Astrophotography diagnosis of Fireworks: Tracking drift and sensor tilt

The PSF panel is modest: FWHM 1.8-2.5 px everywhere, a moderate corner/center ratio of 1.34, but above all an elongation present from the center (1.29 vs a field floor of 1.11) with a very consistent PA orientation from zone to zone (spread 16.3°). This is the classic signature of a tracking component: if it were field optics (tilt, backfocus), the center would stay round. I therefore discard the sensor-tilt candidate despite its measurement: the horizontal asymmetry (0.14) exists but is dominated by the global drift, not by a clear diagonal asymmetry between opposite corners. On 15s the streak stays short and thin, the impact is minor (severity 2).

The background is healthy: radial map at +13% from center to edges (mild optical vignetting, normal before flats, not retained on a sub), a plane of only 2% with no notable light gradient, near-zero anisotropy. No sign of lunar gradient or light pollution despite Bortle 5, consistent with the 74° Moon separation.

The galaxy NGC 6946 and the cluster NGC 6939 are already detected on this sub, confirmed by the DSS. No saturation, no abnormal noise for a luminance frame. The key work remaining is tracking accuracy and accumulating exposures to bring out the target.