Astrophotography diagnosis of NGC 6995: Backfocus error
RawHa300s
The Doc examined this image of NGC 6995 (raw, Ha, 300s). Estimated overall technical quality: 6/10. 1 defect found: Backfocus error (severity 3/5).
&w=1920&q=75)
Info
- Cible
- NGC 6995
- Position
- 20h57m00s · +31°17'06"
Full moon at 100 %, but you are shooting Ha from 2750 m: exactly the case where narrowband rescues the session. The measured background is flat (radial profile 0 %, plane gradient limited to 1 % amplitude), so the moon imposed no usable gradient, the Ha filter did its job. The quoted 2 arcsecond seeing matches the measured central FWHM of about 2 px. Forget OIII or broadband colour that night though: keep full moon for pure Ha like this, and reschedule the blue layers for a waning moon.
Setup
- Type d'image
- Brut
- Télescope
- Lunette 100mm x 740 mm
- Caméra
- IMX 571 MC
- Filtre
- Ha
- Monture
- Taka EM200
- Exposition
- 300s
- Phase de lune
- Pleine lune (100 %)
- Seeing
- 2"
- Notes
- Alt 2750m
- FOV
- 1.82°
Coherent for NGC 6995. With a 1.82 degree field, the eastern Veil is well framed with margin to spare, the subject is not clipped. The sampling of the 740 mm plus IMX571 combination falls in the comfortable range for 2 arcsecond seeing, and the measurements confirm it: 2.16 px central FWHM, stars not oversampled. The Ha filter is the right call on this remnant under full moon. The weak point is not the optics at large but the mechanics: the imaging train spacing, which degrades the corners. The 300s exposure is consistent, the low background to black point margin being expected in narrowband.
The diagnosis in detail
The core of the field is excellent: the PSF sheet shows compact, near round stars at centre (elong 1.14 against a field floor of 1.09), and the 40 degree PA dispersion confirms there is no coherent directional elongation. In other words the EM200 and the guiding are not at fault: no drift, no readable periodic error, no oscillation.
The one structural defect is the FWHM rise toward the periphery, with a corner/centre ratio of 1.61 peaking at the top-left corner (3.74 px against 2.16 at centre). The pattern stays broadly symmetric about the centre, pointing to field curvature tied to spacing rather than outright sensor tilt. The TL corner being worse than the others still deserves an orthogonality check once backfocus is adjusted.
On the background, nothing to report: flat radial profile, only 1 % residual gradient despite the full moon, no clipping in shadows or highlights. The visible nebulosity matches the DSS reference point for point, so it is real signal, not an artefact.
Priority actions
- Correct the optical train backfocus distance and validate on a test sub by inspecting all four corners
- Check focuser orthogonality and tightness if the top-left corner stays softer after adjustment
- Keep the Ha strategy under full moon and save OIII and broadband for moonless nights
