I have been running Kstars 3.6.3 on Ubuntu 20.03, on an Rpi4, for nearly a year and it was working great. I recently added a ZWO EAF and after some discussion with John (in another thread) I decided to upgrade my system to Ubuntu 22.03 and Kstars 3.6.8, the latest stable version in Jan 2024, so I could test and take advantage of the improvements in the focus routine.

I worked out all the system issues, in the garage, and tonight I decided to take it under the stars to try a test focus run. My equipment is a Celestron CGEM mount, a ZWO EAF focuser, a ZWO ASI 533 main camera and a ZWO ASI120 guide/finder camera. I have been imaging Jupiter, so I had a Barlow in the system, which I removed tonight and replaced with a .63x reducer. To do so I had to run the focus, manually, all the way 'in;' re-install the EAF and sync that position to zero. I was going to use Jupiter as an easy target to find, I did a PA (which went great using the Finder scope and the ASI120) and slewed to Jupiter (which went great) and guessed that the focus position might be somewhere above 50000 steps (the guess was good enough to find out of focus Jupiter). I was pleasantly surprised how quickly the EAF went thru those steps (I have a Micro Feather Touch focuser on my SCT and the EAF attaches to the fine shaft). I found close to focus on Jupiter and then moved to a star to run the auto focus routine. It worked very well, I reduced the step size and ran the routine several times and it found a best focus position.

Focus routine Issues:
1) I could not deselect "use all stars"
2) I could not select the options for setting a 'ring' or 'mask' area - I was using either "SEP" or "Centroid" I could not access those boxes ('ring' and 'mask')

From here on the whole process went to C@%P. I realized that my optical train was all wrong (when I looked for 'advice' on focus settings). So, I shut everything down (left the scope pointing where it was, it had a good PA) and started a new equipment profile. I added the mount, the EAF and the two ZWO cameras as my equipment package. Everything started up, I moved to the focus module to set up the new optical trains (with the .63x reducer instead of the Barlow). Kstars crashed. I had not set logs to verbose, for the new setup, so the logs show nothing. I tried this several times with restarts of the entire system (including the RPi itself). The same thing happend when I tried to setup up the optical train. I restarted again, then the Indi Control Panel would not find the ZWO ASI120, it showed a camera, but it came up garbage. I powered down, reseated all the USBs (I have a powered Anker 7 port USB 3.0 hub at the scope) and tried again. If I was able to get Kstars to recognize the second ZWO camera as the ASI120, it would crash when I tried to set the optical train.

I really don't understand how the system could start up and work fine under an existing equipment profile (which had the right equipment but wrong values for the optical train) and then suddenly start acting totally inconsistently. I know the whole software package is very complex and I respect the effort that has gone into it, but the basic functions should be much more robust than the way it is behaving. It could be a power issue, but my power supply (it is a separate 12v regulated supply) has not changed and all this equipment (including the ZWO EAF) has operated without issue over the past month when I used it all to collect videos of Jupiter (however, that was using a different SSD which had Ubuntu 20.03 and Kstars 3.6.3).

I was eventually added logs to the new profile, so I am attaching the last two here. They seem to reference Phd2, which I use, but it has never been an issue in the past to use PHd2 for guiding and I was not guiding tonight at all, so I only opened Phd2 once to see if it would connect to the equipment. It did not, so I shut it down.





I eventually brought the equipment back into the garage, shut it off and restarted it, but the behavior there was no different.

I will try again, soon.

Ok I have an update (sorry for long posts ).Today in the garage the behavior was better (more consistent) but still have issues. I don't know what the issue was last night. I am attempting to monitor the RPI CPU usage and temperature as I operate the system
Two consistent issues:
1) If I start a new profile, with all the same equipment (EAF, CGEM, ASI120, ASI533), I connect to all and save camera configurations (options tab) with the proper scope FL and Aperture. II go to the the focus module, it never asks about the equipment train. In the focus module the train box is empty. I edit it and attempt to add Primary and Secondary and then Kstars crashes. I repeated this several times. This is the new profile fail log below.

2) If I restart Kstars (without a reboot) It will not successfully connect to the ASI120 camera. It does not tell me there is another instance of Indi still running (I have made sure by killing any processes associated with Indi). The log seems to indicate that it found the ASI120 and read in some info, but eventually it would not successfully connect. I have had this failure mode repeated also. The log below for camera fail is associated with this issue.

The GOOD news is that if I use on old profile (one that I ported over from an older version of Kstars) which used the same equipment but had different optical trains. I can connect (on a reboot) and it does ask me to check the optical trains first thing upon connection to the equipment. I can update those numbers accordingly and the program appears to operate normally. The focus module shows both Primary and Secondary. And all the tabs below are working (masking, all stars, SER editing, etc). this log file indicating oldproile is associated with this success.

I hope these help in resolving any issues, but it appears I have a workaround for now.

My experience under the stars tonight was very satisfying. Using a modified older profile the equipment connected and behaved as expected. I performed a PA, then moved to a star field to exercise the focus routine. The focus routine worked well (to the best of my novice ability with it), I was able to connect to Phd2, perform a cal and enable guiding (total 1.12 " until the clouds and wind picked up). I am attaching a log of this session. I monitored CPU activity and temp and both were well within acceptable ranges.

My installation of Kstars 3.6.8 continues to exhibit the same behavior when I try to initiate a new profile of equipment. All the equipment loads and contains all the correct information, however Ekos never requests that I set up an Optical Train and if I proceed to the focus module there are no optical trains to select. If I try to edit the optical train, there are none so I select to add one and then Kstars crashes.

I am attaching a log, but it does not show much, it just stops after the cameras are shown to be online.



As I mentioned earlier the work around is just to modify and old existing equipment profile and then everything works, but I am concerned that this indicates a bug in this version (3.6.8 ) somewhere.

Has anyone else seen this behavior? I do not want to upgrade to 3.6.9 until I find out that what is happening in my instance is understood.