As 1 deg = 60', 1.7 x 1.2 deg is actually equal to 102' x 60'. If the size of the sensor is w x h and the focal length of the scope is f, then the FOV is given by
And by entering the figures given by @gbeaton we get a FOV of 1.698 x 1.273 deg, which is in line with values reported by Ekos and by online calculators. Now, this value may be a little bit different compared to solved images due to the fact that actual focal length may be slightly different from nominal focal length, see also @kbahey post.
Some questions for @gbeaton:
1. Are you using real hardware or simulators in your test?
2. Did you tried Online solver mode? Do you get the same results (failure)?
3. If you use Load & Slew with a real image instead of Capture & Solve, does anything change?
4. If you manually upload your captured images to astrometry.net, do you get a solution?