[evlatests] Deep Imaging of 3C345

[Rick Perley]

    Ken took C-band data on 3C345 a few days ago.  The usual setup:  3 
subbands, each with 128 MHz width, 1 MHz channel resolution, dual 
polarization.  1 second averaging. 

    I've reduced these data, and made a deep image.  Pretty good 
overall:  Dynamic range about 350,000:1 (peak to noise in distant 
places).  About a half dozen obvious background sources, mostly at the 
half-power beam point or beyond, are easily seen. 

    What is interesting are the image artifacts, which are seen near the 
various objects in the field.  The artifacts are all different -- some 
are like rays, some are like arcs (negative and positive), some are 
'crinkly'.  The depths/peaks in these features are at the level of up to 
1 mJy -- these are far above the noise.  There is nothing in any of them 
that tells me of an obvious origin.  The usual tricks were used to 
remove obviously bad data.  We're probably going to need more 
sophisticated tools to dig a little deeper. 

    Some general comments on the data:

    The usual (small) list of non-operating IFs and antennas were seen.

    Antenna 5 is oscillating in phase (by 60 degrees) back and forth, in 
both polarizations.  This looks like the problem we had on many 
antennas, which was corrected by changing some cable length... 
    Many 1 db gain changes were seen.  Ken tells me that this is 
expected, as the 'set and remember' mode (formerly known as 'set and 
forget') was not utilized. 
    Bandpass stability was worse than expected -- with frequency 
structure, typically of 10 -- 30 MHz scale, and amplitude of order 1%, 
slowly changing over the 3 hour interval.  Not all antennas showed 
this.  These 'structures' look to be calibratable (that's good), but I'd 
like to understand their origin. 

    The same baselines which have shown wobbles in recent databases are 
wobbling here also, with the same characteristics.  There are no other 
wobbly baselines (that are detectable with 1 second dumps).