Radio Astronomy

It is only within the twentieth century that the existence of other technologically advanced civilizations in space has become a possibility accepted within the scientific establishment, and far more recently still that the feasibility of detecting such other civilizations has entered mainstream thinking.

The first scientific paper seriously contemplating surveying nearby stars for intelligently directed microwave signals was published in 1959. Unbeknownst to the authors, as they were writing their pivotal paper, a young radio astronomer was preparing to perform the very experiment which they were describing. That scientist, Dr. Frank Drake, launched his Project Ozma search from the National Radio Astronomy Observatory (NRAO) facility at Green Bank, WV in 1960, ushering in the era of modern SETI.

Project Ozma must be considered the very first SETI study. It surveyed two nearby sun-like stars, for just a few weeks, at just one frequency, and detected no extra- terrestrial intelligent signals. Nevertheless, Ozma served as a model for dozens of later SETI projects.

The world's first SETI meeting was convened at Green Bank by Drake in 1961. As the agenda for that conference, Drake drafted an equation for estimating the number of possible communicative technologies in the cosmos. The Drake Equation is today the primary probabilistic tool whereby SETI scientists assess their prospects of success. Drake himself considers it a way of quantifying our ignorance. The exact equation is worthy of a chapter of its own, and in fact whole books have been written about it. Suffice it to say that its seven factors:

Thus SETI is possibly the most interdisciplinary of the sciences.

In the nearly four decades intervening, on the order of fifty different SETI projects have been conducted around the world, with frequency coverage extending throughout the microwave, millimeter-wave, and optical spectra. These searches have been attempted by Government agencies, educational institutions, non-profit scientific organizations, and, more recently, by amateurs.

Although no definitive proof of extra-terrestrial intelligence has yet been received, SETI has achieved scores of tantalizing hints that such signals might indeed exist. Many candidate signals have been attributed to terrestrial, aircraft and satellite interference, others to equipment malfunction and natural astrophysical phenomena, but a few defy explanation. Since these signals have failed to repeat or otherwise eluded our attempts at verification, we can draw no conclusion save that there is much to be learned about the universe we inhabit.

The Sky Survey -- Amateur SETI's Rightful Role

NASA's SETI program consisted of two distinct but complimentary research elements: a targeted search of nearby sun-like stars and an all-sky survey for interesting signals of unknown origin. 

The former, which involves aiming at likely candidate stars for long periods of time, is well suited to large, steerable dishes with their narrow beam widths and high sensitivities. If we guess right as to which stars constitute likely candidates, the targeted search will provide us with the greatest likelihood of immediate success. But since only a limited number of relatively nearby candidate stars is known to us, concentrating our search in their direction may cause us to miss an equally good star of which we happen to be unaware.

An all-sky survey, on the other hand, makes no a prior assumptions as to the most likely direction to explore. The sky survey attempts to sweep out the entire sky which can be seen from a given location. No antenna tracking is required, since it is the entire sky, rather than individual stars, which we seek to scan. While targeted search antennas must be constantly moved, sky survey radio telescopes are operated in meridian transit, or drift- scan mode, in which it is the Earth's rotation which turns them.

NASA's late targeted search has been resurrected by the non-profit California- based SETI Institute. Their Project Phoenix effort employs some of the world's finest radio telescopes, aiming them sequentially at promising targets from a catalog of nearby sun-like stars. But since large antennas have quite narrow beam width, they see only a small portion of the sky at a given time. To sweep out the whole sky with such large antennas would consume inordinate amounts of time. A sky survey effort, by contrast to a targeted search, would be better performed with antennas of moderate size.

Smaller antennas can see more sky within their beam patterns, but have correspondingly less gain. We achieve reasonable sensitivities through digital signal processing, but the antennas need to remain fixed on their targets for relatively long integration times. Fortunately, when used in meridian transit mode, small antennas, with their relatively wide beam widths, provide us with far greater signal acquisition time than do the larger antennas typically used for targeted searches.

The sky survey approach seems ideally suited to the community of amateur radio astronomers desiring to pursue SETI. The non-profit, membership-supported SETI League has designed and initiated just such a sky survey. A grass-roots effort which will ultimately grow to thousands of amateur radio telescopes worldwide, the SETI League's Project Argus sky survey was initiated in April of 1996. When fully deployed early in the next century, it will provide (for the first time ever) real-time full-sky coverage, looking in all directions at once, across all four pi steradians of space.

Selecting the Magic Frequency

Our Earth is currently surrounded by a sphere of microwave radiation roughly fifty light years in radius, which is readily detectable over interstellar distances utilizing technology such as is today available to amateur radio astronomers. This radiation, emanating primarily from our planet's UHF TV transmitters and long range search radars, would mark our planet as inhabited to any similar technological society within fifty light years. Within that range are found hundreds of stars, tens of which are sufficiently sun- like to probably host one or more habitable planets.

