Critique and algorithm suggestions

[jhart]

Hi,

Could I get some critique on my guiding and perhaps suggestions on the RA and DEC algorithms and settings I am using?

I know it is not according to the text book way to guide but I seem to get better guiding, at least in DEC with the Z filter (as opposed to Resist Switch).  If using Z filter is inappropriate for my mount, is there something in the Z filter settings that could be used to implement in Resist Switch to help my guiding with that algorithm?

Regarding my mount sometimes GA is content with amount of backlash it measures in DEC (when it is less than 3000) and sometimes it is not and recommends guiding in one direction.  Sometimes if I accept that one direction guiding it may work for a while but should DEC need a correction in the other direction it doesn't recover and the DEC graph can drift badly off of the graph.

Suggestions on RA would be much appreciated too.

Thanks,

Jeff

https://openphdguiding.org/logs/dl/PHD2_logs_zHDc.zip 

[Ken Self]

Hi Jeff

Z-filter is inappropriate for a mount with backlash. It typically issues very small corrections which have no effetct when you have backlash.

The resist switch algorithm  is your best option. You should keep your exposures between 2 to 3 seconds and use uni-directional guiding in declination as recommended. Make sure you have enough polar alignment error to create a steady drift. 

Ken

[Jeffery Hart]

Hi Ken,

I understand the limitations of Z Filter.  You are right that there are instances where my mount has a large DEC deviation that Z Filter can’t correct (at least not without a bunch of correction tries).  However, most of the time Z filter guides with lower DEC error than anything I can get with Resist Switch, whether using the defaults and the GA recommendations or trying to tweak the settings.  I don’t seem to have luck with “bad” polar alignment to facilitate good DEC guiding.  Again, mostly it seems like I get the best guiding with “excellent” polar alignment (using SharpCap) with the steady but small control of Z Filter.  Is there some set of algorithm settings where PHD2 could guide like Z Filter (~2 seconds with ~ 6 Exposure factor and the default MinMo seem to work) but have the muscle of Resist Switch to step in when a larger DEC excursion happens?

I realize the limitations of my mount and understand autoguiding will not cure all evils but whatever incremental improvement I can find will help.

Thanks again,

Jeff

--
You received this message because you are subscribed to a topic in the Google Groups "Open PHD Guiding" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/open-phd-guiding/GvD3jl2djBo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to open-phd-guidi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/220faacc-398f-498d-87db-7c6239123e02n%40googlegroups.com.

[Ken Self]

Hi Jeff

It could be that with an excellent polar alignment and small corrections from Z-filter you have effectively been operating unguided or close to it. You can see in the guide log sections where there are few dec guide pulses. That is one way to deal with dec backlash.

If you want to use PA offset to get a steady drift you want the PA error to be mostly in azimuth and about 5 arcmin. Otherwise the drift could change direction part way through the session. You could even try a higher PA error if you are not imaging near the pole. I have seen examples where this has been used with very good results. If you have a log where you tried this method unsuccessfully it should be possible to analyze it to see why.

[jhart]

Hi Ken,

Thanks for the suggestion.  I will try the 5 arc' offset as you recommend as see what happens.

Are there other algorithms and/or particular settings I might try to get similar small corrections that Z filter provides, but do not get tripped up by the backlash when a larger correction is needed?  It appears to me that Resist Switch tends to make over-corrections giving a saw-tooth graph while Z filter has small (maybe quicker) corrections that give a flatter graph with a lower DEC error.

Thanks again,

Jeff

[Brian Valente]

you can try lowpass or lowpass2

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 on the web visit https://groups.google.com/d/msgid/open-phd-guiding/6a5424ac-84e8-4dbe-a059-c9461fe529b4n%40googlegroups.com.

--

Brian 

Brian Valente

portfolio brianvalentephotography.com 

[Brian Valente]

PS - the algorithms are described here in some detail:

https://openphdguiding.org/man-dev/Guide_algorithms.htm

[bw_msgboard]

Hi Jeff.  I think you are probably misinterpreting what you’ve seen with your Dec guiding.  The Z-Filter and LowPass2 algorithms are better suited to more accurate mounts and the Resist-switch algorithm is almost certainly what you should be using.  Your guide log doesn’t include a Resist-switch run so I can’t analyze what was going on.  In the log you sent, you can see how slowly the Dec guiding responded to the large excursions you were getting (Dec is green):

 

 

It’s true that near-perfect polar alignment might allow you to run essentially unguided in Dec, and I think that’s what’s happening.  But you have too much Dec backlash and other instability problems to make that a good strategy.  In other words, you need to use an algorithm – the default one – that will respond more aggressively to these excursions.  

 

Although you seem to be focused on the Dec guiding, RA is the constraint on your overall guiding results.  Here’s a typical sequence (RA in red, Dec in green):

 

 

The overall RA guiding RMS was almost 50% higher than Dec and you can how “rough” it looks.  There is a lot of abrupt high-frequency tracking error here, probably a limitation of the RA gearing system.  And this was while pointing at Dec=41 degrees, so the problem would be worse closer to the celestial equator.  So there’s really no point in trying to optimize the Dec guiding in the face of this RA tracking error – you’re just going to get elongated stars.  I suspect you may be up against the limitations of your mount.

 

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.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/d0fdec16-b1d2-47a8-8630-6f805fb23576n%40googlegroups.com.

[jhart]

Hi Bruce,

Thanks so much for the analysis.  Looking back through my logs it appears that on balance I achieved lower RA error using Hysteresis over PPEC.  (Whether is is relevant on not, the log I sent involved 3 iterations of making/loading PEC training on my mount).  I don't know if that is unique to my mount or if I haven't used the PPEC algorithm properly.  I will use Hysteresis and Resist Switch for my next sessions and see whether there can be improvement.

Thanks again,

Jeff