Hi all,

I'm having problems with the Astrometry.net solver in Ekos, while trying both online and offline methods. It usually works perfectly fine on the first image I try in a session, and in less than 5s, but then nothing. On any following image, the solver (both online and offline) will work endlessly without finding a solution. I even tried solving twice in a row the same field ("capture & solve" with the sync option, and "capture & solve" immediately again without moving the telescope) and it didn't work. However, if I disconnect my devices, stop INDI, close Kstars, and restart everything, my first image will solve again in less than 5 seconds.

I attached the log I have for that night but couldn't find anything relevant in it. The options passed by Ekos to the solver are "-O --no-plots --no-verify --resort --no-fits2fits --downsample 2 -L 19.3223 -H 26.6952 -u aw -3 119.209 -4 -90 -5 30". I'm using gPhoto driver (for a Canon EOS 5d Mk II) and Celestron NexStar driver (for a CGE Pro mount). Computer is a laptop running Xubuntu 16.04, Kstars and Ekos. The telescope has a long focal length (3950mm) so my FOV is pretty small.

When the first image is solved, the solver tells me to update the FL to around 5600mm, which I did even if it sounds weird (far from the announced FL, not sure why), but it didn't help. I also checked that the solver options update when the telescope is moved, which they do fine.

So I'm a bit out of options now. Does anyone have an idea of what I could try to troubleshoot this?

Thanks

Ok, so I don't see KStars version in the log which probably means you're using a slightly older KStars, so please update everything.
Next, the difference between 5600 and 3950 is huge so something is wrong. What's the camera's pixel size? You install the 4200 index files?

That's odd, I always 'apt update & apt upgrade' before observing nights. On the Help menu of Kstars, it says I'm using version 2.8.7, Build: 2017-11-09T12:28:19Z, which I thought was the last one!

My camera pixel shift is 8.20 um. I did install 4200 index files through the astrometry Debian packages available on Synaptic.

Resolution of the captured image?

Usually I capture images in native format (.cr2). In that case, my files are 5616 x 3744 pixels, which correspond to the maximul resolution of my camera.

However, I also captures a couple of .fits files, and just realised that some of my fits files have slightly different resolution (4365 x 2910). I'm unsure why. I'm testing now INDI with just gPhoto driver, captured two images, one fit and one raw, and both of them have the same resolution, 5616 x 3744.

On the INDI gPhoto control panel - Image Settings tab, I have the correct Width and Height, while Top and Left are both 0. On the Image Info files, same Width and Height, Pixel X, Y and size is 8.20um, bits per pixel is 8 (I tried to set it to 14, which is the correct number, but it doesn't want to change). Note that on the Image Info tab, I sometimes find a different value for Width and Height: 5634 x 3753, which is above my camera's max resolution.

Also, when I open one of my .cr2 files in PixInsight, it tells me this:Note that apart from Thumb size, all those values are above my camera's max resolution.

I'm trying again at the observatory tonight, so I'm gonna keep a close look at my files resolutions.

Oh and something I forgot to mention as well. In the Capture tab under Ekos, I have weird numbers in the Frame section. W is 5616, which it should be, but H is -1524852282, where it should say 3744. I double checked in the INDI settings and I have the right values so I'm not sure what the problem is.

I had a similar problem last weekend, too.
When I first tried the new option, I did it with the simulation configuration. I uploaded an image I had taken and he solved it without any problems. But I didn't think to try it a second time.
Last Saturday, I used the polar alignment tool. It made and solved the three images of the process without any problems. And that's as far as it goes. I couldn't get it to solve any more images all night long. Not in online or offline mode... But I thought it was a problem of my own...

Camera: ASI 1600MC -> pixel:3,8
Telescope: Newton 150/750
Resolution:1.04"x1.04" per pixel