AZ synscan : strange behavior, can someone explain, please? #49242
See the attached screenshot:
I've turned on "Lock Telescope Centre" so the KStar centre should be aligned to the telescope indicator, but it off by more than a full moon.
On the INDI control panel you can see the coordinates of the scope, it is updated live as the scope tracking.
On the Ekos control panel you can also see the coordinates of the scope, and it is also updated live as the scope tracking.
But the two is differing by more than 2°. Actually the coordinates on the INDI control panel is where the Kstars centre is located on the screen, while the coordinates on the Ekos mount control panel is where the telescope indicator is on the screen. So it seems the "Lock Telescope Centre" locks to the coordinates that is also displayed on the INDI control panel.
But the scope is actually looking at the coordinates that can be seen on the Ekos mount control panel and where the scope indicator is on the screen. So I was able to see the edge of the moon in the scope (quite accurate mount alignment!).
It is not just a visual discrepancy, when I right click on some object on the skymap, and select "Synscan -> goto" it slews the scope off the target about the same offset. It makes the whole system quite unusable.
The offset is somehow connected to the alignment of the synscan controller, it is different each time I repeat the 2 or 3 star alignment process on the mount.
Can someone explain me what is going on here?
ps: I use indi_synscan_telescope driver, my INDI library version is 1.8.3
AZ synscan : strange behavior, can someone explain, please? #52757
I have made some progress in this mystery .
It seems to be related to the sync operation.
I have digged myself into the code of the indi_synscan_telescope driver, and found that it has two type of Goto operation: Goto(double ra, double dec) and GotoAzAlt(double az, double alt), and in case of AltAz mount the Goto operation falls back to GotoAltAz by calculating the Alt/Az coordinates from the requested Ra/Dec coordinates.
In the comment it says:
// For Alt/Az mounts, we must issue Goto Alt/Az
But it is not true, at least not for today's hand controller firmware versions.
The hand controller has two type of Goto commands: 'r' command for Goto Ra/Dec and 'b' command for Goto Alt/Az, and the hand controller happily accepts both for an Alt/Az mount.
Both command works the same as long as the mount is freshly and ideally aligned. But in real life sync operations are needed for correcting imperfect alignments. That works by offsetting the coordinates in the vicinity of the sync star.
And here starts the problems: the Goto command comes in Ra/Dec coordinates from Kstars. The driver converts it into Alt/Az, but the driver does not count for the offset. Sends the Goto to the telescope, the telescope slews to the given Alt/Az position, but the reported Ra/Dec coordinates are offset by the hand controller. So Kstar shows the cross hair for the telescope in an ofset position, which is actually where the telescope points, but not what has been requested to slew to.
I've added a UI switch to the driver by which I can force the driver to issue Ra/Dec goto even for Alt/Az mounts. This way the scope slews to the requested Ra/Dec coordinates, offseted by the sync offset, so it points to the actual right position in the sky, and reports the same Ra/Dec coordinates in the status message, so Kstars shows the cross hair for the telescope in the right position.
What is still not OK and still a mystery is that in Kstars the "Lock Telescope Centre" operation moves the centre cursor to an offset position. I have no idea how Kstars get the coordinates for the Lock operation.