I run a Celestorn C8-N (Newtonian) f/4.9 with a Canon T3i, which gives me a 1.28' x .85' FOV. I routinely use offline plate solving in Ekos with no issues whatsoever, as long as it knows what I'm trying to look at, i.e.in the neighborhood of. Sometimes it solves in only one iteration, other times if I'm further away from the target, six or eight iterations. Rarely does it give up and quit, but it has happened. Moreover, in my experience, I find plate solving in Ekos works just fine with my DSLR, but it does take several minutes - less so with my new RPi 4, 4GB.

Not sure, but I think load&slew does not restrict the FOV. If that solves properly, but not in capture mode then I'd suspect the FOV limits are wrong. They are computed from the setup info (detector size and focal length), and if they don't match it will never find a solution.

OT: I had similar problems when I tested my photo lens, a Canon FD 135. It would refuse to solve, even though I had inserted the proper numbers. Took me a while to realize that I had put the 20mm on, and the limit automatically computed for 135mm was just some 20' or so too high. Corrected that, and it solved in seconds....

Sorry - but I’m wondering what you mean by ‘as long as it knows what I’m trying to look at’?
When I’m attempting to polar align Ekos ‘knows’ that the scope is looking near Polaris. Equally, when plate solving at an arbitrary point, the mount ‘knows’ where it was slewed to....

What I have been trying, and it seems to work but I need to crosscheck it a few times more to assess consistency, is to align using the live view from the Canon and the alignment + polar alignment corrections from the mount control (in my case an AZ GTi from skywatcher with the smartphone app as remote control). Good news with this approach is that it is fast, you see the stars up to mag 3-4 pretty clearly even in the video stream (particularly if you use Magic Lantern and put a high LV gain) and you don't burn the shutter.

After two iterations the app was basically suggesting to move the mount so little that I decided to switch to EKOS to try exposing. My stars were not moving at all in like 2 minutes at 250mm. Maybe at longer exposures or longer FL it is not enough though.

Last night we had exactly this issue with a Canon EOS 77D. Only by putting the file format to JPG were we able to do a polar alignment. When set to RAW the resolver would time out. Very strange...

EDIT: For what it's worth, we are using Jasem's libgphoto2 repository.

Here are the align logs of last night. Unfortunately I didn't save the images used. I'll see if I can do that tonight.

