I'm getting spikes in DEC

[Joel Short]

Hi Andy and Bruce.  I need to pick your brain....again :)!

I have a new to me Mach1GTO that I have had out under the stars for maybe 4 nights (2 full nights).  The first three nights guiding was great to excellent using a 152mm apo tiplet at f/8 (1200mm).  Last night I noticed right away that I was getting large, regular spikes in DEC.  So much so that pretty much every sub over the course of the night was ruined.  Backlash compensation is turned off.  Previous nights I was only imaging one target, NGC3938 (in UMa).  Last night I imaged two targets, Sh2-290 (in Cancer) and then NGC3938 again.  The DEC spikes were more pronounced in the Sh2-290 subs, but still obviously present in the NGC3938 subs.

I confess I do not know enough to properly analyze the log files.  Below is a link to download some logs.  There are two log files, an earlier log from 4/4/16 where all my subs had nice round stars, a log from last night 4/12/16, and the debug log from last night.  Would you mind taking a look and see what you think?  

https://www.dropbox.com/s/pbb5o74vr8fgp25/PHD2.zip?dl=0

joel

[Andy Galasso]

Hi Joel,

I looked at your guide log.   I do not have an explanation, but I will make some comments. Since you are using 2.6.1dev6 we have information in the log about actual directions: up is North for dec and East for RA.

One thing I noticed is the pattern of the spikes was always the same.  First dec guiding would proceed normally with a small amount of southward drift being corrected with occasional north guide pulses (A section).  Then we see the star move off to the north (always north) with PHD2 issuing a series of south pulses trying to get it back (B-C), with the mount not responding for a while (backlash) until it finally responds (C) and overshoots.

Another observation is that in during the Dec spike events RA seems to be spiking as well (correlated with dec). We know you have a good calibration (bad calibration can show up as correlated ra and dec), so it does seem like there is something shifting in a direction that has components in both the RA and dec directions.

I think the fundamental question is what could be causing this northward & eastward shifting (B-C)?  Could there be a cable dragging somewhere?  The eastward drift and Westward RA corrections would be consistent with something dragging (pulling back to the east).

Andy

[Bruce Waddington]

I completely agree with Andy's analysis and I'll just add a few points.  I can see the same behavior in your earlier log, the difference being that the size of the displacements is not as large (2 arc-sec instead of 5 arc-sec).  As with most of these things, there are two things to consider: 1) what is causing the mount to suddenly move north, and 2) how is the mount behaving during the recovery cycle.  The first question is the more important one IMO and to be clear, it is not a guiding problem.  Something is causing the mount to move this way and since it starts with Dec and we know the Dec motor is not running continuously, it is likely something external to the mount drive system.  My first guess is that you might have a cable that's binding such that it tugs on the scope, holds against the resultant guide corrections, then lets go.  I saw your posts on the AP forum so I know you've done through-the-mount cabling and have everything fastened down tight.  Well, maybe too tight.  Any single cable that's attached to something on the scope (i.e. moves with the scope) must be able to move without interference from a stationary surface or any other adjacent cable.  Since you're guiding at 1 a-s/px, it doesn't take much to create guide star displacements.  That's the place I'd start looking.  This sort of thing could also explain why there is always a smaller-magnitude shift in RA shortly after the Dec displacement begins.

 

The second part of the question, how the mount responds, could involve questions of gear mesh and backlash in the mount.  You can measure the backlash directly using the Guiding Assistant.  Since your mount has been shipped multiple times, it wouldn't be surprising if the Dec mesh needs to be adjusted.  But I would first want to eliminate the large displacements.

 

If you can't identify a cable problem or something else that is deflecting the mount, you'll probably need to invest some time to narrow things down.  Here's one approach:

1. Move to the same part of the sky where you had the problem before -same hour angle and Dec.  Run a guiding session to see if the problem recurs.  This might take as long as 30 minutes based on what happened before.

2. Assuming you can duplicate the problem, disable Dec guiding (Dec guide mode = none) and repeat the guiding run.  If it recurs, you can be entirely sure that Dec guiding has nothing to do with it.

3. To be even more thorough, disable all guide output from PHD2 and just watch what the mount does.  How long you can run this way will depend on the size of the tracking region (make it huge), your polar alignment, and how much periodic error you have in RA.

