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

N.I.N.A Autofocus Troubleshooting

astrophotography imaging
  • Please log in to reply
13 replies to this topic

#1 Chaz007

Chaz007

    Explorer 1

  • *****
  • topic starter
  • Posts: 97
  • Joined: 22 Sep 2016
  • Loc: Chicago Suburbs/Tucson AZ

Posted 23 May 2020 - 02:04 PM

Since I started doing astrophotgraphy I've been using APT, but when I got an autofocuser I decided to switch to nina. Last night was first light for the new software and focuser and at first it seemed to go well, but when I checked the frames after the imaging session it was clear that most were out of focus. Running the subs through DSS and plotting the frame number VS score confirmed this suspicion, with only one focus seeming to be successful. Normally I'd mess with settings until I got everything working, but for some reason I'm missing a bunch of settings options in the advanced focuser options. It doesn't appear that I got a corrupt version of nina as everything else seems to be fine, but some really crucial focuser settings are just missing...

 

For the time being I'll do what I can with the information I have. Once the sequence starts it appears that all the autofocusing is done internally (unless I'm missing something) but the first focus I did (manually initiated) gave me a decent V curve. It's clear where backlash is present so hopefully after adjusting backlash settings this problem will be fixed. However, if anyone knew how I could get access to these missing settings it would be a huge help. For example, nina stops guiding when focusing and as a result my object drifts across the field of view over the course of the night. (one of the missing settings would disable this feature)

 

I know nina isn't super common, but any help would be greatly appreciated, even if it's just general autofocusing tips.

 

In case it comes up:

Camera: ZWO ASI 183MM pro

Scope: William optics Zenithstar 73

Focuser: ZWO EAF

Mount: Skywatcher eq35

 

focus1.jpg

focus.png

These are all of the focuser settings I get. Also, the backlash setting I have in there is accounting for the graph.


Edited by Chaz007, 23 May 2020 - 02:07 PM.


#2 3snows

3snows

    Mariner 2

  • *****
  • Posts: 237
  • Joined: 03 Jun 2012
  • Loc: Boston, MA

Posted 23 May 2020 - 02:31 PM

I am also an APT user looking at NINA for autofocusing but quite haven't figured it out yet.  So for now I still use a mask for fine focusing.

 

Try these videos.  They are excellent and I'm trying to duplicate them.

 

https://www.cloudyni.../#entry10161008



#3 overnight

overnight

    Vanguard

  • *****
  • Posts: 2,214
  • Joined: 10 Jan 2017
  • Loc: San Diego, CA

Posted 23 May 2020 - 02:38 PM

Use the beta version (1.10) of NINA instead of 1.9, it has all of the improvements on the autofocus routine including trends+hyperbolic fit.


Edited by overnight, 23 May 2020 - 02:39 PM.

  • scadvice likes this

#4 scadvice

scadvice

    Surveyor 1

  • *****
  • Vendors
  • Posts: 1,976
  • Joined: 20 Feb 2018
  • Loc: Lodi, California

Posted 23 May 2020 - 02:40 PM

APT user also trying out NINA because of the auto focus. I'm watching those same video's by Cuiz The Lazy Geek. He is putting out new video's sometimes twice a week about NINA. There are three on focusing alone.

 

https://www.youtube....rYt-Bb1g/videos



#5 kingjamez

kingjamez

    Vanguard

  • *****
  • Posts: 2,291
  • Joined: 03 Oct 2006
  • Loc: Fairfax, VA

Posted 23 May 2020 - 02:43 PM

Use the beta version (1.10) of NINA instead of 1.9, it has all of the improvements on the autofocus routine including trends+hyperbolic fit.

Do you mean 1.10? It's got all the newest / best settings.

 

OP: Have you read the autofocus portion of the NINA Wiki? That's the first place you should start.

 

You have a lot going on there but here are the things I'd change first.

 

1. Increase your step size, looking at your graph I'd go up to 40.

2. Make sure you start with decent focus, its clear from the graph that you were well out of focus when it started.

 

Try that, read the manual, and you'll be good to go.



