A Possible "Use Case"? A Fiducial Observation Mode Described... S.T. Myers (NRAO Socorro) 6 June 2000 -------------------------------------------------------------------------- Observing Mode: Maximum bandwidth 16 GHz continuum observations at 30 - 46 GHz or 84 - 100 GHz, single polarization, double sideband. Used for: This might be expected to be the main mode for spectral analysis of (unpolarized) nonthermal sources. Both sidebands are used and the bands aligned consecutively (I'm not sure what the gaps between bands, especially between sidebands, will be). In this mode, each intersection XY correlates the sidebands U and L in one polarization (eg. RR, LL) with 256 lag channels for each. The channel resolution is 2 GHz/256 = 7.8 MHz. Note that since this channel resolution implies a velocity width of 78 km/s at 30 GHz and 28 km/s at 84 GHz, this mode would also be useful for searching for redshifted galaxies in CO! Details: Assuming the receivers have wideband horns whose illumination is set to accomdate the low end of the band, then the primary beam will have a width FWHM = 94' x ( 1 GHz / f ) for 12 meter antennas, giving 3.13' FWHM at f = 30 GHz and 1.12' FWHM at f = 84 GHz. If the source is large and we want to OTF scan at a rate of 0.5 degree/sec, then we are moving 10 beams/sec at 30 GHz and 27 beams/sec at 84 GHz, and thus need to sample at roughly 3 per beam or 30 Hz or 80 Hz respectively. This implies maximum accumulation times of 33 ms and 13 ms respectively. Note that at 30 GHz this scan speed allows us to use 32 ms as an accumulator time with no problem - 4 streams per quad (32 x 32 ant) at 160 ns would clear all 524288 results in 21 ns as shown above. On the other hand we see that for 0.5 deg/sec scanning at 84 GHz, even the minimum correlator dump time of 16 ms would be marginal, and the time to clear the data in 4 streams at 160 ns is even further off, though 8 streams per quad would get us clear in 10.5 ms. For a dump time of 16 ms (rate of 62.5 Hz), a scan rate of as much as 20.8 beams/sec at 3 samples/beam could be supported. Note that the ALMA spec for OTF in cross-correlation mode is only 0.05 deg/sec (3'/sec) anyway, which at 30 GHz translates to just under 1 beam/sec, and 2.7 beam/sec at 84 GHz, which are well under the dump limit. Sensitivity Issues: Note that our expected rms noise level at 95 GHz for 32 antennas is expected to be only 7.5 mJy at 30 GHz and 13 mJy at 84 GHz in a 7.8 MHz channel in 1 s (using the sensitivities given in the Butler & Wootten Proto-Memo). If, for example, at 84 GHz you are looking at CO(2-1) at redshift z=1.74, then the expected flux in 7.8 MHz (25 km/s) is about 0.76 mJy (using numbers from Memo 243) and it would take 295 sec to reach S/N of 1 and 78 minutes to reach S/N of 4). Assuming the normal cal cycle of 10 sec on-source, 1 sec slew, 1 sec on calibrator, it is unclear whether anything would be gained by OTF scanning over just integrating on-source for 10 seconds (with probably 1 second integrations). OTF scanning at 0.05 deg/sec is unlikely to be fast enough to help with gain variations, but should help with the atmosphere.