It is far better to have “solidly locked/bolted” tripod parts, and then compensate for any “minor tilt” in the mount set upon it, to get polar aligned using PemPro at the equator. You can even plant the tripod, slightly angled along the side of a hill and not even adjust its legs, so long as you polar align the mount on a tilted platform using that program, for example. Marginally tight screws and slightly loose tripod leg positions can introduce “sag” resulting in minor deviations, during operation.
Just to be clear ... RAPAS isn’t actually centering “right on” Polaris.
That isn’t the centre of rotation, but pretty darned close. RAPAS scale, provides a proper offset of the actual Pole from Polaris. So, judging it on a screen graph, from different Park positions may not quite coincide, and it is still just a good approximation for polar alignment.
The other problem about relying on pointing accuracy based on Polaris sightings, is that it is just a problem of the math involved in converting between celestial and terrestrial coordinate systems (HA-DEC and ALT-AZ), The equations used to convert from sky coordinates to those required to move mount motors, involve not only complex math, but also the TRIG Functions involved (Sines, Tangents, etc. and their inverses), become far less accurate near quadrature angles (zero or 90 degrees, where the results can become indeterminant). Then function approximations are then used, so that the calculated end results seem close enough, for practicality, near either Pole coordinate, and those results are only “fairly adequate”. Worst case is converting coordinates very close to the Alt/Az Pole or the HA/DEC Pole positions. Mathematical accuracy of coordinate conversions, tracking, etc., degrades the higher up you go on either coordinate sphere.
Just as an example for an academic exercise, look up the coordinate conversion equations, and calculate using pencil & paper - since electronic devices/calculators are already trig-biased - to convert Alt=90 and AZ=0 degrees into corresponding HA and DEC – and vice versa. It gets really horrid in either system, north of about 86 degrees (Alt or DEC). Been there, done that for a radio telescope tracking design, many decades ago.
That is why the ideal Polar Alignment is based on “star drift”, at the benign positions of the celestial equator and prime meridian - using PemPro, rather than eyeballing it with something even as excellent as a RAPAS, or maybe just a PoleMaster.
However, all that being said, closely examine just how tight everything is on your Losmandy HD platform, during setup – for a permanent site, or repeatedly on field trips.
You may also want to snug up the two small set screws a bit, (in its rotational channel) on the AP “Rotating pier adapter” - if your mount uses that one. Their factory set looseness provides buttery smooth AZ adjustment, but too loose, I worry that the platform can wobble by a miniscule amount, if the scope weight is unbalanced. But the same would be true even for a massive permanent pier as well.
Of course, as others have advised, there could be several other causes of your “newly developed” problem. I always look at my own Losmandy HD tripod permanent setup for such possibilities, even though (occasional) near hurricane force wind blasts at it, don’t seem to have affected it’s stability, in over more than a decade of service.