I know I'm full of issues, but this is something I tracked down recently. If I use SEP for star detection , Kstars crashes right away.

I've been using Centroid and threshold although they aren't as good, especially with a Newtonian telescope, they don't detect circles as stars if it's out of focus.

Anyway, I also know this is camera independent thing, since it crashes with both of my ASI533 (mc pro and mono) as well as my Meade toupcam.

I also tried remotely from remote-Kstars from windows, controlling the Rasp Pi4, as well as running Kstars on the RPI4 via remote desktop, so it crashes on both Windows or Linux.

If I switch away to another detection algorithm then Kstars won't crash. It's as simple as that.

Hi Frank,

Couple of questions:
1. When exactly does Kstars crash? When you select the SEP option or when you run Autofocus? If the latter, when does it crash, when its analysing the first image, etc, etc?
2. Have you tried using SEP for alignment and / or guiding? Does it give the same problem?
3. Does it crash when you use the camera simulator?
4. It might be helpful to post a log with all focus and fits options checked.

Maybe try shot one frame save it and upload it somewhere?

You can try shot like 10 frames saved it to one directory. Then start profile with CCD simulator and in INDI panel->CCD Simualtor->Simulator Config set Directory so it point where these frames are saved. Then enable it by "Use dir" This way simulator will send previously saved frames. Then you can try run autofocus again to check if it crash with simulator. If yes that would really help.

Ok, so I might have 1 or 2 hours of clear skies (that's all I get lately) tonight and I'm gonna try to reproduce the issue.

1. Kstars crashes just when it starts to analyze the boxed star. So basically taking a picture, boxes in 1 star and analyzes for the shape and either the first time analyzing or at the next focus point, it crashes, I have to restart Kstars and the usual dialog that " previous indi is still running" do I want to turn off and start new, etc. This doesn't happen if I switch to "centroid" or any other detection algorithm.

2. I have the SEP alignment and guiding and the problem doesn't happen.

3. I haven't tried the camera simulator, maybe I'll do that if I won't get any clear skies.
4. Ok, i'll do that.

Ok, so I had several issues, but I ran the problems back to that I was using an external SSD connected to the RPI (custom boot from the SSD drive and not the sd card) .

That's one problem, I guess I tried to fix the slowness of the RPI and it worked , but my choice of SSD is an old drive that was reformatted many times and used plenty of times with large file transfers and it had issues already, including the OS crashing for unknown reasons (The SSD had issues) , when it was used as a primary drive in a PC. Also to add, it does not really fixes things , since now it reads through the USB 3.0 port which can be another bottleneck and also bandwidth issue when 2 other high demand bandwidth hardware (cameras) are also connected to the other USB ports.

Interesting, that I tried to fix one problem (slow reading from micro sd card- an inherent issue with all RPIs ) but created another.