Hi Jasem,
I originally asked the question if this can be done and if there was a workflow to achieve this, and as you mentioned using the plate solved info in kstars I gave it a try a couple of times and could not get the rotations to match between different setups and pack ups. I hence added this to the wishlist if this feature could be added in future release.

The workflow is being able to set a target angle of rotation and when plate solving the result compares to this original target angle and denotes which way and how much to rotate by. The process is repeated until the tolerance is met.
Thanks in advance.

Thanks, I've added it to our issue tracker: invent.kde.org/education/kstars/-/issues/56

I have asked the exact same thing not too long ago! You were probably thinking of me, Jasem!

For the original poster: the workaround I have been using, while the feature is added is this:

- Load and solve an image from the previous session with the exact orientation that you are happy with
- As soon as the Plate solver solves it and reports an angle, take note of it - it will soon be overwritten by the new picture coming from the camera, that will have a different angle
- Go to KStars, set a custom field of view with your focal length and sensor dimension; put the same angle that you saw on your "load and slew" plate solve
- This will show the exact rectangle, with the exact rotation on KStars
- When the new images from the camera are solved, compare the solved rectangle to the custom one
- If they differ, and they'll likely will if you changed the rotation of the camera from session to session, turn the manual rotator CW (or CCW) and solve again
- Check the two rectangles again: is the one from the solver closer or further away from the custom one? If it's closer, your initial guess for direction was correct; if not, reverse direction
- Keep solving until the two rectangles match as closely as you can get them - you can also read the angle computed by the solver and see how much different it is from the one you got from the initial "load and slew" of your reference image

With this method, I have been able to get them as close as a tenth of degree off. In my case, taking dithering into account, it's good enough as I don't have to crop much the edges that don't overlap completely, as I dither pretty aggressively anyway.

Hope this helps!

Thanks Jasem, cant thank enough for the excellent work you do !

Thanks endlessky, It makes sense and will give it a try this week if the sky clears up, have you tried this workflow with mosaics ? many a times I set the job up on scheduler but do the second panel next night and this is when the rotation matters the most, as you would know the difference in rotation between the panels contributes to excessive crop and the only way I know to counteract it is to do more panels.

I worked on this last night, and think I've got a good workflow. This video shows it in action (I'm manually changing the rotation of the CCD simulator in the video). It works off of the "Rotator" option in "Alignment Options".

What about those who don't have a rotator? Would we be able to rotate the camera manually and still use the interface that tells us which direction to turn and how many degrees we are from target?

Thanks!

That's exactly what this change does. If you have the "Rotator" option checked in the "Alignment Options", but don't have an automatic rotator attached, it will pop up this dialog when your rotation is outside the set threshold. If you do have an automatic rotator, this dialog won't show, since it can rotate for you.

Fantastic! Thank you, this is exactly what I needed!

This is great, thank you! This would solve most of the issues with manual rotation and give the user the ability to reduce rotational errors. Just one thing to mention

1.Once the rotation is complete and within tolerance, it would be necessary to check the image center and plate solve again to ensure the user didn't accidentally move the center point. This happens at times and causes the image to shift in the horizontal or vertical directions. [plate solve steps would be : Solve the loaded image send reference co-ordinates and rotation to Ekos --> calculate deviation and center the image and slew to target and be within required accuracy. --> Check rotation and report required rotation [as per your video] --> Once the rotation is within tolerance ---> Plate solve to check center of image and reslew to center and be within tolerance.

Each click of "Take a new image" does a plate solve (it's how we check the rotation). It will ensure we are centered every time.

This is a great feature! Thanks for adding it. I take it that it is working with the load and slew option? I could see this being very useful even when you haven’t taken an image yet. So early in the day, you’re trying to figure out which target for the evening, and how you want it to look in your field of view. It would be great to save the target and the angle you want in Kstars. Then the rotate window is able to help get you framed correctly.

Just a thought. Again, thanks this really helps!

Yeah, this is just for load and slew, which I think gets most of the use case covered. I know when I'm first framing something, I usually just use the preview in the Capture module to get the framing I want. Matching an exact rotation for me is only really needed when coming back to a target.