The distance over which we are detectable is limited only by the time since we first began transmitting sufficiently strong signals in the appropriate frequency range. Fifty years from now, we will be detectable out to 100 light years distance. At that point our signals will have engulfed thousands of stars, including hundreds of potential life sites. With every successive doubling of elapsed time (out to 1,000 years or so), the number of civilizations which our radiation signature can potentially reach goes up by a factor of eight. Sooner or later, our signals may well reach a distant radio telescope.

SETI hypothesizes that other technological civilizations are similarly surrounded by a detectable sphere of microwave radiation, the radius of which will be limited only by the length of time such civilizations have possessed sufficiently advanced radio technology. We depend upon our ability to intercept and recognize (though not necessarily decode) such a radiation signature to achieve the existence proof of other intelligent civilizations which SETI seeks.

The problem with seeking incidental radiation is that the unknown factors exceed the known. We can only guess as to where physically to point our antennas, when to listen, and on what frequency. The time dimension is resolved by starting to look now, and continuing until we detect something noteworthy. A large enough number of coordinated stations, effectively looking in all directions at once, resolves the pointing uncertainty. And we can narrow the search space in the frequency dimension by recognizing the range of frequencies which are least attenuated by planetary atmospheres and the interstellar medium. This, however, leaves us with most of the microwave spectrum, and much of the optical, as likely frequencies.

Since there are no "wrong" frequencies to search,  whatever frequency at which you can assemble a workable radio telescope is fair game. Amateur radio astronomers have long explored the 406 MHz, 610 MHz, 1.42 GHz and 10.6 GHz radio astronomy bands, and I can think of no good reason why they should not pursue SETI in those spectral regions as well.

The foregoing, however, applies only to the problem of scanning for incidental radiation from the distant civilization. What if another intelligent race were making a deliberate, concerted attempt to signal its presence to its interstellar neighbors? Is there a particular frequency, or range of frequencies, which would be self-evident to the receiving civilization? Can we narrow the search space?

Cocconi and Morrison thought so when they published their 1959 Nature article. They reasoned that 1420.405 MHz, the precession frequency of neutral hydrogen atoms, was a good place to start looking for deliberately beamed interstellar beacons. This frequency, which falls in the quietest part of the radio spectrum, is marked for all to see, by Nature herself. There is nothing geocentric about hydrogen radiation; perhaps, they reasoned, selecting it for interstellar communication is a mark of intelligence, in and of itself.

Drake had arrived at the same conclusion independently, and indeed monitored a narrow band of frequencies encompassing the hydrogen line (also known as H1) during his Project Ozma search. Today, nearly four decades later, the hydrogen line region still looks like a good bet to many SETI professionals.

Fortunately for amateur SETI, much amateur and professional radio astronomy already goes on at the hydrogen line. Equipment for this frequency region is abundantly available, and much of it can be readily adapted to SETI use. There are indeed other likely "magic frequencies" which are being scanned for signals of possible intelligent extra- terrestrial origin, and once again, one person's guess is as valid as another's. Nevertheless, many of the world's amateur radio astronomers are already scanning the hydrogen line for natural astrophysical phenomena, and it's a small step to make their receivers search for artificial signals as well. The following sections discuss the hardware, software, and human considerations of amateur SETI.

SETI with a Radio Telescope

"I already own a sensitive radio telescope," many an amateur radio astronomer has noted. "Why can't I use it for SETI?" The short answer is, you can! An antenna and preamplifier adequate for radio astronomy might potentially detect intelligent signals from space. To achieve this adequate sensitivity, we design the preamplifier circuitry so as to generate minimal device and thermal noise. And we design the antenna so as to minimize the noise contributions of our warm planet, instead responding primarily to the cold sky above. Most any successful radio telescope meets these conditions. But for SETI, we also need to pay special attention to the receiver, and the post-detection hardware and/or software, if we are to maximize our admittedly slim chances of success. This section will deal with receiver considerations. Signal processing is addressed in the section which follows.

Any amateur (or professional) radio astronomer pointing a sufficiently sensitive radio telescope at the sky will encounter a wide variety of naturally occurring radio phenomena. Prominent among these will be solar radiation, or sun noise, which spans the spectrum. All stars emit this broadband signal, though it will be most pronounced, and most easily detected, from our nearest stellar neighbor. In addition to solar noise, H1 radiation emanates from the roughly one hydrogen atom found per cubic centimeter of interstellar space. While concentrated at the 21 cm (1420 MHz) line, it is Doppler shifted both up and down in frequency by the random motion of the interstellar medium.

Though hydrogen dominates all of space, countless other atoms and molecules, both inorganic and organic, permeate the interstellar medium, and many emit characteristic signals which are similarly Doppler shifted across the spectrum. These natural emissions, the signals which radio astronomy seeks, are present in receivers pursuing SETI as well, but in this case represent not signals at all, but potential interference.

