Hi Chris,

Thanks for the screenshots. I've pushed a fix for the Full Field / Sub Frame issue.

I've started looking at the focuser position issue, but its more complicated so will take longer.

Yep, that full/sub-frame button fix works. Thanks!

I won't be too sad if somehow linking focuser position to optical train doesn't happen any time soon. It was just a "hm, this would be nice to have..."

Refine curve fit is pretty amazing, and really ought to be enabled by default. Enabling it has dramatically improved focus the last few nights.

Thanks for the feedback Chris. Usually I don’t make new features “default” until they have bedded in. Glad to hear it’s working well for you.

Just looking at your v-curve. Where you get a horizontal line of points on the right (points 1, 2 and 3). That could be un-corrected backlash. If it’s a 1off then fine but if you get that regularly I would suggest increasing your backlash compensation by 2 * your focus step size and seeing if that “fixes” this.

Re-measured my backlash, and spent a while playing with the backlash settings, but that didn't help.

It doesn't seem to be a problem though; it's just a general characteristic of my rig. I've found that when I'm way out of focus there are very few stars detected until the FWHM gets small enough. Playing with all the settings for star detection is probably a project for another night.

OK, good to know.BTW, I'm testing a change to better facilitate the starting position of the focuser being relevant to the optical train. Its not 100% of what you want - the 100% solution is a big change to Optical Trains / Filter Settings database that I'm not going to do just now.

However, this will help. I'm going to save the "Desired Focus Position" field per Optical Train. When you Autofocus this field gets updated to the AF solution so its relevant to the Optical Train configuration...


When you change OTs you'll get the last value for that OT...


So when you change your setup and select the relevant OT, you just need to press the ":Goto Focus Position" button (highlighted) to set the focuser to a good place to start autofocusing.This change is part of other fixes so I need to do more testing before merging iit. Will let you know when its merged.

This change is merging now.

I figured out what that flat line is, but not how to fix it yet.

It's the star detection, and it's not yet clear to me how to tune the detection profiles to allow horribly out of focus stars. The mark-1 eyeball detect terrible, round, 30 pixel stars - it's just not obvious which knobs to twiddle to get stellarsolver to do the same

If you are using SEP then post a screenshot of the parameter config and I'll try and suggest some changes.

Using all the stars


Yep, SEP/FWHM since SEP/HFR was being difficult. I was hoping 2-AllStars would detect all sorts of potential stars


Here's an example of some non-detection on an intentionally out-of focus image

The All Stars profile tries, I think, to detect even the fainter stars - which is the opposite of your issue. I would normally suggest the Focus Default profile to reduce the chances of picking up noise as stars but that isn't so relevant to this problem.

It could be that the larger stars are saturated which may give the star detection algorithm some issues. Although if you are detecting these stars when you are nearer to focus then that wouldn't be the issue.

The ellipse check (set to 2) is quite lenient but you could increase that (or set to 0 to disable).

Other params to look at would be increasing the Conv FWHM to 2 or 3. This works well for my setup but not well for others. Or try the Default Conv Filter.

If none of the above improve the situation, can you send me the focus frames from an Autofocus run and I'll try and recreate the issue. To do this, enable...


The fits image files are stored in a directory called "autofocus/<date>" at the same level as the logs directory. If you turn on the above parameter and focus logging with verbose and Focus set, then do an autofocus run with the problem, then send me the fits images and log I can try and recreate the issue.

You'll probably want to disable Focus Frame logging afterwards to avoid your disk filling up with these fits files.

PM'd you a link to a few focus runs.