This may explain some anomalous results I had with 3.6.9 beta this evening.

For the first time (I'm a new user) I started using the scheduler with all 4 options ticked (track, focus, align, guide). Previously i had slewed and aligned before engaging the scheduler so I only had focus and guide ticked, and I worked well.

But now it was going nuts. I started the evening with M52 because it's a rich star field and i wanted to test some aspects of focusing (there are other threads on why). I selected it on the Kstar chart and slewed there, and aligned. Then I decided to get more subs of the Bubble Nebula, aka NGC 7635 using the Scheduler. NGC 7635 is so close to M52 that it's almost in the same frame. So the scope would slew to NGC 7635, and focus. But the align step would systematically move it back to M52. And then continue the subs of that target instead.

I could solve that problem by removing the Track and Align checkboxes from the Schedule.

Then I decided to try another object as the Bubble Nebula was getting a bit low. So I entered NGC 2023 into the scheduler. This would enable a frame with both the Horsehead and Flame Nebula. Well, the scope slewed there, and focused. But the Align step saw it go back to the Bubble Nebula (presumably because to overcome the constant aligning on M52 I had actually slewed to the Bubble. Again the only way out was to remove the Track and Align steps from the Schedule.

So I think these are real-world consequences of the behaviour you mentioned in the current beta version, so I, for one, am glad it's been changed!

Thanks for fixing it,
Richard

Hi Richard
Thanks for the report. I cannot reproduce this special behaviour and I'm not sure if this could be another issue introduced with the changes for differential slewing. Did you save the scheduler list and the sequences? If so can you provide them for testing? Or can you give detailed information how you entered the target (fitsfile, coordinate, ..)?

@Bryan I'm sorry! The "differential slewing" adaptions I made are not part of the newest 3.6.9 stable anymore. The MR has been reverted.

Hi Toni,

I didn't keep the schedules as I'm still learning how to operate Ekos as a whole and the scheduler in particular. I was having trouble stopping the schedules when the object got too low (it didn't seem to respect the minimum altitude I had set, according to the entry in the schedule, though maybe not in reality). I could stop the sequence but when I tried entering another one, it would just carry on with the one which was too low, so I deleted it. Here is a screenshot of the last version from last night. The sequence called up is a set of 3 captures with SII, Ha and OIII filters, of 300sec each.


Since I don't have the sequences I will describe what I did while it's still (relatively) fresh, and helped by the record of messages in the Schedule window.
The first thing I did in the evening was to use the focus module quite a lot, since I had been chasing some anomalous behaviour there. This is reported in another thread. To exercise that I used the star chart to select M52 and then, from the drop-down menu, told the Paramount to slew there. I then used the Align module to Capture and Slew, with action "slew to target". This was successful.

Having finished with that I decided to carry on a sequence of images of the Bubble Nebula (NGC 7635, also called Sh2-162). This was already loaded in the sequencer as NGC 7635, so I added ticks to the boxes Track and Align, in addition to the Focus and Guide which were already ticked from my previous run.

The mount slewed to the object and focused. I saw the focus images of it. Then Align started and after plate solving it was "corrected" to the last position I had GoTo-ed (ie M52). Guiding was then started.

I said last night that it captured frames of M52 instead, but I could not find any and looking at the messages, it seems I spotted the error before any frame was completed and stopped the schedule:
2024-02-02T21:32:56 Job 'NGC 7635' has not been processed upon scheduler stop, marking aborted.
2024-02-02T21:32:11 Starting guiding procedure for NGC 7635 ...
2024-02-02T21:32:10 Job 'NGC 7635' repositioning is complete.
2024-02-02T21:32:10 Job 'NGC 7635' alignment is complete.
2024-02-02T21:31:25 Job 'NGC 7635' is capturing and plate solving.
2024-02-02T21:31:25 Job 'NGC 7635' focusing is complete.
2024-02-02T21:28:17 Job 'NGC 7635' is focusing.
2024-02-02T21:28:16 Job 'NGC 7635' slew is complete.
2024-02-02T21:28:14 Job 'NGC 7635' is slewing to target.

Not being sure what the problem was I made a new schedule for Sh2-162, another name for the same object, and the same thing happened:
2024-02-02T21:39:10 Job 'Sh2 162' has not been processed upon scheduler stop, marking aborted.
2024-02-02T21:38:23 Starting guiding procedure for Sh2 162 ...
2024-02-02T21:38:22 Job 'Sh2 162' repositioning is complete.
2024-02-02T21:38:22 Job 'Sh2 162' alignment is complete.
2024-02-02T21:37:37 Job 'Sh2 162' is capturing and plate solving.
2024-02-02T21:37:37 Job 'Sh2 162' focusing is complete.
2024-02-02T21:34:52 Job 'Sh2 162' is focusing.
2024-02-02T21:34:51 Job 'Sh2 162' slew is complete.
2024-02-02T21:34:48 Job 'Sh2 162' is slewing to target.

Then, I manually slewed to Sh2 162 using the drop-down menu on the star chart, and aligned successfully. So then I re-ran the scheduler with the Track and Align boxes unticked:
2024-02-02T22:03:29 Job 'Sh2 162' capture is in progress (batch #15)...
2024-02-02T21:47:59 Job 'Sh2 162' capture is in progress (batch #15)...
2024-02-02T21:47:59 Job 'Sh2 162' guiding is in progress.
2024-02-02T21:47:06 Starting guiding procedure for Sh2 162 ...
2024-02-02T21:47:06 Job 'Sh2 162' focusing is complete.
2024-02-02T21:44:16 Job 'Sh2 162' is focusing.

I then added NGC 2023 to the schedule, with a different sequence, this time with LRGB filter, again 300 sec each.

Sh 162 was still above the minimum altitude I had set (20deg) so it would not move to the new object until I deleted the schedule for Sh2 162.
I had track, focus, align and guide all selected. This is what happened:
2024-02-03T00:00:05 Warning: job 'NGC 2023' failed to capture target.
2024-02-02T23:59:02 Job 'NGC 2023' capture is in progress (batch #22)...
2024-02-02T23:59:02 Job 'NGC 2023' guiding is in progress.
2024-02-02T23:58:09 Starting guiding procedure for NGC 2023 ...
2024-02-02T23:58:08 Job 'NGC 2023' repositioning is complete.
2024-02-02T23:58:07 Job 'NGC 2023' alignment is complete.
2024-02-02T23:55:32 Job 'NGC 2023' is capturing and plate solving.
2024-02-02T23:55:32 Job 'NGC 2023' focusing is complete.
2024-02-02T23:52:45 Job 'NGC 2023' is focusing.
2024-02-02T23:52:44 Job 'NGC 2023' slew is complete.
2024-02-02T23:51:41 Job 'NGC 2023' is slewing to target.

Note that it took about 1 minute to slew from Sh2 162 to the new target, and after the Align step, it took another minute to slew back to Sh2 162 (while claiming it was NGC 2023).

Unfortunately the messages in the Align window got deleted before 01:20. I don't know why. I was searching for the dialog box "guide and focus limits" and may have pressed "Clear" instead of "Options" in the Align panel. So none of these problem cases are reported there. what I have is attached as well.

I've just found that I did make a screenshot of the Align module at 21:38:57, which captures some of this anomalous behaviour. It's also attached.

Hope this helps!

Toni, this is the exact behavior in scheduler that I was referring to earlier in the thread. It makes no sense. Richard described what I hadn't taken the time to figure out.

All this really want me to revert to before the change for the rotator. Everything in this regard worked great. Is there really no way to scrap all those changes and figure out a different way to solve the rotator issue? Sorry. Just the system is nonworkable. :/

I really appreciate all your efforts!

@Richard, since you mention a Paramount, did you have broken differential slews on 3.6.8?

Bryan.

