[evlatests] Mysterious RFI at L-band ...

[Rick Perley]

    Perusal of PDif data leads to interesting curiosities ...

    From the 'stability run' taken last Wednesday night, I noted that 
the L-Band 'PDif' values dropped by a relatively small amount (typically 
a few percent), in RCP only, for one scan only, for both subbands (i.e., 
for IFs A and B). 
    The program 'SPFLG' shows that in that particular scan, a moderately 
strong birdie was present at the odd frequency of 1478 MHz -- not a 
multiple of the usual suspects.  The 'birdie' is completely absent in 
LCP -- less than 1% of the RCP value, which makes it really hard to 
explain this as an external source (neither the polarizers nor the 
antenna itself is that purely polarized!).  It is seen on all baselines 
except the longest -- likely due to fringe winding.  The birdie is seen 
in the cross-hands, at a level about 1/10 of that in the RR correlation 
-- notably, it is seen both in RL and LR, and is generally stronger in 
LR and RL.  
    The 'birdie' is barely visible in the autocorrelations, if at all.  
But the 'PSum' monitor changes dramatically in some antennas (usually 
there is less total power in that particular scan), and hardly changes 
at all in others.  Most notably, the PDif and PSum effects are seen 
equally in the 2nd IF (IF 'B'), which tells us that the underlying cause 
is common to the entire 1 GHz spectrum, and is not limited to the 
particular subband where the 'birdie' is visible. 
    Unfortunately,  I utilized the old-fashioned 'OSRO' mode to save 
disk space, so I have no information on what was *really* going on in 
the passband, outside of the two 128 MHz-wide windows. 

    It's rough out there at L-band!