Deep Dive: Timing at the Moon

June 2, 2026 / CAPSTONE Mission, News, Technical Deep Dive

Timing is Infrastructure: How an Atomic Clock on CAPSTONE Is Setting the Baseline for Future Lunar Operations

Advanced Space has been operating an atomic clock on CAPSTONE™ in cislunar space for nearly four years — turning what was once theory into flight-proven data, and laying the foundation for the timing standards, autonomous navigation, and quantum-era infrastructure that the next decade of lunar operations will rely on.

1. Position is really a timing problem

On Earth, we treat knowing where we are as a solved problem. Open a phone, and a position appears in a second or two. What most people never notice is that the answer is built almost entirely out of time. A satellite navigation receiver does not measure distance directly. It measures how long a signal took to arrive, and then converts that tiny interval into a distance. Position, in other words, is a clock problem wearing a map’s clothing.

That fact becomes impossible to ignore the moment a spacecraft leaves Earth’s neighborhood. At the Moon, the quality of a spacecraft’s onboard clock stops being a footnote and becomes one of the key factors determining whether it can find itself at all.

2. Distance is just time, multiplied by the speed of light

Radio signals travel at the speed of light. If you know exactly how long a signal was in flight, you know how far it traveled — multiply the travel time by the speed of light, and you have a distance.

The catch is in the word exactly. Light covers roughly 30 centimeters — about one foot — in a single nanosecond, a billionth of a second. The European Space Agency puts it bluntly in describing its own navigation system: a clock error of more than a few nanoseconds already throws a position off by more than a meter, and a one-second error would leave a receiver guessing at a location somewhere near the Moon [1]. Scale that up and the numbers get unforgiving very quickly: a timing error of one microsecond — a millionth of a second — corresponds to roughly 300 meters of position error [2].

This is why timekeeping and navigation are, at the deepest level, the same discipline. To know where something is to within a meter, you have to know when to within a few billionths of a second.

3. The Moon has no GPS

Satellite navigation on Earth works because dozens of navigation satellites orbit close enough that several are always overhead, beaming a strong, well-conditioned signal at any receiver below. The Moon has nothing comparable. There is no native lunar positioning constellation, no purpose-built lunar timing service, no infrastructure designed for vehicles operating in cislunar space. Put simply, the Moon has no Global Positioning System.

What lunar missions actually rely on day to day is tracking from Earth: large antennas, such as those of NASA’s Deep Space Network, which measure a spacecraft’s distance and speed and relay the answer back. It works, but it has two structural limits. The big dishes are a shared, oversubscribed resource that cannot babysit every vehicle as traffic grows. Also, a spacecraft that depends on the ground to know where it is cannot truly navigate on its own.

The Moon needs its own positioning, navigation, and timing foundation. Building it begins with the clocks.

4. One-way versus two-way ranging — and why the clock decides

There are two ways to turn a radio signal into a distance, and the difference between them is the whole story.

In two-way ranging, a ground station sends a signal, the spacecraft echoes it back, and the ground times the full round trip. The master clock stays on Earth, so the spacecraft can carry a mediocre clock and still be tracked accurately. The price is that the spacecraft is a passive participant — it learns its position only after the ground does the work and tells it.

In one-way ranging, the spacecraft times the signal’s arrival by itself. This is how everyday satellite navigation works, and it is what lets a device compute its own position without asking anyone. But it moves the burden of precise timekeeping onto the vehicle. Now the spacecraft’s own clock has to stay honest to within billionths of a second over long stretches between contacts. Researchers writing in the Institute of Navigation’s journal NAVIGATION put the constraint plainly: ordinary radio oscillators have historically been too unstable for one-way data to be useful for navigation [3]. The clock is the gatekeeper. Get it stable enough, and a spacecraft can navigate autonomously. Fall short, and it stays tethered to the ground.

5. Standard oscillators versus the chip-scale atomic clock

Most spacecraft radios keep time with a quartz crystal oscillator — a small vibrating crystal whose oscillations are counted to mark off the seconds, often with a temperature-compensation circuit to correct for thermal drift. Quartz oscillators are tiny, cheap, rugged, and frugal with power, which is exactly why they are everywhere. What they are not is stable enough for precision one-way navigation over time. Their vibrations change measurably with temperature, age, and operating conditions, and those changes accumulate.

