How GPS Works: Atomic Clocks, Signals and Relativity

Every GPS fix is really a timing measurement. Each navigation satellite broadcasts the exact time it sent a signal along with its own position. A receiver on the ground compares arrival times from several satellites, converts those delays into distances, and solves for the single point that fits all the distances at once. The system is a clock network in space, and the accuracy depends on how well those clocks are kept.

The constellation

The GPS constellation is operated by the United States Space Force. It was designed around 24 operational slots spread across six orbital planes, with spare satellites added over time. The satellites fly in medium Earth orbit at about 20,200 km altitude, tilted about 55 degrees from the equator. Each completes two laps per sidereal day, so the constellation's sky pattern repeats daily.

Because the orbits are high, one satellite is above the horizon for a long stretch of the sky for any user on Earth. A receiver in an open area can usually see more than four satellites at once, which gives it the redundancy needed for a three-dimensional fix plus a timing correction.

Atomic clocks in orbit

Each GPS satellite carries several atomic clocks, typically rubidium and cesium standards. Their stability is the reason the system works. A timing error of one microsecond translates into about 300 meters of distance error, because radio signals travel at roughly 300,000 km per second.

The satellite clocks are tuned before launch to a nominal frequency of 10.23 MHz, which is then scaled up for the carrier signals. Ground control monitors each clock, and the navigation message carries corrections so receivers can estimate the satellite's true time.

What the satellite broadcasts

GPS satellites transmit on several carrier frequencies. The civilian signal on L1 sits at 1575.42 MHz, and the L2 carrier at 1227.60 MHz is used mainly by military and precision applications. Newer civil signals, including L5 at 1176.45 MHz, add redundancy and better performance in difficult conditions.

On top of the carriers, the satellites modulate spreading codes and a navigation message at 50 bits per second. The code lets a receiver lock onto a satellite and measure the signal's travel time. The navigation message includes the satellite's orbital parameters, called the ephemeris, and clock correction terms. A receiver needs this data to know where each satellite is at the moment of transmission.

From time delay to position

Multiply a signal's travel time by the speed of light and the result is a distance. One satellite gives a distance, which places the receiver on a sphere centered on that satellite. A second sphere narrows the position to a circle. A third sphere narrows it to two points, one usually far from Earth's surface and discarded.

The receiver also has its own clock, and that clock is not as accurate as the satellite clocks. Its error adds the same unknown offset to every distance measurement. That makes four unknowns: three coordinates and the clock bias. Four satellites provide four equations, which is why a standard fix needs at least four signals.

Satellites visible	What the receiver can solve	Fix type
1	Distance to one satellite	None
2	Ring of possible positions	None
3	Two points, only if clock is known	Clock not solved
4 or more	Position and clock bias	Full 3D fix

Receivers usually track many more satellites than four. Extra signals improve accuracy and let the receiver discard satellites with poor geometry or bad signal quality.

Relativity must be corrected

GPS is one of the clearest everyday uses of Einstein's theories. Two effects change the satellite clock rate compared with a clock on the ground.

Special relativity says moving clocks run slower. The satellites move at about 3.9 km/s relative to Earth's center, which slows their clocks by roughly 7 microseconds per day compared with ground clocks. General relativity says clocks run faster in weaker gravity. At 20,200 km, the satellites sit far higher in Earth's gravitational potential, and their clocks gain roughly 45 microseconds per day.

The net effect is about 38 microseconds per day, with satellite clocks running fast. Left uncorrected, that offset would accumulate into a position error of roughly 10 to 11 km per day, growing every day. To prevent this, the satellite clocks are set to run slightly slower than their nominal frequency before launch. The correction is part of the system design, not a minor refinement.

Error sources and accuracy

Signals slow down passing through the ionosphere and troposphere, and the delay depends on frequency and weather. Dual-frequency receivers compare L1 and L2 timing to estimate and remove most of the ionospheric error. Signals can also bounce off buildings or terrain before reaching the antenna, a problem called multipath. Poor satellite geometry, where the satellites bunch together in the sky, magnifies all errors. Engineers summarize that effect with a number called dilution of precision.

Typical consumer receivers in open sky give a horizontal position good to a few meters. Dual-frequency receivers and satellite-based augmentation systems such as WAAS and EGNOS can reach better accuracy, often around one meter or less under good conditions. Indoor or urban use is harder, because fewer satellites are visible and multipath is common.

Time transfer and selective availability

The same timing system is useful beyond navigation. Because every receiver can read GPS time with high precision, laboratories and telecommunications networks use it to synchronize clocks across continents. A ground station that tracks GPS time can keep its own clock aligned with the satellites to within tens of nanoseconds under good conditions, which is why GPS time is a common reference for power grids, financial networks and mobile phone base stations.

Accuracy has also changed over time. During the early decades, the U.S. government deliberately degraded the civilian signal through a feature called selective availability. It was switched off in May 2000, which immediately improved civilian positioning. Modern receivers and the augmentation systems described above build on that open signal, so the same satellites now serve far more users with much finer precision.

Cold start and everyday use

When a receiver is switched on after being off for a long time, it does not know where the satellites are. It must search for signals and download the navigation message, which can take anywhere from tens of seconds to several minutes. Receivers speed this up using stored almanac data, assisted data from a network, or a recent saved position. Once locked, a receiver can track satellites continuously and update its fix once per second or faster.

Phones combine GPS with other sensors such as accelerometers and Wi-Fi scanning, but the core fix still comes from timing signals sent from orbit.

Key takeaways

GPS works by timing radio signals from satellites whose clocks are known precisely. Distance comes from travel time, and four signals solve for position and the receiver's own clock error.

Relativity matters. Satellite clocks gain about 38 microseconds per day net, and that offset is corrected by design, or the system would drift by roughly 10 km per day.

Accuracy depends on the signal path and the geometry. Dual-frequency receivers, augmentation systems and clear sky view all improve a fix, while buildings, terrain and atmospheric delay degrade it.