Re: How does the Very Large Array (Telescope) work?

The Very Large Array is a radio interferometer located on the Plains of San Agustin, about 150 km southwest of Albuquerque, NM, and is operated by the National Radio Astronomy Observatory for the National Science Foundation. (This is a particularly auspicious time for me to be answering this question since I just returned from a tour of it.)

An analogy might be useful in explaining how an interferometer works. Suppose one had a large telescope. The resolution of a telescope is related directly to its size. The larger a telescope is, the better angular resolution it produces; angular resolution is simply how well one can resolve fine details. An an example, if one can read the writing on a dime on the other side of a room, one has good angular resolution.

Now imagine covering a tiny portion of the telescope's mirror. If one covers a sufficiently small region, that will leave the telescope's angular resolution essentially unchanged. Now imagine covering three tiny portions of the mirror. Again, if the three regions are sufficiently small, the angular resolution of the telescope is unaffected.

How many small regions can be covered before the telescope's angular resolution is affected significantly? The surprising answer is that most of the telescope's mirror can be covered, yet the telescope would retain essentially all of its angular resolution (provided the telescope is covered randomly; if only the outer portions of the mirror are covered, this is equivalent to making the telescope smaller). Why bother making one large telescope mirror, only to cover most of it? Why not simply make a bunch of small telescope mirrors, then arrange them in the shape of a larger mirror, thereby making a large virtual telescope?

This is what an interferometer is designed to do. An interferometer consists of two antennas separated by some distance, let's call that B. This distance is often termed the interferometer's baseline. By combining the signals from the two antennas, one forms a virtual telescope whose size is about B. For instance, the longest baselines at the VLA are about 35 km in length. Thus, with these baselines, one forms a virtual telescope about 35 km in size. Equivalently, if one observes a celestial source at the wavelength of 21 cm, one produces a virtual telescope with an angular resolution of about 1 arcsecond. (Recall that 1 degree is divided into 60 arcminutes, 1 arcminute is divided into 60 arcseconds, and the unaided human eye can see details no better than about 1 arcminute.)

(There's also no reason to stop at 35 km. One can imagine linking together telescopes on opposite sides of a continent, opposite sides of the planet, or even in space to form a virtual telescope as large or larger than Earth. This technique is called very long baseline interferometry. There's also no reason that interferometry cannot work at wavelengths other than radio. However, it is technically far more difficult. For instance, one has to know the length of the baseline to a fraction of the wavelength of the radiation. That's not too difficult when the wavelength is 21 cm; it's quite a task when the wavelength is 500 nm = 0.0005 mm.)

The VLA consists of 27 antennas arranged in a Y shape. (Actually, 28 antennas, but at any given time one antenna is rotated into the Antenna Assembly Building or "elephant barn" for maintenance.) Some antennas are relatively close together, while others are farther apart. Each (unique) pair of antennas forms an interferometer, and when all of the interferometers are combined together, one forms a reasonably "filled" virtual telescope. (In the case of the VLA, the 27 antennas form 27*26/2 = 351 unique pairs.)

Why does the VLA contain so many antennas? Above I wrote that an interferometer with a 35 km baseline produces a virtual telescope about 35 km in size. That's only partially true. The problem with a single interferometer with a 35 km baseline is that it doesn't have any 25 km baselines or 20 km baselines or 10 kilometer baselines or .... Returning to my example of being able to read the writing on a dime, imagine being able to read only the writing on the dime, but not being able to see the dime itself. The dime is much larger than the writing so it requires less angular resolution to see. By having lots of antennas and a range of baselines, the VLA allows one to build up a more complete picture of a celestial source.

I should point out that interferometers have a cost. In the case of the VLA, the twenty-seven 25-meter antennas are far easier to build than a single 35 km diameter telescope. However, combining the signals from the individual telescopes is a massive computational task.

Other important existing or planned radio interferometers include the Australia Telescope Compact Array, the Atacama Large Millimeter Array, the Giant Metrewave Radio Telescope, a space-based VLBI satellite named HALCA, the Low Frequency Array, the Very Long Baseline Array, the Westerbork Synthesis Radio Telescope , and the Square Kilometer Array.