4. Consider routing some duplicate cables outside the mount just for the purpose of testing.  I don't have any experience with in-mount cabling, so I don't know how hard this would be.  You would only need to have duplicate cables that are attached to devices that are moving around with the scope. Cables going into the CP3 control box are presumably entirely stationary.

 

Once you have clearly demonstrated you don't have a guiding problem but still can't identify the source, you can probably get some help from AP.  And you can always come back to us with additional data as you collect it.

 

Good luck,

Bruce

 

---

--
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.
For more options, visit https://groups.google.com/d/optout.

[Joel Short]

Hi Bruce,

I didn't realize my replies to Andy were not going to the google group, so here are my previous replies:

___________________________

Hi Andy and thanks for the reply.  I really have no idea what could be dragging.  All my cables are routed through the mount and they are secured to the OTA.  Tonight I got things going again and noticed the same general pattern.  So I ran the guiding assistant and enabled backlash compensation as recommended.  Hallelujah!  For the past 30min the Dec spike seems to be gone.  However, now I'm seeing RA spikes just like the DEC ones were last night.  I'm experimenting with RA aggressiveness etc to see if I can cut that down.  

This hobby, which I love, can be maddeningly frustrating!  :)

Thanks again

___________________________

I'm using AP ASCOM 5.08.07 (the latest).  In the driver both RA and DEC backlash comp is set to 0.  

So far tonight the PHD2 DEC backlash comp has completely resolved the DEC spikes.  The RMS is consistent at 0.28 (0.40").  But as mentioned now the RA is doing funky stuff and is not very consistent at all.  I'll post up the log file tomorrow.  I'm at a loss right now as to what could be causing it.

___________________________

The RA spikes seem much more pronounced tonight, but perhaps that's just because the DEC isn't masking it. 

Bruce, thanks for the detailed analysis.  I will thoroughly go over everything on the mount tomorrow in the daylight.  One follow up question, I know that with my old G11 the guide rate was set to 0.5 sidereal.  The AP ASCOM driver is set to 1.00.  Do you have a recommendation for the pulse guide rate?

Joel

[Bruce Waddington]

Backlash comp may improve how the mount responds to a large excursion but it isn't going to eliminate the source of the problem.  I guess I'm skeptical that you've eliminated the problem in Dec.  I've always used 1X sidereal for the guide rate which is what AP recommends.  Your guide pulses are large enough that this shouldn't be a problem.  On the few occasions that I've done back-to-back tests on my mount with 0.5x and 1x guide rates, I've never seen a statistically meaningful difference.  Also, when you're looking at "spikes", you need to pay attention to their size and the extent to which they actually interfere with your imaging results.  I realize you have all your cables secured, but that still leaves open the possibility of binding or rubbing in the cable bundle.

 

Bruce

 

---

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Joel Short
Sent: Tuesday, April 12, 2016 8:20 PM
To: Open PHD Guiding
Cc: andy.g...@gmail.com
Subject: Re: [open-phd-guiding] I'm getting spikes in DEC

 

Hi Bruce,

--

[Joel Short]

When looking at the logs in PHD2 Log Viewer, does the viewer take into account dithers when calculating the peak error?  

It seems apparent to me that my mount has not been tracking well all along.  I went through 17 guide logs from the past 3 months.  The first 9 were with a widefield scope (80mm f/4.5).  All of these showed a peak error of around +/- 5".  The last 8 logs were with a 152mm f/8 refractor.  All of these show peak errors in both DEC and RA of +/- 10" or greater.  It is fairly consistent that the RMS error will show about 0.4" in both RA and DEC, but then these spikes show up rather consistently.  However there are rare instances where it will guide perfectly for a few hours.  Below is a link to a log from 3/28 with the same setup.  In one 3hr guiding session it shows the spikes.  But in another 3hr session the guiding is nearly perfect.  And then after that there is a 1hr session where the spikes reappear and are even more pronounced.  This log is on the same target (NGC3938) all night long.  

https://www.dropbox.com/s/upp9nwmfnmzgpuw/PHD2_GuideLog_2016-03-28_193507.txt?dl=0

So, it seems to me there may be a physical issue with the mount that I did not notice before because of the image scale I was working at with the widefield scope, or the target being nearer the pole an thus masking the error better.  

Does all of this sound reasonable?

