Multi-star guiding

[Eden]

Morning all,

Has anybody had an opportunity to try this out yet? Considering the fact that it was being promoted as one of the more prominent features in the Maxim 6 release, I'm surprised that more hasn't been said on the matter.

The PHD2 folks are currently evaluating an implementation of multi-star guiding (see attachment) and at this early stage it looks quite promising given the technical challenges which have to be overcome, such as field transformation and so on.

I'd like to try this out using my ONAG; I think multi-star guiding at the optical axis is about as good as it gets (aside from having AO to do rapid corrections for wind, PE, etc.). It's a shame that true multi-star is all but out of reach for OAG users.

[RobF]

It would be good to hear more about the PHD2 work on multi-star guiding. Does anyone know how to get hold of the beta code or have updated on how the work is progressing?

[Moon]

I tried it the other night for the first time in Maxim - I got an error 'guide star faded' so I went back to the old way.
Not sure what went wrong - perhaps there is some extra steps that need to be followed?

[Eden]

The PHD2 developer working on this feature provided me with an internal build (as pictured in the OP) which I have done some testing with.

It is still very preliminary but would benefit from widespread testing.

I have tested it with an 80mm external guide scope and with an ONAG on my 120/900 refractor. There is an improvement in guiding with multiple stars but at this stage selected stars must be approximately correlated in SNR. I also observed that selecting stars in close proximity to one another makes little if any difference over a single star, which brings into question whether OAG-equipped systems would benefit.

I suggest contacting the PHD2 developers via the PHD2 Google Group and express your interest in testing this out.

[RobF]

Thank you. Will see if another tester is welcome.

[DavidTrap]

Makes sense that closely related stars might not provide any benefit. I will endeavor to do some testing this weekend of Maxim multistar guiding on a long focal length scope with both OAG and a finder guider.

Regards,

DT

[RobF]

Will be interesting to hear if you do David.
Recently got OAG going on the FSQ and very happy, but multistar guiding has always interested me.

[RickS]

Now that I'm using a guide scope again instead of a tiny off-axis FOV I was thinking about scripting up astrometric guiding - using a plate solve of the guider field as the ultimate in multi star guiding. Of course, I'm working at an image scale where seeing is mostly irrelevant :)

Cheers,
Rick.

[DavidTrap]

Was looking good at dinner time, now cloud...

DT

[RobF]

Looks like high cloud is going to torment all weekend.
Rick, why have you stopped using OAG - surely not using a 2nd guidescope on the Cerovolo at its FL? (AP140 perhaps?)

[RickS]

Hi Rob,

Yes, planning to image with the AP140 for a while. I just need to get my camera back and spend some time at home instead of travelling. Oh, and some clear skies would help too ;)

Cheers,
Rick.

[RobC]

It would be good to hear more about the PHD2 work on multi-star guiding. Does anyone know how to get hold of the beta code or have updated on how the work is progressing?

RobF,

I can download the source for you and build it.

Rob

[ZeroID]

What is the supposed advantage of multistar guiding touted to be ?

[multiweb]

I've recently asked this very question to Mike (Placidus) who's already doing it with his home-made guiding software and he said: "Multi-star guiding increases the accuracy of the measurement of where the scope is pointing and therefore the accuracy of the correction to be applied".

[DavidTrap]

Ok, subjective report on my experience last night.

OAG at 1875mm focal length on a Paramount GT-1100S, lodestar Autoguider, 1 sec exposures with 1second delay after corrections.

With single star guiding, there was much more jumping around on the guiding graph above and below the line, but still within +/- 1-2 arcsec. With multi star guiding enabled, the graph was smoother, and the excursions were within the +/- 1-1.5arcsec range, and usually contained to one side of the graph, not oscillating above and below the line.

In CCD inspector, the aspect reading (roundness measure) on the images taken were 1 or 2 better with multistar guiding, eg. 9 vs 10 or 11 (so pretty good regardless).

The seeing last night was atrocious. The out of focus stars that appeared during a FocusMax run were distorted and varying wildly.

My hunch is that multistar guiding is worthwhile to mitigate the effects of scintillation - not in the same league as AO, but it doesn't cost anything to activate it in software.

DT

[ZeroID]

Wonder if it could be utilised for feild rotation control in a DOB if someone invented the necessary hardware for the OTA. Pattern recognition and rotation control could eliminate the huge EQ mount requirement for a big Newt.

[Paul Haese]

I have Maxim 6 but cannot find where I can enable multistar guiding, either in setup or under the guide tab and settings. Is this a plug in? I thought it was part of Maxim 6.

[Moon]

Paul - It's in the window with the guide box size.
It's also explained in the help file if you are still having problems.

[Paul Haese]

Thanks Brett, never looked there. :thanx:

[g__day]

Short answer (from Cloudy Nights 2013- 2014, MaximDL V5.0 plug-in and I think Criag Stark's analysis) is 50% improvement in guiding when using 3-6 spread guide stars - as averaging 5-6 stars' centroid movement tends to remove very localised seeing fluctations thereby reducing the number of false positive guide commands.

Basically if your guide camera is seeing star movement that is heavily attributable to seeing conditions - you want to remove this. It happens to me a fair bit so I tend to lengthen the duration of my guide shots from 1 - 1.5 seconds up to 2.5 - 4.0 seconds and increase how much drift I must see before a guide pulse is sent. I have found over the years at my location and on my equipment this gives me excellent, long duration, long focal length guiding.

Craig Stark wrote one or two notes over the years that basically had the theme of don't chase the seeing! It hadn't occured to me that in a guide camera's very localised frame that several stars in the frame may move in an unusual manner in relation to each other. Say you had four stars in the frame in a perfect square - well if a frame shows the four stars in a different configuration I take it that one or more of the stars are effected by seeing! So statsistically these changes in geometry (the length and direction of the guide stars from each other) can be identified and used to modify the algorithm of what drift is actually occuring!

[DavidTrap]

Yes,

It's not in a logical spot!

DT

[Moon]

It does kinda make sense - if you enable it you should probably increase the size of the tracking box at the same time - to pick up more stars.

James

[RobF]

Thanks Rob - will send a PM

[DavidTrap]

There are a few things in Maxim that are tucked away in those drop down option screens, like guider settling. Once you've found them it's OK, but they're not in the first place you'd think to look.

Anyway, first world problem really. ;)

DT

[DavidTrap]

Another update.

Tonight, I've been guiding at 750mm via an OAG.

Here are the guiding graphs, first is multi-star and second is normal.

Not much of a difference and most of the corrections are less than +/- 1 arc-sec.

DT

[gregbradley]

I wonder if the improvements from multi star guiding are going to be greater if the seeing is weaker.

Greg.

[ZeroID]

Interesting. that kind of 'error correction' is the same as Trellis Encoding for high speed faxes ( 114.4kbaud upwards ). A pattern of voltage pulses plotted on a sinusoidal AC waveform. Basis being one spot of of kilter does not invaildate the data. Never thought my old expertise would come back to haunt me ... :D