Harmonic mount asymmetric PE, PPEC optimization

[Mustafa Said Cay]

Hello everyone,
this is my first post here so please forgive me for any formal mistakes I may make.
I'm imaging with a Skywatcher Esprit 100EDX, Touptek OAG, SVBbony SV605MC and a non-commercial strainwave gear mount (SkyArrow HD17, videos by the constructor on YT).
As is common with these cheaper SWG mounts the periodic error is rather large. In my case the fact that the PE curve is pretty asymmetric compared to a nice sin function (perhaps because of overlapping PEs by different mechanical parts?) leads to some issues, at least that's what I suspect.

In the attached guidelog

https://openphdguiding.org/logs/dl/PHD2_logs_gfg3.zip

you will see that guiding was pretty good overall throught the night thanks to good seeing and PHD2 beeing generally able to compensate for the huge PE but periodically I have these dips (see attached photo) that correlate to the steepest segment of the PE curve where the guiding can't/doesn't correct despite rather high aggressiveness. 

My current theory is that the PE modell that PPEC generates doesn't account well enough for the asymmetry so my thought was to play around with the PPEC expert parameters. Especially "Approximate data points" and "Signal variance" cought my attention to allow the PE modell to be more flexible. But since I'm no expert I wanted to consult people who are before I made any changes :). 

You'll see that I have to do several rounds of calibration before I hit a good timing where the PE doesn't ruin it as well.

What also seems odd to me at the first sight is that it fails to correct the mount west despite the PE moving west during that steep segment.
0,5s exposures definitely help containing the problem but unfortunately they aren't always feasible (seeing...).

I hope my ramblings make some sense to you and you have some suggestions on how to improve this problem.

Greetings from cloudy Germany

Said

[Bruce Waddington]

Hi Said.  I really can't see what you're worried about here.  The total guiding RMS for your final 2 hour guiding sequence was 0.38 arc-sec which is excellent.  Yes, it's true that the mount has a very large amount of uncorrected periodic error (RA in red), about 46 arc-sec peak-peak:

But this is what the corrected guiding looks like during the same period:

The RA is nicely constrained within an envelope of less than +/- 1 arc-sec.  This is why I don't understand the problem.  At this level of performance, you should only be making judgments based on the results in your final images - namely, FWHM of the stars and their aspect ratios.    Obviously, the guide corrections are being applied in the correction direction, so I don't know what that comment meant either.  In any case, you definitely need to leave the PPEC "expert parameters" alone, those were only put in the UI in the first place for use by the developer.  You also need to at least build a dark library because you are really inviting trouble in the event your sky transparency drops during the night.  Next, you should be using the Calibration Assistant once at the start of the night, trying to get a result that is at least "acceptable".  Re-calibrating on-target as you did here is a poor practice.  Finally, you didn't really follow the help instructions because you didn't include the debug log file so I think we've said as much as we can here.

Regards,

Bruce