Massive guiding excursions with new mount

[Kevin Wisdom]

Hi,

I have recently replaced my reliable Celestron AVX with a Sky-Watcher NEQ6, for which I have Rowan belt modified.

Trouble is, the replacement mount refuses to guide reliably, regardless of payload.

More often than not during a guiding session, massive random guiding excursions (traditionally DEC, but lately RA is affected) are experienced. Often PHD2 does not recover the guide star and if left unattended, the imaging session is ruined.

I have stripped, tuned and rebuilt the mount countless times, dialling out backlash, upgrading bearings. regreasing, but to no avail. I see no obvious mechanical issues to speak of. Cables are neatly secured with no snags.

I've created new profiles from the PHD2 wizard, chopped and changed guiding algorithms, guide cam exposure lengths, run guiding assistant over a worm cycle, calibrated near the meridian or at the point of imaging, swapped from EQMOD to GSS, changed guide rates in RA and DEC but found no cure. 

I use N.I.N.A. for my imaging plans and dither in random pattern every few frames. The dither distance is set correctly for whichever imaging scope I have mounted, as evidenced by past trouble-free AVX imaging sessions.

Needless to say, my polar alignment is more than adequate and I never experienced any of these issues with my AVX. Sometimes HEAVILY unbalancing Dec to be North or South heavy improves things, but not always.

I can't think of any stone I've left unturned, but during my last imaging session I checked the "Reverse Dec output after meridian flip" and "Assume Dec orthogonal to RA" options out of desperation, and it seemed to settle the excursions for the guiding run, but can't be sure if it's a permanent fix, as it was only a single instance.

For reference, I've included some of my most recent guide logs.

Any suggestions, please as I'm at a loss?

Thanks,

Kevin

[Bruce Waddington]

In light of the history of how you've arrived at your current configuration, I think we need to start over.  Please follow all the instructions in the attached document and, even more important, use the PHD2 Upload Logs feature to send us the data we need to help you out.  As an extension to the instructions in the document, we should see a 20-30 minute guiding session on each side of the pier WITH NO CHANGES TO GUIDING PARAMETERS OR ATTEMPTS TO "FIX" SOMETHING.  We're trying to understand what's wrong with your setup before we then determine how to improve it.  You should perform the testing without using NINA or any other session manager, just run PHD2 manually.

Regards,

Bruce

[Kevin Wisdom]

Hi, Bruce

Thanks for your assistance, but I do not see the attached document you speak of?

Unless there's anything else I'm missing, when the skies finally clear I'll upload the log files as instructed.

Regards,

Kevin

--
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/0042f5df-7a6a-419b-b638-4ac36e85785cn%40googlegroups.com.

[Bruce Waddington]

Sorry, this was pretty useless "assistance" I gave you - I forgot to attach the document.  My bad, long day.  The doc is attached to this message now.  Just take your time doing this and don't let yourself start thrashing around.  As I said before, this isn't going to fix anything but it will hopefully tell us what's going wrong.

Good luck,

Bruce

[Kevin Wisdom]

Hi,

Please find my Calibration, Guiding assistant and guide logs each side of the pier as requested.

https://openphdguiding.org/logs/dl/PHD2_logs_Y5tK.zip

Many Thanks,

Kevin

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/0da83789-35ee-4beb-80b3-cdbf0d3cd7can%40googlegroups.com.

[Bruce Waddington]

Unfortunately, it looks to me like there are a lot of mechanical problems you will need to sort out.  I think the starting point is to look at the poor behavior during the calibrations:

In all these graphs, RA is in red, Dec is in Green.  The problem here is that you don't see a smooth, consistent movement of either axis in response to guide commands.  Instead, you see a kind of cogging behavior where the first guide pulse accomplishes very little and then the second one move the axis by an unexpectedly large amount.  You can see it took 12 steps in Dec to move the total distance but 16 steps were required for RA.  This is why you keep getting alert messages and "poor" evaluations from the Calibration Assistant.  In my experience, this looks like a mount that is over-tightened, creating substantial stiction for the low-energy, low amplitude guide corrections.  Obviously, this affects everything down the road.

If we look at the native tracking of the mount during the Guiding Assistant run - in which no guiding is done at all - we see a pretty huge tracking error:

You can see the mount wandered off-track by 50 arc-sec in about an 8-minute period which is quite poor.  

If we look at the Dec guiding behavior, there is further evidence of problems:

At point 1, there is a >3 arc-sec excursion in Dec, a time in which the Dec motor wouldn't have been running at all unless you were using a pointing model of some kind.  This makes me question where you're getting unwanted movement in the guiding assembly that is triggering large guide star excursions.  Once this large error occurs, the recovery doesn't look good.  The first series of south guide pulses have little effect until suddenly you get an over-shoot.  This is followed by another series of non-responsive north guide commands (point 2).  This, again, looks like stiction but it can also be affected by gross imbalance of the payload in Dec.

