MMA Imaging

Mosaicing

• Pointing: Because the emission spans beyond a single primary beam in mosaicing, small antenna pointing errors can have a large effect on the observed flux of a feature which lies near the half power point of the beam. Hence, while pointing of a tenth of a beam width is adequate for interferometers designed for single pointing observations or for single dishes, a mosaicing instrument must have better pointing. The MMA pointing spec is one arcsecond, or about 1/25th of the beam width at 300 GHz. This pointing accuracy will permit mosaics of about 1000:1 dynamic range. Many mosaicing projects will not require such high dynamic range. The effects of the pointing errors on mosaic images scale with frequency.
• Surface Accuracy: Surface errors will scatter radiation into the primary beam sidelobes, and unmodeled primary beam sidelobe structure will limit the quality of mosaic images. While surface accuracy of 1/16th of a wavelength only degrades the dish efficiency by a factor of 2 from Ruze losses, 1000:1 dynamic range mosaics will require surface accuracies of about 1/40th of a wavelength. The effects of the surface errors on mosaic images will scale like frequency squared, so very high frequency mosaics will be limited by surface errors rather than by pointing errors.
• Total Power Measurements: In order to correctly image sources larger than the primary beam, total power observations will be required. In the homogeneous array design, the same elements measure both total power and interferometric data. In many spectral line observations, total power can be measured simultaneously with the visibilities. Using common electronics for the two kinds of data will aid in the cross-calibration of total power and interferometric data. Since continuum total power data requires special consideration for removing atmospheric emission, simultaneous total power and interferometric observations will not be possible. Exactly how many antennas need to be measuring total power is under investigation. However, since total power observations will be scientifically useful in their own right, and since it is clear we need significantly more than order one antenna, we currently state that all MMA antennas should have total power capabilities.
• Dish Illumination Taper: It was initially thought that a uniform dish illumination was required for homogeneous array mosaicing in order to sufficiently cover the gap in effective Fourier plane coverage between the highest spatial frequencies of the total power measurements and the lowest spatial frequencies of the interferometric data. However, uniform illumination requires a lens or a shaped subreflector, which would limit the field of view for beam switching or future focal plane arrays. New antenna illumination results indicate that the MMA can probably still make good mosaic images with a tapered illumination pattern. With tapered illumination, it is even more important to place the antennas as close together as possible.
• Getting Very Short Spacings: The homogeneous array concept requires that the antennas be fairly close together (ie, 1.3 times the dish diameter for zenith observations) to be able to measure spatial frequencies in the range of the dish diameter (see Cornwell, 1988, A&A 202, 316). However, the antennas can actually smack into each other if the separation is less than about 1.5 dish diameters (depending upon the design). He find that if we mechanically restrict the antennas to point above 30 degrees elevation, the antennas can safely be placed 1.3 dish diameters apart for good zenith mosiacing. Low elevation mosaicing can be performed in a somewhat sparser array with no elevation restrictions and a minimum separation of 1.5 dish diameters, as geometrical foreshortening permits the required short baselines to be measured.
• How Many Antennas Do We Need? This issue has been addressed from several points of view (see the MMA Proposal), but mosaicing pushes us towards many antennas. Since mosaiced objects will generally have structure everywhere in the primary beam, good quality images require complete Fourier coverage for each pointing. Since mosaics will often include hundreds of interferometric pointings, we cannot wait for earth rotation to fill out the Fourier plane coverage, but must get complete Fourier plane coverage in each snapshot. With many fewer antennas, the mosaicing capability of the MMA would be severely cut back, requiring long integrations on each pointing.

  Many of the scientific problems the MMA wishes to solve will be sensitivity limited, so we need all the collecting area we can get. Why could we not build fewer antennas and just make them bigger? The larger dishes will have smaller primary beams, and since mosaic image quality degrades linearly with the magnitude of the pointing error as a fraction of the primary beam, the mosaic performance would be drastically curtailed.

• Mosaicing Configurations: There are many issues facing the mosaicing configurations. We need to be aware of shadowing among adjacent antennas, circularity of the beam, maintaining enough very short baselines, and keeping the instantaneous Fourier plane coverage essentially complete. Since each of these issues is highly elevation dependent, we will require three or four different compact arrays to permit mosaicing over the entire declination range observable from Mauna Kea or the Chilean site.
• Why Mosaic? Why not a big Single Dish? Traditionally, large, low brightness objects have been observed with single dishes and small, high brightness objects have been observed with interferometers. The mosaicing array will be about 70~m across. Why not build a 70~m single dish to observe these big sources? Mosaicing is faster than single dish observations, mainly because of the multiple synthesized beams which can be formed within each primary beam.

Very High Frequency Observations