Hi everybody! Has anybody tried to use Polar Alignment Mode with Ekos to measure Az Error and/or Alt Error?
I've been recently trying to use it with no success - each time I get different measurement much beyond acceptable accuracy level, which renders it practically unusable. Either I'm doing something wrong or there is something wrong with the function itself. If you (1) slew to a target near meridian and (2) measure Az Error AND without a touch to your Az settings just repeat the steps (1) and (2) you will end up with two totally different numbers of Az error. I get the feeling that correcting Az error based on these measurements does not get me any closer to the objective. Any ideas?

I've been using it with success and repeated measurements, while yielding slightly different results, are still within acceptable tolerance. Do you live in the north or south hemisphere? It was never tested in south IIRC. After I slew to a relatively bright star on the Meridian, I solve&slew and make sure it's centered in the CCD, then I run the polar alignment routine. This is not necessarily per se for the polar alignment routine, but I always like to have everything in order before I begin.

I live in the northern hemisphere. I noted really strange results when measuring Az errors, jumping from east to west within 1 deg range.

Hi!

I can add to this, concerning the altitude part of it. Trying it out extensively tonight, I found the following:

I slew to a star to the west. Plate solve on it so the scope is synced and spot on. Then I do polar alignment, measuring altitude error. First round, I get around 1 degree error. So I click on correct error, the scope slews, and I turn the knob and recenter the star.
So I do it again. This time it is twice as much, 2 degrees, in the same direction (too high): repeat the process of correcting the error, and recenter the star.
And do it again. This time 4 degrees too high.

Something feels wrong here.

When I manually (that is, without the "correct error slew) adjust the scope in the opposite direction to what Ekos suggests, I find a good alignment after a few iterations. So it seems that Ekos for some reason switches directions here - says too high when it is too low.

I doing this on a Losmandy G-11, so a GEM if that matters. DSLR camera.

Magnus

Ok, it's good that this is getting more testing. I am using a polar alignment error equation by Frank Berret. You can calculate the Alignment error here. I added extensive logging to the polar alignment routine, so please next time you test, enable verbose logging to a file in KStars --> Advanced options. The log file will be saved as ~/.local/share/kstars/kstars.log, please share it.

The correct Alt/Az errors might be broken. So I'll explain here how I do it:

1. Altitude correction
A. Calculate a new target latitude = current latitude + altitude deviation
B. Calculate current mount coordinates corrected for the target latitude. That is, the mount Alt/Az would be different if the latitude changes.
C. Slew to the RA/DEC of the Alt/Az calculated above given the current user-defined latitude and local sidereal time.
D. Ask user to center which should take care of the altitude deviation

2. Azimuth correct
A. Get current mount Alt/Az
B. Add azimuth deviation to current Az.
C. Slew to new coordinates
D. Center star

I don't remember if I ever did the math for any of this or if it was purely experimental, but if anyone can figure out if this is problematic, let me know.

Hi!

Verbose output will come.

Do I understand the Altitude process correctly, that it depends on the geographical latitude set for the mount? (I don't have GPS ) There might perhaps be a source of error, at least for me.

Magnus

Hi!

I didn't manage to produce any debug logs yet, because of clouds... but as I was fiddling to set up, it struck me: how sensitive is this polar aligment procedure (altitude) to the postition of the chosen stars? I have a tree to my west and a house to the east, so I tend to choose something either higher up about the horizon than the suggested 20 degrees, typically more like 30 degrees, or I can choose an arena 20 degrees above the horizon but slightly to the north. How would this impact on the procedure? The formulae are not sufficiently transparent for me to understand this...

Magnus

Hi!

As for testing the polar alignment module in Ekos: Last night I did drift align with PHD2, and then measured the azimuth error with Ekos without any adjustment of the scope in between. See the screenshots below for details. It seems that PHD2 repors a smaller error than Ekos. I did not do a debug file this time, forgot to turn on. Then the clouds came so I never go around to do the same on altitude.








Magnus

KStars Debug was not on? Because that would have been incredibly useful. But Ekos error is actually less, not more, than PHD2. PHD2 reports 29 arcminutes while Ekos ~1 arcminute... unless I am missing something.

Hi!

Yes, I realized too late that debug was not on in Ekos. I'll do that next time.

PHD2 shoes 0.29 arcmins error, doesn't it?

Magnus

oh yeah it's 0.29. But the two methods are different and would normally result in slightly different results.

Hi again!

So a new try. Only with azimuth, but with logfile this time. Here are screenshots from PHD2 and Ekos, no adjustment made between them.






and a logfile attached.

Makes any sense?

Magnus