Hi,

I've been working out autofocus with my ZWO EAF + ZWO ASI183GT + TS 76EDPH scope but I am having some issues.

1. Algorithm: I have tried many combinations. Autoselect start and subframe works fast using SEP and iterative and works well while I am paying attention and manually start from a good rough focus but left alone during an automated sequence it fails. Full field seems to work better when left alone with a step of 25 if focus is close to correct but sometimes it goes off in the wrong direction, loses focus then fails to detect stars (because of going completely out of focus) and it aborts the sequence. Any suggestions as to what algorithm, initial step size, etc.?

2. While watching the curve it looks like the focuser moves further going back than going in. I suspect because the motor has to work against gravity. Will adding some steps to the backlash option help with this?

3. During a sequence the measured HFR with a 5% threshold triggers the autofocus every single frame captured. Should I increase the threshold or there is something else than I am missing? I would like for the autofocus to check every frame for focus but only trigger a focus routine when appropriate and not every time.

4. I have tried to determine an offset for each filter so that I only need to focus using the luminance filter but the values measured for the steps are different every time I run an autofocus sequence. Even for the same filter it will have a different reading every time I perform autofocus. So I can't establish a consistent offset. Is there any work around for this.

Thanks in advance for your help.

Shinken,

I cannot be certain, but this may be due to the same issue I have been looking into in the link below. The autofocus chooses a star and creates a subframe around it but then sometimes uses another star in the subframe to perform the autofocus on, which can fail. If you are not watching it happen you just find it has lost focus.

Autofocus changing to another star to perform the autofocus

Paul

No. I don't think is that but thank you.

After a month of trying INDI/KSTARS/EKOS I think I will move back to the Windows platform. Too many nights trying to make the focus work, the guider work, the sequence to work after already having spend a couple weeks getting the OS loaded and running for someone not familiar with Linux. I find the options in the interface not clear for a new user with not enough knowledge about astronomy and the documentation limited. Too much time and not getting where I want to be which is capturing images without glitches at least most of the time. A bit sad because I really like the idea of using my Raspberry Pi and it simplifies cable management.

We all have been there - at the beginning of a long but rewarding journey. Give yourself some time and come back if you like to.

Clear skies!

Did you try entering a value for backlash? I have changed mine and it is getting better. You are going to have backlash in two spots, the ZWO EAF and the focuser on the scope. You are supposed to measure it(I have not yet) and enter in that value. I believe to measure it you move your scope all the way in until its at ZERO on the scope. Then you move OUT 10 steps at a time until you see the focuser move off rom the zero line, whatever that value is, will be your backlash. So if it took 50 steps until your focuser moved then you enter that value in.

Other platforms are just more limiting. That won't help you improve your imaging.

What you are describing are hardware problems. The ZWO-EAF may be underpowered and thus missing steps. That means your focus will move back further and further in absolute steps.
That's not a problem of EKOS, but of the focuser. You can get around that by adding some rubber bands that support the focusing tube of your scope to compensate for gravity. I have had great success from the days when I used an underpowered analog focuser. In the meantime, I built my own autofocuser and that gave me the option to design the mechanical parameters around the needs of my setup.

Also, backlash is not a problem of EKOS, but of the mechanical nature of your setup. You need to eliminate that by tightening up things.

I have been through all these things. Now I am at the point where I can really exploit the bleeding edge capabilities of Ekos and they are AWESOME!

Hope to see you back here again soon!

Jo

I did try to set a backlash and tried different values to no avail. May be a hardware problem but I just tried NINA last night and despite some issues due to my lack of experience with it I was able to get running after 3 hours only. I could not get my laptop to last long enough as I have no power other than batteries at the location and that was my reason to use a Raspberry Pi and EKOS to begin with. It's so much more portable. I may try it again in the future when I have more knowledge.