At this point I am going to consult with AP and see what they say.  

joel

[Bruce Waddington]

It looks to me like almost all of these large excursions are closely associated with a dither operation.  Why don't you try dithering only in RA to see if the behavior changes.   The LogViewer does exclude dither regions, you can see the gray background around them.  But sometimes I think there can be slight timing issues such that the start of the dither might get included.  I generally look closely at the graph myself if I'm trying to look at peak excursions.

 

Bruce

 

 

---

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Joel Short

Sent: Wednesday, April 13, 2016 5:49 AM
To: Open PHD Guiding

--

[Joel Short]

That's true of the log file from 3/28 that I posted this morning.  But the previous logs in the first post (logs from 4/11 and 4/4) show many other spikes besides those around a dither.  But that is definitely something else to try.  I will turn off dither completely and see how that affects things.  

Thanks again for all your help.

joel

[Joel Short]

Bruce and Andy,

I am happy to report that after tweaking some things on the mount yesterday the guiding last night was nearly perfect and I did not observe large spikes or overcorrections.  It's such a huge change from the previous two nights and I didn't lose one sub all night last night.  Would you mind taking a quick look at my guide log from last night and note anything inconsistent or odd that you see?  To my untrained eyes the guiding appears very good but I trust your knowledge better than mine.  

https://www.dropbox.com/s/6s0mrse0ovotsxi/PHD2_GuideLog_2016-04-13_194541.txt?dl=0

[Bruce Waddington]

Yes, I'd say you've nailed the problem.  I don't see even a hint of the Dec-then-RA excursions from before, so your adjustment of the pre-load appears to have handled that very neatly.  Congratulations!  If I was in your shoes, I'd spend the day with a big smile on my face, seriously. <g>  The guiding now looks excellent with only a very few minor excursions in RA that probably have no significance.  Great work! 

 

Looking back on the dialog, I think we were on the right track by thinking something was binding and not allowing the mount to move freely.  But it was internal to the mount, not an external condition.

 

Have fun imaging,

Bruce

 

---

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Joel Short
Sent: Thursday, April 14, 2016 7:00 AM
To: Open PHD Guiding
Subject: [open-phd-guiding] Re: I'm getting spikes in DEC

 

Bruce and Andy,

I am happy to report that after tweaking some things on the mount yesterday the guiding last night was nearly perfect and I did not observe large spikes or overcorrections.  It's such a huge change from the previous two nights and I didn't lose one sub all night last night.  Would you mind taking a quick look at my guide log from last night and note anything inconsistent or odd that you see?  To my untrained eyes the guiding appears very good but I trust your knowledge better than mine.  

 

https://www.dropbox.com/s/6s0mrse0ovotsxi/PHD2_GuideLog_2016-04-13_194541.txt?dl=0

[Joel Short]

OK fellas, I've had two incredibly clear nights and 99% of the random guiding spikes have disappeared.  I'm tickled pink with it.  However, on each of the last two nights I have had 4 spiking events which have ruined 2hrs of photos each night (only 4 images though - 30min subs).  On both nights I got 2 really wild swings in RA (like 40") and 1-2 lesser swings in DEC.  Just to double check with you, that would clearly indicate something physical in the mount, right?  Or somthing like a cable getting pulled and suddenly released?  If I could find the source of these very few spikes I would throw a party.  

Here's the logs from the last two nights.

joel

[Bruce Waddington]

Yes, spontaneous sudden excursions, mostly in Dec, having nothing to do with PHD2.  This particular example is instructive:

 

 

At the point the guide star started the big Dec excursion, nothing at all was happening on the Dec axis - the Dec motor wasn't even running.  Suggests to me it's something external to the mount and the drive system.

 

Bruce

 

---

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Joel Short
Sent: Sunday, April 17, 2016 11:16 AM
To: Open PHD Guiding
Subject: [open-phd-guiding] Re: I'm getting spikes in DEC

 

OK fellas, I've had two incredibly clear nights and 99% of the random guiding spikes have disappeared.  I'm tickled pink with it.  However, on each of the last two nights I have had 4 spiking events which have ruined 2hrs of photos each night (only 4 images though - 30min subs).  On both nights I got 2 really wild swings in RA (like 40") and 1-2 lesser swings in DEC.  Just to double check with you, that would clearly indicate something physical in the mount, right?  Or somthing like a cable getting pulled and suddenly released?  If I could find the source of these very few spikes I would throw a party.  

Here's the logs from the last two nights.

joel