Astrophysics Group

Cavendish Laboratory


Mullard Radio Astronomy Observatory
The Very Small Array (VSA)

The Very Small Array is a custom-built radio telescope (operating between 26-36 GHz), designed expressly to make high resolution observations of the cosmic microwave background (CMB) radiation on angular scales around and below one degree. This radiation is the relic of the Big Bang fireball, and contains the tiny (one 100,000th of a degree centigrade at most) fluctuations that are the imprints of the very start of the formation of structures, such as galaxies, in the Universe. This means that we are able to use these observations to answer some of the fundamental questions about the origin of such structures, and to determine some of the key cosmolgical parameters that describe our Universe, such as its density, matter content and rate of expansion. The project is a collaboration between the Astrophysics Group at the Cavendish Laboratories (Cambridge University), the Nuffield Radio Astronomy Laboratories (Jodrell Bank, Manchester University) and the Instituto de Astrofisica de Canarias (Tenerife). After initial construction and testing in the UK, the VSA was relocated to a site with prime observing conditions at Mount Teide Observatory at an altitude of 2400 m in Tenerife.

The Very Small Array The VSA has been operated in ‘compact array’ and ‘extended array’ configurations (see the technical description below) and data collected to achieve high sensitivity in both regimes have been analysed. Previous experiments have covered larger areas of sky to lower precision, as well as being based upon distinctly different instruments from the VSA, so the VSA dataset complements these extremely well. The key advantage of this diversity is that if the different results agree, one can be confident that they are correct. It is also noteworthy that the VSA project is unique in being able to robustly remove the unwanted radiation from radio galaxies and quasars that lie between us and the CMB relic radiation and would otherwise contaminate it.

The design of the VSA is strongly based on the Cambridge group’s previous successful CMB imaging experiment, the Cosmic Anisotropy Telescope (CAT). Both are interferometers – that is, they work by combining the radio signals from pairs of individual antennas pointing at the same part of the sky. As each antenna can be paired with every other one, an array of n antennas provides n(n-1)/2 independent pairs, giving us a lot of data to work with. Interferometers have further key advantages over single antenna telescopes. For our purposes, the most important of these are that we can specifically target the fluctuations in the CMB at the angular scales we are interested in (interferometers can be made selectively ‘blind’ to scales that we’re not interested in), and that our telescope has extremely good rejection of signals other than the CMB that single antennas would pick up (such as from the atmosphere, or thermal radiation from the ground). In essence, our telescope is constructed so that it simply ignores a lot of the things we don’t want to see!

Whereas CAT had only 3 antennas, and was sited at sea level in Cambridge, the VSA consists of an array of 14 antennas, and is sited at the (far superior) Teide Observatory site in Tenerife, where both the weather, and the overall contamination from the atmosphere are much smaller problems. The combination of more antennas, better quality receivers, and the better site means that the VSA collects data over 100 times faster than the CAT. It also makes much better quality observations, showing structures ranging in angular size from 0.2-3 degrees.

Technical Description

The VSA is a 14-element radio interferometer, operating at between 26-36 GHz. It can be operated in two complementary configurations, distinguished by a difference of a factor of 2.25 in the dimensions of the collecting horns mounted on the antennas. These two configurations are referred to as the ‘compact array’ (for the smaller horn size) and the ‘extended array’ (for the larger). The extended array gives us a much larger collecting area, and hence flux sensitivity, however the physical size of the larger horns prevents us from packing the antennas close enough together to make some of the observations which are scientifically interesting. Our solution is to perform our first year of observing using the smaller ‘compact array’ horns, which allow the required closeness of packing. This pair of complementary configurations gives us baselines (i.e. antenna separations) covering 15 to 300 wavelengths in length. This allows us to cover a lot more useful science than a single configuration would.


Corrugated horn feeding 90 degree-offset paraboloidal mirror, projected diameter 14 (32) cm. Mirror rotates about horn axis to provide tracking in one coordinate. Forward gain 31 dB (14 cm aperture), sidelobe response < -45 dB relative to main beam at angles > 30 degrees from main beam.


NRAO design 26-36 GHz pseudomorphic HFET amplifiers, system temperature approx 25 K at physical temperature 12 K.


Single-channel analogue phase-switched correlator, 1.5 GHz bandwidth (IF band 0.25 – 1.75 GHz). All independent pairs correlated (91 baselines).


Antennas mounted on tip table, giving zenith angle coverage up to 35 degrees. (Declination range -7 to + 63 from Tenerife, hour angle roughly +/- 3h).


This table compares the VSA sensitivity with that of CAT. Figures in parentheses are for the larger VSA antennas.

Frequency range [bandwidth] (GHz) Number of antennas x polarizations Field of view (degrees) Resolution (degrees) Temperature sensitivity per pixel in 300h (microK) Flux sensitivity per pixel in 300h (mJy) Statistical sensitivity = temp sensitivity/sqrt(no of pixels) (microK)
CAT 13.5 – 16.5 [0.5] 3 x 2 2 0.5 35 7 9
VSA 26-36 [1.5] 14 x 1 4.5 (2) 0.5 (0.2) 7 5 (1.3) 0.7

For information on other MRAO telescopes, follow the links on the left.