Early History of Dipole Arrays

Radiotelescopes with a variety of antennas of different forms have been built to suit the large range of wavelengths over which radio observations are made^7.1. Quasi-optical antennas such as parabolic reflectors are considered more appropriate for milli-meter and centi-meter wavelengths. At the other end of the radio spectrum, multi element arrays of dipole antennas have been preferred for meter and deca-meter wavelengths.

Early observations in radio astronomy were made using one of the two methods, either pencil beam aerials of somewhat lower resolution to investigate the distribution of radio emission over the sky, or interferometers to observe bright sources of small angular size. However, the observations made during the early $1950$'s, showed that to determine the real nature of the radio brightness distribution it is necessary to construct pencil beam radio telescopes having beam widths of the same order as the separation between the lobes of the interferometers then in use $(\sim 1')$. An important step towards such modern high-resolution radiotelescopes was the realisation that in many cases even unfilled apertures, which contain all the relative positions of a filled aperture, (``skeleton telescopes'') can be used to measure the brightness distribution. A cross-type radio telescope, pioneered by Mills was the first to demonstrate the principle of skeleton telescopes.

A cross consists of two long and relatively narrow arrays arranged as a symmetrical cross, usually in the $N-S$ and the $E-W$ directions, intersecting at right angles at their centers (Figure 7.1). Each array has a fan beam response, narrow along its length and wide in a perpendicular direction^7.2. The outputs from both the arrays are amplified and multiplied together; only sources of radiation that lie within the cross hatched portion of Figure 7.1(b) produce a coherent signal. Thus an effective pencil beam is produced of angular size determined solely by the length of the two arrays. A substantial number of telescopes were constructed based on this principle.

Figure 7.1: A cross type telescope. The arrays in Panel (a) produce the fan beams shown in Panel (b). When the outputs of these two arrays are multiplied together, only signals originating from the cross hatched region common to both beams produce a coherent output. The resolution of such a telescope hence depends only on the lengths of the arms.

The Sydney University telescope was constructed as a cross with aerials of overall dimensions approximately $1$ mile long and $40$ ft wide (Mills et al 1963). The mile-long reflectors are in the form of cylindrical parabolas, with a surface of wire mesh. Line feeds for two operating frequencies of $408$MHz and $111.5$MHz were provided at their foci. The $N-S$ arm employs a fixed reflector pointing vertically upwards and the beam is directed in the meridian plane by phasing the dipoles of the feed. The $E-W$ arm is tiltable about its long axis to direct the beam, also in the meridian plane, to intersect the $N-S$ response pattern. No phasing was employed in this aerial. The angular coverage was $55^0$ on either side of the zenith. The $E-W$ aperture is divided into two separate halves through which the continuous $N-S$ arm passes. The total collecting area is $400,000$ sq.ft. This instrument had a resolution of approximately $2'.8$ at 408 MHz. This later came to be known as the ``Mills Cross'' and is one of the earliest cross type radio telescope built. In order to reduce cost, this telescope was built as a meridian transit instrument.

Note that in a cross antenna, one quarter of the antenna provides redundant information, since all element spacings of a filled aperture are still present if half of one array is removed. In fact, it can be shown that the cosine response of a $T$ array is similar to that of a full cross. Thus a survey carried out using a $T$ array has the same resolution as that of a survey carried out using a cross. However it has a collecting area $\sqrt{2}$ times lower than the corresponding cross and hence a lower sensitivity.

Footnotes

... made^7.1

see the illustrations in Chapter 3

... direction^7.2

See Section 6.2.2