I'm having the same issue and I thought I was doing something wrong! So glad it's not me.

I found last night after trying the Ekos PAA that my PA error got progressively worse (starting 5arcmin, ending at 9deg) each time I ran the routine.
In hindsight, when I first used Ekos a couple of years ago I gave up on the PAA after similar symptoms (it was a long cold frustrating night chasing alignment!) and I then switched to a polar drift alignment process using PHD2. I wanted to use the Ekos routine again and had a chance to do so last night.

After several attempts and it progressively getting worse, for an experiment I tried to adjust the star position in the opposite direction by imagining where the target would be if the vector was flipped. To my surprise the PA error reduced after each iteration - I got back down to 5 arcmin. I couldn't get under 3arcmin, it was well passed midnight and I gave up. Maybe it only needs to be flipped on one axis? It was too late to further test this idea.

So... here's the gear in use last night:
- KStars/Ekos/INDI running on UbuntuMate 18.04 LTS with all packages up-to-date as of 28th Jan 2020. Intel i3 4th gen PC, 8GB RAM, 256GB SSD. USB3 connection to Pegasus Astro Ultimate Powerbox v2.
- Skywatcher EQ6-R via EQMod (Bluetooth to PC, using XM-15B module).
- Saxon 200mm F5 (1000m focal length) newtonian, with corrector.
- Canon EOS 650D DSLR, via USB from Powerbox. DIY focus motor via Powerbox.
- QHY5-II-M guide camera on OAG via USB from Powerbox (not used for polar alignment).
- GPS via VK-172 USB GPS dongle (via USB on Powerbox).
- Offline astrometry.net solver
I'm in the southern hemisphere, Victoria, Australia. Latitude is 36.7570° S
The telescope/mount tab on the INDI Control Panel correctly shows location and the 'southern hemisphere' as being selected based on GPS data.
For the Ekos PAA routine, I'd selected 'West', with 30deg steps and a slew of 600x.

For now, I'm happy to know I'm not (a) crazy, or (b) dumb and will go back to doing a PA routine with PHD2 - though I'd really like to be able to manage it all in Ekos as I can't use my DSLR with PHD2, or at least I haven't worked out how to.

I'm happy to try any suggestions you have.

Cheers.

If you have logs with the alignment option checked so the PAA data is shown that may be useful.

I'm looking at the PAA routine, hoping that I can make it work for my situation where I can't see the pole. The code currently sets the declination for the solve hint to 90 so it doesn't work well at 50 or 60 deg. Using the scope position as the solve hint gives a good solve but the polar error increases as the size of the circle increases. I have a fix for that but it's slow going because there is so little documentation.

Currently I am traveling but I'll be back home on Sunday so hopefully I'll be able to provide new logs shortly after that. If you want me to compile and test any code of yours then let me know and I'll gladly do so.

Wouter

I have checked out the code from GitHub and reverted these two lines to what they were before according to the git log. I am recompiling KStars now and will test the polar alignment tonight. Then I will report back so we can see how to investigate this.


Wouter

Hi everyone.
Any news about this issue?
I live in Argentina and I realized that I have to reverse the correction vector to do PA but I did not improve below 4 minutes.
If anyone managed to solve it please share it.
Thanks

Sorry, no progress so far. It has been cloudy in La Serena, Chile, where I live so I have not been able to test my code changes.


Wouter

Oh, La Serena! I visited the place last year for Solar Eclipse. Great place and the sky is just unbelievable! Not mentioning Elqui Valley!
Great move from Spain to Chile!

Could the reason some, but not all southern observers experience this issue come down to different optics and the orientation they flip the image to?

Hello Wouter

I wrote a comment to your cloudy nights issue in thread:

ccd_simulator with transformation to telescope's HA system (Polar Alignment)

Kind regards, wildi

Hi Wildi,

Yes I just saw it. Great stuff! I hope it will help to pin down the source of this issue.

FYI I have cloned the KStars. libgphoto2, indi and indi-3rd party repos and compiled everything. I also modified the align.cpp code in KStars and reverted the fix that Jasem introduced for flipping the pink arrow when in the southern hemisphere. Hopefully I will be able to test it soon.


Wouter

Jasem,

Tonight finally I was able to test reverting those lines and polar alignment for me went back to normal again. Perhaps a checkbox can be introduced to flip the pink arrow for those having issues until this has been sorted out?


Wouter

Hello Wouter and Jasem

Wouter, in the mean time the ccd simulator where you can set the position of the HA axis is available (see ccd_simulator with transformation to telescope's HA system (Polar Alignment) ).

Jasem, I guess I have no commit access any more ) How should I proceed? I described everything necessary in the mentioned thread.

Kind regards, wildi