A few nights ago I tried to use the Polar Alignment module of KStars/Ekos to obtain an accurate polar alignment without the need to spend time drift aligning. Ekos had other ideas though.

Polar Alignment

The way Ekos works is you slew the scope to a star in the south and then east as you would for a normal drift align. Ekos takes an image and plate solves using a local (or online) install of Astrometry.net to determine exactly where the scope is pointing.

Ekos then slews the scope 30 arcminutes in RA only and takes/solves a second image. It uses the result of those two images to calculate1 the polar alignment error. At least, that’s the idea.

The Debian Jessie version of Ekos would not produce a consistent error in either axis. I upgraded to the latest version of Ekos (using a SID install on a spare drive) and whilst the problem remained, the new version included a preview option that displays the image used in the plate solve. Seeing this, the issue became immediately apparent.

Ekos was taking the 2nd image before the scope had finished slewing 30 arcminutes which results in images like:

Ekos failing to wait for slew to end.

Ekos failing to wait for slew to end.

The plate solve then either failed or gave wildly inaccurate results due to the streaks of stars across the image. Fixing this would have to wait for another night and a quick drift align was called for so I could move on to the nights other goal, testing out an off axis guider and Lodestar guide camera.


As the guide graph in the image shows, RA tracking (green line) was fine but the DEC error (blue) accumulated at quite a rate. I had DEC guiding disabled as the poor polar align resulted in DEC corrections making RA even worse and caused the guiding to become too unstable and fail. DEC axis may need looking at in the future to allow its use in guiding, although with a good enough polar alignment it should be possible to guide RA only.

Guide error graph.

Guide error graph.

Despite the DEC issue, I managed to get RA guided exposures of M27 with much rounder stars than in the past, although I opted to limit exposures to 30 seconds.


M27, 30x30s @f6.3

M27, 30x30s @f6.3

Messier discovered this nebula in 1764

July 12, 1764. 27. 19h 49m 27s (297d 21' 41") +22d 04' 00"

Nebula without star, discovered in Vulpecula, between the two forepaws, & very near the star 14 of that constellation, of 5th magnitude according to Flamsteed; one can see it well with an ordinary telescope of 3.5-foot [FL]; it appears of oval shape, & it contains no star. - Charles Messier

At a magnitude of 13.5, Messier was unable to see the nebula’s central star in his telescope. The star is noted for being the largest known white dwarf with a solar radius of 0.055 +- 0.02 2

In terms of exposure duration, I could really do with more data to work with. There’s also a small issue with the stars in the upper right bulging a little. Whether that’s down to the dec drift, slop in the optical train or slight collimation problems, I’m not yet sure. The sky was too unstable to verify collimation other than very roughly which also made focusing difficult.

Despite numerous issues, it’s an improvement on my previous attempts and I’m now considering imaging all the Messier objects. The larger ones may provide an opportunity to try out mosaics.

  1. Determining Polar Axis Alignment Accuracy - Frank Barrett (PDF) ↩︎

  2. Measuring solar radius ↩︎