phd2 goes multistar

[alacant]

If it's anything like as good as EKOS' own internal multistar implementation, phd2 users are in for a treat.

 

 

Edited January 16, 2021 by alacant

[knobby]

Oooh, now that looks interesting !

[rnobleeddy]

Good news. I found the ekos multistar  worked well in terms of guiding accuracy, but unfortunately it crashed kstars every few hours so i went back to phd2.

[rickwayne]

That has been my experience exactly. Well, something was making KStars crash, certainly might have been multistar guiding. Hadn't had issues before StellarMate OS 1.5.5 came out.

[TerryMcK]

To save people looking the developers notes for the PHD2 implementation show this:

“Multi-Star Guiding

Some guiding configurations can benefit from guiding on multiple stars rather than just one.  This results in using a weighted average centroid position of multiple stars rather than just the centroid of a single star.  Multi-star guiding is enabled using a check-box on the Guiding Tab of the Advanced Settings dialog.  When this option is enabled, the Auto-select function will identify up to 12 stars in the field of view that have adequate SNR.  No more than 9 of these will be used at any one time, but the remainder will be used to replace secondary stars that are lost or rejected for some reason.  The "primary" star is the single best candidate, the same one that will be selected if multi-star guiding is disabled.  As guiding proceeds, some of the secondary stars may be rejected because they are too dim or have drifted outside the field of view.  This is of no concern, the multi-star guiding algorithm handles the secondary list automatically.  When multi-star guiding is active, the PHD2 image display will show the usual rectangle around the primary star and circles around the secondary stars.  All other UI features associated with a guide star - brightness properties, SNR, FWHM, etc - relate to the primary star, not the entire list.  Multi-star guiding can be enabled or disabled while guiding is active.  However, enabling the feature will force another 'auto-select' procedure. The multi-star algorithm uses the secondary stars to refine the centroid position and reduce its volatilty, so there is essentially no way for it to degrade guiding performance.  Whether it produces a material benefit to your overall guiding will depend on many factors including image scale, star and background sky brightness, star size, focus, and camera noise.  Because of the way the algorithm is implemented, your best option will be to try it and decide for yourself.”

Edited December 3, 2020 by TerryMcK

[Jamgood]

Looks very interesting. Will give this a run at some point. 👍

[rickwayne]

Plenty of commenters over at Cloudy Nights report a noticeable improvement. I certainly did too, although I'm always trying to tweak the rig and the params for better guiding so it's hard to say PHD multistar was the cause. Still, when you go from around 1.2" total RMS to 0.67"...that ain't bad news.

For Linux users, AFAIK you still have to compile PHD from source to get the new shiny. Fortunately that's easy and straightforward, takes less than half an hour on a Pi 4.

[Jamgood]

Some interesting read over at CN. Cheers @rickwayne 👍

It would be nice to have the chance to have a go with the new build but the weather over here isn't looking great for the foreseeable.

[malc-c]

A video on youtube

https://youtu.be/wDCEjwQNmTE