$> grep align log_22-42-49.txt
[2019-08-27T22:43:11.738 CEST INFO ][ org.kde.kstars.ekos.align] - "Idle."
[2019-08-27T22:43:11.806 CEST DEBG ][ org.kde.kstars.ekos.align] - solve-field Revision "Revision 0.73, date Thu_Nov_16_08:30:44_2017_-0500.\n"
[2019-08-27T22:43:11.807 CEST INFO ][ org.kde.kstars.ekos.align] - "Detected Astrometry.net version 0.73"
[2019-08-27T22:43:12.692 CEST INFO ][ org.kde.kstars.ekos.align] - "World Coordinate System (WCS) is enabled. CCD rotation must be set either manually in the CCD driver or by solving an image before proceeding to capture any further images, otherwise the WCS information may be invalid."
[2019-08-27T22:43:57.739 CEST INFO ][ org.kde.kstars.ekos.align] - "Capturing image..."
[2019-08-27T22:44:07.991 CEST INFO ][ org.kde.kstars.ekos.align] - "Image received."
[2019-08-27T22:44:08.045 CEST INFO ][ org.kde.kstars.ekos.align] - "Starting solver..."
[2019-08-27T22:44:08.048 CEST INFO ][ org.kde.kstars.ekos.align] - "/usr/bin/solve-field -O --no-plots --no-verify --resort --downsample 10 -L 166.25 -H 275.1 -u aw -3 8.27358 -4 87.3053 -5 60 --config /home/wouter/.local/share/kstars/astrometry/astrometry.cfg -W /tmp/solution.wcs /tmp/fitsOL1720.fits"
[2019-08-27T22:44:08.088 CEST INFO ][ org.kde.kstars.ekos.align] - "Reading input file 1 of 1: \"/tmp/fitsOL1720.fits\"..."
[2019-08-27T22:44:10.052 CEST INFO ][ org.kde.kstars.ekos.align] - "Extracting sources..."
[2019-08-27T22:44:10.453 CEST INFO ][ org.kde.kstars.ekos.align] - "Downsampling by 10..."
[2019-08-27T22:44:23.676 CEST INFO ][ org.kde.kstars.ekos.align] - "simplexy: found 170 sources."
[2019-08-27T22:44:26.203 CEST INFO ][ org.kde.kstars.ekos.align] - "/usr/lib/python2.7/dist-packages/astrometry/util/removelines.py:23: RuntimeWarning: divide by zero encountered in log\n logpoisson = k*np.log(mean) - mean - np.array([sum(np.arange(kk)) for kk in k])\n/usr/lib/python2.7/dist-packages/astrometry/util/removelines.py:23: RuntimeWarning: invalid value encountered in multiply\n logpoisson = k*np.log(mean) - mean - np.array([sum(np.arange(kk)) for kk in k])"
[2019-08-27T22:44:26.216 CEST INFO ][ org.kde.kstars.ekos.align] - "/usr/lib/python2.7/dist-packages/astrometry/util/removelines.py:24: RuntimeWarning: invalid value encountered in less\n badbins = occupied[logpoisson < logcut]"
[2019-08-27T22:44:26.428 CEST INFO ][ org.kde.kstars.ekos.align] - "Reading sort column \"FLUX\"\nSorting sort column\nmmapping input file\nCopying table header.\nWriting row 0\nDone"
[2019-08-27T22:44:28.455 CEST INFO ][ org.kde.kstars.ekos.align] - "Solving..."
[2019-08-27T22:44:31.235 CEST INFO ][ org.kde.kstars.ekos.align] - "Reading file \"/tmp/fitsOL1720.axy\"..."
[2019-08-27T22:44:31.241 CEST INFO ][ org.kde.kstars.ekos.align] - "Only searching for solutions within 60 degrees of RA,Dec (8.27358,87.3053)"
[2019-08-27T22:47:07.621 CEST INFO ][ org.kde.kstars.ekos.align] - "Solver timed out."
[2019-08-27T22:47:07.694 CEST INFO ][ org.kde.kstars.ekos.align] - "Capturing image..."
[2019-08-27T22:47:17.869 CEST INFO ][ org.kde.kstars.ekos.align] - "Image received."
[2019-08-27T22:47:17.884 CEST INFO ][ org.kde.kstars.ekos.align] - "Starting solver..."
[2019-08-27T22:47:17.888 CEST INFO ][ org.kde.kstars.ekos.align] - "/usr/bin/solve-field -O --no-plots --no-verify --resort --downsample 10 -L 166.25 -H 275.1 -u aw -3 9.06742 -4 87.3053 -5 60 --config /home/wouter/.local/share/kstars/astrometry/astrometry.cfg -W /tmp/solution.wcs /tmp/fitsGX1720.fits"
[2019-08-27T22:47:18.086 CEST INFO ][ org.kde.kstars.ekos.align] - "Reading input file 1 of 1: \"/tmp/fitsGX1720.fits\"..."
[2019-08-27T22:47:19.792 CEST INFO ][ org.kde.kstars.ekos.align] - "Extracting sources..."
[2019-08-27T22:47:20.214 CEST INFO ][ org.kde.kstars.ekos.align] - "Downsampling by 10..."
[2019-08-27T22:47:33.780 CEST INFO ][ org.kde.kstars.ekos.align] - "simplexy: found 169 sources."
[2019-08-27T22:47:35.833 CEST INFO ][ org.kde.kstars.ekos.align] - "Reading sort column \"FLUX\"\nSorting sort column\nmmapping input file\nCopying table header.\nWriting row 0\nDone"
[2019-08-27T22:47:37.880 CEST INFO ][ org.kde.kstars.ekos.align] - "Solving..."
[2019-08-27T22:47:38.980 CEST INFO ][ org.kde.kstars.ekos.align] - "Reading file \"/tmp/fitsGX1720.axy\"..."
[2019-08-27T22:47:38.992 CEST INFO ][ org.kde.kstars.ekos.align] - "Only searching for solutions within 60 degrees of RA,Dec (9.06742,87.3053)"
[2019-08-27T22:48:01.991 CEST INFO ][ org.kde.kstars.ekos.align] - "Solver aborted after 44 seconds"
[2019-08-27T22:48:12.786 CEST INFO ][ org.kde.kstars.ekos.align] - "Capturing image..."
[2019-08-27T22:48:23.046 CEST INFO ][ org.kde.kstars.ekos.align] - "Image received."
[2019-08-27T22:48:23.062 CEST INFO ][ org.kde.kstars.ekos.align] - "Starting solver..."
[2019-08-27T22:48:23.066 CEST INFO ][ org.kde.kstars.ekos.align] - "/usr/bin/solve-field -O --no-plots --no-verify --resort --downsample 10 -L 166.25 -H 275.1 -u aw -3 9.339 -4 87.3053 -5 20 --config /home/wouter/.local/share/kstars/astrometry/astrometry.cfg -W /tmp/solution.wcs /tmp/fitszA1720.fits"
[2019-08-27T22:48:23.109 CEST INFO ][ org.kde.kstars.ekos.align] - "Reading input file 1 of 1: \"/tmp/fitszA1720.fits\"..."
[2019-08-27T22:48:25.052 CEST INFO ][ org.kde.kstars.ekos.align] - "Extracting sources..."
[2019-08-27T22:48:25.474 CEST INFO ][ org.kde.kstars.ekos.align] - "Downsampling by 10..."
[2019-08-27T22:48:38.867 CEST INFO ][ org.kde.kstars.ekos.align] - "simplexy: found 160 sources."
[2019-08-27T22:48:40.916 CEST INFO ][ org.kde.kstars.ekos.align] - "Reading sort column \"FLUX\"\nSorting sort column\nmmapping input file\nCopying table header.\nWriting row 0\nDone"
[2019-08-27T22:48:42.941 CEST INFO ][ org.kde.kstars.ekos.align] - "Solving..."
[2019-08-27T22:48:44.061 CEST INFO ][ org.kde.kstars.ekos.align] - "Reading file \"/tmp/fitszA1720.axy\"..."
[2019-08-27T22:48:44.071 CEST INFO ][ org.kde.kstars.ekos.align] - "Only searching for solutions within 20 degrees of RA,Dec (9.339,87.3053)"
[2019-08-27T22:49:05.146 CEST INFO ][ org.kde.kstars.ekos.align] - "Solver aborted after 42 seconds"