NASA’s Jet Propulsion Laboratory offers a memorable yardstick: even a high-quality quartz oscillator can drift by about a nanosecond after just an hour, and by something like a millisecond after six weeks — a timing slip that translates into a position error of roughly 300 kilometers [4]. That creates a hopeless situation for someone tracking a deep space probe to within meters.

An atomic clock solves the stability problem by tying its tick to an unchanging physical constant — not to a vibrating crystal that ages and wanders, but to a quantum-mechanical transition inside an atom. In a cesium-based atomic clock, the relevant transition occurs at exactly 9,192,631,770 cycles per second, a frequency so reliable that the international community has used it to define the second itself since 1967. The clock generates microwaves at that frequency, watches the cesium atoms respond, and ticks off seconds against an unchanging atomic constant. That constant gives an atomic clock the long-term stability that vibrating crystals cannot match.

The trouble has always been that atomic clocks were large, heavy, and power-hungry — fine for a navigation satellite, unthinkable on a small spacecraft. The chip-scale atomic clock changed that equation. Built with the same miniaturization techniques used for microchips, it packs atomic-grade timekeeping into a part smaller than 17 cubic centimeters – about the size of a matchbox – that draws under 120 milliwatts, and weighs about 35 grams [5]. The part that flies on CAPSTONE was built by Microchip Technology of Chandler, Arizona, and provided to the mission by Orion Space Solutions.

How clock stability is measured

The gap between a quartz oscillator and a chip-scale atomic clock is easiest to see if you measure clocks the way clock engineers do — not by how accurate a clock is at a single instant, but by how steady its rate stays over a stretch of time. That measure is called the Allan deviation. It is written as a fraction — say, one part in ten billion — quoted at a stated averaging interval, and a smaller deviation means you have a steadier clock. The reason it matters here is that what a coasting spacecraft needs is not a clock that is good for one second, but one that holds its rate over the minutes, hours, or days between corrections. Read at those intervals, the three relevant classes of clock separate cleanly, and the manufacturer and space-agency data tell a consistent story [5][6]:

Clock type	Allan deviation at 1 second	Allan deviation at ~1,000 seconds (about 17 minutes)	Long-term behavior	Size and power
Temperature-compensated quartz oscillator (the usual reference inside a small-spacecraft radio)	~1 × 10⁻⁹	Drift takes over — up to roughly 10,000× less steady than the chip-scale atomic clock over long intervals	Ages and wanders continuously	Tiny, a few milliwatts
Chip-scale atomic clock — CAPSTONE’s class (Microchip SA.45s)	4 × 10⁻¹⁰	1.3 × 10⁻¹¹	Ages under 9 × 10⁻¹⁰ per month	Under 17 cubic centimeters, under 0.12 watt, ~35 grams
Navigation-satellite atomic clock (for example, Galileo’s passive hydrogen maser)	~9 × 10⁻¹³	~2 × 10⁻¹⁴	Reaches a few parts in 10¹⁵ over a day	~15 kilograms, ~45 watts

A quick way to read the table: the chip-scale atomic clock is around a thousand times steadier than the quartz oscillator over the intervals that matter for coasting, and the full navigation-satellite clock is steadier still — but it weighs roughly four hundred times as much and draws several hundred times the power. (NASA’s Jet Propulsion Laboratory has pushed the high end further still with its Deep Space Atomic Clock, a mercury-ion clock built specifically to improve deep-space navigation timing [7]; broader cross-constellation comparisons of navigation-satellite clocks appear in [8].)

The point of the chip-scale clock is not that it beats a full navigation-satellite atomic clock — it does not. The point is that it brings atomic-class stability down to a size, weight, and power budget a small spacecraft can actually afford. That is the door it opens.