#6 georgian82

georgian82

    Viking 1

  • *****
  • Posts: 755
  • Joined: 16 Jan 2016
  • Loc: Healdsburg, California

Posted 23 May 2020 - 03:04 PM

Since I started doing astrophotgraphy I've been using APT, but when I got an autofocuser I decided to switch to nina. Last night was first light for the new software and focuser and at first it seemed to go well, but when I checked the frames after the imaging session it was clear that most were out of focus. Running the subs through DSS and plotting the frame number VS score confirmed this suspicion, with only one focus seeming to be successful. Normally I'd mess with settings until I got everything working, but for some reason I'm missing a bunch of settings options in the advanced focuser options. It doesn't appear that I got a corrupt version of nina as everything else seems to be fine, but some really crucial focuser settings are just missing...

For the time being I'll do what I can with the information I have. Once the sequence starts it appears that all the autofocusing is done internally (unless I'm missing something) but the first focus I did (manually initiated) gave me a decent V curve. It's clear where backlash is present so hopefully after adjusting backlash settings this problem will be fixed. However, if anyone knew how I could get access to these missing settings it would be a huge help. For example, nina stops guiding when focusing and as a result my object drifts across the field of view over the course of the night. (one of the missing settings would disable this feature)

I know nina isn't super common, but any help would be greatly appreciated, even if it's just general autofocusing tips.

In case it comes up:
Camera: ZWO ASI 183MM pro
Scope: William optics Zenithstar 73
Focuser: ZWO EAF
Mount: Skywatcher eq35

focus1.jpg
focus.png
These are all of the focuser settings I get. Also, the backlash setting I have in there is accounting for the graph.

Your curve shows a lot of backlash on the right hand side and it’s probably why you are seeing this issue.

I recommend switching over to beta 1.10 nightly #081 which has a new feature for autofocus called overshoot instead of absolute.

I haven’t tried it myself yet but what I have read seems to indicate that this is a much better way of focusing.

Refer to this thread and watch some of the videos explaining how it works:

https://www.cloudyni.../#entry10207393

Hope this helps

Edited by georgian82, 23 May 2020 - 03:07 PM.


#7 kingjamez

kingjamez

    Vanguard

  • *****
  • Posts: 2,291
  • Joined: 03 Oct 2006
  • Loc: Fairfax, VA

Posted 23 May 2020 - 03:18 PM

Indeed, overshoot is the "right" way to do backlash. The good thing about overshoot is that you choose the direction that goes against gravity then put a reasonably high number in for the backlash value. With overshoot, you really can't have it set wrong as long as your value is more than the actual value of the backlash. 

 

With my C11, I simply have it do a 1/4 turn of the focuser knob for backlash adjustment and it just works. I choose the backlash direction based on having the focus always be approached by pushing the mirror up, against gravity. 

 

I love how NINA is getting better and better each week.

 

-Jim 


Edited by kingjamez, 23 May 2020 - 03:23 PM.


#8 Chaz007

Chaz007

    Explorer 1

  • *****
  • topic starter
  • Posts: 97
  • Joined: 22 Sep 2016
  • Loc: Chicago Suburbs/Tucson AZ

Posted 23 May 2020 - 04:16 PM

This is exactly what I was wondering! I figured the manual would be for the current "official" release version, not the beta for the next version, which is why I was confused at not having those settings.... I got 1.10 and set the mode to overshoot. I changed my step size to 40 (it was 10 last night which is default) and set the curve fitting to hyberbolic which seems to be the best one for what I'm doing according to the manual.

 

Out of curiosity, how close do I have to be to focus to have the AF routine work effectively?

 

Thanks for the help!



#9 kingjamez

kingjamez

    Vanguard

  • *****
  • Posts: 2,291
  • Joined: 03 Oct 2006
  • Loc: Fairfax, VA

Posted 23 May 2020 - 05:05 PM

 

Out of curiosity, how close do I have to be to focus to have the AF routine work effectively?

 

Thanks for the help!

