Autofocus using actual filter being used or auto adjust focus per filter?

Hi - I get good v curves with my Lum filter but not so much with narrowband. I’m shooting with a small refractor around f4.8. I use 2x2 binning and 5 sec exposures with Lum and 25 with ha 5nm. I can’t get a good curve with the ha filter and many times the focuser winds up searching too large of a range and I have to cancel out of the routine.

I was thinking of trying to focus with Lum and using the offsets to get to focus with my narrowband filters. Is this what most people do? I would rather autofocus with the filter I’m imaging with but just doesn’t seem to work. Maybe I need to expose longer? Seems like PE in the mount would start to come into play with longer exposures (I pause auto guiding since I’m OAG). Appreciate any help.

Sorry - couldnt find much information in previous searches but my latest turned up a bunch of threads like mine so looks like I’m not alone! I’ll be going the offset route and focusing with Lum as it seems that is indeed what most folks do.

Filter offsets are exactly what this is for. It’s very hard to autofocus
with narrowband filters. There’s really not much reason to avoid using
offsets, especially in your case using a refractor.

1 Like

I am trying the offset approach tonight and appears to be working well. SGPs autofocus point is a little different than where I think I should be using a bhatinov mask (SGP autofocus results in the focuser being racked out an additional 15-20 steps vs. what I think is an exact focus with the mask). However, when I compare the two approaches, the HFRs look the same or slightly better with SGPs autofocus at least in my test exposures. Maybe the autofocus approach results in a better average focus across the entire frame vs. using a single bright star in the center with the mask?

1 Like


When I set up my offsets, I just slew to a good star field, and run autofocus on each filter. It sucks for the narrowband filters (long exposures) but once you’ve gathered the focus points, it’s pretty accurate.

I wouldn’t rely on a bhatinov mask for it.

If you want to use a bahtinov mask, use something like bahtinov grabber (, so you aren’t relying on visual
judgement. It’s definitely much faster then running the autofocus routine
several times.

1 Like

mads - I was thinking of doing that tonight. How long do you find that you need to expose for narrowband filters?

1 Like

It can be upwards of 60s per frame. Sorry for the late reply.