I’d like to see if some users can test their Celestron GEM mount (CGX / CGEM / VX) with the INDI Celestron AUX driver and post the driver logs here. Here are the instructions:
Put the mount in the home position: Counter-weights down and looking at the celestial pole.
Cycle power to the mount.
Enable mount logging in INDI/Ekos (check signature)
Create a profile with Celestron driver (only VX or CGEM or CGX).
Start profile.
Share logs
This is mostly to know the mount initial encoder position on startup. IIRC, some mounts were reporting 0 encoder values on startup, while other were reporting 90 degrees on startup.
2023-01-28T23:49:51: [ERROR] Cannot continue without connection to motor controllers.
2023-01-28T23:49:51: [ERROR] Got no response from target ALT or AZM.
This is the first time I’ve tried to connect my CGX using the AUX driver and connecting it directly to the mount. I’ve followed the AUX driver directions from the webpage.
I did set “last alignment” on the HC before starting Ekos and had some difficulty getting the driver selected properly, but the “CGX Wired” driver option make it much easier than previous attempts.
The driver seems to be working (in daylight), so if there are outstanding issues, I haven’t yet encountered them
Hopefully, my log is attached. (I did slew around a bit after connecting).
So after a number of days of attempting to use the Celestron AUX driver, it’s not working for my CGX mount connected via the PC port. I’ve tried it with/without the handset connected.
The 1st issue is that the driver assumes that the mount is in the index position when starting. This requires I align / polar align the mount (via the hand set) and then power the mount off / on and choose last alignment, to be able to get to the index position. I don’t see a more reliable way of doing this. It would be better (for me) if it interrogated the mount where it was pointing (as it’s likely still pointing at my polar alignment star).
The 2nd issue is that regardless of where I slew the mount, the Mount tab always reports East (pointing West), even after the mount has obviously just done a meridian flip in reaching the target. If I turn on the “Flip if HA > x” checkbox, then it repeatedly tries to do a meridian flip (while not actually doing anything at the mount).
These may easily be setup/configuration errors on my part, but that is not obvious. If I switch back to using the HC usb interface with the Celestron CGX driver then neither of these issues exist.
Sorry to post this right after a stable release, but it’s not critical as there is a working alternative.
Cheers
●Environment
Mount: AVX
AUX connection: Self-made USB serial, self-made WIFI connection (both connected to HC terminal)
Hand controller: Not connected.
As I reported before, I think that the reference point of the home position differs by 90 degrees between the fork equatorial mount and the German equatorial mount.
I think that it is highly likely that you will get different results if you connect directly to the hand controller. (because the hand controller has alignment information)
Result of connecting
・Self-made USB serial → connection not possible
・Homemade WIFI connection → connection
Homebrew USB serial can be used in Windows, but INDI does not recognize it. (also same as before)
Based on the sniffed information, the idea was to completely omit the HC by sending the same AUX packets from AUX driver to MC
as the HC is sending. See PR: Fix sendingWithoutHC 0x40 0xc0 and logging (#758)
Looking at the information, USB ⇆ Nexstar AUX Relay for Celestron Mounts and the USB connection adapter I made seem to have the same mechanism.
In the case of this circuit, it seems that RTS/CTS is also used, so it seems that permission is required on the driver side. (Although it is recognized by the system on Linux, it cannot be connected with INDI. Perhaps the reason why you can connect with CPWI on Windows is because it supports RTS/CTS.)
The self-made WIFI connection created by the same author can be connected and operated without using HC.
However, when I checked it before, the right ascension was off by 90 degrees when GOTO was used with the German equatorial mount. I’m assuming it’s because the driver only has the coordinates of the forked equatorial mount.
The problem with the correct pier side is indeed a bug, I checked the code and the pierSide parameter is ignored for GEM mounts: celestronaux.cpp#L1779
The code for the GEM mounts is: celestronaux.cpp#L1819
I think the following patch shall solve the problem:
de = LocationN[LOCATION_LATITUDE].value >= 0 ? deEncoder : -deEncoder;
ha = range24(haEncoder / 15.0);
if (deEncoder < 90 || deEncoder > 270)
if (deEncoder < 90.0 || deEncoder > 270.0)
{
de = rangeDec(180 - de);
ha = rangeHA(ha + 12);
}
pierSide = PIER_EAST;
} else {
pierSide = PIER_WEST;
}[/code]
Regarding the index position at startup, we sniffed the communication between the hand controller (HC) and the motor controller (MC). Powering up the mount with attached HC is a cold start
and at some time step one has to confirm at the HC that the mount is at index position, that is counter weights down pointing north (see: 745#issuecomment-1463761161)
Answering the final step on the HC with quick align corresponds to 3 AUX commands send to the MC.
MTR_SET_POSITION
MTR_SET_CORDWRAP_POSITION
MTR_CORDWRAP_ON
The Celestron AUX driver now imitates after connecting the HC cold start.
For what it’s worth, I found that on Windows, using ASCOM, my CGEM-II works best if I connect with the CPWI application and send a “Quick Align” to the mount. Maybe the “Use Last Alignment” would also work if you’re not at the index position?
The key was that the mount seems to need some sort of alignment that Indi wasn’t sending last time I tried it (about 8 months ago).
So with the updated code, as you might expect, I’ve able to verify indoors that the Scope appears to be behaving normally with respect to meridian flips. Thank you!
I’ll be able to try it in 5 hours outside.
With respect to initial starting position, I’ll try to do a polar alignment using the HC (without Kstars), cycle the power to the mount to quickly get back to the index position and choose last alignment before removing the hand controller. If there is a simpler solution for startup, that would be great, but this doesn’t seem onerous. The only nagging question is that I presume once last alignment is chosen, the mount will be begin tracking away from its index position.
Thomas, thank you for figuring this out! I actually just fixed it today and didn’t know about your fix, I wish I checked this post earlier! I also verified with Chris that meridian flip works correctly now.
