After a large telescope is constructed, engineers and astronomers often have to spend months or years tinkering before it finally begins contributing to science in earnest.

But last year, with just one quarter of its construction completed, the Atacama Large Millimeter/submillimeter Array radio telescope—the largest, highest-altitude and most expensive ground-based observatory yet—turned its dishy ears to the skies and began to listen. Already it has begun to sing out its discoveries.

Since astronomers began using the $1.3-billion international observatory—called ALMA for short—a year ago, it has revealed a "death spiral" of gas and dust unwinding around a dying red giant star, giving insight into what our sun's demise may look like. The array has provided a clear picture of a nearby planetary system that had eluded even Hubble. It has also detected sugar molecules floating in gas surrounding a star as well as markers for hundreds of other molecules in space—clues that astrophysicists astrochemists are working furiously to decipher.

And those discoveries are just from ALMA's intentional observations. As scientists test the observatory's ever-growing collection of antennas by pointing them at well-studied objects in space, ALMA is happening on details no other instrument has ever captured. Astronomers, for instance, recently tested the array by aiming it at the Antennae Galaxies, a pair of colliding galaxies that scientists have been studying since 1785, and saw with superlative detail its stellar-nurseries, where billions of new stars are being born. Another test found the poisonous molecule ethyl cyanide floating in a star-forming region in the constellation Orion.

"We are discovering things completely by accident," says Violette Impellizzeri, an astronomer on the observatory's Commissioning Team. "It's to the point that we have to be very careful what we look at," because they are stumbling on discoveries that other scientists were hoping for telescope time to investigate.

The observatory's sensitivity will only improve as construction brings more antennas to its array. Unlike optical telescopes, whose power comes from mirrors and lenses gathering light within a single instrument, modern radio telescopes consist of herds of saucers aimed in unison at a patch of sky. The dishes can be moved around in relationship to one another to get the best angle on a distant domain.  The weak waves detected by each of the dishes are digitally combined to create one strong signal, resulting in a much higher resolution than a single large dish could achieve.

And their aim is exact, says Stefano Stanghellini, antenna project manager for the European Southern Observatory (ESO), which is contributing 25 of ALMA's antennas. A person shooting a gun with the same accuracy would put a bullet through a two-euro coin at a distance of 10 kilometers, he says.

With this exceptional precision and sensitivity, ALMA will look into the darkest and coldest corners of our universe—places where optical telescopes can only grope. For example, a dark cloud surrounding a star may barely be perceived optically, but ALMA can image it in detail, and also determine its composition.

When ALMA first became available to investigators in October 2011, just 16 of its slated 66 antennas were operational, but it already constituted the world's most powerful radio telescope—a distinction that had more than 1,000 scientists salivating for time with it. A lucky 10 percent were selected for ALMA's first research cycle; the second cycle begins in January, with double the number of antennas.

The high-desert telescope is the product of two decades of work. In the 1990s astronomy teams from Europe, North America and Japan were all planning new radio arrays but, in an unusual move, decided to sidestep the normally competitive world of telescope construction and collaborate on one massive project.

Finding a site was the next challenge. Unlike their optical siblings, radio telescopes can operate in daylight. But to collect the  frail radio waves emitted by objects across vast space—particularly the water vapor-vulnerable short wavelengths that defied detection by other observatories but that were the primary aim of the new telescope—ALMA would need an exceptionally parched and transparent atmosphere. Few places fit those criteria better than the 5,000-meter Chajnantor Plateau—a blustery Andean mesa perched over Chile's Atacama Desert, one of the driest places on Earth, averaging just 1.5 centimeters of precipitation a year.

Construction at such altitudes is challenging: Workers need medical monitoring and carry oxygen tanks on every trip from the 3,000-meter elevation operations center to the vertiginous plateau. Also, the strong radiation from the sun at this altitude can distort the dishes' accuracy, an effect space-based telescopes have always battled, says ESO representative Massimo Tarenghi.

"It's like building 66 space telescopes," he says, but "we also have to consider the wind, the snow, the rain, the humidity, the dust."

Of the 43 saucers now roosting on Chajnantor, most are 12 meters in diameter—broad enough to comfortably cap the Washington Monument. As scientists turn their attention to a new celestial target, the dishes roll their gaze in unison across the lucent Andean sky, their synchronized dance witnessed by snow-capped volcanoes.

That they work at all is an impressive feat: If a dish's surface has an imperfection measuring as small as the diameter of a human hair, it will not perform. The antennas' receivers are chilled to just a few degrees above absolute zero—any warmer and their own radio emissions would interfere with those they receive.

To make things more complicated, no fewer than 20 countries contribute financially and scientifically to ALMA, a collaboration that has undoubtedly slowed the project—but also resulted in an excellent final product, Tarenghi says. Because many nations' science budgets are shrinking, such collaborations are on the horizon—including a radio telescope with a square kilometer surface area to be located in South Africa or Australia. That project could be complete by 2024.

In the meantime, astronomers will make due with ALMA, whose entire array should be online by the end of 2013. That's when the really big discoveries will come, Tarenghi says. "Then—I am exaggerating just a little bit—we can point to any part of the sky and see something new," he adds. "We have a window to the universe that's not available any other place in the world."