Hi,
for Windows nightly build, just download from here : binary-factory.kde.org/job/KStars_Nightly_win64/

Regards

PS: only tested with simulator

I compiled the latest build on Astroberry, and the new polar alignment seems to work very well! Thanks a lot! A miss Polaris of about 10 degree so I could not use the old one. I tried pointing near the meridian about 20 degree south of the North Pole and the new procedure allowed me to align to about 2’ quite easily (value confirmed also by phd2).

Great to hear. Out of curiosity, what PHD2 method did you use to confirm the alignment--e.g. Drift Align? Guiding Assistant?

I used Guiding Assistant

Hy, this is great. Like the others, I can't wait to try it.

How sensitive is the decomposition into separate az and alt lines to leveling errors? I.e. if my mount's off level, adjusting az is going to also affect alt as well. Does the math take this into account just from the motion of the scope somehow, or is it always a right triangle?

Just asking, not complaining. At all! In the very worst case, the star wouldn't track the triangle precisely but the overall guidance would still be excellent.

You can try it anytime you like--by using the nightlies. It's there.

Upon reflection, you are right that although it directs you to turn the elevation knob, if your mount is not level,
then that won't quite do the trick and you'd have to improvise.
Clearly best to level the mount as best as possible.

Still, if you iterate, you'd likely get close to a good alignment in any case.

Hy

You may as well level the mount. It only takes 2 mins.

In fact I always do. I'm not asking to shortcut that, I'm curious about how the software mechanisms work.

I will definitely have to try it. Please forgive my ignorance, but are there nightly builds for the Pi? I've downloaded and built it myself on the Pi before, but I'm lazy.

I'm not on top of the nightly builds either, as I always compile from source.
Jasem's ubuntu builds are here:
launchpad.net/~mutlaqja/+archive/ubuntu/...field.series_filter=

The way the alignment works is that the system works out the altitude and azimuth angular offset of the mount's axis from the pole.
The UI then gives you a target position and asks you to adjust the "altitude knob" to correct the altitude offset,
then the "azimuth knob" to correct the azimuth offset.
As you point out, the mount may not be perfectly level, so this isn't entirely true. However, unless you're wildly out-of-level,
I think it would be reasonably effective anyway.

The error in alignment caused by the mount not being level will be a maximum of 1/60th of the level error.
So if your mount is 1 degree out of level a rotation of 1 degree in azimuth will give an error in altitude of up to 1 arc minute.

And in any case the tool also shows where you have to move the guide star to. It just means the star will not travel exactly along the alt and az lines shown. But its the destination that matters in this case, not the path.

I gave the measurement part a try yesterday. The mount is on a permanent pier outside, for adjusting I have to get a monitor etc. out - was too lazy to do that
But I did three runs, one pointing at the pole, then two more at declination 80 and 70. The three measurements were within 30" to each other - very nice!

One thing I noticed: The PA turns on the WCS grid display in the alignment tab. This stayed active also for the normal align afterwards, but the button for toggling it was inactive. When I pressed it, WCS grid was switched off, but now the button was in 'active' state. Not sure if that is always like that, or an outcome of me canceling the PA in the middle. Not a big deal either...

Next step will be actually adjusting PA. But as I said, I'll need the monitor outside, and ATM the conditions (snow and ice) don't encourage me to do that and fiddle with the knobs....