[evlatests] the P band polarization saga

[Ken Sowinski]

I report on work done by Paul and Dan to sort out the polarization
mysteries at P band.  The intent is to clearly establish the relation
between RCP and LCP signal paths in the antenna with the dipole 
orientation considered, as H and V, or X and Y.  The former is the
natural physical interpretation, H is provided by the dipole element
parallel to the horizon when the antenna is pointed at the horizon;
the latter is the convention established by the IAU in which the
element providing X is the one which points to the North Celestial
Pole.  The (H, V) system is natural and for an alt-az antenna and is
easy for technical personal to relate to the hardware.  The (X, Y)
system is appropriate for polarization post processing.

The method was to monitor changes in power measurements along the 
signal path in the T302, T304 and samplers as cables between the
dipoles and the T302 were disconnected, one at a time. all the tests
were carried out at ea22.

The wiring  scheme, as designed,  was to connect  the H dipole to the
LCP signal path in the antenna, and the V dipole to the RCP path.
Thus:

    RCP path =~ V =~ X
    LCP path =~ H =~ Y

We found that when H dipole was disconnected from the receiver
that power was decreased in the RCP signal path.  Similarly
when the receiver H output was disconnected.  And again, when the
V dipole was disconneceted, the power in the LCP signal path decreased.
Everything was recnnected as normally and the input labelled "H"
at the T301 was disconnected.  Now the power in the LCP path decreased,
as expected.  Clearly there is a swap between the receiver output
connector and the T301 input connecter.  Paul and Dan laboriously
traced cables after this test and found that the color coded tags at
the ends one pair of cables were mis-colored, but connected as if they
were correct creating the signal swap we observed.  It is likely that 
this error was made in many antennas.

So, what we really have is:

    RCP path =~ H =~ Y
    LCP path =~ V =~ X

It is proposed that we simply define this as the connectivity and
do not try to make the system agree with previous documentation by
swapping cables in all the antennas.

The final link in the chain is labelling the output of the correlator
in the BDF.  The claim is made that if in the VCI document for a P band
observation we replace R with Y and L with X, the CBE will respect
those labels and order all the polarization products  properly in the 
BDF.  To achieve this a change is needed to model2script.