It is the DDM85.
It probably is hysteresis as the MLPT does make improvement it unguided error.
If you look at the JPEG I posted , the spread in the residual error after superfit shows a non random pattern along the RA.
I wound think this should tell us something . I have not been able to determine a mechanical source yet.
Btw, I tried the mount with Bisque's Tpoint . I automapped about 44 points. It shows no pattern in residual error.
The pointing is no better also.
They have a lot extra terms that autoslew does not use. None of these improve the pointing model.
There is always the possible that the pattern is an artifacts on how the autoslew does superfit.
There is so little explained with the autoslew modelling here.
We should be able to manually adjust the model by removing outliers etc and see if the fit statistics improve.
There is a lot in the advance modeling that is undocumented. I wish we knew more.
Perhaps some day Dr. Keller or ASA will spend some time on this.