ATCA antenna surface accuray
With the advent of observations in the 12 mm band with all six antennas late in 1999 and, more recently in 2004, the installation of the 3 mm receivers on the five ATCA antennas on the eastern 3-km railtrack, the requirements on their surface accuracy have become much more stringent than in the past.
Solid surface panels
The first step towards achieving the required surface accuracy, and thereby antenna efficiency, for high-frequency observations was taken by replacing the two outer rings of perforated panels with solid ones in 1998-1999. The new panels were prefabricated with an rms deviation from a perfect paraboloid form of only a few hundredths of a millimetre.
Each panel, depending on its form, rests on 4 (or 5) support points. These connect to points in the support structure via bolts fixed by two pairs of nuts at both the upper and lower end of the bolts (see photo).
Adjustments of the bolts connecting ATCA antenna panels with the support structure.
In the past antenna surfaces were measured with instruments such as theodolites. However, the accuracy of such measurements is limited and proved to be insufficient to reach sub-mm rms antenna surface accuracies.
The new, currently used method of measuring the shape of antenna surfaces with the required accuracy is holography. (NB: This way not only the surface is imaged, but other parameters of the antenna optics are obtained at the same time).
ATCA antenna surface accuracy
An example of an antenna surface prior to corrections is shown in the figure below. Just below the central "hole", i.e. the area shadowed by the subreflector, one antenna panel is visible as a dark spot. It was raised by 1 mm prior to the measurement to serve as a sanity check of the holography run.
ca01 prior to holographic measurements and subsequent surface adjustments. The grey scale ranges from -1 mm to +1 mm deviations. Contours are placed at intervals of +/-0.2, 0.4, 0.6, 0.8, 1.0 mm. Plot produced by M. Kesteven.
The rms accuracy over the whole primary reflector prior to the upgrade was 0.23 micrometers, with a clearly visible large-scale distortion (which in this case actually comes from the subreflector, not the primary mirror). Compared to the minimum wavelength down to which the ATCA now observes, such inaccuracies are not acceptable.
In the process of adjusting the positions of surface panels a dial gauge is used. The base on which the gauge is fixed rests on one panel while the adjacent panel is moved with respect to it. One person under the dish adjusts the bolts, a second on the dish itself reads the gauge and reports to the one below on the current reading. The repeatability of measurements with such a gauge is good to about 30-50 micrometers (0.03-0.05 mm).
Dial gauge used for measuring the movement of one panel with respect to its neighbour, while adjustments are made below the antenna surface to the supporting bolts (see above).
After 2-3 iterative rounds of holographic measurements and surface panel adjustments, an rms deviation from a perfect paraboloid of about 130 micrometers can be achieved, while also the large-scale deformation was removed. This value, which is about 1/20 of a wavelength at 3 mm, is now acceptable (ideally, one would want the surface accuracy to be no worse than 1/50 of the wavelength [rms]). The power gain of the antenna at 100 GHz was improved from a mere 23% of the theoretical maximum prior to the adjustment to a whopping 67% thereafter! The corresponding holographic image of the antenna, after adjustments, is displayed below.
ca01 after holographic measurements and subsequent surface adjustments in February 2005. Same scale, grey scale and contour levels as above; this final image was used to perform a few more adjustments. Plot produced by M. Kesteven.
Higher accuracies are hard to achieve because of intrinsic uncertainties in the measurements (depending on atmospheric conditions) and the mechanical part of the adjustments (see above). Higher-accuracy corrections are actually not necessarily warranted, because the amount of antenna deformation as a function of elevation is in the same ballpark. Using a geostationary satellite as beacon for holographic measurements, the antennas are optimised for use at an elevation near 55 degrees, which is a good intermediate value (the antenna performance at either end of the elevation range from 12 to 90 degrees would be worse if they were optimised at one end of the range, rather somewhere near the middle).
Mechanical stress and environmental influences, but also wear and tear e.g. due to work on the antennas, lead to a degradation of the antenna surfaces with time, to a level around 250 micrometre rms. Therefore, surface adjustments are repeated about once every two years.