×

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

Bi-monthly release with minor bug fixes and improvements

Internal autoguide in EKOS makes thins worse

  • Posts: 9
  • Thank you received: 1
I tried for the first time to use autoguide feature of EKOS. I'm using the autoguider only for RA (also for calibration).
A strange thing happens. Calibration phase works as intended, meaning that I can see movement on RA axes (on the guiding camera image) and I can hear the mount motor working. But when autoguiding phase starts does the opposite thing that is supposed. Instead of correcting the (periodic) errors of tracking, it amplifies them. Over the time the "corrections" in RA movement become bigger and bigger. My assumption is that the corrections are going in the wrong direction. Instead of correcting the RA axis, they are making axis to go in the opposite direction. But I may be wrong.

Can anybody explain that ?

My setup is:
- OnStep controlled mount
- guiding camera: converted Logitech C270 webcam (driver indi_v4l2_ccd)
- guiding via OnStep driver (directly)
4 years 1 month ago #49487

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

Turn on guide logging and attach the log from it.
4 years 1 month ago #49564

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

  • Posts: 9
  • Thank you received: 1
First evening with clear sky. Here is the log file.

File Attachment:

File Name: log_18-16-...2-22.txt
File Size:64 KB
4 years 1 month ago #49959
Attachments:

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

  • Posts: 152
  • Thank you received: 20
Do you have the guiding log guide_log.txt? On linux it should be located in ~/.local/share/kstars.
Last edit: 4 years 1 month ago by Greg.
4 years 1 month ago #49961

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

  • Posts: 9
  • Thank you received: 1
Found it.

File Attachment:

File Name: guide_log.txt
File Size:1 KB
4 years 1 month ago #50061
Attachments:

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

  • Posts: 984
  • Thank you received: 160
It is important to set control parameters correctly, "proportional gain" in particular. Did you take care of it?
The following user(s) said Thank You: Raul Chirea
4 years 1 month ago #50064

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

  • Posts: 1119
  • Thank you received: 182
And those control parameters vary wildly from mount to mount as I found out with my iOptron SmartEQPro+. There, guide rate has to be set to 1 (not 0.5), and proportional gain in RA to 200, but on DEC to only 67. With that I finally get the mount to stay on target.
Also, try the ALWAYS RECALIBRATE AFTER MOUNT SLEW setting in the guide calibration module. That can make an enormous difference.
The following user(s) said Thank You: Raul Chirea
4 years 1 month ago #50066

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

  • Posts: 152
  • Thank you received: 20
If I graph your RA Error it looks like over correction is happening. Note the end of it:



This is the correction being sent to the mount. Note that it flatlines toward the end. This is probably due to a max move limit.



You probably won't recover from that. Try adjusting the proportional gain for your mount as was suggested.

You can also try to see what the mount is doing unguided by starting guiding with the corrections disabled. This will tell you the "natural" guiding needs of the mount and could help with further tweeks.
The following user(s) said Thank You: Raul Chirea
4 years 1 month ago #50067
Attachments:

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

  • Posts: 9
  • Thank you received: 1
I played around with "Control Parameters"/"Proportional Gain" and, indeed, it makes auto-guiding work as intended. It isn't yet perfect, but there are other factors at play (my guiding hardware is DIY and needs improvement). Anyway, on a longer exposure I can already see the difference.

Many thanks to all of you, guys, for your help !

By the way, never heard about proportional gain before. Is there any documentation that explains what it does ?
4 years 1 month ago #50209

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

  • Posts: 163
  • Thank you received: 26
The autoguider control loop consists of a measurement (the image from the guide scope) and an actuator (the mount that responds to guide pulses) and the controller (Kstars). The autoguider measures the position of a guide star, knows where it should be and based on the difference between the two (the error), sends a signal to the mount to correct this. How large this signal will be, is dependent on a few gain settings.

Proportional, Integral and Derivative (PID) are the gain factors that are used in the commonly used PID control loop.
See:
en.m.wikipedia.org/wiki/PID_controller

The stability of the control loop is strongly dependent on these gain settings. It's easy to get it wrong. The settings are a bit of a trade-off between stability, fast settling time and remaining error.
Last edit: 4 years 1 month ago by Bart. Reason: forgot controller in control loop.
4 years 1 month ago #50211

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

  • Posts: 163
  • Thank you received: 26
...Which makes me think now:
Would it be interesting if there is a PID self-learning button? I wouldn't mind spending some time developing/ implementing a autotuning alghorithm if there is interest (Jasem?).
4 years 1 month ago #50212

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

  • Posts: 1119
  • Thank you received: 182
That would cool! I could definitely use that.

Jo
4 years 1 month ago #50215

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

Time to create page: 1.073 seconds