I’m currently experiencing some issues with my small field instrument. I’m imaging planetary nebulae, and that completely destroys the autofocus routine, since the planetary nebula is treated as a big star, and the HFD is completely wrong.
My petition is to use a predefined near star field to focus the telescope as an option, that would remove any nebula interferences in the routine.
I’ve experienced similar issues when my FOV has a globular cluster inside. Another solution is to exclude the conflicting areas from the focus routine, but it is not trivial because of meridian flips.
One concern about moving away from the target to focus is when you are close to the meridian, in this situation you might trigger a meridian flip when slewing to the focus target. One solution is to specify 2 focusing areas one east and one west to the target image.
The other solution would be to improve the star detection algorithm so that it isn’t fooled by planetary nebulae. It could for example try to cull outliers from the set of detected stars. If there’s a hundred stars with a HFR of 2 pixels and one with a HFR on 40, it’s a fairly obvious outlier.
We already do this, but the AF routine might be tuned further to accommodate other situations… If you want us to look at a particular field in the context of AF, make sure to turn AF Packs on and make them available to us. We will either be able to tune AF to be better for a wider audience or we can explain why that field is difficult to handle.