By LJ du Toit, EMSS Antennas (September 2007)

The 15m reflector antenna of the MeerKAT at the Hartebeesthoek Radio Astronomy Observatory (HartRAO) will soon be configured for two interesting tests, both related to the unwanted multiple bounce of energy between the parabolic reflector surface (especially the area around the vertex) and the feed cluster package at prime focus.

The first test configuration will be an experiment where a circularly polarized floodlight illuminator is mounted at the reflector vertex, with its coupling to the horns at prime focus being the quantity of interest. This experiment is designed to evaluate the possible use of a single noise source to illuminate all the horns in the cluster simultaneously, for one-step calibration of all the receiver chains.

At present the cluster backplane is populated with only two out of a possible seven horns, and since the backplane had to make provision for lower frequency (larger) horns in future, it itself is larger than required for the present 1 414 - 1 670 MHz horns. The large and unpopulated flat part of the metallic surface introduces a large (triple-bounce) standing wave between it and the vertex of the paraboloid approximately 8m away, which degrades the smooth coupling that is required for successful noise calibration.

To reduce this standing wave (by reducing the flat surface area of the feed enclosure), conical covers have been designed and will be substituted for the five missing horns. The floodlight-to-horn coupling will be measured by network analyzer, and will be compared to the design values predicted by our EM analysis software, FEKO.

A computer model of the two horns and five conical covers is shown in Figure 1, and a prediction of the coupling follows in Figure 2. The expected reduction in the ripple, when the five covers are installed, can clearly be seen.

The second test configuration is aimed at normal radio astronomy operation, where a (double-bounce) standing wave will still be present. In this case the quantity of interest is the impedance mismatch of the horn at prime focus, which is strong indicator of the severity of the standing wave problem. To reduce this standing wave over the 17% test bandwidth a small deflector plate has been designed, which is mounted around the floodlight illuminator at the reflector vertex.

This deflector plate and floodlight antenna cavity is depicted in Figure 3, and primary horn mismatch measurements will again be made to validate our electromagnetic design as sound and sufficient.

Figure 1: The five conical covers, two horns, feed cluster backplane (all in yellow), and top sections of the four struts (light blue) are shown on the left. A zoomed-out view is shown on the right, with the floodlight illuminator and deflector plate visible on the reflector surface vertex.

Figure 2: A numerical prediction of coupling between the floodlight illuminator and the center horn. The red trace is without the five conical covers present, and the blue trace is with them installed. The improvement (a smaller ripple in the coupling) is clearly visible, as well as the prediction that we will be within the required ± 2.5dB envelope.

Figure 3: The floodlight antenna cavity and deflector plate (with diameter ~1500mm).

View the latest pictures on the MeerKAT prototype.