How to make guiding more stable and better?

[Quark Coder (Quark)]

I have a DIY mount based on harmonic drives, controlled by OnStep v5. For guiding, I'm using a ZWO ASI120MM camera with an SVBONY SV165 guide scope. My main imaging setup consists of a Sky-Watcher 150/650 telescope with a 0.8x reducer/coma corrector and a ToupTek ATR2600C camera.

Overall, the setup works quite well, but I'm still struggling to fully optimize the guiding. It seems to work reasonably well, yet I have the feeling there's still room for improvement.

My guiding error is typically around 1.0-1.5 arcseconds RMS. Based on reports from several people with similar setups, I would expect something closer to a stable 0.5-0.6 arcseconds RMS.

What improvements would you suggest, other than buying a higher-end mount?

I'm currently using PPEC because the guiding graph clearly shows a significant periodic error. What's even stranger is that the periodic error itself appears to have its own periodic modulation. The peaks of the periodic error seem to ride on top of a much broader wave, almost as if one periodic component is superimposed on another.

I know that the current guide log shows a fairly large polar alignment error. However, DEC performance is actually not too bad, with about 0.63" total RMS in DEC. Because of that, I suspect the main issue is in RA rather than polar alignment itself.

Is this behavior normal for harmonic drive mounts? What mechanical or control-system effects could cause a periodic error to have another longer-period oscillation superimposed on it?

https://openphdguiding.org/logs/dl/PHD2_logs_38tn.zip

[Brian Valente]

Hi Quark

when you say "DIY", is this something you built yourself and are able to make mechanical tweaks?

Your guidelog did not include a calibration or unguided output (guiding assistant) so feedback is a bit limited

On the one run of about 2.5 hours, your total RMS is 1.2" with RA at 0.9 and dec around 0.8" so i wouldn't say Dec is notably better performer than RA. the polar misalignment effectively means you are guiding in one direction, and therefore any reversal performance issues are hidden

Regarding RA performance, your algorithm settings are not ideal for a strainwave mount. You have your gains dialed down pretty far, so you are forever trying to return RA to zero point. here's a good example, note the smaller lines are the guidepulses:

So upping predictive and reactive to something like 0.9 will help

As far as the RA mechanicals, you can also see there are some large and immediate excursions such as the big dip shown above and noted below, nearly 4":

looking at the raw mount performance graphs below, you can see the primary period of 431 seconds is quite high (first graph, green), around 16". this seems pretty typical of strainwave mounts and you just have to basically brute force guide it out. PHD does a good job of this and increasing your aggressiveness should help here. 

It also shows some period spikes that are harmonics of the period, which again point to mechanical shortcomings. 

The second graph with the gold trace shows that guiding does a pretty good job of reducing your errors, but as the periods are shorter, it is more difficult to guide out. You have plenty of them and then top 3 hover around 0.4-0.5". You can try switching to hysteresis and increasing aggressiveness in RA, and reducing your cadence to 0.5sec exposure time. 

Residual (uncorrected) noted in gold (same graph as above, but zoomed in on amplitude):

--
You received this message because you are subscribed to the Google Groups "Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guidi...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/9ea57dad-9abd-4e54-8fae-81e6414067d9n%40googlegroups.com.

--

Brian 

Brian Valente

astronomy portfolio https://www.astrobin.com/users/bvalente/ 

[Quark Coder (Quark)]

Hi. Thanks for the detailed explanation! By DIY I meant that the mount was manually assembled, meaning the housing was built, harmonic drives and the controller were purchased, and everything was integrated into a single system. I don’t think I can really improve the factory harmonic drives themselves to make them perform better =)

Next time I’ll try to get the polar alignment much better, do proper calibration, and test your suggestions with PPEC and hysteresis. I’ll run some tests and share a report.

[Bruce Waddington]

I'll add to Brian's comments by pointing out that your native RA tracking is probably considerably worse than what we're looking at here.  That's because the long guiding session in the log had the scope pointing at Dec=61, so the observed RA excursions we measure on the guider chip are reduced by a factor of 2x.  That's just a consequence of the spherical geometry of the sky.  If you want to perform tests on mount performance, you should do guiding sessions near Dec=0, the same sky area where the Calibration Assistant wants you to be.  Before switching to hysteresis, you should increase both the predictive and reactive settings in the PPEC algorithm.  As Brian's graphs show, you have a lot of unwanted tracking error at low periods relative to the primary period of 430 seconds and those are more difficult to guide out.  For example, you have a 2 arc-sec error with a period of 143 seconds and a 3 arc-sec error with a period of 217 seconds, so those will necessarily add to the RA rms errors.  With this mount, I suspect you will find it hard to suppress all these error sources, so the real question may be whether the guiding you're getting now is sufficient for your imaging objectives - it may be.

Regards,

Bruce

[Quark (Quark)]

 Thanks for the additional clarification! I’ll definitely try checking the mount’s behavior near zero DEC.  

In 300-second single exposures, I can clearly see stars drifting in different directions.  HFR is around 4.3, which is likely a combination of poor guiding, slightly incorrect backfocus, and imperfect collimation. Hopefully I’ll get everything dialed in soon =)

My goal is to squeeze as much as possible out of this mount by properly tuning the guiding first. Later, I’ll consider hardware improvements.

[Quark (Quark)]

I performed the following:

1. Improved the polar alignment significantly. According to the logs, the error is about ±2 arcmin.
2. Performed a calibration at the target location (yes, I know calibration should ideally be done at DEC = 0, but the entire DEC = 0 area was covered by clouds).
3. Used the guiding assistant.
4. DEC=40

The first guiding run used PPEC with both Predictive and Reactive values increased to 0.9.

The second guiding run used Hysteresis with Aggressiveness set to 120.

Overall, guiding accuracy seemed to improve, reaching around 0.8 arcsec RMS. However, it appears that my mount produces frequent spikes/outliers, which degrade the overall performance and make the guiding results look worse than they otherwise would be.

https://openphdguiding.org/logs/dl/PHD2_logs_xU4f.zip

[Brian Valente]

Quark

I think you summed it up pretty well. The PPEC performed slightly better than hysteresis. Your primary PE is nearly 23" and is well corrected by PHD, but it's still a peak error that doesn't get below 0.30" 

Comparing raw RA (green line) and corrected RA (yellow), you can see guiding is making substantial improvements:

The sky position isn't the greatest for assessing mount performance (Dec 0 is, as you know). but the spikes are not uncommon with these types of mounts. 500ms cadence probably also helped things here. 

overall i'm not sure there's a lot more you can do guiding-wise to improve things. you can certainly experiment but i suspect you aren't going to see any significant improvements on guiding parameters alone.

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/5a18cf37-dd23-40fb-a0c2-3f3c32efa667n%40googlegroups.com.

[Quark (Quark)]

Okay. Thank you for your help!  When the mount is placed on the pier, I’ll try experimenting a bit more with the settings.

[Phil Bordelon]

Brian, is the Raw vs Residual RA graph available in PHD2Log Viewer?

Phil Bordelon
Advantage Software Technologies, Inc.
pbor...@gmail.com
337-298-1015 (cell)

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/CAJa45i4a-jFkz9sRBKZne15%3D_H%2Ba2t51u7Xf3Nv%2BV_%2B8MpDR9g%40mail.gmail.com.

[Brian Valente]

Phil

I will contact you off list. It's not an official release but i can make it available to you

Brian

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/CA%2BHBbzARsqyAah8UxUcD0eLYjra2LD%2Bgpuh9X7rTMCCb%2BzmbOQ%40mail.gmail.com.