[evlatests] Modcomp-free: Line frequency check.

[Vivek Dhawan]

Goal: to check if the EVLA and VLA are singing the same
          song, in the correct pitch.

Observation mode: stokes RR and LL, 2 IFs, 15chan per IF.
          Fixed frequency, preset for Doppler.
          BW was 12.5 MHz at K and Q band, intended to give
          some sensitivity to delay slope while showing the
          lines as well. L band I intended to use 0.78MHz
          but mistakenly left in the same BW - very coarse
          for OH.

Result: Good(ish).  This dataset was on Sunday afternoon,
         riddled with the non-closing effects on many baselines.
         Enough data survived to show clean fringes between (some)
         antennas in every subset V-V, E-V, E-E, on both line and
         continuum, at K, Q, L bands.

Next: Repeat test with better frequency resolution, after the
       dust settles on CMP errors, baseline errors, modcomp
       retirement etc. Is Doppler expected to work then?

A bit more detail, in time order:

       K band, W3OH H2O maser alternating with continuum calibrator,
       2 min each, 24 min total. EVLA started on time, VLA late by
       6-7 minutes (more than 2-3 scans on some antennas).

       L band: W3OH, OH maser 1665MHz, same calibrator, 16 minutes
       total. This observation block started right on time (BD IFs).

       Q band: Orion SiO maser + cal, 20min total. This block had
       a funny start: the first 8 minutes were labelled with a
       single source name, and only the last 90sec of the scan had
       fringes on that source. There was a long slew at the start,
       but the 2-min source cycle seems to have been ignored while
       the antennas were slewing. I should investigate this, when
       time permits; or just get more data.

V.