In addition to these basic mechanical issues, you were also affected by a recurring bug in the GSServer driver in which the driver simply stops responding to guide commands.  We have seen this many times with this driver but not with any others.  This happened 10 times during the course of the guiding sessions you sent us.  You might consider using the EQMOD driver which doesn't have these issues.  If you try that, be sure to create a new PHD2 profile and leave the "Reverse Dec on meridian flip" option un-checked.

Finally, you said initially that you were happy with your dithering specifications but I really wonder about them.  In these logs you were at times  generating dithers of over 1 arc-minute on both axes!  This is huge and with mount performance like you have, the recovery is poor and what little stability you might have achieved is lost.  To determine the size of the dithers you need, you need to calculate the difference in image scales between the main camera - where the dither size matters - and the guiding image scale which is likely to be quite different.

Regards,

Bruce

[Kevin Wisdom]

Hmm, that is odd. I have adjusted the tension in the mount over and over again with the repeated rebuilds. Can't say I feel any stiction whatsoever (apart from the very first Rowan belt retrofit where I didn't open up the clearance slot for the RA belt - remedied with the subsequent rebuild).

Upon slew, there is absolutely no hint of any audible motor strain or binding and the axes move more freely than when I initially got the mount. I made concerted efforts to dial out backlash but not to the detriment of the axes being overtightened. With no payload, they move freely. I can loosen off the Dec thrust bearing counterweight bar nut a tiny amount, but this will introduce a slight end float on this axis, due to the tolerance of my tensioning.

Thanks to limited clear skies, calibration was undertaken in late twilight, so not truly dark. Mount was equally balanced and neither east nor north/south heavy. Would I be best served waiting for darker skies (gonna be a while with the weather here)?

I switched back from GSS to eqmod and sure enough, the pulse guide warnings seemed to lessen, so I'll at least stick with that.

Thanks,

Kevin

On Mon, 10 Nov 2025, 23:17 Bryan, <bryan....@gmail.com> wrote:

Roy

It's essentially impossible to respond to a request for help with no supporting information.  All the detailed guidance that Bruce provided came because Kevin posted his logs.

The following is also posted at the top of the forum home page.

To get help with PHD2, please upload your log files - here's how: How to ask for help with PHD2. 

Bryan

On Monday, November 10, 2025 at 11:05:47 AM UTC-7 mypat...@gmail.com wrote:

Can someone help me too

Roy

On Nov 8, 2025, at 4:56 PM, Bruce Waddington <bw_m...@earthlink.net> wrote:



Unfortunately, it looks to me like there are a lot of mechanical problems you will need to sort out.  I think the starting point is to look at the poor behavior during the calibrations:

<Bad_Calibrations.jpg>

In all these graphs, RA is in red, Dec is in Green.  The problem here is that you don't see a smooth, consistent movement of either axis in response to guide commands.  Instead, you see a kind of cogging behavior where the first guide pulse accomplishes very little and then the second one move the axis by an unexpectedly large amount.  You can see it took 12 steps in Dec to move the total distance but 16 steps were required for RA.  This is why you keep getting alert messages and "poor" evaluations from the Calibration Assistant.  In my experience, this looks like a mount that is over-tightened, creating substantial stiction for the low-energy, low amplitude guide corrections.  Obviously, this affects everything down the road.

If we look at the native tracking of the mount during the Guiding Assistant run - in which no guiding is done at all - we see a pretty huge tracking error:

<GA_Native_Tracking.jpg>

You can see the mount wandered off-track by 50 arc-sec in about an 8-minute period which is quite poor.  

If we look at the Dec guiding behavior, there is further evidence of problems:

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/cc7281ae-0168-4b5f-9c19-96121c609b5en%40googlegroups.com.

<Bad_Dec.jpg>

<GA_Native_Tracking.jpg>

<Bad_Calibrations.jpg>

--
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/05125142-9c26-485a-b079-517288179db3n%40googlegroups.com.

[Bruce Waddington]

I wonder if you've internalized the incredibly small movements and amounts of energy that are associated with guiding.  Over the years, I've found that many people who are mechanically inclined (not me!) have some trouble with that.  I say that because you've been talking about how things feel and sound during slewing.  Most of these mounts typically slew at rates of several-hundred times the sidereal rate.  Your mount is guiding at 0.9x sidereal, so the amounts of energy being applied and the movements requested are smaller by probably 400-600x.  Slewing will simply blow through most kinds of static resistance but that is not the case for guiding - slewing pretty much doesn't tell you anything useful for guiding.  I'm attaching a link to a Mount Exerciser app that may help you with diagnosis:

https://1drv.ms/u/s!Amy5FkXK3OuQgnKd1-7Er-5P3Iad?e=LwPrme

The app can mimic the things that PHD2 does in terms of pulse guiding and calibration, and it can be used indoors or in the daytime with no stars.  If you expose the gears on your mount axes and tell the app to start pulse-guiding, you may be surprised at how little appears to happen.  If you look very closely, you may even see that successive guide pulses don't move the drive system by equal amounts.  Instructions for using the app are included in the folder.

