Dear Paolo, you are correct, with an external focuser you systematically and intentionally violate the BF specification of the Edge series. As
@John Hayes pointed out, this is less of a problem if you're BF is within say 100 mu, the bad thing is, that owing to the optical design of an SCT, a BF mismatch of 100 mu corresponds to an error on the mirror position of only 1/25 (5^2, 5 being the magnifying factor of the secondary) of this length, i.e. only 4 mu. For typical SCTs with an aluminum tube this corresponds to the thermal contraction of the tube by only a fraction of a degree C, while for a typical mid European climate (where I am observing), the temperature difference between the beginning and the end of the night is more like 5C (and for a dry desert climate it can be 2x as large and more). Will say: even when you fine tune the focus with a Bahtinov mask and the mirror focuser at say the beginning of the night, the secondary focuser will move the sensor out of BF fairly soon. Solution could be to use a carbon tube (ie open heart surgery on your SCT) or the Optec focuser shifting the secondary. The later is probably conceptionally the best solution. I have not used it (yet), but
@John Hayes has excellent experience. My main problem is that for an Edge 11 (my scope), the cost of the focuser would be 2/3 of the cost of the OTA - which appears a bit disproportionate. For an Edge 14 (listen, Santa!), I probably would go that route.
However, the 100 mu tolerance above may be a bit to stringent, as we usually don't use our Celestrons in space, and most of us not even on Paranal or Mauna Kea. There is a fairly detailed analysis by Innovations foresight
https://www.innovationsforesight.com/support/celestron-edgehd-back-focus-tolerance/ who basically come to the conclusion that a much larger offset is tolerable, they are actually arguing (and giving the derivation for it) that for a 2" seeing even a BF shift of 20 mm (sic!) would only result in a 10% increase of the FWHM on a full frame camera (and for APSC one could even tolerate 50mm). Personally I consider these estimates far too optimistic, as they assume a spherical geometry for the PSF and also for the aberration, while in praxis stars are deformed radially/tangentially if your BF is too small/too large, and your eye is quite sensitive it detecting deviations from circular. But the BF requirement seems to be considerably weaker than mentioned above.
Having said this, I also think that many of the problems with the internal focuser can be mitigated by making sure that the last motion of the focuser is always CCW, ie against the gravity, and that while focussing you always run CCW (which for the commonly used ZWO EAF means to have it set to reverse). This allows good V curves when focussing and prevents mirror flop except possibly after substantial repointing of the telescope like a meridian flip - but under that conditions it is good practice to refocus anyway. What remains is a image shift when focussing - which is a real nuisance when switching filters in combination with an OAG, and my experience is also that the V curves with a good external focuser are better (smoother, more monotonic).
My own experience is consistent with the BF tolerances. Since I am using an external focuser (and I set the mirror position at the beginning of the night such that at the extrapolated middle of the night temperature I am in perfect BF) I am getting better (sharper) PSFs, even though the focus position can shift by almost a mm over a night, compared to focusing with the mirror/and OAG. So from a pragmatic point of view: all good, but the purist/physicist in me is not fully satisfied.
Matthias
PS: one could of course think of having two focuser over the night, the standard mirror one every hour or so to position the mirror (maybe with a temperature predictor) and the external one for fine tuning and filter shifts - but I haven't found a way to tell NINA how to operate (and switch between) two focusers ... (in case a NINA hacker reads this ...)
