Over the years in this column I’ve discussed some of my passions in dealing with radio signals (talking to astronauts aboard the ISS, tracking aircraft, decoding signals from near-earth weather satellites are among some of the topics I’ve covered).
However, for several years now I’ve wanted to tackle the decoding of signals from a geostationary weather satellite. This is an order of magnitude more complex because the satellites involved are far from Earth, at an altitude of almost 36,000 km above the equator.
Such geostationary satellites appear fixed with respect to an observer on Earth because at this great altitude their orbital period matches the rotational period of our planet. And therein lies both a challenge and an opportunity for a hobbyist such as me. Because the satellite appears fixed in place it should be possible to aim a suitable antenna in its direction and perhaps receive, and decode, signals from it.
Indeed, this is possible, more or less as I’ve described it. And a Canadian-American company, NooElec, has taken some of the complexity involved and simplified it quite dramatically. Their simplification is on the hardware side, packaging a dish antenna and receiver mount in three parts, and coupling these with two parts for which NooElec was already well known, a small software defined radio (SDR), and an equally small low noise amplifier (LNA).
Putting the physical components together was completely straightforward. More challenging was the data processing required to turn radio signals from the satellite, as received by the dish, amplified by the LNA, and tuned by the SDR, into images.
Here in western North America, the geostationary satellite of choice is named GOES-18, Geostationary Operational Environmental Satellite number 18, operated by the National Oceanic and Atmospheric Administration (NOAA). 18 is part of a group of four GOES satellites known collectively as GOES-R. This group has a collective price tag of almost $12 billion.
GOES-18 has particular strengths in four areas: atmospheric weather, environmental hazard monitoring, ocean observations, and space weather. GOES-R program director Pam Sullivan summarized the capabilities of the three existing, and one to come, satellites: “NOAA’s geostationary satellites provide the only continuous coverage of weather and hazardous environmental conditions in the Western Hemisphere, protecting the lives and properties of the 1 billion people who live and work there.”
Fast forward about a year. I had all the components assembled, but had put off tackling the software installation needed to bring everything together. This past weekend I did just that. I already had a small Raspberry Pi computer to handle the task. An enthusiast in Holland had put together a set of applications, named GOEStools, in the Linux environment for just this satellite group. Other enthusiasts had improved on this over time, chief among them weather satellite enthusiast Carl Reinemann in Wisconsin.
After installing an operating system on the Pi computer I proceeded to install the GOEStools software, carefully following the steps set out by my fellow aficionado in Wisconsin, more generally known by his Twitter name, USRadioGuy. After about an hour I was ready to begin the actual process of testing the entire assembly.
Part of the reason for my year-long trepidation was a concern that I might not be able to aim at the satellite, that it might be just out of reach behind a neighbour’s house. Using a site called DishPointer I was able to select GOES-18 as the satellite I was interested in, enter my exact location, and see a line plotted on a visual representation of my neighbourhood. Phew, just enough clearance as it turned out.
With sighting not being an issue it was time to get a signal. I activated the Pi computer, and the software component to test for signal strength. I had one problem: no screen on the Raspberry Pi. The solution was a phone application to connect to the Pi by its local address, then run the application while looking at the phone.
With RaspController showing the signal “quality” as processed on the Pi computer I began nudging the antenna rack a centimetre at a time. Once the error rate, known as a viterbi value, began dropping, I knew I was generally in the right direction. Once I had a low point, under 400, I took on the vertical angle. That required a small wrench. Slightly up, slightly down. Eventually that too had the viterbi value drop. With subsequent small adjustments I managed to get below 100.
Time to see if there was actual data. I issued the data collection command and sat back. About half an hour later I saw the first “writing” statement and shortly after that I was able to see the first image from a geostationary satellite as processed with my own ground station.
It was an exhilarating moment and I immediately posted it to X/Twitter: “First light! Signal that is. GOES-16 shot delivered via GOES-18. First image from my @Nooelec dish setup with SMArTee XTR tuner and software on a Raspberry Pi.” The tweet quickly drew several thousand views.
However, within the tweet I was disguising some disappointment. The image was from GOES-16. I was aiming at 18, also known as GOES-west. The software I was using was geared to users in eastern North America. Adaptation was needed to have the software directly process the images from 18. This took several additional steps and the assistance of USradioguy.
As I write, my GOES-18 station is working beautifully, serving up dozens of stunning images every day. The signal also delivers relayed content from GOES-east, Japanese meteorological satellite himawari, and weather maps from the American National Weather Service.
Amazing indeed to be able to pull in and decode signals from a billion-dollar satellite with inexpensive equipment costing no more than a few hundred dollars.
Follow me on Facebook (facebook.com/PeterVogelCA), or on Twitter (@PeterVogel)
Share your thoughts and contribute to the ongoing conversation by sending us a Letter to the Editor.