[evlatests] Slow spins

[Rick Perley]

    I've reviewed the spectral line data.  These add nothing
but confusion (to me, at least) to the saga of the phase spin. 

    First:  The spectral bandpasses for 7 of the 8 IFs (at
X-band, over 12.5 MHz BW) show a 40 degree phase
wind.  For one IF (14B), it is 100 degrees, corresponding
to a 22 nsec delay error. 
    For the 7 IFs with ~40 degrees phase winds, the slope
is not seen in the overall 50 MHz BP.  This is thus the
residual after the digital filter correction.
    For 14B, the slope is seen in the 50 MHz BP, and
Ken had already found and removed it. 

    Moving on the phase spins and amplitude drops:

    Phase spins were seen in one scan of 3C48, and
two scans of 3C84.  The rates are very slow compared to
the continuum:  about 1.5 degrees per second.  The shape
is parabolic, in the sense that the phase rate starts at zero
at the beginning, and continuously increases throughout the
scan.  This is the same behavior as in Thursday night's
data. 

    Phases are not connecting well between scans, for those
scans without the phase spin.  Offsets of ~50 degrees are
seen. 

    There is a peculiar secular change in antenna amplitude
gain.  for antenna 16, the rise is small -- about 4% in amplitude
over 2.5 hours.  For antenna 14, it is large -- rising by 40%
over this same time range.  This cannot be due to pointing. 
(Note -- this is antenna gain correction factor, so the fringe
amplitude loss is given by the inverse). 

    The familiar dropouts are seen  -- but *only* in the
scans in which there is a phase spin.  This exclusive
association is almost seen in the continuum data:  In that
dataset, dropouts are seen only on phase spinning scans
for 3C84.  A single 3C48 scan had dropouts, but as noted
before, there was no phase wind. 

    I hope somebody can see a plausible cause here -- for
I cannot.