×

INDI Library v1.9.7 Released (29 Jul 2022)

Bi-monthly INDI Library released with new drivers and bug fixes.

Re:Re:Re:Weird guide calibration

  • Posts: 826
  • Thank you received: 139
As I had reported earlier, I continue to have difficulties getting the guider to calibrate. Actually to such an extent that these difficulties prevented me from guiding altogether. The setup is a Losmandy G-11 with a TS107-S main scope (107/700) and Celestron C4-R guide scope (102/1000). Both scopes are equipped with ASI 294 cams and stepper motor focusers. Cable tension/flexure is not an issue as all cables are kept within a cable tube that moves in lockstep with the scopes. This setup worked flawlessly for years, calibration included.

Recentyl though I experience with what I consider "impossible" calibration results. The most recent night started with calibration via the main scope and delivered this result:


Ugh! Another try with the guide scope: 


I aborted. WTF! Earlier I also had calibration attempts where the RA backward movement went exactly in a 90° angle vs the forward movement. This IMO is totally impossible to do (particularly for 10 pulses in a row, all going exactly in the same direction). So in order to SEE what my mount is actually doing I captured a video of RA and DE slews (the DE slew is not included in the video).  


As you can see movement is perfect. No hint of any irregularities. So KNOWING the mount was doing exactly what it should, I immediately had another calibration attempt. Outcome: 


Check out DE backward movement. Can you believe it? This may look like a loose DE axis but I double checked, it was locked firmly. The next thing I did was take another video illustrating RA forward and backward movment utilizing guiding speed (0.5x). The video once again shows the star behaves perfectly. RA forward and backward move in parallel, same with DE forward and backward. 


Right after this I again started a calibration attempt. Outcome:


At that point "candid camera" came to mind. I almost ran out of ideas what else I could try. I ended up searching for my old ST4 cable (Yes, I was desperate indeed). I found it, changed guiding from "Via: Losmandy Gemini" to "Via: ZWO ASI 294MC" and started calibrating again. 
Last edit: 1 year 3 months ago by Alfred.
1 year 3 months ago #72342
Attachments:

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

  • Posts: 826
  • Thank you received: 139

Replied by Alfred on topic Weird guide calibration

Now I got the 90° angle between RA forward and RA backward movement again. Frustrating! What I noticed is there were just 3 green dots (4 at earlier attempts, RA forward) and two blue ones (DE forward) while in fact 5 interations had been configured in guiding options. At least I had learned pulse guiding was not to blame. So I reverted to "Via: Losmandy Gemini" and had another go:

 
Huh!? Not exactly the most beautiful calibration plot ever but it ended with a success. Now I was able to guide, test the new code and collect logs.

I'm at a loss as to what causes these calibration failures and in particular the ridiculous plots. I am pretty sure the Losmandy performs as it should, the calibration plot however goes bonkers. Any suggestion as to what might go wrong or what I could try next would be much appreciated!
 
1 year 3 months ago #72345
Attachments:

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

  • Posts: 382
  • Thank you received: 95

Replied by Kevin Ross on topic Weird guide calibration

Wow! Weird is an understatement!

The first thing I'd do is look for something mechanical. Slew to a star, and either use an eyepiece or a camera quickly looping short exposures. Set the slew speed to something slow, like 2x or 4x, and bump the direction buttons, up, down, left, right, and see if the mount moves as you expect it to.

Edit: I see you did that already, your links to videos were hidden between all those crazy graphs!
Last edit: 1 year 3 months ago by Kevin Ross.
1 year 3 months ago #72355

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

  • Posts: 382
  • Thank you received: 95

Replied by Kevin Ross on topic Weird guide calibration

Since no one else is offering any words of wisdom, even though it looks like a mechanical cause, have you tried eliminating the software as a possible cause? Have you tried doing a calibration with PHD2, either with INDI as the backend, or preferably, on Windows with ASCOM (to completely eliminate INDI as a possible cause)?
1 year 3 months ago #72378

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

  • Posts: 1029
  • Thank you received: 301

Replied by Eric on topic Re:Weird guide calibration

Yes, try to do a manual star cross test by running a long exposure and moving the star with slew calls. Then try the same with guide pulses. This should remove some parameters from the equation.

