Achieving smaller star sizes: focus, guiding, or atmospheric seeing?

Sky-Watcher Star Adventurer GTi ZWO ASIAIR Mini Seeing Conditions
8 replies•87 views
Igor Fulvi avatar
Igor Fulvi avatar

I’m trying to better understand a star size issue with my current setup:

  • Sky-Watcher Star Adventurer GTi

  • Sky-Watcher Evostar 72ED

  • ZWO ASI585MC Pro

  • image scale ~1.42"/px

  • guiding RMS typically between 0.9" and 1.3"

I consistently get an average star size around 3.5 px in ASIAIR, even with:

  • focus checked carefully with a Bahtinov mask,

  • fairly round stars,

  • apparently stable guiding,

  • both 120s and 180s exposures.

What makes me think this could be more related to seeing / atmospheric conditions than focus or guiding is the following:

Some time ago I got a session with star size around ~2.5 px. That session happened right after a long rainy period with very clean air, and it was also around the time I had just changed camera and filter.

However, enough time has passed that I’m honestly not 100% sure that the ~2.5 px session was already with my current full setup (ASI585MC Pro + current filter). At the time I was probably still using my previous setup:

  • non-cooled ASI585MC

  • Optolong L-Pro filter

Now I’m using:

  • ASI585MC Pro cooled

  • a different filter

So the comparison may not be completely fair, and I’m trying not to jump to conclusions too quickly. Human memory is apparently another source of noise in astrophotography.

Over the last few nights, conditions here in central Italy (San Gemini, Umbria) have been much more unstable, with:

  • nighttime humidity around ~75-90%;

  • variable cloud coverage / thin clouds;

  • moderate wind, with gusts around 25-35 km/h during the evening/night;

  • generally unstable atmosphere.

At the same time, my star size has stayed consistently around ~3.5 px.

What I also find interesting is:

  • going from 180s to 120s exposures changes the star size very little;

  • I don’t see obvious trailing;

  • during live focus I can visibly see brightness fluctuations on brighter stars.

From PixInsight BatchStatistics I can also see that:

  • sky background improves during the night;

  • noise decreases progressively;

  • but star size still remains relatively high.

This is making me wonder how much of the limit is actually coming from:

  • seeing / turbulence,

  • the relatively aggressive sampling of the 585MC at 420 mm,

  • or simply the physiological limit of the GTi at this focal length.

For people using similar setups:

  • what RMS would you realistically expect?

  • and what average star size would you consider normal under “good but not perfect” conditions?

At this point I’m starting to suspect that the issue may be less about exposure length and more about the combined effect of seeing + sampling + mount limitations. Because apparently photons enjoy chaos almost as much as the atmosphere does.

Well written Helpful Insightful Respectful Engaging
andrea tasselli avatar
andrea tasselli avatar

Takes a set of 1 to 5s exposures pointing the scope at Polaris with tracking disabled (i.e. no tracking at all) and see what you get in terms of FWHM. If that doesn’t show improvements then you know you’re limiting factor isn’t the seeing but the scope. Vice versa if the FWHM shows marked improvements, the limiting factor isn’t the scope but the seeing. If your stars are round during exposures then mount is operating decently too.

Well written Helpful Concise Engaging Supportive
Igor Fulvi avatar
Igor Fulvi avatar

Thanks for the suggestion, it’s actually a very interesting test and I’ll probably try something similar.

Unfortunately I don’t really have Polaris / the northern sky visible from my location, so reproducing the exact test may be difficult, but I could probably try the same idea on another high-altitude star using very short exposures.

One thing I was wondering about is whether, if the FWHM improves significantly in very short exposures, the limiting factor might not only be seeing but also some time-integrated effects such as:

  • guiding/tracking,

  • mount limitations,

  • microdrift,

  • vibrations,

  • or seeing integrated over longer exposures.

So maybe:

  • if short exposures are already bloated → likely seeing / optics / focus;

  • if short exposures improve a lot → then tracking/guiding effects could also be contributing.

Still, I really like the idea of the test itself because it isolates several variables quite nicely.

Well written Helpful Insightful Respectful Engaging
Scott Badger avatar
Scott Badger avatar

andrea tasselli · May 23, 2026, 10:16 AM

Takes a set of 1 to 5s exposures pointing the scope at Polaris with tracking disabled (i.e. no tracking at all) and see what you get in terms of FWHM. If that doesn’t show improvements then you know you’re limiting factor isn’t the seeing but the scope.

Why would pointing at Polaris, untracked or tracked, mitigate the effects of seeing?

Cheers,

Scott

Well written Respectful
Igor Fulvi avatar
Igor Fulvi avatar

It doesn't reduce the effects of seeing, but it does reduce the variables at play, since we've disabled tracking and autoguiding; Polaris is a suggestion because it's a good star and practically stable on long exposures, that's all... (at least that's how I interpreted it).

Scott Badger avatar
Scott Badger avatar

Igor Fulvi · May 23, 2026, 11:20 AM

It doesn't reduce the effects of seeing, but it does reduce the variables at play, since we've disabled tracking and autoguiding; Polaris is a suggestion because it's a good star and practically stable on long exposures, that's all... (at least that's how I interpreted it).

