Andy Garcia · freelance dev
FRENES
Back to the blog
Space

BepiColombo Is Scheduled to Orbit Mercury on November 21. About 10% Less Thrust Cost It 11 Months.

October 20, 2026 8 min read

On November 21, 2026, BepiColombo is scheduled to enter orbit around Mercury. The mission, led by the European Space Agency with the Japan Aerospace Exploration Agency (JAXA), launched on an Ariane 5 on October 20, 2018. By September 3, 2026, it had travelled 9.9 billion kilometers and completed nine planetary flybys. ESA says it will be the first mission to orbit Mercury with two spacecraft, and only the second to orbit the planet at all.

The dates depend on which ESA page you read. ESA's live updates page gives November 21 for orbit entry of the two orbiters, still stacked, and December 9 to 10 for their separation. The mission's main page lists "December 2026" as the arrival, which matches ESA's other phrasing: two spacecraft in orbit around Mercury in December 2026. Science is planned to start in April 2027. All of these are plans. When I read the live page on October 20, its latest entry was dated October 6.

From launch until 2024, the plan was orbit insertion in December 2025. A fault in April 2024 moved it, and how ESA handled that fault is what I wanted to read closely. I do not build spacecraft. I build software, and the questions are the same ones I ask after any incident: what broke, what did it cost, and what did the team test before the next risky step?

Why getting to Mercury is hard

ESA gives two reasons. As the spacecraft gets closer to the Sun, the Sun's gravity accelerates it. And it left Earth with a lot of energy, far too fast to be captured by a planet as small as Mercury. So it has to brake against the Sun's pull for years. ESA's own comparison: placing a spacecraft in a stable orbit around Mercury takes even more energy than sending a mission to Pluto.

Its Mercury Transfer Module (MTM) carried two 15-meter solar wings and four ion thrusters. They turn xenon into plasma and throw the ions out at 50,000 meters per second, 15 times the exhaust speed of a conventional chemical rocket, per ESA. The price is force. When the first thruster reached full throttle in 2018, ESA compared its 125 millinewtons to the weight of a AAA battery held at sea level. ESA's 2020 fact sheet gives up to 290 millinewtons for the module.

The arithmetic shows why electric propulsion is measured in years. 290 millinewtons pushing the 4,100 kg launch stack is an acceleration of about 0.00007 m/s². Over a full day of thrusting, that changes the velocity by about 6 meters per second. So BepiColombo combined years of thrust arcs with nine gravity assists, one at Earth, two at Venus and six at Mercury, to lose energy and slow down.

The fault, and the redesign

On April 26, 2024, as a thrust manoeuvre was about to begin, the transfer module failed to deliver enough electrical power to its thrusters. By May 7, a combined team from ESA and industry had restored thrust to approximately 90% of its previous level. ESA's update of September 2, 2024 named what engineers had identified: unexpected electric currents between the solar array and the unit that extracts and distributes power. Mission manager Santa Martinez gave the conclusion: the thrusters "will remain operating below the minimum thrust required for an insertion into orbit around Mercury in December 2025."

Read that literally. About 10% less thrust was enough to break the date.

It did not break the mission. ESA's flight dynamics team designed a new trajectory that works with lower thrust and keeps the baseline science. The three remaining Mercury flybys went ahead as planned, ESA said. The fourth, on September 4, 2024, was rerouted about 35 km closer than originally planned, to reduce the propulsion needed to reach the fifth; ESA reported afterwards that the spacecraft came down to around 165 km above the surface. The sixth, in January 2025, was to branch onto the new route.

The bill was paid in time. Orbit insertion moved from December 2025 to November 21, 2026. The final science orbit for ESA's Mercury Planetary Orbiter (MPO) moved from March 2026 on the 2020 fact sheet to March 2027. Routine science moved from spring 2026 to April 2027. I found no deeper ESA account of what caused those currents.

A checkout in stages

The ion engines were switched off for the last time on June 15, 2026, at 15:24 CEST, after commands sent well in advance. From then until separation, the stack followed what ESA calls a ballistic, free-falling trajectory.

