When the twin Voyager spacecraft were launched from Cape Canaveral in 1977, no one expected them still to be operating in the 2020s. They had been designed for a four-year primary mission to Jupiter and Saturn, taking advantage of an unusually favourable alignment of the outer planets. Almost half a century later, both probes continue to return data from beyond the heliopause, the boundary that marks where the Sun's influence gives way to interstellar space. The story of how a pair of 1970s machines came to outlive their creators' expectations is partly a story of careful engineering and partly one of patient, sometimes ingenious, operations.

The first design decision that made the long life possible was redundancy. Every major subsystem on each Voyager was built in duplicate, with the second copy ready to take over if the first failed. This was not a matter of caution alone; deep-space missions cannot be repaired, and a single point of failure in a critical system could end the mission many years early. The redundancy extended to the onboard computers, which were primitive even by the standards of the 1970s, with far less memory than a pocket calculator today, but were shielded against radiation and designed to recover from errors automatically.

The second decision was power. Voyager does not rely on solar panels, which would be useless in the outer solar system, where sunlight is weak. Instead, each probe carries three radioisotope thermoelectric generators, or RTGs, which convert the heat from decaying plutonium into electricity. The power output of these generators falls slowly but steadily over time, by about four watts per year. By 2020, the available power had dropped to the point where engineers had to begin switching off instruments that had run continuously for decades. Each decision about which instrument to sacrifice was made jointly by mission scientists and operators, based on the scientific priorities that remain at Voyager's current distance.

That distance is itself difficult to grasp. Voyager 1 is now more than 23 billion kilometres from Earth, and a radio signal sent at the speed of light takes over 22 hours to reach it and another 22 to return. Commands are therefore sent with long lead times, and mistakes cannot be corrected quickly. Dr. Suzanne Dodd, who led the Voyager programme at NASA's Jet Propulsion Laboratory for many years, has described the operations team as 'working with a 50-year-old machine whose instruction manual was partly lost'. Some of the engineers who originally designed the probes have retired, and their successors have had to reconstruct details of subsystems from incomplete documents.

Occasionally, Voyager has given its operators reason to worry. In 2022 Voyager 1 began sending back garbled telemetry: the numbers describing its orientation and internal state stopped making sense, even though the probe itself was clearly still functioning. After months of analysis, engineers traced the fault to a memory chip in one of the computers. They were able to work around the damage by storing the relevant data in a different location, a fix that required uploading new instructions to a machine whose operating system had not been updated in any substantial way since its launch. The fact that such a repair was possible is itself a testament to the flexibility of the original design.

Scientifically, the most important results of Voyager's extended mission have come from outside the heliosphere, the bubble of charged particles blown by the Sun into the surrounding medium. Voyager 1 crossed the heliopause in 2012, Voyager 2 in 2018. Beyond it, the probes have measured the temperature and density of the interstellar medium directly, rather than inferring them from astronomical observations. The data have confirmed some predictions, such as the sudden increase in cosmic-ray intensity, and overturned others, including the expected decrease in plasma density. Astrophysicist Dr. Raji Mehta has argued that the Voyager measurements make it possible, for the first time, to 'calibrate' models of the interstellar environment that had previously rested on indirect evidence.

At some point in the 2030s, the RTGs will no longer be able to power any instrument, and the two probes will go silent. They will continue on their trajectories, drifting through the galaxy at roughly 17 kilometres per second. Voyager 1 is expected to pass within about 1.6 light years of the star Gliese 445 in around 40,000 years; Voyager 2 will pass a similar distance from another star, Ross 248, at about the same time. Neither encounter will be observable from Earth, and neither probe is likely to encounter anything else in the vast emptiness between. Their continued existence, beyond the end of active operations, is a reminder that a well-made machine can survive far longer than its designers dared to plan for.