[Jens Scheidtmann]

  Hallo list, hallo Bruce,

https://openphdguiding.org/logs/dl/PHD2_logs_k4iQ.zip

I have an AZ-EQ6GT and have recently switched to a larger focal length on my main scope.

Both PEC and encoders are disabled, when I start the mount using the hand controller. I use the USB port and directly connect the mount to my computer.  

My main imaging train now is a 1000mm Newton with coma-corrector and ASI294mc. The resolution of that is 0.96 arcsec / px. I have William Optics Guidescope 50/200 with an ASI224mc for guiding. This has a resolution of 3.87 arcsec / px. I have EQMOD setup as per documentation on the PHD2 website (https://github.com/OpenPHDGuiding/phd2/wiki/EQASCOM-Settings). Guiding is set to 0.9 sidereal on both axes. 

I have tried to establish a guiding "baseline" as per the dropbox document that has been shared here on this list. The two differences are that I have used Nina's Three Point Polar Alignment and I am using the "stable" version of PHD2 (2.6.11). I fixed the guide camera's exposure to 2.0s. 

The relevant guide sections are as follows, after creating a new profile:

1) Calibration attempt with open clutch on DEC
2) Calibration 

3) Guide Assistant session, with 10 s (!) backlash detected, see below. I accepted the proposals for min-moves.

5) 33 min guiding west of pier.

6) 27 min guiding east of pier (after meridian flip, same target as 4)

7) 15 min guiding with guide algorithm changed to "Predictive PEC" (new target, west of pier)

Backlash Graph: 

Looking at the graphs of the guiding sessions, it seems that my mount seems not to react to guide pulses or only very, very slowly.

 

Is there another setting in the EQMOD I may have missed or that has been added recently, so that my mount does not recognize guide pulses? (I do not even own an ST4 cable.) 

Then at least on Dec, when the mount finally reacts, it overshoots at lot. So, do I have a stiction problem instead of a backlash problem? 

Do I have that stiction problem on RA, too?

How can I make the mount react "better" to guide pulses?

Thanks in advance, & Clear Skies!

Jens

[Bruce Waddington]

Hi Jens.  The guide pulses are behaving normally and I don't see any solid evidence for stiction - the axes aren't "over-shooting" consistently.  Here's what I do see though.  You have a lot of periodic error in the mount and RA performance is likely to be your limiting factor.  Here's a look at the native tracking error in the mount:

Depending on where you are in the worm cycle, the periodic error could be as high as 50 arc-sec peak-peak.  PHD2 is doing a good job of damping this behavior but there is still a residual sinusoidal error in RA of about 2 arc-sec.  The contributors to this error occur at two periods:

The big spike occurs at approximately the native worm period of 480 sec and the other occurs at about 120 seconds.  If you can do a permanent periodic error correction with your mount that would be best.  If you can't, you can use the PHD2 PPEC algorithm that you tried - but get it locked on a specific problem frequency.  I suggest setting the period length to 120 sec and disabling the auto-adjust-period checkbox in PPEC.  Because the sidereal tracking has this much error, you should probably try using shorter exposure times - just be sure that multi-star guiding has multiple stars to work with.  Try 1-second exposures to start.

The Dec axis has a large amount of backlash as you've seen.  Probably more importantly, you're getting large guide star excursions that are coming from an external source. You can see two of those events here (Dec is green):

These are probably caused by something shifting around with your guiding assembly or perhaps cable routing problems.  When these large excursions happen, it's going to take a long time recover because of the large Dec backlash.  You should probably try to improve the backlash situation by adjusting the gear mesh but don't over-tighten things and create worse problems.

You may want to study this document if you're interesting the analyzing the log data:

https://openphdguiding.org/tutorial-analyzing-phd2-guiding-results/

Good luck,

Bruce

[Jens Scheidtmann]

Thank you Bruce for your analysis. Much appreciated!

While checking the gear mesh on DEC, which had a tiny wobble, I noticed a large wobble in RA. I therefore verschlimmbessered gear mesh on both axes.  

I will also reroute my cabling, in order to eliminate swings and movements and in order to reduce wind drag. 

I will get back once I am able to test the changed set-up (as soon as clouds allow). 

If necessary I'll also train the PEC beforehand. 

I read the tutorial, but after reading it I had a lot of questions left.

For me the tutorial zoomed in to some common problems too fast, with-out providing the big picture first.  

What would help me as a tutorial, is a series of questions I'd need to answer when conducting an analysis and then how to answer them. 

E.g.:

The first thing that you looked at was 'raw performance', i.e. "What were the properties of the mount, if it tracked the target with-out guiding?"

Ok, you choose a session with some data, maybe eliminate some outliers caused by wind gust, or some such (I don't know if you usually do that), then right-click and choose "Analyze selected, raw RA"

Then which things do you need to look at: "What is the peak-to-peak in this graph?", maybe other things, that I don't recognize yet, that need attention, if they are present...

Next thing was: "What are the dominant frequencies that play a role in the log?" -> Click on "Frequency Analysis" -> What to look for in the graph? 

You get the picture. I believe that there are a couple of more questions that come to your mind, but that I as a newbie just don't know. 

I wanted to bootstrap the approach from already answered questions, but got stuck, as for the most relevant questions that I found answered here on the list (AZ-EQ6, large backlash), I could not download the logs (already deleted from the website) and a lot of those answers discussed the data, but had no pictures. So I could not correlate with what was displayed to me and I was unsure whether people had similar problems.  

I would be happy to offer some newbie eyes and some technical writing skills and help write something down. It would also help my learning process. 

So which questions do you answer in what order? And why are you doing them in this order?

Thanks again,

Jens

P.S.: "Verschlimmbessern", verb, german, making something worse, while trying to make it better (having best intentions, but failing). 

Am Sa., 12. Aug. 2023 um 19:22 Uhr schrieb Bruce Waddington <bw_m...@earthlink.net>:

Hi Jens.  The guide pulses are behaving normally and I don't see any solid evidence for stiction - the axes aren't "over-shooting" consistently.  Here's what I do see though.  You have a lot of periodic error in the mount and RA performance is likely to be your limiting factor.  Here's a look at the native tracking error in the mount:

[...] 

The big spike occurs at approximately the native worm period of 480 sec and the other occurs at about 120 seconds.  If you can do a permanent periodic error correction with your mount that would be best.  If you can't, you can use the PHD2 PPEC algorithm that you tried - but get it locked on a specific problem frequency.  I suggest setting the period length to 120 sec and disabling the auto-adjust-period checkbox in PPEC.  Because the sidereal tracking has this much error, you should probably try using shorter exposure times - just be sure that multi-star guiding has multiple stars to work with.  Try 1-second exposures to start.

The Dec axis has a large amount of backlash as you've seen.  Probably more importantly, you're getting large guide star excursions that are coming from an external source. You can see two of those events here (Dec is green):

These are probably caused by something shifting around with your guiding assembly or perhaps cable routing problems.  When these large excursions happen, it's going to take a long time recover because of the large Dec backlash.  You should probably try to improve the backlash situation by adjusting the gear mesh but don't over-tighten things and create worse problems.

You may want to study this document if you're interesting the analyzing the log data:

https://openphdguiding.org/tutorial-analyzing-phd2-guiding-results/

Good luck,

Bruce

On Saturday, August 12, 2023 at 6:42:03 AM UTC-7 jens.sch...@gmail.com wrote

[...]