One of Jasem's videos suggests that when you start out to set your FOV to 0' x 0' and let the software figure out your exact FOV. Reasoning is, although it takes longer for the first run, there are so many variables that can change the FOV even by a little. Thanks for posting this thread, I've had similar thoughts. I've got a pier and I rarely have to adjust my PA, but even a tiny correction can make a big improvement in guiding.
Part 9: Ekos Alignment Module

I have watched that video. The FOV figure I am using is the one that was returned by ‘Load and Slew’ process so it looks like it is pretty accurate.
In my case I set up the mount anew each evening, so performing a reasonably speedy and accurate PA is crucial. I do have a fresh software installation to test out, on the next clear night.......

I have not tried the PA tool, but really want to. I was under the opinion that it requires an open view of the sky. I have a fairly narrow view between trees and the house... oh well.

What I've been using the app "PS Align Pro" on my phone and my Skywatcher EQ6-Pro mount seems to do a really great job, but I would like to really nail the PA with the the EKOS module.

Oh and I understand your "next clear night"... it's been terrible here in Maryland.

I don’t have a polarscope on my mount, and I am still hopeful that I may not need to buy one! The FOV of my scope is a little over two degrees which I suspect would still be rather narrow in terms of the gap between ‘the trees and the house’ in your situation.
I am hopeful that an earlier posting about using Jpeg rather than RAW images may have some significance.......

Like I wrote, for us changing to JPEG at least made the polar alignment routine work. And I think using JPEG for slewing also should work. However, this is a temporary workaround until it is clear what's wrong. I have used CR2 RAW files (converted automatically to FITS by Ekos) for polar alignment and slewing LOTS of times and it always used to work. Meanwhile just don't forget to set it back to RAW when you actually start imaging because we forgot to do that...

We are trying to solve some of the FITS files on the command line on Linux and the solve-field command times out all the time. Next I hooked up the DSLR camera to my MacBook and tried the solver again. The solver completes in 5 seconds. In the Image Settings tab of GPhoto CCD in the INDI Control Panel, both on Linux and Mac Image is set to RAW and Capture Format is set to RAW. Only when we set it to one of the JPEG formats on Linux, the solver will succeed and it succeeds within a few seconds.

Did anything change in the RAW to FITS conversion in KStars recently?