2 Observations and data reduction

The April 1990 session was a global 100 GHz mm-VLBI session involving the following telescopes: Hat Creek (U.S.A.), Kitt Peak (U.S.A.), Nobeyama (Japan), Onsala (Sweden), Owens Valley (U.S.A.), and SEST (Chile). The data was recorded on MarkIII-tapes using 14 IF-channels each with a 4 MHz bandwidth, and later correlated at the Haystack correlator. The raw output from the correlator were read into the AIPS package using specially written software (Bååth et al. 1992). The fringe finding and the necessary special post-processing to handle the interferometric mm-VLBI data are described by Bååth et al. (1992), in which a full description is given of the initial data processing stages done at the correlator. The 1990 and 1993 data were processed in exactly the same way. The amplitude calibration was done by applying the measured source flux density at each station, correcting for gain and atmospheric effects. We believe that the amplitudes are correct to within 30%. The discrepancy between the measured flux density in the maps and the estimated flux density at the shortest baselines (Table2) is attributed to the generally poor uv-coverage. The lack of short uv-spacings not only causes additional noise in the maps, but can lead to spurious features in the maps which may be misidentified as components.

The later hybrid mapping and uv data model fitting were done in parallel in the AIPS and DIFMAP packages in order to have two different approaches to making the maps from the UV data and thus minimize the effects of sparse uv-sampling and mapping algorithms. We have been very careful to constrain the placement of source components by careful use of windowing techniques, to minimize the appearance of spurious features discussed previously. If the two methods yielded markedly different hybrid maps then both maps were discarded and we present only a model fit to the UV data.

The 1993 April observations were a global mm-VLBI experiment similar to the previous epoch but with the receivers tuned to 86GHz. Telescopes participating in addition to those at the 1990 April session were: Effelsberg (Germany), Pico Veleta (Spain), and Quabbin (U.S.A.). The fringe finding, post-processing, and hybrid mapping followed the same procedures as for the 1990 April session. Information on the participating antennas can be found in Table1.

  

Table 1: Telescope information. D is the diameter in meters, $\mu$ is the sensitivity in Jy/K, $T_{\rm sys}$ the single-sideband system equivalent noise in Kelvin referred to outside the atmosphere, and the final column gives the sessions in which the telescope participated

The following sources yielded no fringes in either the 1990 April or the 1993 April experiments: 0016+731, 0133+476, 0212+735, 0235+164, 0234+285, 0300+471, 0458-020, 0528+134, 0642+449, 0716+714, 0748+126, 1012+232, 1044+719, 1116+128, 1156+295, 1308+326, 1334-127, 1354+196, 1413+135, 1418+546, 1611+343, 1633+382, 1637+574, 1642+690, 1716+686, 1749+096, 1802+784, 1803+784, 1823+568, 1928+738, 1954+513, 2005+403, 2007+777, 2021+317, 2201+315, 2234+282, 3C286, 3C274, 3C371, 3C380, 4C39.25, 4C67.05, 4C71.07, Centaurus A, Cygnus A, Cygnus X-3, DA193, NRAO140, NRAO150, NRAO512, NRAOO530, OA129, and Saggitarius A. Since the observations were limited to only one or a few scans it is not surprising that later observations with more scans have found fringes to some of these sources (Krichbaum et al. 1996a; Alberdi et al. 1997; Krichbaum et al. 1994; Standke et al. 1994; Schalinski et al. 1994).