Hello together,
yesterday (after years of using SGPro) I wanted to start to implement the autofocus routine in my imaging.sessions.
The first problem I encountered is probably not an issue regarding SGPro. But however I would be happy, if I could get some advice or a confirmation of my assumption.
The following behaviour was reproducable over the hole night:
After getting the rough focus by a manually procedure, I started the autofocus routine. But allways the autofocus needs too much steps until coming to the lowes HFR value. And after finishing, the validation focus frame was always very off the expected result. I “fixed” that by moving the focusser 100-steps-wise inwards, take an image and evaluate manually the HFR until I came near the expected HFR value.
A typical autofocus graph looked like this:
After that I manually moved inwards:
23163: → HFR: 1,61
23063: → HFR: 1,38
22963: → HFR: 1,31
22863: → HFR: 1,23 (stopped moving inwards, as it is around the lowest value shown in the autofocus graph)
The only explanation for this behaviour is, that on moving out, the focusser actually moves more than indicated by the focuser position numbers. May it be possible, that the focusser is slipping a little bit?
Or may it be that it has to do with the backlash compensating settings? To be honest, it is very good possible that I did not understand that correctly. From my understanding the observed behaviour has nothing to do with backlash. In my opinion fhis would be the case, if the final out move to the estimated focus position would be too short. But the opposit is the case. On moving out, the focuser is moving too much. Please correct me, if I am wrong.
Unfortunately I couldn’t test the autofocus with my old camera (Moravian G2-8300). Now I have attached a G3-16200 with enhanced cooling and an external filter wheel. In addition a tilt adapter is within the image train.
I have a FSQ85 with original focusser and a HSM35 stepper motor controlled by Focus Boss II.
Thanks a lot in advance for input and best regards
Reinhard
PS:
Here is the link to the log file of that night: