[evlatests] Correlator Amplitude Jumps, and Antenna 5

[Rick Perley]

    Two more issues need to be mentioned -- both of which were present 
in the Thursday tests, and are present in tests taken late Friday afternoon.

    1)  Antenna 5 is misbehaving on two of its IFs:

             5C is showing a distinct square wave in its correlation 
coefficients.  The amplitude is about 20%, and the period is about 1.1 
minutes. 
             5B is showing the infamous 'shredded levels' problem -- 5 
different amplitude states, with no coherence between successive 
observations. 
             Both look like DTS issues. 

    2) Correlator amplitude jumps.  When in spectral line mode, we are 
seeing a very small (less than .1%) of the correlations with kilojansky 
amplitudes.  These always affect all channels of an individual spectrum, 
and occur on single baselines for a single record.  There is no 
correlation in the occurrence amongst different polarization products.  
(That is, a bad visibility on an RR baseline is not accompanied by a bad 
visibility on any other correlation product, nor on any other baseline). 
       Ken rebooted the system controller this afternoon, hoping this 
might have a beneficial effect.  It does not. 

    In addition, it seems that on IF 'C', when in 'PA' mode, another, 
lower-level correlator amplitude problem occurs.  Here, it seems that an 
antenna gives a modestly higher visibility amplitude against a few, but 
not all, of the other antennas -- again for a single visibility record.  
This is also a very uncommon phenomenon, but easily recognizable. 

    This problem is completely absent in continuum mode.