Techniques for Phasing an Array

The basic requirement for phasing an array is to combine the signals from the elements with proper delay and phase adjustments so that the beam can be pointed or steered in the chosen direction. Some of the earliest methods employed techniques for mechanically switching in different lengths of cables between each element and the summing point, to introduce the delays required to phase the array for different directions. The job became somewhat less cumbersome with the use of electronic switches, such as PIN diodes. However, the complexity of the cabling and switching network increases enormously with the increase in number of elements and the number of directions for which phasing is required.

Another method of phasing involves the use of phase shifters at each element of the array. For example, this can be achieved by using ferrite devices or by switching in incremental lengths of cable (or microstrip delay lines), using electronic switches. The phase increments are usually implemented in binary steps (for example $\lambda/2,\, \lambda/4,\, \lambda/8, \ldots \,$). In this scheme, the value of the smallest incremental phase difference controls the accuracy of the phasing that can be achieved.

In most modern radio telescopes, digital electronic techniques are used for processing the signals. The output from an antenna is usually down-converted to a baseband frequency in a heterodyne receiver after which it is Nyquist sampled for further processing. Techniques for introducing delays and phase changes in the signal in the digital domain, using computers or special purpose hardware, are fairly easy to implement and flexible.

The description of phasing techniques given above applies when the delay compensation of the signals from the different elements of the array is carried out at the radio frequency of observation. When this delay compensation is carried out at the intermediate or baseband frequency of the heterodyne receiver, the signals pick up an extra phase term of $2\pi\,\nu_{LO}\,\tau_{g}$, where $\nu_{LO}$ is the local oscillator frequency used for the down conversion and $\tau_{g}$ is the delay (with respect to the phase centre of the array) suffered for the element (see for example Thompson, Moran & Swenson, 1986). To obtain the optimum phased array signal, these phase terms have to be compensated before the signals from array elements with different values of $\tau_{g}$ are added. Furthermore, $\tau_{g}$ for an array element varies with time for observations of a given source and this also needs to be compensated.

For an array with similar elements, the amplitude of the signals from the elements is usually kept constant at a common value, while the phase is varied to phase the array. However, in the most general case, the amplitude of the signals from different elements can be adjusted to enhance some features of the array response. This is most often used to reduce the sidelobe levels of the telescope or shift the nulls of the array pattern to desired locations, such as directions from which unwanted interference signals may be coming. Arrays where such adjustments are easily and dynamically possible are called adaptive beam-forming arrays, and are discussed further in Chapter 7.