-day report (updated once daily) can be found at http://www.coral.noaa.gov/static/data_lciy2_Web_72.html.

The station's feed to NDBC is once again online and can be monitored at http://www.ndbc.noaa.gov/station_page.php?station=lciy2.

The station's "ecoforecast" project page, which is again loading the near-real time data, can be found at http://ecoforecast.coral.noaa.gov/index/0/LCIY2/station-home.

The state of station communications is therefore as follows: there is a near-real time means of access to station data. This communication link depends on the continued performance of the RF401 radios (both in the pylon and on land), the continued good health of LCRC's land-based computer, the power supply to LCRC, and LCRC's network connectivity. Any one of these elements might be subject to interruption, some to frequent interruption. However, it is now the case that if any of these elements experience temporary downtime, then all missing data will become available once the connection resumes. This is in stark contrast to the GOES communications link for most CREWS stations, where dropped transmissions may wait up to a year to be recovered.

Regarding the failure of the new TX312 satellite transmitter, we started a dialogue with the support people at Campbell Scientific, wherein we described the unit's failure and shared its error codes. We have been told that there may be a failure with the transmitter's OCXO crystal oscillator. This diagnosis could only be confirmed by retrieving the transmitter and downloading its logs through a direct connection to its diagnostics serial port. If confirmed, CSI would provide us with an RMA number for returning/repairing the transmitter.

The bottom line is that it would be preferable to have both means of communication, satellite and radio, operational. However, the station survived for a year with satellite but no radio, and it could easily survive for another year with radio but no satellite.

Mike J+