First experience with Quadratic Fit best focus evaluation method (SGP v.3.1.0.295)

Thanks Bernd.
Basically, I’m happy with the result. Worked straight away.

I used the SGP Image Grader to analyse the data from yesterday and compare it some data from a week ago. The old data displayed an average HFR of 1,32 in comparison to the average of 1,48 yesterday (the weather was better yesterday, same target though.). Not sure if this means that the old method delivered sharper results and I know that the comparison is very rudimentary and statistically irrelevant. So, I’ll hold off with a judgement, till I have more experience with the new method and have processed some of the data.

Just wanted to add; my impression in the preliminary discussions was that the quadratic method would make it easier to auto focus, even if the initial focus is a little off the ideal focus point. To me that did not seem to be the case yesterday.I still had to “manually” focus to close to the ideal focus point (as was necessary with the previous method) before the auto focus would deliver decent results.

This is true. The new method only deals with how to determine the best focus based on the image curve that has just been made. Has nothing to do with where the focuser is positioned prior to running the focus routine. Initial positioning is a new feature that has been proposed.

@Ken

Why is everyone being forced to use the dark theme? Perhaps it helps if you are looking through a sheet of red plexiglass but the dark theme is of no value when operating remotely – at least for me. I find the loss of readability to be a problem.

Charlie

Had first light with the new focus routine. Screen grabs below. The red one was grabbed before the end of the routine and SGP went on to move the start point. These are with a Tak Baby Q, usually pretty well behaved, but the conditions were not great.

[added] It will be interesting to see how it works with an RCT that is slightly out of collimation. If my recollection is right, the quadratic function is symmetrical but with a small amount of mis-collimation the HFR graph is asymmetrical.

SNAG-0000 SNAG-0004 SNAG-0003 SNAG-0002 SNAG-0001

By the way - a big thank you to everyone who made this possible. I suggested a quadratic fit as far back as Feb 15. How time flies!

I think this episode epitomizes the very best of astrophotography, where everyone collaborates and the developers are receptive to new ideas.

Last night I used v3.1.0.313 for the first time to experiment with the new AF routine. Very impressive! I did about 20 AF runs to build new filter offset values and all my runs pretty much looked like this sample:

image

My Meade ACF operating at f/6 usually has somewhat erratic HFR calculations so, in addition to the new quadratic curve fitting, there must have been an improvement to the HFR calculation routines to get such consistent HFR values across the entire focusing range.

Well done, guys!!

Charlie

Thank you Buzz. It has been fun working on this project and we all really appreciate all the new enhancements that are going into SGP. Ken and Jarod have been working like crazy putting out this new stuff and deserve a lot of kudos
.
Your red graph shows the routine selected the lowest data point at 3902, which I assume is where it started the rerun from. Is that correct? It also looks like it would have been much better to pick 3962 which looks like a really good focus position. Would that have been a good focus position?
We do not have a display of R^2 for that graph but it is was < .9 which the program therefore rejects and it then automatically selects the lowest data point for the rerun. The end points of the run are often very bad and I have code in there to reject one bad point. I think it should have rejected this one and given a much better result. Could you please send me your logs? Thanks.

1 Like

Charlie, could you please send me your logs for this and some future runs with 313+?. I would like to run them through our analysis process. Looks to me like you have selected a very effective step range that is giving great results. Also what gear are you using?

There are several logs. I was setting up a new camera - I started with focusing the main camera and then tried the same with the OAG, having worked out the microns/step etc.

here are a few more - where the quadratic fit probably works better than a classical approachSNAG-0010 SNAG-0011

If you want log files, I can sort them out for you

Thanks Buzz. To send more logs you might wait until you have quite a few. Then we can do a thorough analysis of your focusing. For the moment, it looks like it is working well for you.

@jmacon

I will save the next set of auto focus frames and the SGP log and get them to you.

Charlie

I had a gremlin one after this but missed the screen grab. It may have been linked to the smart focus option which you talk about in another thread. I cannot recall if it was set of not.

Here is a full collection of SGP log file and the auto focus packs from a 6 hr imaging run I completed a few nights ago. About 150 megs:

SGP Dropbox Link

Charlie

buzz & jmacon,

Do you have any results thus far on how well the new quadratic fit is performing when used with longer focal length setups? I’m currently using a 10"RC at either its native 2000 mm f/8 or with a FR taking it down to 1370 mm at about f/6. As yet, unless I’ve missed it, I’ve not seen a response to buzz’s question, especially in the context of any slight collimation issues added to the mix.

Great work though and I look forward to the next major release build on this beta stream.
Thanks

I am using an AG12" at 1360mm with an ASI1600mm 3.8 microns, resolution = 0.70 and the TV127 has resolution 0.75 arcsec. New Quadratic focus working great with those specs. I have not seen any reports for longer FLs. Looking forward to getting some.

My imaging all last night on both scopes with release 335 was perfect. Two targets with two flips, not an issue.

Not yet - I’m finishing off a project with my Baby Q right now and the weather is absolutely hopeless.

Thanks Guys.
It appears Jerry that you’re AG12" FL is very near mine when I use the FR, which is good to know.
Buzz, you seem to be in the same waiting game as we are here. I’m on the other side of London to you, so am generally getting this rainy/cloudy rubbish first
Cheers

I slapped the RC on this week and unbelievably, it is approximately in collimation. It probably only needs a small tweak (famous last words). I ran a few AF runs before it clouded over - nothing untoward. The AF range was quite small and the Q-fit was good over the shallow curve. I ran out of clear sky to expand the range to see what happens when the stars break up into small donuts.