Uneven curves and flat across the bottom


I have tried auto focus last night and although it seems to work well, my curves are longer on the inside of focus than the outside and they tend to have 3 very close points across the bottom.

I am going to rerun the setup again tonight and see if I can improve it. My scope is very temperature variable and I was thinking of focusing for each filter rather than offsets. What is the use ratio of each method, offsets, lum only or each filter?




Three very close points on the bottom usually mean your increment between focus points is too small. I think most people focus on L and use offsets.


if you’re not seeing much movement between points then it’s likely that your points are too close together, like Jerry mentioned. You can find more info on setting up auto focus here:


As for using the actual filter vs just lum. You’ll likely find that folks that image LRGB just use the filter and those that do narrowband tend to almost always use Lum with offsets. I do both.



Thank you for the feedback, much appreciated.
I did use the SGP setup method last night and I got 46.6 steps if I remember correctly. After I had a fiddle with it I worked best at 4 for star size, 40 steps and 11 points. I’d say 6/10 for focused frames. Any thoughts on redoing it or fiddling instead. Sjould i retry 9 steps, that would give bigger gaps? I seem to need to refocus every 10mins or so. Might need another scope!
Temp drop is too quick at the moment to do offsets. It dropped 6°C in 2 hrs. Or should I go temp compensation.


I managed to get out again last night and rerun the set up for my focuser. This time I used 9 steps again and 6 for star size and ended up with a step width of 56. Although this has seen a marked improvement in the curves, I still get the odd one with a flat bottom. Across the 4 filters, as I tried focusing with each filter, I ended up with HFD’s between 1.2ish and 1.5ish. I could not get lower than 1.18 focusing with a Bahtinov mask. Are they ok figures? My images look ok.



The rule for determining how many focuser steps you should have between data points is that the biggest HFR reading you get should be about 3x the smallest HFR. So, if you are seeing an HFR of 1.2 with good focus, adjust your steps between data points so that the first data point that is calculated has an HFR around 3.6 - 4.0.

Also, HFR is not an absolute – it’s totally dependent on your equipment. Some systems get perfect focus when the HFR is 3.5 and other get it at 0.5. HFR is in pixels, BTW, not arc seconds.

Also, the V curve will only be symmetrical when the camera was already in perfect focus when the AF routine started. The more asymmetry you see, the more the camera was out of focus to start.



Hi chasmiller46,

Great explanation, thank you. I’ll check this up next time I’m out.