[evlatests] Switched Power Characteristics at L, S, C, and X Bands

[Rick Perley]

    I've reviewed all the switched power tables from the latest Herc A 
Demo Project.  Quite a zoo, but most of the beasts are in fact tame... 

    Much of what needs to be revealed is best done through plots, as 
attempts at description will only tire the gentle reader.  But some of 
these characteristics are easily summarized:

    A)  Switched Power = 0, or very low (not noise!)

    Some antennas show whole scans during which the switched power is 
nearly zero.  But not integer zero, and no noise -- the 'PDif' values 
are typically 1 microcount, the PSum values are in tens of millicounts.  
In all cases, the visibilities during these events are also zero, (but 
not integer zero), with random phase.  The opposite polarizations always 
share this characteristic, but the opposite IF pair have a curious 
relation.  Read on ...
    All four bands showed this problem, but it is by far most prevalent 
at C-band.  In all cases, the behavior started and ended with a band 
change.  The relation between the AC and BD IFs is best illustrated by 
the events at C-Band:
    Six antennas, 12, 15, 16, 19, 20, and 22 all showed this problem.  
For all of these, the characteristics described above applied to the A 
and C IFs.  A total of 7 'tunings' were affected -- one each for 
antennas 12, 16, 19, 20, 22 and two times for antenna 15.  For these, 
the B and D IFs behaved as followed:
              In antenna 15, the B and D IFs also showed 'zero' switched 
power and 'zero' visibility.
              In antennas 19 and 20, the B and D IFs were not affected 
at all.
              In antennas 12, 16, and 22, the B and D IFs showed a 
reduction of 30% in the switched power values for the tuning of 7 -- 8 
GHz, but an *increase* of the same factor for the 5 -- 6 GHz tuning.  
For both tunings, the visibilities were not affected. 
    At the other bands:
             In X band, antennas 9, 19, and 23 showed the zero effect, 
only on the B+D IFs.  For 9 and 19, only the first (initial) scan was 
affected.  Antenna 23 had the problem three separate times. 
             In S-band, only antenna 6 showed the problem, and only in 
the B and D IFs.  The A and C IFs showed an increased Psum and Pdif.
             In L-band, only antenna 10 showed the problem, one time, on 
IFs A and C.  IFs B and D showed an increased Psum and PDif at the same 
time. 

    B) In general, C-band shows by far the most stable switched power.  
The other bands, particularly X and S, show all sorts of 'step' changes 
(corresponding to band changes), typically 2 -- 7%, or weird 
instabilities, where the power monitors are changing by tens of percent 
over a few seconds to a minute.  It's hard to blame RFI for these, as 
the other antennas don't show the same change at that same time, nor do 
the opposing polarizations.  L-band is generally quite messy, but it's 
easy to explain this in most cases by the external RFI environment. 

    C) At L-band, we still have no, or very little switched power on:

       Antenna 9, in both polarizations.  If memory serves me right, 
this antenna has *never* had switched power working! 
       Antenna 6, in LCP (IFs C and D):  the PDif is far too low, but is 
not zero.  The corresponding visibilities are also far too low, so it 
looks like the power into the sampler is far too low. 
       Antenna 7, in RCP.  Exactly the same as antenna 6 in LCP. 

       Furthermore, antenna 18 at L-band, in LCP, is showing two 
different PDif values, differing by a factor of about 2.  These two 
states exist throughout the run.  Meanwhile, in RCP, there are also two 
states, alternating back and forth, but now the power difference is only 
about 25%. 

    D) At S-band, some other odd effects are noted:

            Antenna 6 in RCP, and antenna 27 in LCP, are both returning 
what appears to be reasonable values of PSum and PDif, but in fact they 
are not.  The derived Tsys shows no variations at all -- all other 
antennas, at all IFs, and all bands, can easily detect the change in 
system temperature as we move from source to source, or over time as the 
antennas change elevation.  This predictable variation is not seen in 
the two antenna-IFs nominated above, where there are no spatial/temporal 
changes seen at all.  That can't be right ...
       



   
       -- Antennas 12, 15, 16, 19, 20, and 22 all had periods of time 
where the PSum and PDif values were essentially zero.  (Not quite -- 
PDifs were generally ~ 1 microcount, and PSums were near 30 
millicounts).  I checked the recorded visibilities at these times -- 
they all were zero at the same times for those antennas.   Both 
polarizations within an IF are affected at the same time, but the two 
IFs are independent.