Classical nova in Sagitta, discovered 2026 Aug 25 at V = 7.4, from CrowdSky Seestar data.
Next automatic update in 5h 59m 52s · due 14:00 UTC on 23 Sep · this page rebuilds every 6h
1 · Light curve of V488 Sge
Latest V = 11.35 on 2026-09-19T20:09:47.154765 — 32 calibrated IRCUT epochs from 4 Seestar units (0 new this run). A further 16 epoch(s) in LP are plotted as a separate series and excluded from the decline fit.
Highlight Seestar: scroll to zoom · drag to pan
S30 Pro_6d547766S30 Pro_ee0d83c2a54a17f6d9ed61a4
diamond = LP filter (16 of 48 epochs). A different bandpass from the primary IRCUT series: this nova has strong Hα emission and the zero point is calibrated on continuum stars, so the two series can carry a constant offset with respect to each other. Plotted for coverage — excluded from the maximum and the t2/t3 fit.
Hover a point for the Seestar serial, date, V and that frame's ZP uncertainty (its calibration RMS against Gaia). The Grey out ZP uncertainty > slider sets the cut, from 0.10 to 0.40 mag (default 0.25): frames above it are drawn as faint hollow marks rather than removed, so every contribution stays on the plot. The grey star is the reported discovery magnitude. Highlight a single Seestar with the dropdown; scroll to zoom and drag to pan (Reset zoom restores the full view).
2 · Decline rate and speed class
A classical nova is characterised less by how bright it got than by how fast it fades. The conventional measures are t2 and t3 — the days from maximum to 2 and 3 magnitudes below it — which place the nova on the Payne-Gaposchkin speed scale (very fast < 10 d, fast 11–25 d, moderately fast 26–80 d, slow 81–150 d, very slow 151–250 d). This is precisely what a dense amateur time series measures well: for a fast nova t2 is days to weeks, so cadence matters far more than aperture.
Fit from the IRCUT series only (32 epochs). The LP epochs (16) are a different bandpass and are excluded: combining filters on an emission-line source would offset the curve and bias the rate.
Measured maximum
V = ≥ 8.24
2026-08-28 (MJD 61280.85)
t2
10.0 d
max → max + 2 mag
t3
17.0 d
max → max + 3 mag
Speed class
Very fast or slower
Payne-Gaposchkin — an upper bound on speed, since t2 is a lower limit
Mean decline
0.141 mag/d
over 22.0 d since maximum
32 unsaturated epoch(s) used.
The reported maximum (V = 7.42) is brighter than the brightest epoch measured here (V = 8.24), so coverage of this field began after maximum. t2/t3 below are measured from our own peak and are therefore LOWER LIMITS; the speed class may read too fast. The mean decline rate is the robust number to compare, since it does not depend on catching the peak.
Per-epoch zero point & calibration quality
Highlight Seestar: scroll to zoom · drag to pan
Each point is one frame's own zero point (error bar = that frame's calibration RMS against Gaia), coloured by Seestar — every epoch is calibrated independently, across nights, units and sites. Drag the slider to set the ZP-uncertainty cut (0.10–0.40 mag, default 0.25): frames above it are shown as faint hollow circles, never dropped. You can also highlight one Seestar, or zoom (scroll) and pan (drag).
3 · Calibration and data quality
Each stack is reduced independently, so heterogeneous Seestars self-calibrate:
Source extraction — SEP background subtraction + a chromatic fixed aperture (per-band 90%-enclosed radius) on the R/G/B planes.
Astrometry — the CrowdSky-solved WCS is used directly (re-solving with astrometry.net is available as a fallback).
Zero point — green instrumental mags are fit independently per frame to Gaia DR3 synthetic Johnson V over V∈[10,14]: V = m_inst + ZP + k·(B−R).
Forced photometry — the fixed aperture is measured at the nova's known position, whether or not a source is detected there, so the series never goes ragged.
Saturation test — the aperture's core is checked for a flat top; clipped epochs are flagged and excluded from the decline fit.
Two systematics worth stating plainly.
Saturation. The nova reached V ≈ 7.4, roughly 6 mag brighter than the Gaia stars (V∈[10,14]) that calibrate each frame, and brighter than a Seestar stack can record without clipping. A clipped core loses flux, so a saturated epoch reads too faint — a failure that looks like data rather than like an error. Those epochs are detected from the shape of the core (see the pixel cut below), drawn as hollow triangles, and held out of the t2/t3 fit. They are never silently deleted.
Emission lines in a broad filter. A classical nova is not a blackbody: this one shows strong Hα and Hβ P Cygni profiles, and the IRCUT filter passes Hα. The zero point and colour term are fit on ordinary continuum stars, so applying them to a strong emission-line source carries a systematic that the quoted errors do not include. Treat these magnitudes as a well-sampled relative light curve — excellent for decline rates and shape — rather than as absolute spectrophotometry.
Zero-point calibration for the latest epoch — a representative example only. Every epoch is calibrated independently against Gaia; each epoch's own ZP, star count and RMS are in the per-epoch strip above and the credits table.Left: the aperture on the nova. Right: a pixel cut through the core — the evidence behind the saturation flag. A sharp peak is a good measurement; a plateau means the sensor clipped and the magnitude is too faint.
Latest epoch (2026-09-19T20:09:47.154765, S30 Pro_6d547766): ZP = 21.915, calibration RMS = 0.405 mag over 2261 stars; nearest neighbour 18.7 px away; core unsaturated (1 px within 2% of the aperture peak).
4 · Credits — contributing Seestars
Contributing Seestars (4 units, 48 frames) — click to collapse; click a column header to sort