-Eric
HEQ5-Pro - Atik 314E - Orion ED80T - DMK21 on Orion 50mm
DIY 3D-printed Moonlite and FWheel RGB/LPR
KStars and indiserver on two Atom 1.6GHz 1GB RAM Linux, VPN remote access
1 year 3 months ago #72382

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

  • Posts: 826
  • Thank you received: 139

Replied by Alfred on topic Re:Weird guide calibration

Thanks a lot, Kevin and Eric! @Eric, I'm unsure whether I do understand you correctly. I think the two videos that I posted proof the fact that the mount performs as expected, at guide and slew rates. I did it by video as opposed to long exposure but the results should be the same IMO. @Kevin: No windows in my house! So the ascom route is blocked but PHD2 is something that I should try, indeed.

I plan to disasseble and check the worm gears later today but a preliminary test didn't reveal any backlash or other problems.
1 year 3 months ago #72406

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

  • Posts: 1029
  • Thank you received: 301

Replied by Eric on topic Re:Re:Weird guide calibration

Sorry I wasn't very clear in my comment. The calibration plot shows a drift of about 20 pixels. Given the compression on your videos, I am unable to verify what happens when you slew. Honestly I don't think there is anything wrong with the mount, but exposing for a star cross test would allow you to zoom and check the movement more precisely. Then, as far as I read your tests, you didn't run the cross star with guide pulses. Imagine your configuration suddenly switched to timed pulses (saying weird things here of course). Also, use the INDI panel to remove part of Ekos from the equation.

-Eric
HEQ5-Pro - Atik 314E - Orion ED80T - DMK21 on Orion 50mm
DIY 3D-printed Moonlite and FWheel RGB/LPR
KStars and indiserver on two Atom 1.6GHz 1GB RAM Linux, VPN remote access
1 year 3 months ago #72453

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

  • Posts: 826
  • Thank you received: 139
Indeed, a mechanical issue is highly unlikely as I got perfect calibration plots in my previous session just 4 weeks ago.

      

The mount hasn't been touched since.Got it. Indeed, a long exposure with guide pulses in all directions could much better show what's really going on. I'll report back what happens. Thanks a lot for your suggestion, much appreciated!

CS!
 
The following user(s) said Thank You: Eric
Last edit: 1 year 3 months ago by Alfred.
1 year 3 months ago #72547
Attachments:

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

  • Posts: 826
  • Thank you received: 139
Eric, in my reply I quoted certain parts of your message but for whatever reason the quotes are not displayed in my posting.
1 year 3 months ago #72548

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

  • Posts: 826
  • Thank you received: 139
Recently I got the chance to follow Eric's advice and produce a 50s exposure while sending guide pulses from the Indi panel. 5 pulses of 1000ms in DE+ , RA+, DE- and RA- each. Crop:

 

[The shot was taken through a closed window thus the double DE lines] This IMO underpins my previous assumption that the G-11 works OK and cannot be responsible for the weird calibration plots.  RA+/- movement is perfectly parallel, as is DE+/-. No "curves" or other unexpected moves whatsoever.

What I do not understand though is this: All 20 pulses were of 1000ms duration. I would expect star displacement to be of the same extent in all 4 directions. However, RA+ is roughly 2x that of all others.

I then tried another guide calibration.



OMG. It works! But again, overall RA+ displacement is ~18 pixels while it's ~9 pixels for DE+. Both comprise of 5x 1000ms pulses. What's the reason? I double checked, the mount is well balanced.

Now it looks like I had two distinct problems:
1. Weird calibration plots that come and go as they wish.
2. RA movement that is inconsistent with the pulses that have been sent to the mount.
Last edit: 1 year 3 months ago by Alfred.
1 year 3 months ago #72764
Attachments:

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

  • Posts: 382
  • Thank you received: 95
Could it be because you are too well balanced? That the mount is kinda wandering around within the backlash? What if you purposely misbalance a bit, east-heavy on the RA axis (if northern hemisphere, west-heavy if southern hemisphere), and tail heavy on the Dec axis? I looked back through this thread, and balance was never mentioned until the last post, where you said it was "well balanced".
The following user(s) said Thank You: Alfred
1 year 3 months ago #72766

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

  • Posts: 1029
  • Thank you received: 301
I did not understand what you captured in your screenshot. If it was a static object, a lamp post would be a better target for the guide test to avoid the double lines. And if it was a static object and you were tracking when you did the test, it's normal that declination move differently from ascension, and that the final position be different.

Now unless the star detection algorithm during your calibration was jumping between confusing sources, we still don't have a clue what happened.

-Eric
HEQ5-Pro - Atik 314E - Orion ED80T - DMK21 on Orion 50mm
DIY 3D-printed Moonlite and FWheel RGB/LPR
KStars and indiserver on two Atom 1.6GHz 1GB RAM Linux, VPN remote access
The following user(s) said Thank You: Alfred
1 year 2 months ago #72849

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

Time to create page: 1.654 seconds