[evlatests] S-band polarizers

[rperley]

Eric and I are putting together an EVLA memo on the state of the EVLA 
polarizers from L through Ku bands.  The data come from the 'receiver 
rotation' method to determine absolute D-terms.

As I work through the bands, some interesting trends, and differences 
have been noted.  While the L-band polarizers, as a function of 
frequency, are very similar amongst all antennas (varying from 2 to 6% 
voltage cross-coupling), the situation at S-band is quite different.

At S-band, there are three clear 'families':  Convex, Concave, and Flat.

Convex antennas (1 2 4 6 9 11 13 1 20 and 28) have lower polarization at 
the frequency extreme, and higher in the middle.  They are all very 
similar in shape (with frequency), and good polarization -- 2 to 6%.

Flat antennas (3 12 14 15 21 23 and 26) are just that.  No trends with 
frequency, at about 4%.  Frequency undulations are similar to the convex 
group.

Concave antennas (5 7 10 16 19 22 24 27) are higher at the extrema, and 
lower in the middle (3 GHz).  Their cross polarizations are higher than 
members of either of the other two groups, typically 10% at the low end, 
and 8% at the high frequency end.  .  Ea07 is is much the worst -- over 
14% cross-polarization at 2 GHz, and 6 -- 10% near 3 GHz.

The first two groups are sufficiently similar that we should probably 
not consider them separate.  But the last group is clearly different 
(and worse) than the others.  Any ideas on why these are 'special'?