Poor RA Guiding

[Tim Stevenson]

Bruce and Brian ... I just keep coming back with more issues.  My setup is the EQ6R Pro with an AT8RC telescope and an OAG/ASI220 mini guide system.  My last issue is now gone when I disabled Dec backlash and in fact the Dec guiding is very good IMO.  But, the RA is twice that of Dec.   I'm using Predictive PEC with the standard settings with the only exception being the period is set to 480 period length.  The RA motion seems very erratic, i.e. back and forth, while the Dec is much more stable.  I'm wondering if it's stiction causing the back and forth.  I have it in Dec as it showed up in the backlash chart on my last post.  If so, then i probably will need to work both axis worm gear adjustments.

Thank you again

Tim Stevenson

https://openphdguiding.org/logs/dl/PHD2_logs_9Ek7.zip

[Tim Stevenson]

Adding a little more information... Segment 6 leading up to flip, guiding was RA 0.56" / Dec 0.48" with a Total of 0.74".  Not bad.  The frequency analysis show all peaks gone thanks to Predictive PEC algorithm.  In segment 7 after the flip, guiding was RA 0.67" / Dec 0.29" with a Total of 0.73".  This segment is still has reasonable guiding but this is were I noticed the RA 2x Dec.  Also the frequency analysis shows the 181 and 121 Hz peaks.

Before I show you some of the screenshots.  I noticed a large difference between the Dec MinMo (0.35) and RA MinMo (0.23).  Also if I were using  Hysteresis algorithm and saw this bouncing I would reduce aggressiveness.

Thought I would show you the "bouncing" around of the RA guiding as seen in Segment 6...

Segment 6 frequency analysis.

and here is Segment 7 frequency analysis.

Here is the Frequency analysis with Hysteresis algorithm (several days earlier).

[Bruce Waddington]

I think we're probably reaching the limit for how much more we can tell you.  Guiding data is only a crude tool, at best, for trying to identify mechanical problems.  One point I'll make is to respond to your "instinct" that the RA guiding is simply over-shooting and probably warrants a lower aggressiveness.  That's a very common mistake people make and it happens because they don't look closely at the individual guide commands.  Here's an example from your data:

Look at each of  the guide commands - the rectangles - and notice how many consecutive pulses it often takes to get the intended response from the mount.  Reducing the aggressiveness factor will make this problem worse, not better.  I will agree there are a few time periods when it does look like over-shooting but those are in the minority and there's certainly no consistency.  Your problem is not with the guiding parameters, it's with the mount.  You identified a time period where the RA RMS was 2x that of Dec, and it looks like this:

Where are all these big RA excursions coming from?  They were completely absent in the guiding session you were happy with which, by the way, was done on the opposite side of the pier.  I don't know why this is happening.  Is the gear well-balanced in RA?  Are these excursions originating with the actual mount tracking or from something in the payload?  While this remains unresolved, I think you may need to revert to hysteresis for RA, and you probably need to reduce your guide camera exposure time in order to guide at a faster cadence.   The PPEC algorithm necessarily assumes that the mount is reasonably well-behaved and that the RA tracking errors are periodic and comparatively small.  Having all these wild excursions get baked into a PEC model is not what you want to happen.  If you can't figure out what's going wrong, you can probably isolate whether the side-of-pier dependency is actually coming from the RA drive.  By loosening the RA clutch and manually moving the scope around from one side of the pier to the other, you can effectively rotate the worm wheel by something close to 180 degrees.  You have to be careful about this because the scope will no longer know where it's pointing, so a plate solve and synch will be required right away.  It's just a suggestion, I wouldn't do it unless I had convinced myself that the problem isn't coming from a different source.

Sorry I can't offer any better explanations,

Bruce