Variability class · EW · EA · EB

Eclipsing

Geometry, not physics — two stars taking turns in front of each other.

Eclipsing binaries don't actually change their light output. They orbit so that, from our line of sight, each star periodically hides the other — and the shape of those dips reads out the geometry of an entire stellar system.

412,803 active stars EW · EA · EB

Phase φ 0.00 · Δm 0.00

Drag the phase scrubber through one orbit. Switch between a contact pair (two near-equal minima) and a detached Algol (flat, with one sharp deep eclipse).

What it is

A pair of stars bound in orbit, oriented edge-on enough that each passes in front of the other once per cycle. The deeper drop — the primary eclipse — happens when the hotter star is hidden; the shallower secondary half an orbit later. Because it is pure geometry, the light curve is exactly periodic and the period is the orbital period (or, for contact systems, half of it — see below).

The physics

Eclipse depths encode the ratio of surface brightnesses; eclipse widths encode the stellar radii relative to the orbit; the spacing between primary and secondary reveals orbital eccentricity. A contact binary shares a common envelope, so its light varies continuously and its two minima are nearly equal. A detached pair sits quietly out of eclipse and drops sharply only during the brief transits.

The families within

Subtypes

W UMa contact

Stars touching, sharing an envelope. Continuous variation, near-equal minima, periods typically 0.2–0.8 d.

Algol / detached

Well-separated stars: flat out of eclipse, then sharp — often unequal — eclipses. Periods from hours to years.

β Lyrae / semi-detached

One star tidally distorted by its companion. Continuous, ellipsoidal variation with clearly unequal minima.

The varchive method

Finding the period

The same science code runs for every star. Here is how it behaves for this class — and where it can be fooled.

A contact binary shows two near-equal minima per orbit, so a floating-mean GLS fit latches onto the P/2 harmonic — by design. Class-aware adoption doubles it back to the true orbital period unless the fold's odd/even minima depths say the minima really are identical. Detached systems (EA) are nearly flat between eclipses, where a sinusoid sees almost nothing, so varchive runs BLS on every eclipsing candidate and lets it win for Algols. Every eclipser's epoch is then refined from our own fold to the primary minimum — never trusted from VSX — and used to predict the next eclipse.

What to watch for

The top GLS peak for an EW is almost always P/2; never adopt it blindly.
A shallow secondary can vanish into the noise, making a genuine binary look like a single-dip transit.
A disagreement with the catalogued VSX period (a vsx_period_mismatch flag) usually means VSX listed the half-period — our fold is the arbiter.
Faint, brief eclipses push the magnitude distribution to positive skew and high kurtosis — exactly the eclipse hint that selects a star for BLS.

From the archive

Worked examples

WISE J193632.2−794139

EW · the worked example — GLS finds P/2, adoption doubles it

#421 →

Gaia DR3 6349775999409480192

EA · a 14.4-day detached Algol, BLS-adopted

#3142 →

NSV 6231

EA · flagged: our period disagrees with VSX

#1171 →

Keep exploring

Browse the Eclipsing population

Browse 412,803 stars → See them in the explorer