Radical Radio Visions

Radio SETI may no longer be the only game in town, but it's still the game to which most researchers belly up. That's because the odds of a jackpot, though quite unknown, are unquestionably getting better all the time — because the instruments are growing more capable by leaps and bounds.

The ideal SETI radio telescope can only be imagined. It would monitor every point on the sky, in every radio channel from one end of the microwave window to the other (about 1,000 to 11,000 megahertz), all the time — a true Omnidirectional Search System, or OSS.

Unfortunately, this ideal is a very long way off. But it's no longer impossible to work toward. The STWG team considered what it would take to build a reasonable interim OSS. They were seduced by the thought of a telescope able to find powerful but intermittent signals, the kind that none of the current large SETI experiments has a hope of detecting.

A baby prototype OSS has already been developed by radio astronomer Robert Dixon at Ohio State University. Known as Argus (a reference to the hundred-eyed monster in Greek mythology), Dixon's prototype uses an array of small, fixed antennas to synthesize many beams, or pixels, on the sky. Synthesis is a well-established radio-astronomy technique. Its most famous embodiment is at the Very Large Array (VLA), where 27 dish antennas are splayed across the New Mexico desert. By meticulously combining data from each of the array elements, one can artificially construct, or synthesize, the beam of a very large antenna, though the computing demands are tremendous.

If even greater computer horsepower is on tap, many different beams can be constructed simultaneously. In this way a field of simple antennas can form a simultaneous radio image composed of hundreds of sky beams, or pixels. That's what Dixon's Argus prototype does.

Still, the OSS wouldn't be just a souped-up VLA. Unlike a conventional radio telescope, a SETI instrument also needs to divide the radio spectrum into very many narrow channels, each of which has to be examined separately. This is because E.T. hasn't sent us a fax revealing which hailing frequency he's using. Unless we make an incredibly lucky guess, we have to check them all.

A bare starting point for an OSS, the working group decided, would be to have 1,000 simultaneous channels per beam. Making a receiver capable of that much spectral resolution is primarily an exercise in computation, and computers are rapidly getting cheaper. But alas, the bit-busting doesn't stop there. Synthesizing about 4,000 pixels to cover the sky (each one a big 2 degrees across, and each with those 1,000 channels) requires galloping gigaflops. The imposing total is 1.5 million giga-operations per second, which at today's prices means about $3 million worth of computer.

Multiply that by 1,000 if you want a million simultaneous channels. (Even today's SETI receivers resolve tens of millions of channels at once.) If you want billions of narrowband channels to cover the whole microwave window, you're budgeting the gross national product of the United States for your computers. And that's before the electric bills arrive.

It gets worse. Unlike the VLA, whose 25-meter antennas typically restrict its imaging field to a patch of sky roughly 1/2 degree wide, the OSS would require very small antennas (about 0.2 meter) if it wants to extend its peripheral vision to two-thirds of the sky. But such diminutive antenna elements impose a price: even with an array of 4,096 of them, the OSS would sport no more collecting area than a single dish 7 meters across. This paltry aperture means that only very strong signals will get our OSS's attention.

Despite its poor sensitivity and mammoth computing requirements, a preliminary OSS is still an attractive SETI option. Even a 1,000-channel unit could be used to march through a billion channels, step by step, over a period of years. If we make a brilliant guess about some preferred frequency, the OSS will open a new region of search space by looking everywhere for radio blasts that persist for only an hour, a day, or a week. As an example, it could detect comet- and asteroid-tracking radars used by a prudent alien society 1,000 light-years away to check its planetary system for dangerous rubble. That's a seductive prospect.

But what's the price tag? The bill for the OSS is entirely dominated by computer costs, and these continue to plummet. So the STWG recommended investing a few hundred thousand dollars a year in OSS research and development for now, and awaiting much cheaper computing.

As everyone in the computer world knows, "Moore's Law" suggests that the computing power you get per dollar doubles every 18 months. That means a thousandfold improvement in 15 years, at least for computation that can be spread among a lot of chips. The OSS planners are betting that Moore's Law will keep holding true; they hope to have an actual instrument on the air by 2015. If computing power continues advancing beyond that, "all sky, all frequencies, all the time" will no longer be a SETI pipe dream.

Continued; click "Next Page" below.