After a flip, it does re-align itself, but I don't think the issue is occurring after the flip for me... it might just be the random dithering is getting stuck in one direction.

I changed my dither to 1 pixel, but it did it again last night. Had to throw away about a dozen 4 minute images.

Unfortunately I had to get my rig in quickly at 3am as rain was minutes away. I woke up and gave my images a look-see on my phone and saw the Cigar Galaxy going lower and lower in the images.
I wish I had done the view in the FITS Viewer recent images tab, because you can see the target dithering and following one direction. You can see it in my previous post with the attached MP4 ( #51825 )

I wish I could change the title of this thread... it's not after the flip that is the issue. It a while after the flip.

Thank you for the investigation David. The simple dither algorithm is implemented here

Perhaps there is a reason it's going in one direction, but the base math checks out unless I'm not accounting for vector math perhaps or even something more fundamental .. if you spot an issue, let me know!

Sorry it's taken so long, but I've been spoiled in the past by being able to walk through code watching the variable values.

Wondering why Random is used vs. a fixed direction? default_random_engine generator(seed); Can I see what value it might return?

Why not just do a Right/Down/Left/Up... in order? Or even Right/Right/Left/Left. Eliminating the possibility of a Down/Down/Down/Down.

Another bunch of photos lost to what looks like dithering in one direction. (only guessing on that)
Here is my log from last night.

We've had dithering always going in one direction a year ago but it was resumed and worked nicely ever since. indilib.org/forum/ekos/3466-lost-star-wh....html?start=12#26775
Is this a regression now or something new? My last observing session was 03/25 and dithering performed well.

I had the same problem once last week, and noticed that my Alignment module was set to "sync" only, not "slew to target". After the flip, EKOS moved the mount well, but didn't try to re-allign the scope, and instead just simply synced it. It appears that it does not go back into the alignment module and adjust the settings to "slew to target", but simply leaves the settings as they were.

Note (I changed the title of this thread)

Here is another video showing the drift while dithering...

I wonder if this is a problem with polar alignment. I have been making sure my PA has been within a few Arc Seconds and noticed my drift virtually went away.
Last night I let my system run for 4 hours, taking images every minute and dithering every 4 frames. Four hours later the target was no more than a minute off.

Peter

My guiding seems fine. It's the shift during the dither process. Which we expect of course. But it's dithering in a direction as opposed to moving around the original target.