Auto Focus Offsets problem

Last night I attempted to find the focal offsets for each of my filters, L,R,G,B,Ha,OIII and SII and I got strange results…

My method was to use a Bahtinov mask to get L pretty much perfect and then I ran auto focus for each filter concurrently in the order above, each time noting the position given, which is in the table below. The temperature readout from the sensor on the focuser stayed at 12.5 degrees Celsius. I have also added the offsets. As you can see they are all over the place. I have a feeling that backlash may be the issue, I don’t know how to counteract this though, I have read several posts and am none the wiser. I have included screenshots of my settings in SGP.

Another thought is that my method may not have been the best, should I have focused L each time for every filter? For example L and then R, 4 times in a row?Autofocus%20Settings Filter%20Offsets Focuser%20Settings

You should give a generous amount for your backlash setting. Extra does not hurt anything, except a little extra time for moving the focuser out the extra backlash amount. One indication of not enough backlash is if the first focus data point is not in a straight line with the ones that follow. My experience is it is usually lower than it should be. In other words, closer to focus because not all the backlash has been removed.

When I determine filter offsets, I like to run L before each filter just to confirm that I have a correct value for it, and double up on the filters. Like this: LRR LGG LBB etc. Or; you can just do LR LG LB… then repeat a time or two. Using a Bahtinov mask may make all the Ls not so necessary. But you might do a few. Or better yet, use the Bahtinov mask with the filters.

I got a fair spread of values when doing mine, just like you have. It comes with the variability of focusing. Make lots of runs to get a more valid average. May also be an indication of some hardware issue.


Try with “Compensation step size” = 100 - 200.

In my case, with a refractor with Feathertouch fosuser I use 150 to be safe and I do have very consistent data…


1 Like

Thank you for the help both of you, apologies for taking a while to reply, have been away visiting family.

jmacon, yes I did notice in some of the graphs that the first focus data point did not follow the line that the others were part of. I shall try the method that you mention.

rws, I have seen these values in other threads, would they not be specific to each system or do they work as generic values?


Not specific to a system. Two important things, one, “Comp step size” should be larger tan the steps it takes your focuser to remove backlash. If you use a larger number is ok. In my case I use more than what my system requires.

And two, the “Compensation direction” has to be in accordance with the instructions shown above and considering gravity. For a refractor and an SCT it should be IN, for a Newtonian it will depend on where is your focuser in relation to gravity.



Thank you both, backlash was indeed the problem, I used JMacon’s technique for the offsets and now have much more consistent results. I have had several successful nights of imaging RGBRGB etc… with a Lum autofocus every hour.