Hi all - please can someone help… I have a Celestron SCT focuser, and by watching very carefully when I move focus forward and backward, I observe that it takes a move of 9 clicks to take out the backlash. IE - I’ve been moving the focuser OUT, and want to move the focuser IN, I need to apply 9 clicks before the actual focus knob moves on the telescope and starts to refocus the scope.

So I guess the ‘backlash’ amount is 9 on my setup.

Question then is - do I tick the ‘AF Backlash’ tick- box on or off, and/or do I enter 9 in the Mechanics/Backlash inbox, or do I do neither?

Thanks very much - I’m sure there’s a quick answer.

Please forgive me if this has been answered elsewhere - I’ve read through much of the forum on this general topic and I’m afraid I’m still confused.

Thanks very much - much appreciated.

Hi John,

I don't have a Celstron focuser so maybe someone with direct experience can give a better answer but here are my thoughts.

I had a quick look at the Celestron driver and it looks like the driver supports backlash so the simplest way to go is to put 9 ticks in the backlash field and turn off AF Backlash. The driver will manage the backlash for all moves of the focus for all focus algorithms.

The main aim of the AF backlash is to cover focusers for which the driver does not provide a backlash option or if you have a focuser where the backlash is somewhat unpredictable (sometimes 9 ticks, sometimes 15 ticks, etc). In this situation the Linear and Linear 1 Pass algorithms provide a mechanism (AF Backlash) outside of the driver to cope with backlash. Note that manual moves of the focuser using in / out and goto buttons do not use AF Backlash at this time.

So if your backlash is always 9 ticks there is no benefit to using AF Backlash. If you turned it on then it probably won't have a negative effect, and if your backlash were to be a bit unpredictable using AF Backlash on top of focuser driver backlash would be a little better. So personally I would try this.

Please be aware though, I have seen issues with AF Backlash conflicting with certain focusers own backlash, which is why the AF Backlash can be turned off. However, I am not aware of any issues with the Celestron.

Thank you John - thank you for such a considered and thoughtful reply. Very helpful It makes much sense now.

I’ve just realised that I’d forgotten about the ‘slack’ inherent in the internal workings of the telescope focussing mechanism … which I guess also manifests as backlash - so after my 9 clicks to take out the focusser gearing backlash, an additional unknown number of clicks are needed before the mirror in the OTA actually starts to move.

I guess this requirement can be handled by the ‘AF Backlash’ tick box?

Thanks again

Hi John,

Yes Backlash from Kstars/Ekos' perspective is the total of:
- Focuser backlash (e.g. gearing). In your case = 9
- Telescope focus mechanism backlash (gearing, etc). In your case = x
- Binding of focuser to telescope. In your case = y

So make sure you set backlash = 9 + x + y

Hi,

now I'm a bit confused about driver's and Linear 1pass' backlash. I don't want to use the driver's backlash feature, so I disabled it there. In the Ekos AF tab (Linear 1pass) I enabled "AF backlash comp" and entered a value certainly larger than the physical backlash in the "backlash" field in the "Mechanics" tab.

I did this for all my 3 optical trains (all EAF focusers, all different amounts of backlash). When I switch from one train to another, it always enables backlash compensation again in the EAF INDI driver (where I disabled and saved it before), and the backlash value in the driver is updated with the value of the backlash field in the "Mechanics" tab. So switching trains writes these parameters into the EAF INDI driver.

Should I simply disable backlash compensation again in the driver after switching trains? Or would that switch off backlash comp of the Linear 1pass as well?

Thanks for any help,
Bernd

Hi Bernd,

So if don't want focuser driver backlash you can set it to 0 in the Mechanics tab or Indi driver screen (these 2 fields are linked so are the same thing). In this case the focuser driver will have backlash of zero, so disabled.

If you have AF Backlash ticked then:
1. If backlash > 0 then AF will use the entered value. For example, if you set backlash to 9 then the driver (if it supports backlash) will adjust by 9 steps. If you use Linear or Linear 1 Pass it will also adjust the outward moves by a further 9 steps.
2. If backlash = 0 then AF backlash is set to 5 * Initial Step Size (in the mechanics tab). This is fine if the real backlash value is <= 5 * Initial Step Size.

This setup isn't ideal so there is an unmerged MR that does the following:
1. The backlash field behaves as it does now.
2. The AF Backlash is now an input field called AF Overscan. Set to zero to disable or set > 0 to enable.

Backlash and Overscan will no longer be connected so you can set both off, either one can be enabled or both can be enabled. Each can have whatever value the user chooses.

Hope this helps.

Hi John,

thanks for the clarification that the BL value in the Mechanics tab and in the driver screen are the same. So I will set the BL value to zero and have an intrinsic AF BL of 5*Initial Step Size and see how that works.

But I have one EAF where the total backlash is around 140 steps (might be a bad EAF unit...), so there I would have to set the Initial Step Size to 28 to cover the BL distance. Have to check whether this is too large a step size. In this case, your mentioned MR with the AF Overscan field would be the "natural" solution, so I hope it will make it into the master branch soon .

Helped indeed, thanks
Bernd

Hi Bernd,

MR 795 "Auto focus overscan" was just merged.

Hi John,

thanks for the info, I just built it and will test later today!

@John Pillar.

I have a celestron c6 with the celestron focuser. I’m using an astroberry so on the slightly older version of ekos before AF backlash was introduced.

I use a backlash setting of 100 which swamps what I believe is my backlash of approx 40 (I manually stepped while watching the screen on the back of my sony a7rii and took note of when the image started moving). I use an initial step size of 30 and the polynomial algorithm. This gets a solution in 6 or 7 steps and seems quite reliable.

Remember for sct’s you ideally want the last movement to raise the mirror. I don’t know if the latest one pass algorithm and the splitting of the backlash enables you to do this but the old version I am on it doesn’t appear to which is why i’m using polynomial.

@Nigel, thanks - very useful to hear of parameters that work. I’ll try with a larger Backlash value and see how I get on. I’ve actually had some luck recently using a step size of 20, Linear 1 pass, and quadratic. I have found that I need to be very close to optimal focus before I start the autofocus process, if the stars go to donuts the process tends to fail quickly.

Thanks again - will keep on experimenting.