×

INDI Library v1.9.2 Released (14 Sep 2021)

Bimonthly Stable INDI Library release introduces new drivers and fixes for existing ones.

Crazy slews with Celestron CGEM and Capture and Solve

  • Posts: 10
  • Thank you received: 1
Hi, I hope you can help me solve this problem!

I have a Celestron CGEM mount with StarSense. I connect to the mount using Kstars/Ekos with the Celestron GPS driver, and essentially all of the functions work correctly, including GOTOs from Kstars, focus commands within Ekos, etc. However, the one feature that is not working for me is Capture and Solve under the alignment tab.

Here is the problem. I perform a Capture and Solve using an Atik One camera. The Star Solver reports a valid solution and some modest offset to point the telescope, usually between two and five arcminutes. Then, the telescope slews like crazy, winding in RA and Dec, sometimes orienting the telescope dangerously close to the mount and straining the camera cables.

After some frustrating experimentation, I found that these crazy slews occur for about 7 out of 10 Capture and Solve attempts. Sometimes they work, and the telescope performs an appropriate small slew, but mostly Capture and Solve with Slew to Target sends the telescope spinning.

In one experiment, the telescope finally did manage to recover pointing after a crazy slew. However, I usually have to power down the telescope to avoid collision with the pier.

For comparison, I also have StarSense on a SkyWatcher mount (EQ8-Rh), but Capture and Solve works correctly (perfectly, really) there. I have also reverted back to the NexStar+ HC on the Celestron CGEM, and Capture and Solve worked fine with NexStar+. The problem seems to be StarSense on the Celestron CGEM DX mount.

On the Celestron CGEM mount, both StarSense and the motor controllers have been updated to the latest versions (1.20 for StarSense, 6.51 for the motor controllers). I am running Kstars 3.5.0 Stable on Ubuntu 20.04.

I did some experiments last night, and I've attached the log files that were generated.
9 months 3 weeks ago #65039

Please Log in or Create an account to join the conversation.

If I had to guess, I think StarSense applies its own handling of the sync points (since it effectively is a built-in solver). So this might explain the wild swings with it connected.
Jasem Mutlaq
Support INDI & Ekos; Get StellarMate Astrophotography Gadget.
How to Submit Logs when you have problems?
Add your observatory info
9 months 3 weeks ago #65083

Please Log in or Create an account to join the conversation.

  • Posts: 10
  • Thank you received: 1
Just to emphasize the point: Starsense on the Skywatcher mount works flawlessly with Slew to Target. I guess the communication between (Celestron) Starsense and Skywatcher somehow differs from the communication between (Celestron) Starsense and Celestron CGEM.

FWIW - I tested Slew to Target with other Celestron CGEM mounts (it is a small college observatory), and so the issue seems to be systemic. I'd like to stay with Starsense if possible. I'll see if I can find a workaround.
9 months 3 weeks ago #65090

Please Log in or Create an account to join the conversation.

  • Posts: 10
  • Thank you received: 1
A workaround: It seems using "Sync" on the Align tab works well with Celestron Starsense. I was able to use "Sync" on a field and then slew using KStars. For most targets that I tried, this procedure got the pointing to within a few arcseconds, which is extraordinary for a Celestron CGEM mount. One target, NGC 40, failed, and the resulting slew ran the camera into the pier. I suspect the Starsense pointing model may have been poor at high declination; another Starsense field nearer the pole might have helped.

It's a mystery to me why Sync + Kstars slew works but Slew to Target is problematic. Nevertheless, the workaround should be fine for my needs!
9 months 2 weeks ago #65170

Please Log in or Create an account to join the conversation.

  • Posts: 554
  • Thank you received: 138
It's not possible to say what is going on without good information. We need logs of the communication between the mount and EKOS. Something that shows the positions read before the sync, the sync command and positions and slew commands afterwards.
9 months 2 weeks ago #65180

Please Log in or Create an account to join the conversation.

  • Posts: 10
  • Thank you received: 1
Just a quick question - I attached logs to my first post, are you asking for information that is not in the logs?

EDIT: I couldn't see the attachments to my first post. Weird. I've attached the logs again to this post.

File Attachment:

File Name: astrometry...2-30.txt
File Size:3,393 KB

File Attachment:

