I have that very same lens. It performs very well when it is stopped down one step. I personally rather use step down rings on the front of the lens, gives more natural looking stars. 37mm ring stopping it down to F5.4 does the trick. The residual CA is so small that it can easily be removed in post processing.
Great image btw.
Thanks. And thanks to everyone.
I do have the requisite rings to step down to 37mm and it does exactly what you say. I chose not to install them for this shot for two reasons. First., there are really only two stars in the FOV that are bright enough to show obvious diffraction spikes. If there were more bright stars or brighter stars period I would have used step-down rings. Second, I have found that fall-off (vignetting) is considerably worse when stopping down using rings. Whereas I can get away without flats at f/5.6 achieved with the internal iris, I definitely would need flats at f/5.4 achieved using step-down rings. I haven't got around to shooting flats for any of my lenses yet but that is high on my AP to-do list. (I realize that such long-after-the-fact flats would only correct for vignetting. But I don't yet have any obvious dust motes.)
Re the small residual CA, the reason it has not been entirely corrected here is that I did the CA corrections after stacking. With Astrotracer, the target moves in the field of view between shots. So every few shots (in this case four to six) I recenter the target. This means the target moves around in the FOV and hence the CA is a little different for any given star from shot to shot (because it will be in a different position in the frame from shot to shot). Properly, I should remove CA for each shot individually (can be done in a batch process since each frame will require the exact same settings) after raw development but before stacking. But that all takes a lot more effort than what I did here, which was to use Sequator to stack from the raw files and as a later step do CA correction.
Edited by Alen K, 18 August 2019 - 09:45 PM.