I am using the Celestron SCT focuser with the dedicated driver, that allows for backlash compensation. However, I am not quite clear what that setting actually does e.g. during an autofocus run.
Will the focuser "overshoot" by that amount and come back to the target setting if the direction is reversed e.g. if I want to go from 1000 to 1100 it would go to 1150 first and then move back to 1100 (assuming BL is set to 50)?
Or will it just "add" the backlash to the difference e.g. just move 150 ticks in the above example.
I guess if it is the first variant it would be sufficient to set a BL value that is just high enough to compensate. The second variant would require exact measurement of the the actual backlash.
thanks for responding.
I am not sure I fully understand the answer. Are you saying that, with backlash 50, if I am going from 1000 to 1100 it will move by +150? And if I after that go to 1000 it will move by -100 (in spite of direction reversal)?
Or is it -150 in the second move as we changed direction?
This would mean there is no overshoot and I need to know the exact actual backlash value?
No there is no overshoot. Oh sorry I thought 150 is backlash value. So if you issue +100 +100 +100 -100 -100 moves commands it will apply -50 backlash compensation before that first negative one. So in reality first one will move -150.
Thanks, that clarifies it for me. So when leveraging this setting it is crucial to first measure the backlash with the same accuracy as desired for the final focus value. Is there a really good way of doing that, while not adding all other random focus measurement errors to the process?
From searching about this focuser on the net I am getting the impression that users of other platforms are taking a different approach. E.g. NINA users are setting a 100 ticks overshoot, the Celestron focuser tool using a 500 ticks overshoot and when controlling the focuser directly with SkySafari that appears to be using an overshoot in one direction as well.
Is there any consideration to implement something similar on Indi? I imagine that may be relevant for users of other focuser models as well.
Thanks again for your patience explaining all that. If I understand that approach correctly it requires the ability to measure the actual focus position, which is pretty hard to do on an SCT. I may be able to determine that backlash of the focuser motor unit, but that would not include any backlash of the focus unit within the SCT itself (which is probably the main contribution to the total backlash).
I guess the main other option may be to try assessing the backlash from the number of ticks that don‘t have an impact to a stars hfr. Perhaps I will try that with a somewhat defocussed bright star, where changes should be very visible.
I don‘t know if an overshoot approach actually works better. It is just what I had seen users of some other software platforms recommending. If I don’t succeed in finding the proper backlash setting I probably can still fall back to the "linear" algorithm, which appears to be the most recommended approach for ekos.