New AF routine a bust for me


I spent some time last night trying out the new routine. My gear is challenging - EdgeHD 14 at 2737mm. The routine still does not select donut stars for me. Oddly, for some of the more out-of-focus images, it would elect the Poisson spot at the center of the donut. It only selects real stars when they are tight enough to not show donuts, and then the HFR difference from step to step was not enough to make for a nice curve. I tried various binning settings and exposure lengths. I didn’t think to get screen shots of the curves, but I did get 10 focus packs for you to look at. I’m not sure how valuable they will be though, since in most cases, I canceled the focus run in just about every case when it was clear that it was going nowhere. For now, I’m going back to the old routine.

Link to focus packs:



OK… So I took a look through most of this data. Some changes I have made make it work a bit better, but honestly, it’s not great.

A couple potential issues / notes:

  • The new AF method expects stars (even out of focus ones to be circles). They don’t need to be perfect circles, but yours seem to have enough coma, or drift or something else to make them more blobby than circular. This one here looks like a heart.
  • You seem to have a fair amount of drift between AF frames… this may be contributing to the less than circular shape of your stars… maybe not? I’m not sure.

I’m not entirely sure how this helps… based on your comments you can get the new AF to behave if you avoid donuts, but cannot get sufficient distance between AF frames to create a reasonable V curve. Going to the 2.5.0 AF would require the same thing (avoiding donuts, decreasing travel range and thus creating the same issue?). Can you clarify?


The “drift” is mirror shift from focusing. At 2737mm, even a small amount of mirror shift means a fairly large image displacement. One of the things I think I noticed last night is that the AF routine may not be waiting long enough for the focuser to finish moving before starting the frame. Does the focuser report an ASCOM state that it is finished moving that SGP uses? If not, how does SGP know that the focuser is stopped before starting a frame?

However, all of this may be moot, because I also noticed last night that the drive gear on my focuser was loose. There was huge slop and it makes me think that the whole data set is suspect. I’ll try again tonight and report back.



Do you have dew heaters on High?


I have also a problem with the new AF version. Simply it does not work for me, it detects everything except stars. Using different binning or exposure does not change anything, the previous version worked very fine with the same settings. My telescope is a Newtonian and my camera an atik 414, smart focus disabled. If you have some ideas, the are welcome.


Pascal, please see this post (unfortunately we unable to be of any assistance with AFPack feedback):


I just thought I would give a brief update. I found the problem with my star shape problem. My OAG was not square, resulting in the camera being tilted. I have a run going right now, and AF is performing much better. Still not perfect, but much better than the other night. I’ll post the focus packs tomorrow.



Thank you Ken, sorry, so my post is just a confirmation that AF is still in beta version :wink:
Good sky.


I noticed this again the other night. Is it possible that the focuser is still moving when SGP starts a frame capture? Is there an ASCOM property that the focuser driver use to report the completion of a move?



There is. We use that and the ensure the focuser is at the requested position before moving on.