[evlatests] More on 'an extraordinary event'

[Rick Perley]

    At the EVLA test meeting last week, I showed some remarkable 
'fringes', which affected only the RCP correlations for about two 
minutes during the long L-band observation on 3C147.   All four subbands 
were affected, but in distinctly different ways.  The overall effect was 
to decrease the correlations.  Some subbands showed rapid oscillations 
within the two minute period, one subband showed no oscillations, and 
only a smooth decrease. 

    I've looked more carefully at this subband (#3 in the database, 
centered at 1820 MHz), and find some more remarkable characteristics.  
The effect varies greatly by antenna, with no spatial relationships 
apparent.  The following table shows the magnitude of the effect 
(Vis_min/Vis_normal), the antenna number, and its location. 

Ant  Ratio Location
-----------------
25   1.13   N2
3     1.08   E9
24   1.07   W5
27   1.06   E3
5     1.06   W8
19   1.05   W4
8     1.05   N1
15   1.04   W6
4     1.02   W1
28   1.02   E2
2     1.01   E2
9     1.00   E6
-------------------

The effect is completely invisible on antenna 9, and barely detectable 
on antenna 2.  And as noted before, there is utterly no effect in LCP. 

There were some speculations that the effect might be due to compression 
from a passing satellite.  I don't believe this for two reasons -- (1) 
The satellite's polarization would have to perfectly match the (variable 
and imperfect) polarizations of all the EVLA antennas, and (2) The 
satellite would have to have a truly remarkable forward gain capability 
in order to illuminate some EVLA antennas, and not others located a few 
meters away. 

Are there any other ideas for this?