I pulled the data out of the working directory of fbins

 from             to               flux  b180  phdiff
 2016-02-28_09:30 2016-02-28_10:30 0.08  99.2  114.9
 2016-02-28_09:45 2016-02-28_10:45 0.08 109.9  134.3
 2016-02-28_10:00 2016-02-28_11:00 0.09 124.8  138.3
 2016-02-28_10:15 2016-02-28_11:15 0.08 142.8  138.5
 2016-02-28_10:30 2016-02-28_11:30 0.10 125.3  125.6
 2016-02-28_10:45 2016-02-28_11:45 0.07 115.4  108.2
 2016-02-28_11:00 2016-02-28_12:00 0.06  74.5   65.1
 2016-02-28_11:15 2016-02-28_12:15 0.04  73.1   51.9

Flux density is fT, b180 the apparent bearing, and phdiff is
the phase angle between the E/W and N/S loops which gives an
idea of the polarisation distortion.  For signals in the South-
East compass quadrant the phdiff should nominally be 180 the
way my loops are wired.

The above wouldn't show up any phase glitches on the signal
because only the loop phase difference goes into the goniometry
calculation.  Below is the absolute phase lead in 1 hour
averages captured along bearing 120.

 from             phase
 2016-02-28_09:30 -73.9
 2016-02-28_09:45 -37.4
 2016-02-28_10:00 -22.4
 2016-02-28_10:15 -14.8
 2016-02-28_10:30 -30.8
 2016-02-28_10:45 -32.0
 2016-02-28_11:00 -59.0

That looks pretty steady, no problem for a 3 hour coherent
BPSK message.

--
Paul Nicholson
--