Hi Rick. It looks to me like the biggest problem is that
the mount performs poorly whenever the direction of Dec guiding is
reversed. Here is one of many examples:

You can see that a sizable Dec excursion begins near the left
and PHD2 issues 9 consecutive (south) guide pulses trying to correct
it. The mount shows no reaction during this time but on the 10th guide
command, it abruply over-shoots and thus triggers another series of guide pulses
in the opposite (north) direction. This pattern happens repeatedly and the
Dec guiding never stabilizes for any significant length of time. This
behavior can come from multiple interacting problems: 1) a backlash period when
the drive system isn't fully engaged and isn't turning the Dec axis and 2) some
form of resistance (stiction) on the axis that prevents it from moving
until the resistance is overcome. Stiction can be present because the axis
doesn't rotate completely smoothly, or because there is a significant weight
imbalance in Dec or because there are cables or other external components that
are interfering - just to name a few reasons. If the gears are fully
engaged but the axis can't rotate, the
motor is essentially pumping energy into the system that
is then stored elsewhere - often by deforming or compressing other
parts of the mechanical system. Once the static resistance is overcome,
that stored energy is released and the axis then rotates more than the desired
amount - and you get over-shoots like this. These are mechanical problems
that need mechanical correction, they're not something that can be "fixed" by
guiding. You can, however, improve the responsiveness of the Dec system by
increasing the guide speed setting in the mount. You're currently guiding
at 0.5x sidereal so I'd suggest bumping that up to 1.0x sidereal or slightly
less. That will improve responsiveness for any backlash component and may
improve the ability of the system to push through small amounts of static
resistance. You should also be sure the OTA is well-balanced in Dec for
all pointing positions and that the cables are well-routed so they can't tug on
any of the gear and interfere with its movement.
There is one other significant limitation to your overall
system - your reliance on a tiny guide scope arrangement and the
associated huge guider image scale of 6.45 arc-sec/px. This puts you in a
precarious position at best. If the guide camera sensor moves by as
little as 4 microns, you will see a very large guide star excursion of 6.5
arc-sec and probably a ruined image. 4 microns is less than 10% of the
thickness of a human hair - so ask yourself if that little guider assembly is
really rigid enough to not move that much as it traverses a 180 degree arc
across the sky. In terms of operational procedures, you should also be
using multi-star guiding. That means setting the option to do so and
letting PHD2 select the guide stars based on the min-HFD and saturation values
you specify - details are in the manual.
Good luck,
Bruce