[evlatests] Stellar high-frequency data

[Rick Perley]

    I ran a test last night, in preparation for the observations of the 
'pretty-picture' of Cyg A, for the eventual commissioning. 

    In this test, I set up a trial run, 2 hours, cycling through K, Ka 
and Q bands, applying referenced pointing, and observing the 
flux/polarization calibrator 3C286, the local calibrator 2007+4029, and 
various positions along Cyg A (since its extend of 2 arcminutes makes it 
larger than the antenna primary beam at Q-band).   These observations 
were taken in OSRO1 mode. 

    The short 'executive summary' is that the data are marvelous.  Two 
2-minute snapshots on Cyg A at K band has provided images with over 
200:1 dynamic range, using only a few central channels. 
    I shorted the durations for each observation to 2 minutes -- in no 
case did more than the initial 15 seconds of any of these had to be 
flagged.  So there appear to be no 'setup' problems. 

    But things are not perfect.  I have analyzed only the K-band data so 
far (taken at 19 and 25 GHz).  Below I list the issues -- some of which 
are being addressed already.

    1) Due to an oversight, the 'BD' frequencies were not observed in 
'set and remember' mode.  This meant the table values were used, which 
caused effective loss of fringes (very weak in general)  on a number of 
antennas.  I'm using  frequency tunings near the edges of the bands, so 
there is no surprise in the result. 

    2) Antenna 20 is clearly not pointing correctly, with a cyclic 
response of a few minutes.  I think this is known to be a subreflector 
or encoder problem -- it would be good to hear this is being addressed 
as it affects all bands (even L-band!). 

    3) Antenna 22 was stowed due to an issue with IF 'A'. 

    4) 5.5% of the visibilities were flagged as integer 0 -- all of 
these are near the beginning of the scans. 

    5) There were a very few 'failures to tune'. 

    All in all, a very encouraging test.