6. A quantum sensor at the Moon

The chip-scale atomic clock is not just a smaller version of a familiar instrument. It is a quantum device in the strict physical sense: the cesium hyperfine transition that anchors its tick is the same atomic-level phenomenon that defines the second in the International System of Units. That measurement underpins the broader fields of quantum sensing, quantum metrology, and the time-and-frequency-transfer architectures the quantum networking community is now building toward. Operating one in lunar orbit is, in those terms, operating a quantum sensor in cislunar space — among the first publicly known instances, and quite possibly the first.

The point is not the label. The point is that the same physics that produces a stable timing reference also underpins the precision time and frequency transfer between distant nodes, the coherent measurement networks, and the eventual entanglement-based links that will be required to extend quantum communication across cislunar distances. Advanced Space is not theorizing about that future. We have a quantum-class timing device operating at the Moon today, generating data that will inform how the next generation of these architectures gets designed — and we are continuing to evaluate other quantum-class instruments, references, and sensors for missions still to fly.

7. CAPSTONE: atomic timing in a CubeSat, at the Moon

Advanced Space’s CAPSTONE mission set out to walk through the door this technology opens. CAPSTONE is a twelve-unit CubeSat about the size of a microwave oven; it launched on June 28, 2022, and has now been operating in cislunar space for nearly four years. It became the first commercially owned and operated spacecraft to fly in the orbital region around the Moon [9][10]. It operates in an elongated halo orbit swinging from as close as about 1,600 kilometers over one lunar pole to roughly 70,000 kilometers over the other every approximately seven days [10].

CAPSTONE was designed as a navigation pathfinder. It carries two demonstrations whose significance is difficult to overstate: the first peer-to-peer ranging between two spacecraft at the Moon, and an in-flight test of one-way ranging anchored by the onboard chip-scale atomic clock.

8. The crosslink with the Lunar Reconnaissance Orbiter

The first demonstration was peer-to-peer navigation between two spacecraft. Working with the team behind NASA’s Lunar Reconnaissance Orbiter (LRO), which has circled the Moon since 2009, CAPSTONE exchanged ranging signals directly with that orbiter to measure how far apart they were and how fast that distance was changing — and its onboard Cislunar Autonomous Positioning System (CAPS) software used those measurements to estimate the position and velocity of both vehicles [9]. The first such measurements were collected on May 9, 2023, in what we described as a first-of-its-kind crosslink at the Moon; a later pass ran 66 minutes and produced 200 measurements [9][12]. Achieving this result took more than two years of joint work with the Lunar Reconnaissance Orbiter team and a great deal of on-orbit refinement by our payload team to get the most out of every pass [9].

The idea behind the LRO crosslink experiment is quietly radical. Instead of every lunar spacecraft individually competing for time on Earth’s tracking antennas, vehicles can help locate one another, and the network grows more accurate and more resilient as more participants join. It is the architectural seed of an inter-spacecraft positioning fabric that the Moon will need as traffic grows.

9. One-way ranging from Earth — and the reality of doing it on a live spacecraft

The second demonstration is where the chip-scale atomic clock earns its place. CAPSTONE flew one of these clocks specifically to test one-way ranging with NASA’s Deep Space Network — the global network of tracking antennas operated by the Jet Propulsion Laboratory — using a flight radio derived from JPL’s Iris transponder [10][13][14]. The experiment is straightforward, in principle: a Deep Space Network station on Earth sends a signal at a known time; CAPSTONE receives that signal and time-tags it using the onboard chip-scale atomic clock; the spacecraft compares the “send” time on Earth with the “receive” time onboard and removes the duration of time when the signal was being processed by either radio. The result yields distance and relative motion — the core ingredients of navigation.

On a whiteboard, it is neat and clean. On a live spacecraft, it is anything but.

The atomic clock is exquisitely sensitive to its environment. Thermal variations from lunar eclipses, changes in spacecraft attitude, and even heat radiated by the radio itself shift the clock’s behavior in measurable ways. The clock does not tick at the same rate that it would in a laboratory on Earth, either: relativistic effects from CAPSTONE’s velocity and from its position within the lunar gravity field alter the clock’s tick rate relative to its Earth-based counterparts. Every one of these effects has to be characterized in flight and modeled out to extract a clean navigation answer.

Pre-flight testing of the pairing between the spacecraft’s radio and the atomic clock had already shown one-way ranging precision in the range of a fraction of a meter to a couple of meters, with relative-speed measurements good to about 11 millimeters per second [3] — performance simply not reachable with a crystal oscillator alone. Demonstrating it in flight, in deep space, on a small commercial spacecraft, is what turns a laboratory result into operational data. The flight characterization is already informing peer-reviewed work across the navigation community, including a full orbit-determination demonstration using onboard one-way radiometric data from CAPSTONE (see Related papers and presentations, below).

10. The Moon keeps its own time

There is a deeper reason all of this is hard, and it reaches past engineering into physics. A clock is only useful for navigation if it agrees with the other clocks in the system. At the Moon, they cannot agree perfectly — because time itself runs at a different rate there.

Einstein’s general theory of relativity predicts that clocks in weaker gravity tick faster. The Moon’s gravity is weaker than Earth’s, so a clock on the lunar surface runs faster than an identical clock at home; once the effects of motion are folded in as well, the net result is that a lunar clock runs roughly 56 microseconds per day ahead of an Earth clock [15]. That is not a small number for a navigation system. Run it through the same arithmetic that governs everything in this story — roughly 300 meters of range error for every microsecond of timing error — and an uncorrected day at the Moon would compound into something on the order of seventeen kilometers of accumulating drift. At the Moon, getting Einstein right is not an academic nicety. It is a hard navigation requirement.

The policy world has been moving to catch up with the physics. In April 2024, the U.S. government directed NASA — working with other federal agencies — to establish a Coordinated Lunar Time by the end of 2026, and the International Astronomical Union has called on the world’s space agencies to do the same [16][17]. As NASA’s space communications and navigation chief has put it, when you travel to another body it makes sense that each one gets its own heartbeat [18]. Coordinated Lunar Time is sometimes described informally as a “time zone for the Moon,” but the analogy breaks down on inspection. It will not be a time zone in the everyday sense; it will be a time scale — a reference coordinate, the zero point from which everything else at the Moon is measured, maintained much the way Coordinated Universal Time is maintained on Earth: by an ensemble of atomic clocks compared and weighted together [17].

This is precisely where doing the work in flight, rather than on paper, becomes decisive. A lunar time standard cannot be settled in the abstract. It has to be anchored in real atomic clocks operating in the real lunar environment, measured against real tracking from Earth — the kind of operational data that exists only when someone actually flies it. That is the work we are doing. CAPSTONE has put an atomic clock to work in orbit at the Moon and ranged it against both NASA’s Deep Space Network and a partner spacecraft, generating exactly this class of measurement under real conditions [9][13]. We are doing it in concert with the Jet Propulsion Laboratory and others across NASA, and our aim is direct: to help establish the empirical baseline on which future lunar time standards, networks such as LunaNet, and the broader cislunar infrastructure that growing traffic will demand can be built. Leadership in this field is not a matter of having the best diagram. It is a matter of being the team with hardware on orbit, generating the data, while the standard is still being written.

11. Flight is the only honest test

A model of a clock is a set of assumptions. A clock in deep space is a fact. The two are not the same, and the distance between them is exactly where missions succeed or fail.

On the ground you can predict how a chip-scale atomic clock should behave — how steadily its rate should hold, how radiation should nudge it, how thermal swings and launch vibration should perturb it. But the real environment does things no model fully anticipates: the actual radiation dose over months in cislunar space, the real thermal cycling through lunar eclipses, the true behavior of a commercially derived part operating far from the near-Earth conditions it was first built for, and the operational surprises that appear only when a spacecraft is genuinely out of contact and has to fend for itself. Flying the clock is how assumptions turn into measurements.

CAPSTONE, in that sense, is a learning machine. The one-way ranging performance was characterized in the laboratory first and then demonstrated for real in flight, against both NASA’s Deep Space Network and the Lunar Reconnaissance Orbiter [3][13][18]. The first-of-its-kind crosslink took more than two years of joint work with the orbiter’s team to achieve, and our payload team kept refining the radio’s configuration on orbit, pass after pass, to pull more out of every measurement [9]. When a missed ground contact once triggered a spacecraft reset, that was not merely an anomaly — it was a real-world demonstration that the vehicle could carry on when no one was talking to it, and a concrete lesson in designing for that case [9]. None of that knowledge existed before the mission flew.

And the value of a pathfinder is that the knowledge does not stay with the pathfinder. Every model coefficient we tighten with flight data, every operational procedure we harden, every surprise we catalog becomes an input to the next mission, and the one after that. This is how a real capability compounds — not by drawing a better diagram, but by flying, measuring, and feeding what is learned forward into the next design. It is also why flight heritage cannot be shortcut. An architecture can be copied on paper; what cannot be copied are the years of operating data that reveal how that architecture actually behaves at the Moon. We have been accumulating that record since 2022, and each mission we fly builds on it rather than starting over.

12. What this means for what’s coming

The coming decade will put more spacecraft on and around the Moon than have flown there in all of history combined: landers, relays, science orbiters, crewed vehicles, and the infrastructure to support them. Earth’s tracking antennas cannot scale to meet that demand one vehicle at a time, and waiting hours for a position update is incompatible with vehicles that need to make decisions in real time. The Moon’s missing positioning, navigation, and timing foundation has to be built — and the precision instruments we are operating today, including the chip-scale atomic clock on CAPSTONE, are the building blocks.

The way out is autonomy. Autonomy at these distances rests on a foundation that is easy to overlook precisely because it is so fundamental: a spacecraft has to keep excellent time, and that time has to be reconciled with a standard the Moon does not yet have. The chip-scale atomic clock is what makes that foundation portable enough for the small, affordable spacecraft that will do much of the coming work. CAPSTONE has shown it can be done. The clocks are no longer the bottleneck — and the work of turning that proof into a shared timing and navigation backbone for the Moon is happening now, with us in the middle of it, one stable tick at a time.

It is tempting to treat time as a background detail — just another field in a telemetry packet. But at the Moon, and throughout the broader space environment, timing is critical infrastructure.

