The real beauty of the iPolar is that you don't have to shoot three images, rotating the mount each time, and then adjust the alt/az. Once you have it calibrated to your mount (and, I suppose, to your rough geographic position) by a shoot-and-rotate session, subsequently the software just does a live display of RA axis and celestial pole, you put the cross on the dot, and you're done.
Of course that end stage is very similar to the polar alignment assistant.
Is there any chance that some bright spark among us will figure out a way to save that calibration so that Ekos could shoot a single iPolar image, solve it, and jump straight to the "Refresh" stage of the PAA?
The iPolar camera needs be mounted where the optical polarscope would go or an equivalent place on the RA axis. It also needs to be well collimated with the RA axis (just like an optical polar scope).
As I understand it, the reason the mount needs to be rotated in the EKOS algorithm is to allow for the camera not being collimated with the RA axis. The rotartion establishes where on the sensor the RA axis is.
Without rotation you would need the iPolar hardware or equivalent for it to work.
It is possible to reduce the number of images needed in the EKOS algorithm to just two provided it knows: which hemisphere, direction of rotation, whether the camera image is flipped or not. The first and third can be determined from the plate solve. Direction of rotation could be determined as long as the mount is rotated using the EKOS/INDI system.
RPi3 Ubuntu Server 20.04, Windows 10 AMD64, AAEON UP Core Ubuntu Desktop 20.04
Avalon M-Uno, EQ6 Pro, Atik420, ASI1600MM-C, ASI120MM-S, DBK21AU04, ZWO EFW, Optec TCFSi
Vixen R150S, GSO RC8, ST80
Sorry, I wasn't clear -- I was hoping for a streamlined Ekos solution for those of us who do indeed have the iPolar hardware (or at least a hope of having it, if iOptron ever get their COVID production headaches sorted).
For a permanently mounted polar-axis scope like the iPolar, the collimation/cone error should be fixed, and calculable from one set of rotation images, right? I mean, obviously if you bang it all bets are off, but absent that, once you've made the calculations, they should be good for quite awhile. That wouldn't necessarily be true of a telescope on an equatorial mount. It might not get placed in exactly the same orientation in the saddle, or the DEC or RA axis might not be starting from the same spot, even sag in the imaging train might affect things. But a small rigid camera system bolted permanently into the end of the RA axle should give repeatable results.
After all, that's exactly what the iPolar app is doing, right? It stores the results of the computation from taking and solving two images rotated by at least 45 degrees, and derives from that exactly where in the camera's field of view the RA axis is pointing. On the second and subsequent alignments, it just jumps directly to the equivalent of the "Refresh" stage in the Ekos PAA, with the correction vector already computed and a synthetic "star" (the RA axis cross) already at its base. The target reticle (the red dot) is at the celestial pole. You just switch it on and it solves and puts up the cross and the dot, and you adjust the screws to make them coincide.
It makes me wonder what iPolar has got that Polemaster hasn't got -- unless I misunderstand the PM process, that always makes you reshoot the calibration every time too. Is iOptron just more tolerant of error? Or is it just a matter of having the software? ("Just", he says. As a software developer, I know how many sins are encompassed in that little "just".)
I suspect that PoleMaster have discovered that most people can't keep the camera aligned between sessions while iOptron haven't discovered this yet, or are not admitting it. Maintaining alignment to the sort of precision people expect is not trivial.