Fortunately, all known natural radio phenomena emanating from space are inherently broadband in nature, none being narrower than a few hundred kHz. Intelligently generated microwave signals, on the other hand, are characterized by their relative spectral purity or coherence, and (depending upon their modulation mode and information content) might be very narrow band indeed. So spectral coherence is one of the hallmarks of artificiality which SETI seeks, and which helps us to distinguish between a SETI signal and natural "noise."

Most microwave receivers used for classical radio astronomy tend to be relatively broadband. If the signal energy we seek represents a natural astrophysical phenomenon (which we can expect to occupy a broad slice of spectrum), then it makes good sense to employ broadband receivers, so as to intercept as much as possible of the signal energy. Such is not the case for SETI.

SETI tends to utilize extremely narrow-band receivers, if only at the post-detection level. That is, our radio frequency (rf) circuitry might scan wide spectral expanses, but we process the received signals in some way, into very narrow channels or "bins," in search of artificial phenomena. These bins tend to be tens of Hertz to tenths of Hertz wide. This has significant implications if we try to adapt existing (presumably broadband) radio telescopes to SETI.

We could, for example, modify any superheterodyne radio astronomy receiver for narrow-band reception, simply by adding a narrow IF (intermediate frequency) filter. But unless the LO (local oscillator) used to downconvert the incoming signal frequency is sufficiently stable, the signal may not stay within the IF passband long enough to process. Thus the radio astronomy receivers which hold most promise for SETI applications will be those with crystal-controlled LO chains. And to reduce thermal drift, an oven-stabilized crystal is highly desirable.

Many of the more capable microwave receivers employ digital frequency synthesis of the local oscillator frequency. Synthesizers generally provide us with ample frequency stability, in that they involve phase-locking a free-running oscillator to a highly regulated, temperature-controlled crystal reference oscillator. Unfortunately, all but the most sophisticated synthesizers suffer from marginal spectral purity. This is because synthesizers tend to generate a plethora of phase-noise sidebands only a few tens of dB weaker than the desired LO frequency.

Phase noise limits the SETI receiver's ultimate sensitivity, by adding noise prior to the detector. But it has an additional detrimental effect, in that noisy LO's might generate spurious receiver responses, giving us multiple opportunities for a false indication of a coherent signal where none is in fact present. A high level of falsing can be expected for SETI anyway, due to the polluted nature of our planet's RF environment. Why complicate the situation with receiver-generated false hits? It is probably better to avoid synthesized receivers, unless they have been designed for the lowest possible phase noise.

Another LO concern deals with long-term stability. In order to maximize the sensitivity of a SETI receiver, it might be necessary to integrate the signal (in either hardware or software) for many minutes. The LO must hold still so that the received signal remains in the bin width for the entire integration period. All but the most carefully designed oscillator circuits will exhibit excessive long-term drift.

In summary, radio telescope receivers may prove useful for SETI, with modification. A narrower bandwidth IF filter is usually called for, and it is often necessary to employ an external, crystal-controlled and temperature regulated LO chain exhibiting the very highest possible frequency stability, and the very lowest possible phase noise. Such an LO is the most critical element of a suitable SETI receiver.

Signal Processing Considerations

OK, so we've come up with a radio telescope which employs an acceptable LO, ample IF filtering, and adequate sensitivity to recover the weakest of signals. We're still not done. We now need to process the recovered signals into narrow bins, and identify within them those signals which might emanate from distant technological civilizations.

The earliest SETI receivers employed filter-bank technology. That is, the IF was split into multiple filters, each with a bandwidth of a few kHz, on adjacent frequencies. Each filter drove its own square-law detector circuit, and any signal which appeared at the output of one filter channel, but not the adjacent ones, was considered narrow enough in bandwidth to constitute a SETI candidate. This is the very scheme employed at the Ohio State University Radio Observatory in 1977, when the so-called "Wow!" signal (the most tantalizing SETI candidate signal to date) was detected.

Fortunately, our technology has advanced significantly since then. Today the favored tool for SETI signal analysis is digital signal processing (DSP), employing computers executing fast Fourier transform (FFT) algorithms. Implementing such techniques in custom, dedicated DSP microcircuits, the SETI research community has for some time concentrated on developing sophisticated multi-channel spectrum analyzers (MCSAs) capable of scanning millions of bins, over hundreds of MegaHertz of spectrum, in real time. The current state of the art in MCSA technology is probably BETA, developed at Harvard University by physicist Dr. Paul Horowitz, with funding from the Planetary Society and other private and corporate donors. BETA now analyzes several hundred million bins, each less than one Hertz wide. Such technology is, unfortunately, well beyond the reach of the amateur SETI community at the present time. But we can learn from it, and emulate it on a small scale.

