[evlatests] Phase vs. Elevation fix

[Vivek Dhawan]

The fix (digital /4 dividers) which was put in on antennas 11
and 19 has been verified to work well in the field - limited by
weather, as always. So this fix should now propagate to the
rest of the EVLA. Assuming we do the worst offenders first,
these are:  ea2, 4, 9, 19, 21, 23, 24, 26.

Some more background before I forget it:

Once upon a time, O best beloved, there were evil phase changes
with elevation... The terse version of the bedtime story is that
the L305 makes 128 MHz and 512 MHz. The 512 is made from a 128 MHz
VCXO by a x4 multiplier. The 512 is phase locked after x4, the 128
is not. The L301 depends on the 512 phase, the L302 on the 128.
This is thought to be the cause of most of our woes. More details
are available from Jim Jackson or the LO engineers (** 1).

A digital divider with low phase noise, unavailable at the time
of the previous design, has now been installed in antennas 11 &
19, to derive the 128 from the locked 512. The L301 and L302 are
now more nearly ideal and identical. Antennas 11 & 19 show much
reduced sensitivity of phase vs. elevation.

Vivek.

(** 1) the phase noise of the 128 MHz is reportedly within spec.
Could engineering please clarify if this is a lab measurement by
NRAO or a claim from the spec sheet - my memory is already hazy
on this point.