Astrophotography diagnosis of Caldwell 30: Sensor tilt
RawIr-Cut20s04 sept. 2026
The Doc examined this image of Caldwell 30 (raw, Ir-Cut, 20s). Estimated overall technical quality: 4/10. 1 defect found: Sensor tilt (severity 3/5).
&w=1920&q=75)
Info
- Cible
- Caldwell 30
- Date
- 04 sept. 2026, 19:08
- Lune
- Gibbeuse croissante 55.6% (sous l'horizon)
- Site
- Bortle 6 · banlieue dense (VIIRS)
- Position
- 22h37m03s · +34°25'18"
Moon conditions are favourable: it was below the horizon at the time of the exposure and 80.8 degrees from the target, so no lunar contribution is expected in this background. Bortle 6 remains penalising for a broadband galaxy with a plain Ir-Cut filter: the sky background rises fast and contrast on the faint spiral arms of NGC 7331 will suffer. With seeing quoted at 2.5 arcsec, the atmosphere is not the limiting factor here, the 0.44 arcsec/px sampling is far more so. Under this sky the winning strategy is volume: many subs, dithering, and a total integration of several hours to push noise below the light-pollution floor.
Setup
- Type d'image
- Brut
- Télescope
- C11
- Caméra
- ZWO ASI533MC Pro
- Filtre
- Ir-Cut
- Monture
- EQ6r pro
- Exposition
- 20s
- Phase de lune
- Dernier quartier (46 %)
- Seeing
- 2.5
- Notes
- Réducteur Celestron F=6.3
- FOV
- 22.1'
A C11 reduced to f/6.3 on an EQ6R-Pro with an ASI533MC is coherent for a compact galaxy like NGC 7331, which fits comfortably in the 0.37 degree field with room for the Deer Lick group. The debatable point is sampling: 0.44 arcsec/px under 2.5 arcsec seeing is heavily oversampled, spreading signal over roughly 6 pixels of FWHM without gaining detail. Binning 2x2 at capture or in processing would give back a substantial apparent signal-to-noise gain. On settings, gain 100 and offset 20 are healthy (no significant dark clipping), but two points need fixing: the 20s exposure is too short (e_margin 2.83, you are not sky-limited) and above all the 19.5C sensor temperature shows cooling was not active, which inflates dark current.
The diagnosis in detail
The dominant diagnosis is instrumental and it is measured, not assumed. The aberration inspector puts the centre among the sharpest zones of the field (FWHM 3.77 px) while the top-left corner reaches 5.85 px and the bottom-left 5.25 px, with the right side staying at 4.3 px. The asymmetry is almost purely horizontal (0.262 against 0.03 vertically): this is not field curvature, which would hit all four corners equally, but a genuine tilt of the sensor plane relative to the focal plane, with the left side too long or too short. On a C11 with an f/6.3 reducer the train is heavy and backfocus is critical, the two most common causes.
Tracking is not at fault: a PA dispersion of 32.7 degrees between zones and a central elongation of 1.18, barely above the debayered OSC floor, indicate round stars at centre. In 20s on an EQ6R-Pro that is expected. This matters because it clearly points the effort at the optics rather than the mount.
The second limitation is photometric. With e_margin at 2.83 you are still in the read-noise-limited regime: each sub contributes little signal and the galaxy stays a diffuse blob with no structure. Add the 19.5C sensor temperature, which loads the background with dark current and hot pixels. Fixing the tilt, cooling, guiding and moving to 60-120s would transform the result on this target.
Priority actions
- Diagnose and correct the imaging train tilt: repeat the test with the camera rotated 90 degrees to isolate camera/spacers from telescope, then retighten evenly and verify the f/6.3 reducer backfocus
- Enable sensor cooling at -10C and wait for stabilisation before any run
- Move to 60-120s exposures with autoguiding to become sky-limited (target e_margin at least 5)
- Bin 2x2 to bring sampling near 0.9 arcsec/px, matched to 2.5 arcsec seeing
- Build darks and flats at the acquisition temperature and gain, and enable dithering between subs
