Astrophotography diagnosis of Sh2 155: Sensor tilt and background chroma noise
Masterrgb ultra 22340×10s03 sept. 2026
The Doc examined this image of Sh2 155 (master, rgb ultra 2, 2340×10s). Estimated overall technical quality: 6/10. 2 defects found: Sensor tilt (severity 3/5), Background chroma noise (severity 3/5).
&w=1920&q=75)
Info
- Cible
- Sh2 155
- Date
- 03 sept. 2026, 20:03
- Lune
- Croissant croissant 44.6% (sous l'horizon)
- Site
- Bortle 4 · transition rural/banlieue (saisi)
- Position
- 22h57m10s · +62°35'26"
Favorable conditions. The Moon was below the horizon during the session, so no lunar pollution was possible despite a 44.6 percent phase, and the Bortle 4 sky is decent for a faint emission nebula like Sh2-155. The measured background is in fact clean on the gradient side: the +102 percent variation recorded corresponds to the nebulosity itself, confirmed by the DSS, not to an artifact. Under these conditions nothing stopped you from using much longer subs: it is the settings, not the sky, that limit your SNR here.
Setup
- Type d'image
- Master
- Télescope
- Skywatcher250-Mbx
- Caméra
- Uranus-C
- Filtre
- rgb ultra 2
- Exposition
- 2340×10s
- Phase de lune
- Dernier quartier (56 %)
- FOV
- 38.5'
A sound setup in principle but poorly configured. The 0.64 degree field frames the central part of Sh2-155 tightly, which works visually, the target fitting inside the frame. On the other hand the 0.498 arcsec/px sampling is heavily oversampled for typical seeing: the 6.5 px FWHM at center is about 3.2 arcsec, you can bin 2x2 without losing anything and gain SNR. On the settings side, 10 s at gain 210 with the sensor at 22.7 C is the real handicap: 17.5 percent dark clipping and a Bowley skew of -0.97 point to very poor subs, only compensated by the 2340 frames. Cool the camera and go longer.
The diagnosis in detail
The diagnosis comes down to two independent causes. The first is mechanical: the PSF panel shows a flawless center (FWHM 6.5 px, elong 1.01) while the periphery degrades in a clearly asymmetric way, the top-left corner heavily bloated at 9.9 px while the top-right stays at 7.4 px and perfectly round. This single-axis asymmetry, with vertical asymmetry of 0.184 against 0.046 horizontal, is the signature of a tilted focal plane, not a backfocus error which would spoil all four corners equally. On an f/4.7 Newtonian the tolerance is tiny, a few tens of microns are enough to produce this.
The second is an acquisition strategy problem. Stacking 2340 subs of 10 s at gain 210 on an uncooled sensor at 22.7 C yields frames dominated by read noise and dark current: the result is a background with obvious chroma mottling that even massive integration does not fully erase. The field edges, covered by far fewer subs because of the framing drift visible in the stepped borders, are the worst affected.
Good news: the background carries no instrumental gradient, and the red nebulosity at center is genuine, corroborated by the DSS.
Priority actions
- Fix the sensor tilt: fully threaded imaging train and tilt adapter adjustment, iterating on the aberration inspector until the corner/center ratio drops below 1.15
- Lengthen sub exposure to 60-120 s and lower the gain toward unity gain to become sky-limited
- Enable sensor cooling, target -10 C instead of the current 22.7 C
- Bin 2x2 at capture or in processing, 0.498 arcsec/px sampling brings nothing under typical seeing
- Crop the low-coverage borders before integration to remove the edge chroma noise
