Auto Focus vs. Bahtinov mask

Hi all, I have two questions about auto focus. just learning how to use it and a friend suggested backlash of 250 then we brought it down to 100. It does not seem to choose it’s own perfect value so I’m a bit confused.

I set focus with the Bat mask, then I do an auto focus routine and it chooses a slightly hight value, see PDF.

Last night, it appears at 1:23am that it’s attempting to focus and ended up dying at that time. I have disabled “smart” focus but see it in the log. Can you help me out?

Focus pic:

SGP Logfile 1:23AM

Celestron Edge HD 11"
Mount CGX-L Pier mount
Shooting at F10
Imaging camera ASI2600 mcPro
OAG camera ASI294 mcPro
SGP 3.1.0 479
PHD2 guiding software - 2.6.8

BACKLASH: backlash is never going to be determined automatically by SGP, you must figure it out yourself by trial and error. When in doubt just use a larger value. The only drawback is a larger value will take a very slightly longer time to perform the backlash compensation, usually not noticeable.

AUTOFOCUS: you mention you turned off “smart” focus but see it in the log. I don’t see it in the log. You should always post the entire log.

You say the focus routine chooses a slightly high value, I assume by that you mean the selected value is slightly higher than the lower data point to the left. The software routine that produces this is called a Quadratic best fit focus routine, and uses a mathematical quadratic equation to determine the lowest point on the quadratic curve determined by all the data points, not just the lowest one. In practice this gives, by far, the most reliable best focus value.

Thanks Jmacon,

BACKLASH - Ok, I have seen different values based upon changing the backlash from 250, to 100 to 60. Does that make sense?

AUTOFOCUS - I provided the whole log, so if it’s not in there then I mis-read it. Good to know.

QUADRATIC EQUATION - Ok, should I be able to get repeatable results? I’m shooting at F10 with average sky conditions- don’t know if that impacts things. AF is set to 4x4 bin with 30 second exposure.

What do you mean by different values? Different values of what?

Repeatable results? Depends on what you mean by repeatable. Will you get the exact same value? Probably not.
Will you get something close? Yes. How close? Depends.
Depends on:
Focal length, seeing, equipment precision, software settings.
On my OO 12" newtonian with ASI183mm running at .44 arcsec/pixel I see about a 10 point spread over several runs.
On my TV NP127is refractor with ASI 1600mm running at 1.18 arcsec/pixel I see about a 2 point spread over several runs.

Different focus values. So I increased exposure time and the focus value came closer to what I had with the bat mask. Your explanation is very helpful and gives me more confidence that I’m on the right path. I have 11" 2800mm with a .29arcsec/pixel. My difference was about 20 steps.

I may be be complicating things by doing AF with a L-eNhance filter in place, versus no filter and doing an offset.

If a specific backlash value consistently gives you a “hook” on the right side data point (in other words the right most data point is too low), that probably means your backlash is not big enough. Double it and see if the hook goes away.
Focusing only with your L or no filter is usually the best approach because you can get enough exposure for good results with the shortest possible exposure time. Particularly if you will be using NB filters.

Ok, I’m going to focus using no filter, then add an offset for my L-eNhance filter. My goal is to get AF to work for me.

HOOK: I think I’ve seen this on the “V” the right and left top portion of the V is flat or hooking down. I think that’s what you’re describing… An I have seen that in prior attempts.

Is there an issue is setting backlash too high other than time? If not, I’ll get more aggressive with 250 - 500… This is seconds for me, not a big issue if it gets me hours of good images.

The only issue is the time required to perform the backlash compensation. Always make it large enough to avoid the hook.

Great, thanks Jmacon for your help with this. Will try it out hopefully in a few days time

A couple of thoughts on focusing an SCT (I have a C925 that I occasionally use for imaging)

If you are getting the ends of the focus curve going bad, try also reducing the number of steps, I found with mine that if you used the formula provided to determine the number of steps, at the ends of the focus run the out of focus doughnut can fool the HFR/HFD measurement, less steps helped mine there.

Secondly and on backlash, use lots of steps! The backlash compensation is usually used to cover any backlash in your focus motor and or R&P focuser (There shouldn’t be any in a crayford) but with an SCT you also need to cover off the fact that it is moving a big mirror up and down a greased baffle using a threaded rod offset from the middle! I use backlash compensation that equates to at least half a turn of the focus knob with my C925. 500 steps using a Celestron focus motor.

Third and not knowing what focus motor you are using. The Celestron focus motor sees “In” as moving the mirror towards the rear cell. If you are using that one certainly and other motors possibly, you need to reverse the focuser direction in SGP and then do some test movements to make sure that “In” moves the mirror up the baffle (Against gravity) and that your backlash compensation ends with an “In” movement. Otherwise if your backlash compensation is set correctly all focus moves except the initial one and the final one to the calculated position will use backlash comp. If the direction and backlash compensation is set wrong the mirror will probably still be moving when the focus exposure is taken.

1 Like

The_bluester, thanks for the tips, yes, first scope/SCT!! I have Edge HD 11" with Celestron motor focus. I confirmed the IN/OUT movements and changed the setting in SGP to reverse focuser direction. Now IN = UP or going away from rear cell).

I also increased my back lash to 500; I recall each turn was about 1,000 steps (about 40 turns end to end if I recall). I decided to keep # steps at 7, didn’t want to change too many things. AF now working fine.

Thanks for the Celestron focus motor tip! I do have another question about focusing. I have the Optolong L-eNhance dual bandpass filter. Difficult if not impossible to use SGP’s AF routine with that filter. I would like to use the filter and actually had to read the manual to figure it out…

So, am I correct in the following: I determined the offset between no-filter and the L-eNhance is 75 steps OUT with my Bat mask. I do not have a filter wheel yet, and believe I’ll need one. If I did have a filter wheel, my question is would I use AF with clear or no-filter to get accurate focus, then move to the L-eNhance filter and apply the 75 steps OUT adjustment? Is that how it’s done? I’m still a newbie and have not explored filters much.


Sorry for reviving an old topic but I was wondering if you’d share your settings for auto-focus. I’m setting up the same OTA and focus motor and having those settings would give me a solid starting point to get dialed in when the weather finally improves. Thanks!