[evlatests] One problem gone, another remains

[Rick Perley]

    Michael took an hour of data at L-band on our favorite northern 
source, split between the calibrator and a 'blank' field a few degrees 
away.  He had pity on me, and set things up so only 256 channels per 
sub-band, 1 second averaging.  The four subbands were set at 1308, 1436, 
1820 and 1948 MHz.  The first and fourth of these had lots of RFI 
(particularly subband 1, which has HUMONGOUS RFI near 1250 MHz...)  
Subband 4 is littered with cellphone transmission, according to Dan M. 

    I concentrated in subbands 2 and 3, which are nearly perfectly 
'clean'. 

    These data look very nice indeed.  There is no sign of these 
peculiar sinusoidal oscillations on any baseline. The phase and 
amplitude stability is superb.  Bandpass solutions are fine.  Everything 
looks normal. 

    But the maps are terrible.  After the standard self-calibration on 
the calibrator, the field containing this calibrator has no background 
sources at all -- just the calibrator shows up.  The rms noise is 1/10 
of the peak background source.  What can be wrong here?
    The blank field does show 3 or 4 objects -- but the appearance of 
the map just isn't right.  Furthermore, the peak background source is 
only 7 mJy -- far too low for a randomly selected field at this band.  
Both fields show the same 'lumpy bumpy' characteristics we've noted 
before. 
   
    Michael informs me he used the remaining antennas (non-WIDAR) to 
observe simultaneously the field with the VLA correlator.  I'll reduce 
these data now, to see what that gets us.