File Name: log_20-17-...2-30.txt
File Size:1,773 KB
Last edit: 9 months 2 weeks ago by Jack Gallimore. Reason: Added logs
9 months 2 weeks ago #65195
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 554
  • Thank you received: 138
The log-29-... file shows communication with the mount but the relevent area is buried among a lot of things that are irrelevant to this.
I've found a sync command at 20:24:23, it's close to where the mount is pointing. There's a slew which completes at 20:24:30 which sems file but there is a message saying that the target accuracy is not met.
One thing is that the last position check is before the mount has finished slewing.
[2020-12-26T20:24:30.719 EST DEBG ][           org.kde.kstars.indi] - Celestron GPS : "[DEBUG] CMD <e> "
[2020-12-26T20:24:30.810 EST DEBG ][           org.kde.kstars.indi] - Celestron GPS : "[DEBUG] RES <41B5B100,09F0BB00#> "
[2020-12-26T20:24:30.810 EST DEBG ][           org.kde.kstars.indi] - Celestron GPS : "[SCOPE] RA-DEC ( 6:09:37,13:58:43) "
[2020-12-26T20:24:30.810 EST DEBG ][           org.kde.kstars.indi] - Celestron GPS : "[DEBUG] CMD <p> "
[2020-12-26T20:24:30.857 EST DEBG ][           org.kde.kstars.indi] - Atik One : "[DEBUG] ParDeviceLibUSB::In - OK!! "
[2020-12-26T20:24:30.876 EST DEBG ][           org.kde.kstars.indi] - Celestron GPS : "[DEBUG] RES <E#> "
[2020-12-26T20:24:30.877 EST DEBG ][           org.kde.kstars.indi] - Celestron GPS : "[DEBUG] latitude 40.9511, sop E, PierSide W "
[2020-12-26T20:24:30.877 EST DEBG ][           org.kde.kstars.indi] - Celestron GPS : "[DEBUG] CMD <L> "
[2020-12-26T20:24:30.880 EST DEBG ][           org.kde.kstars.indi] - Celestron GPS : "[DEBUG] RES <0#> "
[2020-12-26T20:24:30.882 EST INFO ][           org.kde.kstars.indi] - Celestron GPS :  "[INFO] Slew complete, tracking... "
[2020-12-26T20:24:30.882 EST DEBG ][           org.kde.kstars.indi] - Celestron GPS : "[DEBUG] raoffset 38.1, SlewOffsetRa 35.6 arcsec "
[2020-12-26T20:24:30.885 EST INFO ][     org.kde.kstars.ekos.align] - "Slew complete. Target accuracy is not met, running solver again..."

That's all I've found, No obvious problem.
One thing, what sync tolerance are you using? It seems to be an arc minute or less. Are you expecting too much from your mount?
9 months 2 weeks ago #65283

Please Log in or Create an account to join the conversation.

  • Posts: 10
  • Thank you received: 1
I activated logging on all devices because my concern was that a dev would respond that "there isn't enough information," and I don't get many clear nights. Better too much information than too little.

I usually set the tolerance to 60 arcseconds, but sometimes I forget, and the default is 30 arcseconds. If there is a way to change the default, I would love to know it.

OK, so what I'm hearing is that there is no information in the log file that can be used to debug the crazy slews with Capture and Solve. As I wrote above, if I let the crazy slew finish (that is, if it looks like there is no risk of running the camera into the pier), the telescope will return to the target, usually with improved pointing.

If there is no way to see the behavior in the log file, then the best I can do is perhaps try to capture the behavior on video. I'm open to other ideas!
9 months 2 weeks ago #65286

Please Log in or Create an account to join the conversation.

  • Posts: 554
  • Thank you received: 138
There may be a crazy slew in the log data but it's dificult to tell with no information about where in the log data to look.
If you can give the time of a crazy slew then it might help.


The trouble isn't the quantity of the data it's the quality. There's 15,000 to 20,000 lines in the log file and finding the relevant place in that lot is going to be time consuming. Most of it is irrelevant.

Try turning the debug logging off for the cameras, most of the junk is from them.

A video won't provide any more information. I already believe you.
9 months 2 weeks ago #65289

Please Log in or Create an account to join the conversation.

  • Posts: 10
  • Thank you received: 1
We finally had a clear night, and I was able to capture a crazy slew in the log. Some information:

I picked some random star on KStars to align on. I did a quick focus, and somewhere between 18:05 and 18:06 I initiated an alignment with the "Slew to Target" option. I set the accuracy to 60 arcseconds. The reported offset was something like 9 arcminutes; I could see the star near the corner of the field-of-view. The telescope began to slew rapidly on the dec axis, and just after one rotation I had to start pulling cables because they were starting to wind up. I eventually turned off power to the telescope (it was still spinning) and tried to shutdown Ekos as gracefully as possible. As a result, the alignment procedure never got a second image.

I removed all of the setup lines from the log to just before the start of the alignment. I hope this is easier to read!

File Attachment:

File Name: log_17-55-27.txt
File Size:275 KB
9 months 1 week ago #65723
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 554
  • Thank you received: 138
Thanks, that shows the crazy slew.

I get the impression this is a firmware issue in the HC. All the commands look fine except that after the mount has slewed through +90 in Dec the focus position is reported incorrectly.

I see it is Starsense version 1.20 but what is the exact version? There are three or four more numbers available on the HC Version display.

Chris
9 months 1 week ago #65798

Please Log in or Create an account to join the conversation.

  • Posts: 10
  • Thank you received: 1
Sure, from the hand controller, I find
Ver: 01.20.20244
Bld: Sep 1 2020
9 months 1 week ago #65799

Please Log in or Create an account to join the conversation.

Time to create page: 1.316 seconds