Play around with the step size such that the first measurement that NINA takes ends up being around ~2-2.5 times +/- the lowest measurement. So if NINA's best focus is an HFR of 3, then you want your highest HFR point up around 6-8ish. 

 

The reason you want to be reasonably in focus is that the HFR measurement algorithm gets weird on larger unfocused stars. To that end, you don't want to start with very large stars. Anywhere on the V of your graph would be fine.

 

Keep in mind that Auto-focus really isn't for getting initial focus, it's for keeping focus as you image through the night. 



#10 Eric Horton

Eric Horton

    Mariner 2

  • *****
  • Posts: 216
  • Joined: 04 Dec 2011

Posted 24 May 2020 - 04:25 AM

Nina has a utility to measure backlash and then insert the values it determines. Once I ran that and a couple additional tweaks to backlash, my curve that looked like yours started to look quite normal.

#11 Eric Horton

Eric Horton

    Mariner 2

  • *****
  • Posts: 216
  • Joined: 04 Dec 2011

Posted 24 May 2020 - 07:02 AM

Additionally, I have found the following settings to be helpful.  Focuser Settle Time = 1, AF Crop Ration = .75



#12 Chaz007

Chaz007

    Explorer 1

  • *****
  • topic starter
  • Posts: 97
  • Joined: 22 Sep 2016
  • Loc: Chicago Suburbs/Tucson AZ

Posted 24 May 2020 - 02:26 PM

I was able to do some more experimenting last night with overshoot and a backlash value of 250. The V curve looked pretty normal but upon inspecting the frames today the stars still aren't quite as sharp as I was hoping, though things have definitely improved. I think maybe I need to decrease the step size a bit? The HFR hovered around 4 the entire night. I also wonder how binning might help...

 

focus2.png



#13 dghent

dghent

    Messenger

  • *****
  • Vendors
  • Posts: 447
  • Joined: 10 Jun 2007

Posted 24 May 2020 - 02:29 PM

I would say a smaller step size would be a good thing to iterate on next.

 

Think of step size as a sampling rate. Too big of a distance between the samples causes the curve to be less defined. Too small, and the star HFR analysis might not be able to discern a meaningful difference between two successive frames or be able to "see" past poor seeing.


Edited by dghent, 24 May 2020 - 02:44 PM.


#14 kingjamez

kingjamez

    Vanguard

  • *****
  • Posts: 2,291
  • Joined: 03 Oct 2006
  • Loc: Fairfax, VA

Posted 24 May 2020 - 02:50 PM

I was able to do some more experimenting last night with overshoot and a backlash value of 250. The V curve looked pretty normal but upon inspecting the frames today the stars still aren't quite as sharp as I was hoping, though things have definitely improved. I think maybe I need to decrease the step size a bit? The HFR hovered around 4 the entire night. I also wonder how binning might help...

 

attachicon.giffocus2.png

That really doesn't look bad, I'd bet that 31685 is pretty close to optimal. Keep in mind that all scopes (even my F/2 RASA) will have a small zone where the focus is as small as it's going to be. This is called depth of focus and it defines the critical focus zone. If you know how much your physical movement per step of your focuser is, and the depth of focus (based on aperture and focal ratio) of your telescope, you can calculate how many steps of your focuser will be inside the CFZ. It's not required to do, this, but it will give you more confidence that you've got optimal sampling. You want to make sure that your step size is not larger than your CFZ. 

 

A smaller step size sounds right, you want to avoid the extremes of your curve turning flat/down as they are in your screen shot. Ideally, the curve would be smooth and relatively consistent in slope. This is very much a tune it till it works like you want process. 

 

Some folks have made suggestions for tuning based on their setup, but each telescope and focuser will be different. What works for me has no bearing on what will work for you. So tune it with that in mind. 

 

One way of making sure that you've got it tuned well is to get a Bahtinov mask. Use autofocus then put the mask on. If the mask agrees that you are in focus, you know you are good to stop tuning. 

 

-Jim


Edited by kingjamez, 24 May 2020 - 02:52 PM.



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





Also tagged with one or more of these keywords: astrophotography, imaging



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics