The incredibly low computing power guiding Voyager into interstellar space
Launched in 1977, the Voyager space probes have traveled farther from Earth than any other human-made objects. Their onboard computer systems possess just 69 kilobytes of memory—millions of times less than a typical modern smartphone. Yet, through hyper-efficient assembly programming, these ancient computers continue to manage power, instruments, and communications from beyond our solar system.
Architecture of a 1970s Deep Space Computer
When Voyager 1 and Voyager 2 departed Earth in 1977, their computing architecture represented the cutting edge of aerospace engineering. Rather than relying on a single central microprocessor, each probe was built with three separate, dual-redundant computer systems: the Computer Command System (CCS), the Flight Data System (FDS), and the Attitude and Articulation Control System (AACS). Each pair of systems functioned as an active primary and a backup, meaning six distinct computers operated aboard each craft.
The total memory capacity distributed across these systems was roughly 69 kilobytes. To put this in perspective, the Computer Command System held about 4,096 words of memory per unit, and the Flight Data System held approximately 8,192 words per unit. These systems used magnetic core memory, which relies on tiny magnetized ferrite rings threaded onto wires, alongside early solid-state components. Unlike volatile modern RAM, magnetic core memory retained its data even if the system suffered an unexpected drop in electrical power.
Each subsystem had a tightly scoped operational mandate. The CCS was responsible for parsing instructions sent from Earth, sequencing mission events, and monitoring the spacecraft for onboard faults. The FDS collected scientific data and engineering telemetry from onboard instruments, formatting and compressing it before routing it to the communications array or to an onboard magnetic digital tape recorder. The AACS managed spacecraft orientation, pointing the high-gain antenna toward Earth and stabilizing the scientific scan platform.
The Voyager mission took advantage of a rare celestial alignment that occurs only once every 175 years. In the late 1970s, the outer planets—Jupiter, Saturn, Uranus, and Neptune—were positioned along an arc that allowed a single spacecraft to visit multiple worlds sequentially using gravity assists, also known as gravitational slingshots.
By flying close to a massive planet, a spacecraft can exchange momentum with the planet's orbital motion, gaining velocity relative to the Sun without firing propulsion engines. This technique allowed Voyager 2 to complete the historic 'Grand Tour,' flying past Jupiter in 1979, Saturn in 1981, Uranus in 1986, and Neptune in 1989. Voyager 1 used Jupiter and Saturn for gravity assists, choosing a trajectory that prioritized a close flyby of Saturn's large moon Titan, which deflected the probe out of the plane of the planets and upward toward deep space.
Executing these complex orbital trajectories required absolute precision from the onboard computers. If the guidance routines or trajectory timing were off by even a tiny fraction, the spacecraft would miss the narrow gravitational corridors required to reach subsequent outer planets, bringing an early end to the mission.
Programming and Troubleshooting Across Light-Hours
Operating computers across billions of miles requires entirely different protocols than conventional terrestrial software engineering. Ground teams wrote code primarily in assembly language and early versions of Fortran, painstakingly verifying every byte and memory address before transmitting instructions. Because the onboard memory was so scarce, software routines had to be reused and constantly overwritten depending on the active mission phase.
As the spacecraft moved farther from the Sun, the light-travel time for radio communications increased from minutes to hours. This latency made real-time control impossible. If an onboard anomaly occurred, the spacecraft could not wait for engineers on Earth to diagnose the problem; the Computer Command System had to detect the fault, isolate failing hardware, switch to a redundant unit, and place the probe into a safe operating state independently.
Over the decades, engineers developed clever workarounds to overcome hardware failures. For example, when specific memory chips degraded due to cosmic radiation or age, mission controllers uploaded patches that mapped instructions around bad memory sectors. Engineers rewrote the spacecraft's FDS code in flight to allow new data compression algorithms, enabling the probes to transmit higher-quality imaging data from Uranus and Neptune than their original 1977 configurations were designed to handle.
Power Depletion and the Cold Interstellar Medium
Voyager's computers and scientific payloads do not run on solar panels, because solar irradiance decreases drastically in the outer solar system. Instead, the probes are powered by radioisotope thermoelectric generators (RTGs). These devices convert the decay heat of plutonium-238 into direct-current electricity using thermocouples.
Because plutonium-238 has a half-life of roughly 87.7 years, and because the thermocouples themselves degrade over time, the electrical output of the RTGs drops by a small amount each year. At launch, the generators produced around 470 watts of electrical power per probe. Decades later, that power output has fallen by more than half, forcing mission managers to implement severe energy-saving measures.
To keep the primary computers and essential transmitters operating, engineers systematically turned off non-essential systems, scientific sensors, and internal heaters. Each decision represents a calculated trade-off: shutting down an instrument's heater risks freezing the mechanism permanently, but leaving it powered could trigger an under-voltage fault that resets the Computer Command System and causes the probe to lose its Earth-pointing lock.
Reaching Interstellar Space
Despite their diminutive memory by modern standards, the Voyager computers successfully carried both probes across the boundary of the solar system. In August 2012, Voyager 1 crossed the heliopause—the boundary where the solar wind is turned back by the interstellar medium—becoming the first human-made object to enter interstellar space. Voyager 2 followed, crossing the heliopause in November 2018 at a different location along the solar bubble.
Data gathered by the probes' remaining instruments provided humanity with its first direct measurements of the interstellar environment. The instruments recorded a sharp drop in solar wind particles and a corresponding surge in galactic cosmic rays, confirming the transition out of the heliosphere. The Flight Data System packaged these measurements and fed them to the radio transmitters, sending signals across more than a dozen billion miles.
Receiving these faint signals requires the massive parabolic dishes of the Deep Space Network (DSN), located in California, Spain, and Australia. The radio transmissions arriving at Earth carry an imperceptible fraction of a watt of power, requiring cryogenic receivers and digital signal integration to reconstruct the telemetry originally assembled by Voyager's 69-kilobyte computing core.
Key takeaways
•The Voyager probes divide their ~69 kilobytes of memory across three dual-redundant computer subsystems: CCS, FDS, and AACS.
•Engineers repeatedly reprogrammed and patched the probes' computers in flight to bypass failing hardware and improve data compression over immense distances.
•Voyager 1 and 2 operate on decaying plutonium-238 RTG power sources, requiring controllers to shut down heaters and instruments to keep essential computing and communications online.
•Voyager 1 crossed into interstellar space in 2012, followed by Voyager 2 in 2018, both transmitting direct scientific measurements of the interstellar medium back to Earth.