I think Andrea may have meant it removes guiding/tracking as a variable, and if you’ve verified focus with a b mask, then a couple second exposure should give you your seeing (and resolution limit). Polaris (or the celestial pole) because the stars close to it are moving slower than anywhere else in the sky, so less star trailing. Unless you live in the arctic, a ‘high-altitude’ star won’t work.

Note though, seeing not only varies night to night and throughout the night, but can vary throughout the sky, both according to how high you’re pointing/thickness of the atmosphere, but also because of more local differences, like the terrain, a warm roof on a cool night, etc. since thermal turbulence is the main offender.

For guiding/tracking, I generally find that so long as my total RMS is a third or less of my seeing (or best possible resolution) and RA RMS is mostly equal to DEC RMS, then guiding is sufficient.

Cheers,
Scott

Well written Helpful Insightful Respectful Engaging
Igor Fulvi avatar
Igor Fulvi avatar

Thanks everyone for the suggestions and explanations, they’ve actually been very helpful.

I think I’m starting to realize that my issue is probably not caused by a single factor, but rather by a combination of:

  • seeing / atmospheric turbulence,

  • relatively aggressive sampling (~1.42"/px),

  • GTi limitations at 420 mm,

  • and possibly some DEC backlash/stiction behavior.

The comment about comparing guiding RMS to seeing also made me realize that one of the core problems is that… I don’t actually know my real seeing conditions.

With RMS around 0.9"-1.3", the interpretation changes completely depending on whether the seeing is:

  • 2",

  • 3",

  • or 4".

And at the moment I have no reliable estimate for that.

The short exposure idea is very interesting because it removes guiding/tracking from the equation and could help isolate what is already present “instantaneously” versus what accumulates over time.

Unfortunately I don’t have Polaris / the northern celestial pole visible from my location, so I can’t reproduce the exact test properly. But I may still try a similar experiment with very short exposures just to compare FWHM evolution versus exposure length.

One thing I also noticed while reviewing my logs is that during guiding calibration, the last DEC reversal step does not fully return to the initial position. The mount seems to struggle a bit when reversing DEC direction, suggesting some backlash or delayed response there.

So now I’m wondering whether what I’m seeing is something like:

  • seeing already limiting the minimum achievable FWHM,

  • with guiding/tracking/DEC behavior adding a bit more blur on top over longer exposures.

What keeps confusing me is that:

  • stars are generally still fairly round,

  • 120s and 180s exposures produce very similar star sizes,

  • but the overall star size still stays around ~3.5 px.

So at this point I’m starting to think this may simply be the realistic combined limit of:

  • current atmospheric conditions,

  • sampling,

  • and a small portable mount working at 420 mm.

Astrophotography really is the art of discovering new ways the atmosphere can ruin photons.

Well written Helpful Insightful Respectful Engaging
Igor Fulvi avatar
Igor Fulvi avatar

One thing I noticed while reviewing my guiding calibration logs is that during the final DEC reversal phase, the mount struggles to return fully to the initial position.

The calibration North movement looks reasonably fine, but when it has to reverse direction and move back South, the response becomes noticeably slower and the calibration ends before the star fully returns to the starting point.

During that calibration the telescope was pointing South, so this seems consistent with some backlash/stiction appearing during the North → South DEC reversal.

Based on this, I’ll probably try a very slight DEC imbalance (likely a tiny bit nose-heavy) just to keep the gears more consistently loaded in that direction and see whether:

  • calibration improves,

  • settle behavior improves,

  • or DEC RMS becomes more stable.

Well written Helpful Insightful Respectful Engaging
Tony Gondola avatar
Tony Gondola avatar

The first thing to realize it you are working with a very small aperture. In general, the diffraction limit for your telescope is about 1.6 arc/sec. This is the best it can deliver even if it was in space.

Here are the things that in my experience will reduce star size in refractors:

Blue and IR bloat. Makes sure you are filtering the extremes of the spectrum out. If no other filters are used you’ll need at least a UV-IR cut filter and depending on your optics you my need something stronger like a fringe-killer. Instead if using something like the L-pro, you’ll get tighter results with a dual band Ha-Oiii filter, certainly you will get smaller stars.

Shorten exposure times. The longer you expose, the more your stars will bloat. I use a 585 and my exposure times are typically anything from 15 sec. to a maximum of 60 sec. Shooting the 585 in HCG mode brings read noise down to very low values. Going long with a 585 isn’t doing anything other than allowing more time for guiding and the atmosphere to blur your images.

Focus matters. Get an EAF. You are probably doing a good job if getting a decent focus to start with using the mask but refractors change focus pretty rapidly as the temp. changes. If you want sharp results, it’s critical that you maintain perfect focus and during a long session, the use of an EAF is the only practical way to do it.

Processing. with a small sensor like a 585, star reduction/sharpening is not an option, it’s a must. BlurX in Pi is the gold standard and will transform your images. If you want to stay on the free software side then Cosmic Clarity is the way to go.

Hope all that helps!

Helpful Engaging