@Richard, since you mention a Paramount, did you have broken differential slews on 3.6.8?

Yep! And as 3.6.8 was my introduction to Kstars/Ekos it was a bit frustrating.

Oh man, I feel your pain! Sadly Ive been really re-thinking the paramount because of this. Not that the two are specifically related, it's just that I have a strong dislike for TSX, being a monolithic beast of buggy legacy software. Bisque really needs to make a telescope standalone module. Now I realize that is never going to happen since they started their business with TSX long before making mounts! Looking at a Trident P75 friction mount with 24bit encoders.

What really bugs me Is turning off t-point and differential slewing does not allow the paramount to act as a standard EQ mount in kstars. (Maybe you could test that) maybe I missed something Seems like it should behave as any other mount and work fine. I don't really need a model because I can guide and plate solve like the "old days" before I had the paramount. I need to guide anyway due to being at 2450mm. Even with a 300 point model and spot on polar alignment I could never reliably best 120 second unguided exposures. PEC should take care of a lot of the guiding hassles.

Bryan.

Toni, I realise I didn't answer your question:
Or can you give detailed information how you entered the target (fitsfile, coordinate, ..)?

Though perhaps it was implicit in my description.
For the first object, M52/NGC 7654 I selected it on the star chart and did a goto
For the next, Bubble/NGC 7635 I relied on the fact it was already named in the Schedule module and had worked previously (though I had GoTo-ed first). There was no GoTo on this evening.
For the third time I entered Sh2 162 (i.e. Bubble) directly in the Schedule module). Again, no GoTo
Each of these times it Aligned back to the GoTo object, number 1.

For the fourth time I used the star chart to select Sh2 162 and did a GoTo, then Aligned successfully.
For the Fifth time I entered NGC 2023 by name into the Scheduler. No GoTo.
It Aligned back to the GoTo object.

All subsequent use that evening were based on GoTo and no Track nor Align in the Scheduler.

Clearly this is unworkable for automated switching between objects.

Bryan, I have only recently made the switch from TSX to Kstars/Ekos so it's still a learning curve for me. I very much like what I'm seeing, both in capabilities and in the speed and transparency of problem resolution.

I am suffering from an inability to put together a build environment on my macOS system as I want to build device drivers to fit with my own observatory systems. With TSX I can build my own X2 drivers, so that's something I miss.

Bisque really needs to make a telescope standalone module.
I don't really have a problem with this. At the moment I leave it running on a second screen, and I've also thought of just putting it on a Raspberry Pi next to the scope. I actually have 2 Paramounts and the small one was running off a Raspberry Pi mounted on its scope until last year. In that way the Pi becomes an extension of the MKS board and it almost becomes a "standalone module".

Cheers,
Richard

This is why I chose a 10u over bisque or AP… I wanted the modeling contained to the scope and a choice if i wanted to use it.

I wish this was a priority for the rest of the team. You’d think having a broken align module would be an issue.

It sure is an issue for me!

But if I stick to the 3.6.9 beta I downloaded the other day, and avoid multiple objects in a Schedule, I can live with it for a while.

Oh my, I finally was able to reproduce the problem!!!
Many, many thanks to Richard for his detailed summary and to all for the abundance of patience! I'm so sorry for the struggle with this issue about "differential slewing".

Is there really no way to scrap all those changes and figure out a different way to solve the rotator issue?
Indeed the whole story began with an error report concerning the rotator in 3.6.7 and first I tried to go back from 3.6.8 to 3.6.7, but the situation resulted even worse. I did not find a starting point to solve the problems which are also present there. ( I actually discovered some more flaws in version 3.6.7!) Now I'm confident to find a much better solution.

I wish this was a priority for the rest of the team. You’d think having a broken align module would be an issue.
The KStars team is very small one and all of them are very busy at the moment. I wouldn't blame anyone of them regarding the present issue. The align module has a howling complexity and I dived into it rather grudgingly and only because I wanted to preserve the improvements of the rotator module.