How long should my subs be?
As long as your sky and your mount let you, and no longer. On a smart telescope that usually means the app's default from town, and 30 to 60 s subs in EQ mode once you are under a dark sky or behind a dual-band filter.
What a sub length actually changes
Every sub pays the camera's read noise once. Everything else in the photon budget, the target's light and the sky's light, scales with total time and does not care how it is sliced. So sub length only matters when read noise is a noticeable part of the noise in each frame. From a bright sky it is not: the skyglow in a 10 s frame already swamps a modern sensor's 1 e⁻ of read noise. From a dark sky, or with a filter that removes most of the skyglow, the sky signal per frame drops until read noise is a real share of the total, and longer subs start to pay.
The model says it in one line: total time T = SNR² · (S + B + RN²/t + D) / S². The sub length t appears only in the read-noise term. When the sky term B is far bigger than RN²/t, changing t changes nothing.
Worked example: Andromeda with the Seestar S50
| Sky | 10 s subs | 30 s subs | 60 s subs (EQ) |
|---|---|---|---|
| Bortle 8 | 4 h 04 min1,465 subs | 3 h 57 min475 subs | 3 h 56 min236 subs |
| Bortle 6 | 1 h 27 min525 subs | 1 h 21 min162 subs | 1 h 19 min80 subs |
| Bortle 4 | 42 min254 subs | 36 min72 subs | 34 min34 subs |
| Bortle 3 | 34 min204 subs | 27 min55 subs | 30 min30 subs |
From Bortle 8 the three columns are almost identical: the sky sets the noise and the sub length is irrelevant. From Bortle 4 the 30 s column is about a fifth shorter, because there the read noise from hundreds of 10 s frames is a real cost. At Bortle 3 the 60 s figure stops falling: a solid image needs at least 30 frames for stacking to reject satellites and bad tracking, so 60 s subs cannot finish in under 30 minutes. Longer subs stop paying once you run out of frames.
Worked example: the Veil with the dual-band filter
| Sky | 10 s subs | 30 s subs | 60 s subs (EQ) |
|---|---|---|---|
| Bortle 8 | 2 h 56 min1,054 subs | 2 h 32 min305 subs | 2 h 26 min147 subs |
| Bortle 6 | 1 h 33 min557 subs | 1 h 09 min139 subs | 1 h 03 min64 subs |
| Bortle 4 | 1 h 09 min414 subs | 45 min91 subs | 39 min40 subs |
The filter throws away most of the skyglow, which is the point, and that makes even a suburban sky read-noise limited. This is why the two things go together: a dual-band filter on an emission nebula is the situation where longer subs help most, even from town.
What longer subs cost
- Field rotation. On an Alt-Az mount the field turns during each sub. The apps cap sub length for that reason (Seestar allows 30 s in Alt-Az and 60 s only in EQ mode). See EQ mode and field rotation.
- Rejected frames. A satellite, a gust of wind or a knock throws away the whole sub. Losing a 60 s frame costs six times more than losing a 10 s one, and dark-site sessions on a light tripod lose plenty.
- Saturated stars. Bright stars clip sooner in long subs, which costs star colour. On a nebula that rarely matters; on a cluster it does.
- Tracking. The little mounts in these scopes are good, not perfect. Polar alignment in EQ mode has to be reasonable for 60 s frames to stay round.
What each scope allows
| Telescope | Sub lengths | Longest, and how |
|---|---|---|
| ZWO Seestar S50 | 10 s, 20 s, 30 s, 60 s | 60 s in EQ mode (ZWO Seestar TH10 fluid tripod head) |
| ZWO Seestar S50 Pro | 10 s, 20 s, 30 s, 60 s | 60 s in EQ mode (ZWO Seestar TH10 fluid tripod head) |
| ZWO Seestar S30 | 10 s, 20 s, 30 s, 60 s | 60 s in EQ mode (ZWO Seestar TH10 fluid tripod head) |
| ZWO Seestar S30 Pro | 10 s, 20 s, 30 s, 60 s | 60 s in EQ mode (ZWO Seestar TH10 fluid tripod head) |
| DwarfLab DWARF 3 | 15 s, 30 s, 60 s | 60 s in EQ mode (DWARFLAB mini hydraulic tripod DWARF 3) |
| DwarfLab DWARF mini | 15 s, 30 s, 60 s, 90 s | 90 s in EQ mode (DWARFLAB mini hydraulic tripod) |
| DwarfLab Draco | 30 s, 60 s, 120 s, 300 s | Pro Mode, 300 s ceiling, no EQ mode |
| DwarfLab DWARF II | 15 s | fixed 15 s, nothing to choose |
| Vaonis Vespera II | 10 s, 20 s, 30 s | Expert Mode, 30 s ceiling, no EQ mode |
| Vaonis Vespera Pro | 10 s, 20 s, 30 s | Expert Mode, 30 s ceiling, no EQ mode |
| Vaonis Vespera 3 | 10 s, 20 s, 30 s | Expert Mode, 30 s ceiling, no EQ mode |
| Vaonis Vespera Pro 2 | 10 s, 20 s, 30 s | Expert Mode, 30 s ceiling, no EQ mode |
| Unistellar eQuinox 2 | 4 s | fixed 4 s, nothing to choose |
| Unistellar Odyssey | 4 s | fixed 4 s, nothing to choose |
| Celestron Origin | 10 s, 15 s, 30 s | 30 s in EQ mode (Celestron 93665 equatorial wedge NexStar Evolution Origin) |
| Celestron Origin Mark II | 10 s, 15 s, 30 s | 30 s in EQ mode (Celestron 93665 equatorial wedge NexStar Evolution Origin) |
Rule of thumb
- Town, no filter: the default. Longer subs just cost you frames.
- Any sky, dual-band filter on an emission nebula: 20 to 30 s in Alt-Az, 60 s if you have EQ mode.
- Dark site (Bortle 4 or better): EQ mode and the longest sub the app offers, once your polar alignment gives round stars.
- Not sure: run the target through the calculator at each sub length. If the times are within 10% of each other, stay short.
Every target page carries this comparison for its scope, for example Andromeda Galaxy with the Seestar S50.
More guides
- Which smart telescope should you buy?
- Integration time explained
- The Bortle scale for smart telescope owners
- EQ mode and field rotation on smart telescopes
- Dual-band filters on smart telescopes: when they help and when they hurt
- Filters explained: VIS, Astro, UV/IR-cut, dual-band, CLS and solar
Updated 2026-08-27. Corrections: about page.