[evlatests] More from L-band data

[Rick Perley]

    Some more 'top-level' observations from the L-band data taken 
yesterday afternoon:

    The number of 'zero record' visibilities (not associated with the 
antenna issues identified earlier) is quite small -- CLIP reports 0.003% 
of the visibilities lay outside the range of 1e6 to 1e-6. 

    Antenna 24 in RCP is undergoing 1dB power step changes every 30 
seconds:  at 22:58:40, 22:59:10, 22:59:40, ...

    There are big wobbles on some baselines, most easily visible in 
phase.  The wobbles appear to be much stronger in the first IF (1180 MHz 
center) than in the third (1436).  Baselines particularly affected:

    1 x 5   (W9 x W8) -- in RCP only.  SPFLG shows frequency-dependent 
phase-time structure -- like broad-band RFI beating with the 
astronomical signal.  LCP is clean. 
    2 x 18 (E2 x N9).  Both polarizations, amplitude and phase. 
    All other easily identified wobbles (IF3 ) are only seen easily in 
phase.  There are many baselines -- for some of them, the phase plot 
shows 'chatter' -- the periodicity is ~2 seconds (averaging time 1 
sec).  Baselines most obvious are 7 x 21, 8 x 24, 4 x 8, 1 x 25, 2 x 8.  
Both polarizations show apparently identical effects.  A proper FT would 
likely find more periodicities.  It would be good to get a faster dump ...

    Other than this, the visibility plots (amplitude vs U-V distance) 
are very nice, showing the 3C84 halo quite clearly.  I'm fairly sure 
that eliminating the large delay will return us to a closure regime like 
we had before -- residual closures will be due to 2nd order polarizer 
leakages, which we can (somewhat crudely) fix in BLCAL.