Notes and sources

European Space Agency — Galileo: no way without time. Source for the figure that light travels about 30 centimeters per nanosecond, and that a clock error of more than a few nanoseconds already exceeds a meter of position error. https://www.esa.int/Applications/Satellite_navigation/Galileo_no_way_without_time
Scientific Research Publishing — Satellite Clock Error and Orbital Solution Error Estimation. Source for the relationship that a timing error of one microsecond corresponds to roughly 300 meters of range error. https://www.scirp.org/html/3-8501083_42845.htm
Institute of Navigation, NAVIGATION: Journal of the ION — Formulation and Characterization of One-Way Radiometric Tracking with the Iris Radio Using a Chip-Scale Atomic Clock (Ely, Towfic, Sorensen). Source for ordinary radio oscillators being too unstable for one-way navigation, and for the pre-flight one-way ranging precision (roughly 0.38–2.21 meters, with relative speed to about 11 millimeters per second). https://navi.ion.org/content/71/1/navi.633
NASA Jet Propulsion Laboratory — What Is an Atomic Clock? Source for a high-quality quartz oscillator drifting about a nanosecond in an hour and about a millisecond (≈300 kilometers) over six weeks. https://www.jpl.nasa.gov/news/what-is-an-atomic-clock/
Microchip Technology — CSAC-SA.45s Chip-Scale Atomic Clock datasheet. Source for the chip-scale clock’s Allan deviation (4 × 10⁻¹⁰ at 1 second; 1.3 × 10⁻¹¹ at 1,000 seconds), aging (under 9 × 10⁻¹⁰ per month), and size, weight, and power (under 17 cubic centimeters, under 120 milliwatts, about 35 grams). https://ww1.microchip.com/downloads/en/DeviceDoc/Microchip_CSAC_008_Datasheet_900-00744-008_C.pdf
Observatory of Neuchâtel / European Space Agency (Wang, Busca et al.) — The On-Board Galileo Clocks: Rubidium Standard and Passive Hydrogen Maser. Source for the navigation-satellite passive hydrogen maser figures used in the table (8.8 × 10⁻¹³ at 1 second; flicker floor near 7 × 10⁻¹⁵; about 2 × 10⁻¹⁵ over a day; about 15 kilograms and 45 watts). https://www.academia.edu/6068472/The_onboard_galileo_rubidium_and_passive_maser_status_and_performance
Ely et al., Radio Science (2025) — The Benefit of Space Clocks for the Deep Space Network. Source for the side-by-side Allan deviation of chip-scale clocks, rubidium standards, hydrogen masers, and NASA’s Deep Space Atomic Clock. https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2025RS008244
Griggs et al., Radio Science — Short-term GNSS satellite clock stability. Source for the cross-constellation (GPS, GLONASS, Galileo, BeiDou) characterization of navigation-satellite clock stability via the Allan deviation. https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1002/2015RS005667
Advanced Space — CAPSTONE mission and Cislunar Autonomous Positioning System (CAPS). Source for the peer-to-peer crosslink with the Lunar Reconnaissance Orbiter, the first-of-its-kind crosslink and the multi-year effort to achieve it, the on-orbit refinement of the radio, and the spacecraft’s autonomous operation through a missed ground contact. https://advancedspace.com/missions/capstone/ and https://advancedspace.com/caps/
Inside GNSS — CAPSTONE spacecraft details. Source for CAPSTONE as the first commercially owned spacecraft to operate at the Moon. https://insidegnss.com/advanced-spaces-capstone-mission-surpasses-440-days-of-lunar-operations-testing-pnt-technologies/
Space News – NASA’s tiny CAPSTONE cubesat launches on pioneering moon mission. Its halo-orbit geometry (about 1,600 kilometers over one pole to about 70,000 kilometers), and the one-way ranging experiment with the Deep Space Network
Inside GNSS — CAPSTONE Mission Surpasses 440 Days of Lunar Operations. Source for the 66-minute crosslink pass that produced 200 crosslink measurements. https://insidegnss.com/advanced-spaces-capstone-mission-surpasses-440-days-of-lunar-operations-testing-pnt-technologies/
SpaceNews — NASA’s CAPSTONE is testing “Autopilot” software suite for cislunar operations. Source for the in-flight demonstration of one-way uplink ranging with the Jet Propulsion Laboratory’s Deep Space Network. https://spacenews.com/nasa-capstone-testing-autopilot-software-suite-cislunar-operations/
AIAA / ASCEND — CAPSTONE: A Unique CubeSat Platform for a Navigation Demonstration in Cislunar Space.Source for the description of CAPSTONE’s one-way ranging demonstration using a chip-scale atomic clock. https://arc.aiaa.org/doi/10.2514/6.2022-4382
Ashby & Patla (National Institute of Standards and Technology), arXiv — A Relativistic Framework to Establish Coordinate Time on the Moon and Beyond. Source for the relativistic clock-rate difference between the Moon and Earth — roughly 56 microseconds per day. https://arxiv.org/pdf/2402.11150
Reuters (via AOL) — White House directs NASA to create time standard for the moon. Source for the April 2024 directive to establish Coordinated Lunar Time by the end of 2026 and for the “heartbeat” framing from NASA’s space communications and navigation chief. https://www.aol.com/news/exclusive-white-house-directs-nasa-183828258.html
American Institute of Physics, FYI — Lunar Time Standard Taking Shape. Source for Coordinated Lunar Time being defined as a reference coordinate (a time scale, not a time zone), the 2026 deadline, the International Astronomical Union’s call for an international standard, and the proposal for a lunar clock ensemble maintained much the way Coordinated Universal Time is maintained on Earth. https://www.aip.org/fyi/lunar-time-standard-taking-shape
Ely, Zara, Sorensen, Towfic, Ott, Forsman, Baker — Orbit Determination Demonstration Using Onboard One-Way Radiometrics From the Iris Radio on the CAPSTONE Mission (2024). Space Dynamics Laboratory Publications, Paper 286. Source for the in-flight one-way radiometric orbit determination demonstration on CAPSTONE. https://digitalcommons.usu.edu/sdl_pubs/286