Both sites, or it didn’t happen

Folded to fit the fairing

Voyager, on a few hundred watts and falling

The engineering problem is simple: a fixed inventory of plutonium, a growing distance, and instruments that cannot run on nothing.

A deep space network dish against a night sky
Deep Space NetworkA dish this size is needed at this end because the spacecraft answers on tens of watts.

The power budget that governs everything

Both Voyager spacecraft — Voyager 1, launched on 5 September 1977, and Voyager 2, launched sixteen days earlier on 20 August — carry a set of radioisotope thermoelectric generators, RTGs, that convert the heat of decaying plutonium-238 into electricity. At launch, each spacecraft's three RTGs together produced roughly 470 watts of electrical power. That number has been falling ever since, for two compounding reasons: plutonium-238 has a half-life of 87.7 years, so its heat output declines on a fixed curve; and the thermocouples that convert heat to electricity degrade independently, at their own rate. The combined effect is a power loss of roughly four watts per year. In 2025, each spacecraft is generating somewhere in the neighbourhood of 220 to 240 watts — less than half the launch endowment, and declining on a schedule NASA's Jet Propulsion Laboratory tracks in detail.

A spacecraft launched in 1977 is still returning data on decaying plutonium, with instruments switched off one at a time to make it last.

That number has to cover everything: keeping the electronics above their minimum operating temperature in an environment that is effectively 2.7 kelvin outside the spacecraft body, running the attitude-control thrusters, operating the communications hardware needed to reach the Deep Space Network, and powering whatever science instruments remain active. It is a zero-sum constraint, and it has been managed, since the 1990s, by a sequence of deliberate shutdowns.

Switching things off one at a time

The instruments Voyager carries were designed to characterise the outer planets — their magnetospheres, atmospheric compositions, rings, and moons. Most of that work was done by the mid-1980s, and the active scientific mission since then has been the study of the heliosphere and, eventually, the interstellar medium. That narrower scope meant some instruments could go dark without loss of mission-critical data.

A large radio dish on dry scrub plateau with engineers in hi-vis at its base
Antenna fieldA phased array trades one large dish for many small ones and leaves the rest to the correlator.Photo: West arm of UTR-2 radio telescope phased array antenna · Wikimedia Commons

The process is not simple. Each instrument draws power, but it also produces heat that contributes to the thermal balance of the spacecraft. Switch off a heater without accounting for what it was warming, and the result can be a bus or circuit that drops below its survival temperature. Every shutdown decision therefore involves thermal modelling, not just a power subtraction. JPL engineers work from a model of the spacecraft's thermal environment that has been maintained and refined for decades, updated as the actual temperature readings — still returned by both spacecraft — reveal how the system is behaving in practice.

Key numbers

RTG power at launch: ~470 W per spacecraft (both Voyager 1 and 2)
RTG power circa 2025: ~220–240 W per spacecraft (estimated)
Rate of power decline: ~4 W per year
Plutonium-238 half-life: 87.7 years
Transmitter RF output: ~23 W
Downlink data rate (Voyager 1, current): ~160 bits per second
Current distance, Voyager 1: >160 astronomical units
Round-trip signal time, Voyager 1: >23 hours
Voyager 1 heliopause crossing: August 2012, confirmed September 2013
Voyager 2 heliopause crossing: November 2018

By the early 2020s, Voyager 1 retained five functioning science instruments; Voyager 2, on a slightly different trajectory through the heliosphere, also retained a small complement of active instruments. In 2023, NASA announced that a heater serving one of Voyager 2's instruments had been turned off — a deliberate sacrifice of thermal margin to extend the power runway for science operations further into the decade. The logic is straightforward: each instrument kept alive buys another year or two of data from a region of space no other spacecraft has reached or is scheduled to reach. The cost is an irreversible one, because turning instruments back on is not necessarily possible once thermal cycles have stressed the hardware.

What interstellar space has to teach, and at what data rate

Voyager 1 crossed the heliopause — the boundary where the solar wind's pressure yields to the interstellar medium — in August 2012, as confirmed in a Science paper published in September 2013. Voyager 2 followed in November 2018. The confirmation in each case rested on plasma wave measurements showing the density of the surrounding medium had jumped sharply, consistent with the denser interstellar plasma, and on the behaviour of the magnetic field. These are the first direct measurements of the local interstellar medium by an in-situ probe, and no follow-on mission is currently in development that would replicate them.

A crane lowers a large dish-shaped telescope reflector onto its support structure in a desert setting
KarooA dish arrives in panels and is assembled where it stands, which is why the crane is a permanent line in the schedule.Photo: SKA-MPI-2 · Wikimedia Commons

The data comes back at rates that would seem vanishingly slow by terrestrial standards — on the order of 160 bits per second from Voyager 1 at its current distance of more than 160 astronomical units from the Sun, transmitted at a frequency in the X-band, received by the 70-metre dishes of the Deep Space Network at Goldstone, Madrid and Canberra. The one-way light-travel time for a signal to Voyager 1 is now more than 22 hours. Commands sent from JPL today arrive tomorrow; confirmation of their effect arrives the day after.

Chronology

  1. 20 August 1977Voyager 2 launches
  2. 5 September 1977Voyager 1 launches
  3. Mid-1980splanetary encounter phase complete; mission shifts to heliospheric science
  4. August 2012Voyager 1 crosses the heliopause
  5. September 2013heliopause crossing confirmed in Science
  6. November 2018Voyager 2 crosses the heliopause
  7. 2023heater switched off on Voyager 2 to preserve power budget
  8. Late 2020sprojected end of science instrument operations
  9. 2030sprojected loss of transmitter capability

The transmitter on each spacecraft generates roughly 23 watts of radio-frequency power — a number that is itself fixed, because it has not changed since 1977. What has changed is the free-space path loss, which grows with the square of the distance. Keeping the link usable requires the largest ground antennas available, and on occasion multiple dishes have been arrayed together to improve the effective collecting area.

The arithmetic of remaining years

The formal expectation at JPL is that both spacecraft will retain at least one functioning science instrument, and the capability to transmit its data, into the late 2020s. The 2030s represent a harder boundary: below some threshold of electrical power, the transmitter itself cannot be kept warm enough, and the spacecraft will fall silent regardless of how carefully the remaining instruments are managed.

A gravitational-wave observatory control room at night, several people at desks, screens the only light
Night shiftA control room is where a candidate becomes a candidate. The crew’s first move is to check the other site.Photo: Control Room of Rubin Observatory (6V8A0446-CC) · Wikimedia Commons

The Voyager program is the longest continuously operating planetary and interstellar mission in history. The engineering achievement is not the 1977 launch — it is the sustained, constraint-driven management of a system built with 1970s components, on power levels that were already shrinking when the planetary encounters were complete, across distances that introduce a full day of latency into every diagnostic exchange. The physics results — plasma density, magnetic field orientation, particle fluxes at and beyond the heliopause — are inseparable from that engineering context. A measurement 160 astronomical units from the Sun is only possible because someone, four decades ago, chose plutonium-238 for a reason, and because someone now is deciding, carefully, which heater to turn off next.

Read next