Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Recommended settings for AF in SGpro

  • Please log in to reply
9 replies to this topic

#1 ChristopherGem

ChristopherGem

    Lift Off

  • -----
  • topic starter
  • Posts: 20
  • Joined: 20 May 2018
  • Loc: Thüringen, Germany

Posted 27 June 2019 - 12:47 AM

Hi fellow imaging enthusiasts! 
First off, I'd like to thank you for your excellent support so far! 

Recently, I switched to SGpro in order to improve my general productivity and make mosaics a little less hit or miss and I must say, after the first run, I am in love with this piece of software engineering! I reckon I have come to grips with the general operation (not suggesting I have actually understood everything) and would like to improve my parameters. 

I am using a TS Optics 80/480 Apo at f6, an Atik One 6.0 (ICX694 with 4,54µm px) and a Sesto Senso focus motor. 

My autofocus parameters are the following: 

data points:10

stepsize: 10

min star size at 1x1: 8

I chose these parameters rather randomly after a cursory read of the routine (should have been less cursory) to see if it works at all. The V-Curve was rather flat with fluctuations in the HFR (i.e. the HFR was sometimes lower then higher and then lower again when the focuser moved in the same direction, didn't take a screengrab though). Stars were still kinda shark (final HFR around 1.0) I believe those parameters are part of the problem and would really appreciate your help! 
Thanks in advance! 
Cheers from Germany! 
Chris

 



#2 jerahian

jerahian

    Viking 1

  • *****
  • Posts: 667
  • Joined: 02 Aug 2018
  • Loc: Maine

Posted 27 June 2019 - 01:12 AM

Welcome to CN, Chris!

 

First off, I would recommend an odd step size, such as 9 or 11 instead of 10.

 

Then, you can follow the often prescribed approach to calculating an appropriate step size by referring to this documentation on the SGP website:  Understanding Auto Focus.  It will run you through the process of calculating your SGP AF settings properly.

 

Good luck,

-Ara



#3 ChristopherGem

ChristopherGem

    Lift Off

  • -----
  • topic starter
  • Posts: 20
  • Joined: 20 May 2018
  • Loc: Thüringen, Germany

Posted 27 June 2019 - 01:24 AM

Hi Ara, 

thank you very much, that is very helpful! Will try it as soon as possible. Thanks again! 


  • jerahian likes this

#4 ChrisWhite

ChrisWhite

    Aurora

  • *****
  • Posts: 4769
  • Joined: 28 Feb 2015
  • Loc: Colchester, VT

Posted 27 June 2019 - 05:03 AM

Ara,  I think you are confusing step size with sample number. 

 

To the OP:  If your v-curve is flat, you should increase your step size.  There is a process in the documentation to determine your step size and I have found it to be pretty accurate.   It's possible you are getting a sawtooth pattern in the curve because of the number of stars being analyzed. Pay attention to how many stars the software is grabbing for the hfr calculation. 

 

Try increasing your step size to 20 or 30 and see if that helps, then fool around with star size parameter to try to get more stars if you see only a few stars being looked at. 



#5 Alex McConahay

Alex McConahay

    Cosmos

  • *****
  • Posts: 8250
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 27 June 2019 - 08:45 AM

Use a Bhat mask or your eyeball to get as good a focus as you can. (What I call a "human" focus as opposed to autofocus). 

 

Note the numerical position of the focuser at this point and the HFR (using the statistics tool). 

 

Now, move the focuser out some number of steps (try 100) and take another test shot. Use the Statistics tool again to measure the HFR. Continue this (maybe moving in or out as needed, and then rechecking) until your HFR reading is twice what it was when at focus. 

 

When you get a HFR twice the best focus HFR, note the numerical focuser position. 

 

Subtract the at-focus position from the twice HFR position. 

 

Divide this remainder by 4.

 

This answer will be your step size. (Or at least a good approximation for your step size. A good place to start)

 

Now, set your number of samples to 7. 

 

Return your focuser to the best focus position, and run the autofocus routine. 

 

If you get a reasonable looking "V" you got it. 

 

If you get a "J" (or a reversed "J") you probably have the step size right but were not in perfect focus when you started. Simply rerun the autofocus routine from where SGP puts it, and the J should turn into a V. 

 

If you get "gull wings" (a bit of a drop in the middle, but at the extremes just weird data and not much change in HFR) your step size is too big. 

 

If you get a slanted line that never reverses slope, you were too far out of focus in the first place. You must start roughly in focus. 

 

Experiment a bit. But the key is that the HFR in your first and last samples should be about twice or so more than at perfect focus. 

 

 

Alex


  • 2ghouls likes this

#6 H-Alfa

H-Alfa

    Messenger

  • -----
  • Posts: 424
  • Joined: 21 Sep 2006
  • Loc: Spain

Posted 27 June 2019 - 12:18 PM

The official manual suggests HFR x3-x5. But if you get a nice V with HFR x2, as Alex suggest, the autofocus routine will be quicker (something useful if you plan to imaging sequences with several filter changes as RRRGGGBBBLLL).

Enviado desde mi ANE-LX1 mediante Tapatalk

#7 Alex McConahay

Alex McConahay

    Cosmos

  • *****
  • Posts: 8250
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 27 June 2019 - 12:48 PM

>>>> The official manual suggests HFR x3-x5. But if you get a nice V with HFR x2, as Alex suggest, the autofocus routine will be quicker (something useful if you plan to imaging sequences with several filter changes as RRRGGGBBBLLL).

 

Yes, please do not take my 2x as gospel. It will work, and work well. Many people find other solutions. YOu have a choice of how many samples. The more you have, the longer it takes. And it may or may not get you more accuracy. You judge. YOu also have a choice of step size. Again, the fewer steps per sample, the longer it will take to move from out of focus to in to back out. (Not nearly as significant timewise as the number of steps, but still something.) So, just get it to work, and then later mess with configuring it perfectly.

 

I tell people starting out to take 9 steps, usually. Just for the heck of it. But, after it is working, cut that down to 7. 

Alex



#8 H-Alfa

H-Alfa

    Messenger

  • -----
  • Posts: 424
  • Joined: 21 Sep 2006
  • Loc: Spain

Posted 27 June 2019 - 05:05 PM

"...But, after it is working, cut that down to 7."

Oh! I will try this. Thanks!

Enviado desde mi MI 5X mediante Tapatalk

#9 jerahian

jerahian

    Viking 1

  • *****
  • Posts: 667
  • Joined: 02 Aug 2018
  • Loc: Maine

Posted 28 June 2019 - 04:56 AM

Ara,  I think you are confusing step size with sample number. 

Thanks, you are correct.  In my first paragraph, I should have typed “I would recommend an odd sample number...”

 

My apologies...



#10 ChristopherGem

ChristopherGem

    Lift Off

  • -----
  • topic starter
  • Posts: 20
  • Joined: 20 May 2018
  • Loc: Thüringen, Germany

Posted 28 June 2019 - 10:58 AM

Hi fellows! 
Thank you all for your replies! I feel well prepared for the next imaging session and will procede as recommended. Thank you!  

Warm regards 

Chris




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics