Data Center
No Snow Job
by Jerry Robaidek
Two things will be missing from our 11-meter Orchard Street antenna this winter: students with snow rakes, and the snow in the dish. Although increasing student unemployment was not a goal, a clever solution to a perennial problem will improve the reliability of our satellite data reception here at SSEC.
To an outside observer, collecting data from over 20 different satellites has seemed to become quite routine at SSEC. But, from antenna to server, and everything in between, each link in the chain has its own challenges. When 8 out of 10 of the antennas are located 210 feet up and squeezed onto our rooftop, the challenges are multiplied.
Over the years Dave Jones, the SSEC antenna engineer and technician, has had to come up with creative engineering solutions to keep the data flowing. Whether it is tracking a drifting geostationary satellite, or swapping a feed horn in freezing or windy conditions 210 feet up, or designing a makeshift feed horn cover from a plastic garbage can, Dave is responsible for assuring the data flows.
Over the last several years a new problem began to impact the data. More and more data are being relayed via C-Band. The C-Band is the part of the electromagnetic spectrum from 4 GHz to 8 GHz. Unlike L-Band, which the GOES, and older NOAA polar orbiters used as a downlink frequency, C-band is much more susceptible to attenuation due to liquid water. We rarely get a rain heavy enough to impact the signal, but at our latitude, heavy wet snow accumulating in the dish will completely obliterate the signal.
In the last several years we have dealt with snow in a pretty manual way. Dave modified a standard snow rake by adding an extended handle to pull snow from the dish. This method worked well, but usually some data loss occurred before a person could get to the antenna to pull the snow out. Since it doesn't only snow during business hours, 5 months of the year students were paid to be on-call over weekends and holidays to clean snow out of the antennas. Unfortunately, this strategy still did not prevent the middle of the night outages due to snow. Some of those outages could last several hours. Because many of our data users depend on the reliability of our data feeds, we decided a more automated solution was needed.
Dave did some research into automated heaters for our largest antenna (the 11-meter Orchard Street antenna). Options were limited, and early solutions proposed by commercial providers had price tags over $100,000. Dave had some ideas for less costly alternatives, and decided to design his own heating system. The material alone for the solution Dave came up with (an electric forced air closed chamber) was less than 1/10 the cost of the complete commercial solutions offered.
The installation of the heating system design was completed in November of 2012. Its first test came with the heavy wet snow on Sunday, December 9, 2012. As designed, the bottom third of the antenna was kept clear of snow, and not a bit of data was lost in that event. A bigger test came with the snowstorm that shutdown the campus on December 21. Again, no data was lost and no manual intervention was needed. February and March provided several more opportunities to test the system, and in every case, the antenna heating system worked perfectly.
We did have one data outage last season because of snow, that even our new antenna heating system could not protect us against. That outage resulted in a several hour loss of Meteosat-9 data on January 13, 2013. But that outage was due to snow in the *uplink* antenna at Wallops Island, Virginia.
The 11-meter Orchard Street antenna during the January 30, 2013, snow event.