Good luck,

Bruce

[Kevin Wisdom]

I've finally managed to record a guide simulation using Mount Exerciser, as instructed.

Here is my log, run over 500 pulses for your perusal.

Please ignore first few pulses, as I didn't check the sidereal box.

To my untrained eye, with this program, Dec seems fine. What's your opinion?

Cheers,

Kevin

To view this discussion visit https://groups.google.com/d/msgid/open-phd-guiding/b6641092-8e84-4a3e-a4a4-f75fde24c68an%40googlegroups.com.

[Bruce Waddington]

Well, I don't think this tells me anything.  The idea with the MountExerciser is to try to reproduce a problem and then closely examine the gear system - visually -  to understand the cause.  What you did with the MountExerciser doesn't really address any of that.  To go back to the intent, we want to explain two things (RA is always in red):

Why does your mount take so long to correct for large excursions in Dec:

And why do the calibrations exhibit these non-uniform moves on both axes even with identical-length guide pulses:

So the test you ran with the MountExerciser (ME) doesn't really target any of this.  To get something useful from the test, you should do the following:

1.  Reposition the scope to the approximate position of the calibration - that means Dec at around 5 degrees and about 5 degrees east of the central meridian with sidereal tracking on.

2.  Choose a guide pulse length that matches the calibrations - 700ms

3.  Expose the gear system on both axes so you can closely examine what the gears/belts are doing

4.  Choose the ME option for 'uni-directional' guiding.

5.  Run the RA for at least 15 guide pulses West, then run it back for at least 15 guide pulses East.  See if the drive system is moving consistently and smoothly.

6  Do the same thing in Dec moving first North, then back South.  Pay particular attention to the first few guide pulses going South.

The ME app and log aren't going to jump up and down and cry 'Aha' - that will need to come from your close eyeball inspection of what's going on.

Good luck,

Bruce

[Kevin Wisdom]

Hello, Bruce

Come to think of it, we've had a cold snap around here and that has now led me to believe the mount itself has physically contracted since being rebuilt, therefore causing the possible binding present. 

It did not exist when the mount was rebuilt during warmer weather.

After slacking off the worm mesh in Dec then later Ra, sur enough calibration was reported to be good, the Dec excursions seem to be absent and i used the suggested guiding assistant parameters. Good call.

Last night's imaging run was far better, apart from one more wild Dec excursion which was solved by running the Meridian Flip Calibration tool, which advised me the disable "Reverse Dec output after Meridian Flip".

If you inspect my last unattended 5+ hour imaging session, it is now beset with periodic Ra excursions, where when questioned, chatGPT has not put down to any residual binding, but rather mount PE.

It's advised changing algorithm from Hysteresis to the Predictive Pec model and for me to diagnose guiding (without dithers) for a few worm cycles. It also advised changing the Ra Minimum Move from 0.13 to around 0.4 to better match my 3.75 micron, 180mm focal length finder-guider setup at 4.3 arcseconds/pixel, as to avoid guiding around small jumps in seeing.

Instead of putting my faith in AI, what's your expert opinion, please?

Regards,

Kevin

[Bruce Waddington]

I think it's clearly better and the Dec stability is about where you want it.  I think the RA data show some unusual behaviors.  First, look at the amount of drift in RA while guiding was disabled (RA in red):

This is about 4x the amount of drift in Dec so that's a bit unusual.  You might want to check balance in RA, make sure the clutches are tight, and see if there is any "sag" in the guiding assembly.  PHD2 does a good job of guiding this out but it would be better to reduce the amount of guiding needed.  Looking at the frequency components of the RA tracking error, we get this:

The primary component at the right is probably the native worm period, around 480 seconds.  It's actually pretty good for this kind of mount, it's only an 8 arc-sec peak-peak error.  But you can see there are also higher frequency components to the left, at about 180 and 240 seconds.  None of these readily explain the big RA excursions that are causing you trouble.  These could come from instability in the guiding assembly or - wait for it - the RA mesh might still be too tight.  Remember that some backlash in RA is of no consequence for guiding because the RA motor is always running and fully engaged to drive the RA axis west - it doesn't have to reverse direction except for slewing. You could try using the PPEC algorithm in RA,  but I probably wouldn't do that because of all the RA spikes that are not periodic.  You don't want those being included in the correction model.

The AI suggestion that you should increase the Min-Move values is probably rubbish.  The Guiding Assistant directly measures the amount of guide star movement with guiding disabled and makes Min-Move recommendations accordingly.  An independent check of this shows that the guiding cadence in Dec is low, occurring less than 20% of the time.  That's a good target, definitely not chasing the seeing.  If you set the Min-Move to 0.4, you would be saying uncorrected tracking errors of 1.6 arc-sec would be fine with you.  Does that make any sense to you?  It doesn't to me.

Regards,

Bruce