On September 3, the transfer module separated. ESA gave the go around noon CEST and separation was planned for 14:00. A first hint came from a preliminary Doppler signal, a change in the radio link's frequency caused by the change in velocity: 14:20 CEST in ESA's story, 14:41 in its live log. The acquisition of signal that confirmed it came nearly two hours after separation: 15:49 CEST in the story, 15:52 in the log. At 200 million kilometers, light takes about 11 minutes each way, so the distance explains a small part of the wait. The rest was planned. After separation, ESA writes, the stack entered safe mode, readjusted its attitude and reconfigured itself before reporting to Earth. Days earlier, ESA had published 15:53 CEST as the earliest expected acquisition of signal. Waiting for telemetry until then was the plan, not an incident.

On September 15, every instrument on the MPO was switched on together for the first time since separation. ESA reports that first analyses show them in good health. The BELA laser altimeter fired more than 700,000 pulses into open space over 21.5 hours, and the tests show it in excellent health since its last full test on Earth, in 2016. The MERTIS instrument, whose planet-facing port had been blocked by the transfer module until separation, returned about 1 gigabit of data in a 22-hour test, and a first analysis shows little to no loss in its spectrometer's sensitivity since its last low-temperature ground calibration, in 2013. The SIMBIO-SYS camera took its first image in space.

On September 24, the MPO fired its axial thrusters for the first time in this configuration: 8 minutes and 42 seconds, a push of 11.4 m/s, with mission control watching power, temperature, pressure, acceleration and Doppler data. That burn averaged about 0.022 m/s², a few hundred times the ion engines' full-thrust acceleration on the launch stack (the masses differ, so read that as an order of magnitude). That is the trade: chemical engines for the moments that need force now, ion engines for the years that need efficiency. ESA says more small correction slots are planned before the November insertion.

What is scheduled, and what is not known

ESA's June 2026 page calls November 21 "the critical Mercury orbit insertion manoeuvre." The MPO's chemical engines will then lower the stack toward the elliptical polar orbit planned for JAXA's Mio, 590 by 11,640 km. Mio is scheduled to separate on December 9 to 10. The MPO is then to descend to its own orbit, planned at 480 by 1,500 km, by March 2027. Space.com, covering the September 3 separation, reported that the November 21 arrival begins a sequence of 16 manoeuvres, and quoted operations manager Elsa Montagnon: "In our case, it's going to be six months of continuous activities."

What the sources do not show: the insertion burn itself. ESA's operations page, last updated in 2021 for the December 2025 plan, described a "weak stability boundary" capture needing only a small manoeuvre. I found no 2026 ESA description of the burn's size or duration on the new trajectory, so I am not assuming the old numbers carry over. The instrument results are first analyses. And a scheduled date is not an outcome.

What I take from it

I orchestrate much of my work with Claude Code, and long agent runs have the same shape on a far smaller scale: a configuration change, a stretch with no feedback, then a step that is costly to undo. What I would copy, as analogies only:

  1. Check your margin against the date, not only the goal. On about 90% thrust, a new route still keeps the baseline science, according to ESA. It could not keep December 2025. A plan with slack only on the outcome will spend the calendar first.
  2. When a component degrades, move work to what is still free. The team could not get full thrust back. It asked more of gravity and accepted time as the cost. In software: reroute around the degraded service and let latency absorb the loss.
  3. Test small and real before the big step. A burn of under nine minutes before the critical one. Every instrument switched on, and two of them compared with measurements taken on the ground years before. After a migration, that is one real write and a comparison with a baseline you saved beforehand.
  4. Write down when silence is expected. A published earliest time for telemetry turns nearly two hours without it into a plan rather than an alarm. A safe state by default after a risky step means the system reports before it continues.

On November 21, the plan meets Mercury. Whatever happens, the instructive part is already on ESA's pages: a team that lost about 10% of its thrust, redesigned the route around it, then tested each piece in turn before the big one.

Sources

A project like this one?

I design and deploy products like this. Let's talk.

Let's talk