Personal computer technology today makes it possible for the amateur radio astronomer to scan thousands of bins, over tens of kiloHertz, at virtually negligible cost. The audio output from a SETI receiver must first be digitized for signal analysis, and this is accomplished in any of a number of inexpensive computer sound cards. SETI League members have developed a variety of shareware FFT programs to sort this audio output into bins, and display the results on the computer monitor as histograms, waterfall displays, or any number of alternative formats.

Early amateur SETI systems are digitizing a 12.5 kHz audio bandwidth, and applying DSP software to break it down into 1024 individual bins, each about 12 Hz wide. It remains to be seen whether these values are optimal, but the beauty of the PC-driven DSP approach is that the search parameters are readily changed in software. As faster personal computers and more advanced sound cards become available, it becomes possible to reduce the width of individual bins, increase the total number of bins scanned, or increase the bandwidth of the audio spectrum which is being monitored.

Since sensitivity of radio telescopes increases with the square root of integration time, small-aperture amateur instruments generally time-average a very large number of observations to achieve reasonable performance. Long integration would similarly improve the sensitivity of amateur SETI systems, but with a complication. We are observing the heavens from a rotating platform, which imposes on all received signals a characteristic Doppler shift related to the Earth's motion. Depending on frequency and declination angle, this Doppler shift can be ten to hundreds of Hertz during the time it takes a signal to transit the antenna's beamwidth. For wideband radio telescopes, the Doppler shift is minute compared to the signal bandwidth, hence we can integrate for the entire transit time.

Narrow-band SETI receivers, on the other hand, are integration- limited by Doppler to the time it takes the signal to drift between bins. Given, for example, a 10 Hz bin width, and a Doppler rate due to the Earth's rotation of 10 Hz/min, we would be limited to only one minute integration periods. Beyond that, the signal would find itself in the next bin of the digital signal processor. This Doppler phenomenon significantly limits the maximum integration time we can utilize, hence the maximum sensitivity we can achieve.

There is a partial solution to the above problem. The same computer which performs signal analysis can compute the Doppler rate, as a function of the frequency scanned and the coordinates of the antenna. Many microwave receivers can be tuned under computer control. If the receiver's local oscillator is properly chirped (that is, tuned slowly in frequency) at exactly the Doppler rate, the effects of the Earth's rotation can be nullified, and longer integration becomes possible.

Unfortunately, chirping the receiver's LO only compensates for the rotation of our own planet. A valid SETI signal would most likely be emanating from a similarly rotating planet, which would impose a Doppler shift on the transmitter which we can in no way predict. It is hypothesized that any civilization producing a deliberately beamed interstellar beacon would solve the problem for us, by drifting their transmitter's frequency so as to compensate for their own Doppler. However, we can expect no such assistance in the case of intercepting a civilization's leakage radiation, hence our practical integration times are likely to be limited.

Software is currently under development to automate the signal analysis and verification process, by alerting the operator (and through the Internet, other SETI participants) when a signal meets a set of user-programmed criteria. Terrestrial and satellite interference have already generated false alarms for our early participants. But through the application of artificial intelligence (AI) techniques, it is expected that the system will ultimately learn from its false detections, so that in time, it will only respond to those signals which represent the most viable candidates for SETI success.

Conclusions

The world's amateur radio astronomers are in a unique position to make major contributions to the ongoing Search for Extra-Terrestrial Intelligence (SETI). Their radio telescopes already contain much of the hardware and software which comprises a credible SETI station. By paying careful attention to LO stability, IF filtering and DSP techniques, they can achieve sensitivities adequate to detect signals of likely power level out to perhaps several hundred light years.

Our signal analysis capabilities are presently limited primarily by the power of our computers. But that's a very good place to be limited. Computer power has been roughly doubling every year for the past few decades. If the technological trend continues, within ten years our available computers will be about 1,000 times as powerful as the ones we use today. At that point, there may well be no place in the Milky Way galaxy which evades our gaze.

Lacking a concentrated, Government-sponsored SETI program, success will most likely require thousands of individual stations in a coordinated effort. The SETI League is one organization willing to provide the needed coordination. But discipline on the part of the participants is also crucial. Fortunately, the optical astronomy community has already showed us that amateurs have the discipline necessary to make significant scientific contributions. Why should it be otherwise in the radio spectrum?

Those amateur astronomers interested in pursuing the SETI challenge are invited to join the non-profit, membership-supported SETI League, Inc. The SETI League maintains an extensive Internet presence; publishes quarterly newsletters, how-to manuals, and other technical documents; assists its members in locating equipment and software, as well as setting up their SETI stations; provides coordination of frequency and sky coverage; and provides a medium of communications for participants in its Project Argus all-sky survey. 

Details available on the World Wide Web at http://www.setileague.org/