[evlatests] P-Band cross-polarization

[Rick Perley]

     A recent test has allowed a proper calibration of the 'leakages' 
for P-band.

     We are all fairly sure that the dominant term in this comes from 
misorientation of the dipoles.  The cross-polarization solutions are all 
very constant in frequency, with little frequency structure -- this is 
consistent with a rotation origin.

     The solutions are made with ea18 as the reference.  So the 
resulting cross-polarizations reflect (mostly) the misorientation of the 
dipoles w.r.t. this antenna.

     I find there are two 'families' -- about 2/3 of the antennas have 
cross-polarizations of 3% or less.  This is very good.  The other 1/3 
are at the 8 -- 12% level.  This corresponds to a rotation of about 5 
degrees -- not much, but noticeable.

     The 'high polarization/rotation' antennas are:

     ea01, 4, 8, 9, 12, 20, 25, 26 and 28.

     The current (AIPS) software seems to correctly remove these 
leakages from the cross-polarized correlations.