Thanks for the info, seems like the Focus profile made no difference.

So I think the last thing we can try is for you to capture the focus images during a 1x1 and 3x3 run and send them to me along with the log (sorry to have to do this again). I'll then use your images to debug the focus routine on my system.

To do this switch Focus image logging on:


The images will be stored in a directory at. the same level at the logs directory on your system in a directory called "autofocus". Within that will a directory per date and then the fits images taken by focus. I can tie the timestamp in the image name to the timestamps in the log (verbose with Focus selected as per last time please).

So what I should be able to do is to run the successful 3x3 images through a focus run and replicate the curve your get, and then again for the 1x1 images.

Just be careful with leaving Focus Image logging enabled as it will consume your disk space.

I'm away this weekend so may take a few days to do this, but hopefully I'll be able to replicate the issue and see what the problem is.

Hi John,


Focus images binning 1x1:



Pictures in attach.


Focus images binning 3x3:



Pictures in attach.

Thanks very much for the focus images. I think there is one missing in the 1x1 set (10 and I was expecting 11). It doesn't really matter - I duplicated point 6 and use it as the 11th (in focus) frame.

So here is more or less what you get with bin 1x1...


Here, I've played around with the SEP settings...


What I've done is increase the noise threshold and increase the number of pixels that need to be above the threshold in order for star extraction to consider a group of pixels as a star. So, the effect of this is that I'm identifying less stars and the ones I'm identifying are bigger and as a consequence I'm excluding more noise.

I believe that by binning 3x3 its achieving the same thing; its boosting the S/N so you get less stars but also less noise masquerading as stars, which gives you better results. Here are my changed SEP parameters run on 3x3 bin data...


It works OK, but the number of stars is too low IMHO (<10 on most datapoints) so I would suggest that you match the SEP parameters to the bining level.

I'm not saying that these settings are necessarily the best as I've only got 1 set of bin 1x1 focus data to play with but it should point you in the right direction.

There is one other adjustment I would like to play with (it may or may not make much difference) but I need to know your image scale (have many arc seconds per pixel). If you don't know this just tell me your camera and telescope (and reducer if you have one) and I can look it up.

Hope this helps.