[evlatests] Nothing but noise!

[Rick Perley]

    I used 4.5 hours of free 'Dynamic Time' last night to do a deep 
integration at L-band on a random background field. 

    Setup:     BW = 12.5 MHz.  Mode = '4'.  This gives 16 channels, each 
of 781 kHz, for each of the four parallel hand correlations.   
Integration time was set to 3.33 seconds. 
    Calibration:  I observed 3C286 once every 15 minutes, for one minute. 

    Bandpass solutions (with Hanning smoothing on, to reduce the Gibb's 
oscillations) were done for every calibrator observation. 

    Straightforward gain calibration (2-point) was then done.  The usual 
flagging was done to remove bad antennas and visibilites.  In general, 
the data were of excellent quality.  27 antennas worked (but not all of 
them well). 

    The randomly chosen field was at:  RA = 13 30 00, Dec = 28 00 00.  
(The strong sources 3C286 and 3C287 were each about 2.5 degrees away).  
Afterwards, I checked the NVSS map archive, to find that I had put the 
beam in a remarkable 'hole' -- virtually no sources within the primary 
beam!  (I checked because my first image returned rather fewer sources 
than I expected to see!) 

    Imaging Results:

    I made many images, with single channels, and blocks of channels, 
with one IF and both IFs,  of selected fields, and of the whole primary 
beam (this takes a long time -- my machine is still laboring away ...).  
But one result is very clear: 

    *********  There are no artifacts of any kind, at any place, in any 
channel, in any images  ********

    The full-beam, both-IF, 13 channel (channels 2 through 14) image is 
showing dozens of sources at the few-mJy level (and less), with rms 
noise levels below 100 microJy.  The sources are clean and sharp.  Tests 
with EVLA only and VLA only subsets show the same sky, with noise levels 
more or less in inverse proportion to the number of antennas available. 

    All looks completely normal to me. 

    Specific EVLA problems (and there are a few!) will be described in a 
subsequent report.