Rotation curves with a small telescope

An asteroid is a rigid body spinning freely in a vacuum. Nothing damps it and nothing perturbs it on human timescales, so its light curve repeats with a precision closer to a clock than to a star. Point a telescope at the same asteroid a decade later and the same shape returns.

But the shape itself is not a sine wave. It is a projection of the object's actual silhouette, and that makes it far more informative than a simple oscillation. The four curves below are real, archived measurements — every point was made by a named observer and deposited in a NASA archive.

Almost every asteroid is closer to a potato than a ball. As an elongated body turns, it shows you its broad side twice and its narrow end twice per rotation, so you get two brightness maxima and two minima per revolution — which has a practical consequence that catches people out constantly: the strongest period in your data is half the rotation period, and you fold at twice the obvious answer. If asteroids were smooth triaxial ellipsoids the two maxima would match exactly and there would be no way to tell a full rotation from a half, but they do not match. Concavities, flat facets, craters and darker patches make one maximum higher than the other and one minimum deeper, and that asymmetry is what proves you have the full period. It is visible in the Vesta and Ariadne panels above.

How large the variation gets depends on where you happen to be standing. Amplitude is set by the angle between your line of sight and the asteroid's spin axis: viewed along the pole an elongated body barely varies at all, viewed across the equator it varies maximally. So the same asteroid shows a different amplitude at each apparition, and that is not noise but signal — observe several apparitions and the pattern of amplitudes solves for the pole direction and a shape model, a multi-year project achievable from one back garden. Amplitude also grows with phase angle, because the further the Sun is off to one side the longer the shadows and the more the shape is exaggerated; the effect is a few per cent per degree, small but enough that amplitudes are only comparable when the geometry is.

Some curves are genuinely strange

Binaries. A companion produces something no shape can: discrete, sharp mutual eclipse events superimposed on the ordinary rotational variation, on a completely different period. The 90 Antiope panel above shows this — two components of nearly equal size, eclipsing each other. Amateur photometrists have discovered many binary asteroids exactly this way, by noticing dips that the rotation could not explain. Tumblers are stranger still: an object rotating about a non-principal axis does not repeat with any single period at all, because the curve is the beat of two incommensurate rhythms, so it looks chaotic. Tumbling damps out over time, so these are almost always slow rotators, and each one is a clue about internal structure and collision history.

The spin barrier. Plot period against size and a wall appears at about 2.2 hours. A loose pile of rubble held together by its own weight cannot spin faster without flying apart. So the objects that do — and the catalog holds 1,817 of them — must be either single coherent rocks or genuinely held together by cohesion. Finding another one is a real result.

This is the fact that makes small-telescope rotation work viable. Across 29,970 asteroids with a measured amplitude, the median peak-to-peak variation is 0.35 magnitudes — half of them fall between 0.21 and 0.56 mag, and 7% exceed 0.8 mag. Only 4% vary by less than 0.1 mag.

A rig already delivering 5–10 mmag is therefore working with a signal-to-noise ratio on the astrophysics of roughly 35 to 70. You do not need better precision. You need photons from a fainter object, and enough hours to cover a cycle.

The rig this project uses is a 60 mm f/6.2 refractor with a cooled CMOS camera — the same instrument, and the same photometry pipeline, as The Variable Zoo Project, where it has produced light curves of variable stars between 9th and 12th magnitude at 5–10 mmag. The rig and the method are described there →

The geometry happens to suit moving targets very well:

• 1.29 arcsec per pixel, over a 3.0° × 2.0° field — hundreds of usable comparison stars in every frame.
• A main-belt asteroid near opposition drifts about 35 arcsec per hour. In a 30-second exposure that is 0.29 arcsec; even at 120 seconds it is 1.17 arcsec, or 0.9 pixel. No trailing.
• Over a six-hour session the target moves about 3.5 arcmin — a couple of per cent of the field width, so the same comparison stars stay in frame all night.

Brightnesses below are quoted at opposition, when an object is closest and visible all night. Working from the demonstrated performance: relaxing the target precision from 7 to about 25 mmag — still a 10:1 measurement of a typical amplitude — buys roughly 1.4 magnitudes of depth. Going from 30 to 120-second exposures adds about 1.5 more. Binning three or four frames into one photometric point, which costs nothing when the period is seven hours, adds most of another. That lands the practical limit around V = 15.

A reasonable first assumption is that everything bright was measured decades ago. Check it against the catalogs and the picture is more interesting. Almost every asteroid brighter than about 13th magnitude does have a published period — but having a period is not the same as having a reliable one.

The Asteroid Lightcurve Data Base grades every period on a quality scale. Across all 31,907 graded entries, 79% are not securely determined — they are single-night results, or fits that leave the period ambiguous, and they are flagged as such by the people who compiled them. Cross-matching those grades against our own orbits shows where that leaves a small telescope:

There are 243 asteroids that reach 15th magnitude or brighter at an ordinary opposition and whose published rotation period is not secure. That is the target list. It does not require pushing the equipment to its limit, and confirming or correcting an uncertain period is a publishable result.

The sequence matters more than the ambition. A pipeline that has never reproduced a known answer cannot be trusted with an unknown one.