Hi guys,
Can we have the ability to select our own star for the EAF routine? With a newtonian, or anything with a central obstruction, the EAF seems to have a tougher time focusing accurately and I am able to focus better manually (which kind of defeats the purpose). It looks like it can't determine a quality star to use (the small preview of the star that's being focused sometimes doesn't have any stars, (ive also played with exp & gain & filter being used) so hoping that if we can get the option to manually select the star for focusing then it should be able to do a more accurate focus routine perhaps.

    a month later
    4 days later

    I would very much like to see this ability added. There are times I have seen the autofocus get confused, like when attempting for focus when imaging a globular cluster, and the star is actually several stars and the centroid is never quite right for finding focus by the algorithm.

    Improve101 the EAF works only occasionally in my experience.

    Hi,
    You are doing this on ASIAIR or PC? Do you want us just add this ability to ASIAIR? Have you tried on PC to check how it works? I mean we should first find out this is about ASIAIR or EAF or something else.

      @tech@zwo#51597 Regardless of whether OP has a lemon EAF or not can we get this option anyway? I've had several sessions where the AAP will select a double star or stars in a star cluster as the "suitable" star and make the focus even worse than before. In those cases I re-run the Auto Focus routine only for it select the exact same stars again so it was pointless. I had to slew the telescope away from the target, focus, plate solve, synch, the go back to my target. Needles to say that in those cases setting AF as part of a Plan or Autorun was pointless.

      • w7ay replied to this.

        WalterT AAP will select a double star or stars in a star cluster as the "suitable" star

        Walter,

        The problem is not as simple as that.

        ZWO uses multiple stars to obtain the average star size. The small inset image that it shows is just one of the "star" it has chosen.

        So, giving the user the choice is not easy, since you need to give it many stars.

        That being said, because they are using multiple stars from various parts of the frame, a system that is not perfectly flat will give inconsistent results. Most of us old timers pick on stars about 1/3 of the way away from the center to focus on; I don't think ZWO does that.

        The second big flaw is that the ASIAIR is very poor at rejecting hot pixels. The ZWO algorithm would pick hot pixels over dim stars -- that means it is really bad with telescopes with large f numbers.

        Worse still, it is using parabolic matching instead of hyperbolic matching.

        Just use a Bahtinov mask on a star about 1/3 of the way from the center if you want good focus. Since it works on diffraction patterns, the Bahtinov mask is hardly affected by atmospheric turbulence, and do not need multiple stars. It is also much more sensitive than depending on HFD. The Bahtinov mask can see changes in focus when the HFD changes by only 10%.

        Chen

          Hey Chen, you got me there, I had no clue they used multiple stars for the AF, that's actually pretty cool and maybe they should mention that somewhere (like in these comments not some obscure documentation). Still pretty baffling when the star in the preview window is obviously not suitable and the curve looks like a shotgun blast instead of a V or U.

          I agree with you on the Bahtinov mask. I use it on my scopes and camera lenses everytime to get perfect focus at the beginning. But I also like to go inside to stay warm and let the scope do its thing and refocus at certain points and definitely every time a filter changes. And in all fairness the EAF works pretty well most of the time, but when it doesn't it makes me want to run outside and yell at it in front of all it's friends until it cries :-)

            WalterT @w7ay Also good to know it uses multiple stars, because sometimes the star in the image is not at all in focus and the image just gets worse and worse and the EAF eventually gives up. I'm using ASIAIR only by the way.. no pc. Id he happy to select multiple stars if needed. At the end of the day I've noticed the EAF doesnt seem to do a reliable & repeatable enough job for me to got to bed & just trust it. I don't trust the EAF at all, and thats the issue. However if we can tweak the EAF so that it works every time that would be amazing. So I personally should not be able to do a better job of focusing than the EAF, but i consistently can by just selecting a star and moving the focuser in slow steps and just waiting to see what the new FWHM is, or even just with a bahtinov mask. Something needs tweaking anyway. For a start, can we see all the stars selected, like a grid of 9 stars or something, and also have an idea from where in the frame they were selected from?

              3 years later

              On several recent occasions ZWO EAF routine selects a pair of stars close to each other. It then gets confused--in one step using two stars as if one and getting a huge number, then isolating one of the two stars getting a much smaller number, then using the two stars again. End result is that EAF never achieves focus, or says it has when the focus is actually lousy. It would be nice if we could tell it what star to isolate. I don't understand the "selects multiple stars" business. This makes no sense to me. It always displays a single star it is measuring.
              Another nice feature would be if EAF allowed offsets. I know how much to change the focus when changing from Ha to Siii, for example. It would be nice if I could tell EAF this information.

                viperduude What is the software you use to control the EAF? if via ASIAIR, could you please submit a bug report for further analysis? we calcuate the FWHM via many stars average, a pair of stars closed should not cause this problem, thank you.

                    2 months later

                    Tech@ZWO
                    My apology. It's been awhile now since this problem so I do not know which log to submit. Yes, controlled with ASIAIR. If it happens again I'll immediately send you the log file.

                        Write a Reply...