The 0.1 Millimeter That Keeps a Satellite in Orbit: A Primer on Star Trackers
Archi3 min read
Contents
The Part That Weighs Less Than a Paper Clip#
A star tracker is a small camera that looks at the sky and compares what it sees to a catalog of known stars. Most units weigh under 500 grams and draw a few watts. They are not the biggest line item on a spacecraft, and they rarely appear in press releases about new constellations.
They also determine where the satellite is pointing. That number sets everything downstream: where an antenna aims, which patch of ground a camera captures, how much propellant gets burned correcting a drift.
Why pointing is harder than it sounds#
A communications satellite in geostationary orbit hovers about 35,786 kilometers above the equator. A pointing error of 0.01 degrees at that altitude works out to roughly 6 kilometers of ground error. For a spot beam serving a dense city, that is the difference between covering it and missing it.
Earth sensors help, but they are coarse. A horizon sensor might resolve heading to a fraction of a degree. Star trackers push that into arcseconds, units where one arcsecond is 1/3600 of a degree. The best flight units reach a few arcseconds of cross-axis accuracy, which is enough for optical imaging at high resolution.
The smallest detail: optical alignment#
Inside a star tracker, a lens focuses starlight onto a detector. Light arrives at the detector at an angle. The distance between the lens and the detector, often a few millimeters, sits on a mount that may be made of titanium or a composite.
The problem is thermal expansion. Titanium expands by about 8.6 parts per million per degree Celsius. A composite housing might expand by less than 1 part per million, but it can also change shape unevenly depending on fiber orientation.
Over a temperature swing of 60 degrees Celsius, a 10 millimeter spacing can shift by several micrometers. That sounds trivial. But at the pixel level of a modern star tracker detector, a few micrometers is a meaningful fraction of a pixel. The software can calibrate around some of this, but only if the drift is predictable and slow. Fast, uneven shifts are harder to model.
Engineers spend months characterizing how each unit behaves across its operating temperature range. Sometimes they shim the lens mount. Sometimes they add heaters. Sometimes they simply reject parts that drift too much. A single misaligned unit on a satellite that cannot be serviced becomes a permanent pointing bias.
The catalog problem#
The star tracker is only as good as the map it uses. Flight software carries a star catalog, typically a few thousand entries bright enough to be seen reliably without confusing them with noise or debris. If the catalog has an old position or a misidentified star, the tracker can lock onto the wrong reference and produce a confidently wrong attitude.
That is why missions run what is called a lost in space recovery routine. If the tracker cannot match its view to the catalog within a timeout, the spacecraft drops to a safe mode and waits for a ground command. The delay can cost days of observing time or a missed imaging window. It can also force a propellant burn that shortens the mission.
What usually fails first#
Star trackers are optoelectronic devices. The detector, often a CMOS sensor, degrades under radiation. Hot pixels accumulate. Lenses darken slightly as coatings and materials are exposed to ultraviolet light and atomic oxygen in low Earth orbit.
Most of these effects are slow. The bigger risk is a single event upset: a cosmic ray flipping a bit in memory. Designers handle this with error correction and periodic self-checks. A tracker that misses its checks can be reset from the ground, but each reset costs time.
A practical checklist for mission designers looks like this. Characterize thermal drift across the full range, not just room temperature. Keep at least two trackers, or one tracker plus a capable backup attitude source. Budget time for lost in space recovery. Track detector degradation over the mission life and plan a recalibration schedule.
The cost of getting it wrong#
The hardware is cheap relative to the consequences. A pointing error that costs a few percent of a satellite's imaging capacity compounds over years. On a remote sensing mission, that is revenue. On a scientific mission, it is data that never gets collected.
The smallest part on the spacecraft is often the one that decides whether the rest of it does its job.
More articles