SETI Signal Detectors See First Light on Radio Telescope

October 23,2008
by Jill Tarter, Director, Center for SETI Research, SETI Institute

Jane Jordan runs the software team at the Center for SETI Research. She is an avid birdwatcher who displays lists of the birds she has been able to observe in the Bay Area and around the world. In the cubicle down the hall, where we conduct remote observations with our SETI detectors installed on the Allen Telescope Array (ATA) at the Hat Creek Radio Observatory in Northern California, Tom Kilsdonk has been keeping another list of ‘birds’ for the past few months. Tom’s list contains the distant spacecraft whose signals he has been able to detect with our new/old Prelude detection system working in concert with the beamformers built by our consultant Billy Barott, along with Oren Milgrome, Matt Dexter, Dave MacMahon, and others from the Radio Astronomy Lab at UC Berkeley. This list of ‘scored’ satellites is part of the changeover from our old observing programs, using two widely separated, large single-dish antennas as a pseudo-interferometer during Project Phoenix, to a new set of strategies using the real deal i.e.many small antennas linked together interferometrically to make up the ATA-42. It’s a small cubicle, but both days and nights, it’s full of people like Peter Backus (our Observing Projects Manager) and Gerry Harp (our resident astrophysicist, who has also written much of the code that commands the array) who assist Tom, and cheer with excitement and pride as a new check mark goes on the list to signify another captured ‘bird’.

Over the years, our near-real time Project Phoenix SETI observations have routinely used some of these spacecraft to verify the correct operation of all our hardware and software detectors. Until NASA stopped transmitting to the distant Pioneer 10 spacecraft shortly after its 30 th birthday in 2004, we checked in on its transmitter every day. Twice during our decade-long Project Phoenix exploration, Pioneer 10 didn’t show up --- on one occasion the clock at our remote telescope in Australia had mysteriously gained 21 seconds of time, and later the equipment at another remote site in the UK became ‘electron-challenged’ when someone accidentally disconnected the power cord. When Pioneer 10 retired, we finished Project Phoenix with the help of the SETI League and east coast radio amateurs who bounced signals off the Moon for us to find.

As we begin our SETI observing projects on the ATA, we face new challenges and even more need for these fiducials on the sky. We no longer have large single dishes as our collectors for the radio signals (for lots of really good reasons). So instead of big pieces of aluminum focusing the incoming radio waves onto a detector, we must first electronically combine the signals (introducing time delays and phase shifts) from all the individual antennas in just the right way to point the ATA at a specific star before we send those signals to our detectors. This is done electronically with a beamformer. One of the good things about the ATA is that there are likely to be many stars that are visible at any one time within its large field of view, so with multiple beamformers, and multiple detectors, we can explore multiple stars simultaneously, at different frequencies if we want. Furthermore, we can do this while our astronomy colleagues are mapping the sky for hydrogen gas, or large biogenic molecules, or other phenomena of scientific interest to them. This multiplexing potential is a new and exciting innovation that will speed up the SETI searching in the next decades.

While beamforming may sound easy, it’s difficult in detail. It’s necessary to calibrate all the required phase and time delays using astronomical sources like quasars, to keep all the signals aligned to a part in 10 11 (since we are now searching for ETI signals and DXing these distant spacecraft at higher X-band frequencies [8 GHz] in addition to L and S-band). It’s also necessary to completely remove the Walsh functions that are introduced to modulate the individual antenna voltages, all before the SETI detectors can begin their job. Our beamformers are built out of FPGA-based modules developed by the Berkeley Wireless Research Center and the CASPER efforts at UC Berkeley. On July 12, 2008 the first beamformer combined 12 antennas together, and SETI Prelude system detected the faint carrier signal from the Voyager 1 spacecraft, that has recently passed through the termination shock in the solar wind to move beyond the edge of our solar system. This is the most distant man-made object – the signal we detected was transmitted from a distance of 106 AU (106 times the average distance of the Earth from the Sun) or 9.85 billion miles. Figure 1 shows the detected signal in a ‘waterfall plot’ of time vs. frequency. Although it appears very faint to the human eye, the SETI Prelude detector integrated all the power along the track of the signal and made a reliable detection with a very high signal to noise ratio. On October 16, 2008 our beamformer testing had added another three antennas from the array, and even though the Voyager 1 spacecraft had moved another 2 AU farther away, to top the 10 billion mile mark, the detected signal in Figure 2 is more discernible to the naked eye --- just think how easy it will be to find with all 42 antennas at work! If you are still not convinced that the signal is there, take a look at Figure 3, which shows the results of integrating the power from the spacecraft carrier over the ~3 minute observation.


Fig 1.
Detection of Voyager 1 at 106 AU with 10 antennas.


Fig 2.
Detection of Voyager 1 at 108 AU with 15 antennas.

In addition to forming beams on particular stars in the sky, beamformers can also form nulls in other directions at the same time. Peter Backus and Gerry Harp are now exercising this null capability to design observing strategies that are least likely to be fooled by interference from terrestrial technologies that find their way into the side-lobes of the array (like your eyes, radio telescopes have a kind of peripheral vision).

During our recent testing, Ken Smolek, who consults for the Center for SETI Research from his home in Oregon, has also been on the phone with the folks in our observing cubicle. Ken is helping to build the next generation of SETI signal detector called SonATA (SETI on the ATA) to replace the Prelude detectors. SonATA does everything that Prelude does, but it is a software-only detector, capable of operating on commodity servers, without the special-purpose hardware accelerators that had to be built into Prelude to make it run in near-real time. A demonstration version of SonATA has now reached its own milestone. On October 9, 2008 the X-band signal from the Rosetta spacecraft was detected by the SonATA demo system! Figure 4 shows that detection. The Rosetta X-band signal is much stronger than the signals from Voyager 1, because it is only a few AU away, having recently flown-by asteroid Steins on its way to rendezvous with comet 67P/Churyumov-Gerasimenkoin 2014.

SonATA will be one of the first software clients to operate on a new platform being developed at the SETI Institute with the help of private donors – a Software-Defined Radio Telescope (SDRT). The SDRT is a multi-year project to develop, operate and support SETI observations for a simple, open and scalable, software-centric digital processing back-end for the Allen Telescope Array (ATA). The SDRT goals are to:

  1. Sustain exploitation of exponential "Moore's Law" improvements in commodity microprocessors, programmable logic, memory, storage and networking.
  2. Encourage new waves of innovation through software on general-purpose systems and by opening up the instrument and its simulacrum to a much wider community of users, and to
  3. Support the SETI Institute mission for full-time observing and detection of radio signals from extraterrestrials.


Fig 3
– Integrated power from Voyager 1 spacecraft over 192 second observation.


Fig 4
. – Detection of Rosetta carrier at a few AU