×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

Guiding problem after recent update to 2.9.4

  • Posts: 1119
  • Thank you received: 182
Hi Jasem,
It looks like the scheduler is working now, as far as I have tested it so far. Thanks to you and Eric for that hard work!

But I wonder whether something happened to the Guide Module during the update. That seems to be out of whack now.

I ran my Messier Marathon schedule last night and it started picking off one target after the other rather nicely, tracking, focusing, aligning all went fine. However, when it came to initiating autoguiding (using the internal guider, not PhD2), it would fail to clear the calibration and instead use the previous guide star window (the green box, set to 32) for guiding, although that would now point to empty space. Only after about a dozen failed attempts would the module then select a new guide star and guiding would work fine from then. However, by that time the object that was previously perfectly aligned in the center of the frame would have drifted uncontrollably, usually now to the edge of the frame, basically destroying the usability of that image sequence. To recalibrate the mount, I had to manually clear the calibration data every time, it no longer happened automatically.

Moreover, instead of waiting for the guide module to settle on a new guide star, the capture module now starts to expose prematurely while the mount is still drifting, making for interesting zigzags in the first image of the sequence.

Then after performing a meridian flip another problem surfaces: Instead of swapping the calibration or recalibrating, the guide module now apparently uses the old calibration data, which then sends the mount way off course and results in huge errors.

The log file is too large to attach, I have put it into my Dropbox. You can download it here:

www.dropbox.com/s/884hw5e8b0vw3oc/log_21-28-55.txt?dl=0

It looks pretty chaotic to me. I would have thought that the guide module would work independent of the scheduler?

Thanks for looking into it. Let me know if I can do anything to help.

Jo
5 years 11 months ago #24978

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

Sorry for the regression. Clearing calibration was disabled which now that I think about it a big mistake. I'm trying to figure out now the best to know when to clear calibration. Using pier side property is one option, but only a handful of mounts define that. Doing it like before for each job, even if they're on the same side of the meridian, is quite inefficient.

Perhaps we can simply have two groups. Group E (0 to +180) and group W (+180 to +360), when tracking from one side to another, then we clear calibration.. now this is NOT the same thing as pier-side, but it's generic enough for all mounts. Therefore, even if we occasionally clear calibration, it's not as bad as clearing it on each scheduler object.

Any thoughts?
The following user(s) said Thank You: Jim, Jose Corazon
5 years 11 months ago #24979

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

  • Posts: 1119
  • Thank you received: 182
It may be a little inefficient to clear calibration when switching to a new object, but frankly I never considered that a major issue and it feels almost academic. It really only comes into play when running a list like the Messier Marathon, but that was designed only to test how fast the scheduler could go through it and whether the scheduler could reassess the visible objects and restart after reaching the end of the list.

If it were just for me, I would say change the guide module back to exactly the way it was before. That worked perfectly and the minor inefficiency would only come into play if one wanted to rapidly move through a list of many objects. Realistically, what most users would probably want to do is just image a few objects every night and then have the scheduler work them off, so that small inefficiency doesn't matter. And if they only want to use it for GoTo and observing, then there is not even a need to autoguide, so calibration is a moot point anyway.

Along the same lines, in order to speed the calibration up I reduced the iterations from 5 to 3 with no ill effects whatsoever. In fact, the guiding seemed to improve after that. With that change, calibration in both axes usually completes in less than 1 minute. Most users would not even notice.

Also, a little redundancy wouldn't necessarily be bad, especially when switching to objects with different declinations.

Best regards and thanks for looking into this so promptly,

Jo

PS: One more thought regarding the E vs W option for clearing calibration: Would that wreak havoc when the mount tracks past the meridian but has not yet reached the hourly angle at which a meridian flip has been scheduled to be initiated? The default settings are HA +0.1h. If the guide module were to recalibrate starting at HA of 0 then those frames would be ruined and would have to be manually discarded afterwards.

It feels simpler to just accept the little redundancy and go back to the tried and proven configuration.
Last edit: 5 years 11 months ago by Jose Corazon. Reason: Added the PS
5 years 11 months ago #24981

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

Thanks for the feedback. I now added an option which is enabled by default that would reset the calibration. A new build should be out in 1-2 hours.
The following user(s) said Thank You: Teseo, Tom, Jose Corazon
5 years 11 months ago #24982

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

  • Posts: 1119
  • Thank you received: 182
I tried it last night and everything worked perfectly.

Thanks, Jasem!
5 years 11 months ago #25022

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

Time to create page: 0.372 seconds