Radio astronomy, satellite communication, incoherent-scatter radar and other applications make use of large parabolic antennas to provide directional sensitivity to radio sources in the sky. The accuracy to which an antenna can be directed to a desired location is referred to as the antenna pointing. However, numerous effects can degrade the performance: manufacturing tolerances, alignment, gravitational and thermal distortion, to name but a few. It is possible to model these effects and then use the results to correct the nominal position. However, in order for such a calibration to be made, a series of measurements needs to be carried out against radio sources on the sky with known coordinates. From these, each axis positional error, (θ requested − θ actual ) can be determined, as a function of the antenna axes (typically azimuth and elevation). One technique that can be used is scanning the antenna across a known target source and fitting the beampattern to radio power measurements. This works well for large telescopes with good sensitivity.

For moderate-to-small sized telescopes, there are numerous challenges which make obtaining such measurements difficult. The small collecting areas makes their sensitivity lower. The smaller physical size also precludes larger front-end systems. This means cryogenic receivers are not possible and even thermal insulation often needs to be reduced due to the restricted size. Smaller radio telescopes are often limited in budget and staff resources, which results in additional challenges in finding cost-effective calibration solutions. Limited budget may impact aspects such as receiver bandwidth, site selection, or mechanical tolerances on the structure. The lower sensitivity, then means fewer usable target sources for obtaining positional error measurements. It also means spending longer integrating on sources in order to detect them in the first place. This introduces a challenge of atmospheric stability, making traditional scanning of radio power measurements prohibitive. The method presented here is the use of spectral-and-beam-scanning of astrophysical masers. The telescope used was MCA #1 of the Metsähovi Compact Array, Finland, which is a 5.5-m diameter ratio telescope, with azimuth-elevation mount and C-band receiver.

Software-defined radio provides a low-cost means of obtaining radio spectra. Although the bandwidth is limited compared to larger systems, this is not needed due to the narrow width of the spectral line being used for the observation. The masers also have the advantage of being point-sources, with the Doppler velocity being an additional means for validation. The astrophysical maser positions and relative velocities (in this case, methanol masers at 6.7 GHz) were obtained from an establish catalogue.

Spectra are measured for different relative offsets from the nominal position of the maser. The order of the offsets is randomised to avoid systematic effects. A model of the spectral bandpass is fitted and subtracted, thus removing the bulk of the sky noise incurred from a variable atmosphere — even at low elevations and under poor conditions. A Gaussian is fitted to the remaining spectral power, from whence the positional error can be obtained.

And example of such an observation is shown in Figure 1. With positional errors, an all-sky pointing correction model can be determined using established linear algebra techniques. Although the application here is radio astronomy, the technique is relevant to any directional antenna system which has sidereal tracking capability. This may include, for instance, satellite ground stations or radar tracking antennas. Indeed, the MCA antenna used in the work presented here is an ex-satellite ground station, now redeployed to astronomical research and student training. By obtaining higher-accuracy antenna pointing, the overall sensitivity and quality of results is substantially improved.