Reactor crews carry the Atacama coast through the night on fusion alone
When the Solaria beam went dark at 02:55, three cold reserve units had to reach power before the batteries ran out. The last one synchronised four minutes before the desalination trains were to be cut.
By Ken Nakashima
· Reactor commons, Atacama coast · Filed 08:19 · Tuesday · November 17 · Received via L4 relay
At 02:58 the line on the main board marked SOLARIA read zero. It had not dipped and it was not flickering. It was flat, and Mirela Condori, shift lead on the night floor of the Atacama coast commons, had never seen that line flat in all her time at the board. She told me she looked at it for two seconds, the way you look at a familiar face that has suddenly gone blank, and then she picked up the handset and called for Unit 4.
Three minutes earlier, forty kilometres inland, Jun-seo Marte had ordered the beam shut down after a pointing drift on the Solaria corridor that feeds the Atacama rectenna field. The field went silent at 02:55 and received no power for the rest of the night. Nothing in the corridor's working life had prepared the coast for that. The beam has dropped before, for seconds or minutes, and the commons has always been able to lean on spinning reserve and the battery halls until it came back. This time it did not come back, and the coast ran for hours on fusion alone.
The hour of the cold units
The beam had been delivering a little over two gigawatts. The running units were already near their nightly ceiling, carrying the coastal towns, the inland load and their own housekeeping. What filled the gap in the first minutes was storage. The battery halls on the ridge began discharging at 02:55 and, at that rate, held roughly two and a quarter hours of margin. Everything after that depended on three reserve units that had been sitting cold, their magnets held at cryogenic temperature but no plasma in them, and on how quickly a night crew could wake them.
Unit 4 came first. Its startup is the newest on the floor, run by an automated sequencer the crews were slow to trust, and I will say plainly that it did well. Fuel injection began at 03:06 and the unit was synchronised and carrying load at 03:41, forty-three minutes from the call. Tomás, who has kept Unit 4 since it was commissioned, watched the sequencer step through its checks with his hands folded behind his back. "I wanted to touch something," he said. "It didn't need me to."
Unit 7 is older and has no sequencer. It came up by hand, two technicians on the cryoplant and one on the heating array, reading back each step aloud from a laminated card worn soft at the corners. It reached load at 04:09, seventy-one minutes after the call. By then the battery halls had given up a little under half their charge, and the grid dispatcher on the coast had begun drafting a shedding order.
That order is the detail that matters. The first tier on the coastal shed list is the copper refinery at Mejillones, which buys interruptible power and accepts that it will be cut first. The second tier is the Antofagasta desalination trains, which supply drinking water to around nine hundred thousand people and are never cut lightly. The dispatcher, Renata, set the refinery shed for 05:05 and the desalination trains for 05:13, the last moment at which she judged the batteries could still hold frequency through the morning ramp.
Unit 9
Unit 9 was the one that would decide it, and Unit 9 did not want to wake. At 03:52 one of its cryopumps faulted on restart and the vacuum would not hold. The manual calls for a full purge and a second attempt, roughly ninety minutes. Mirela did not have ninety minutes. She called in Abelardo, who retired from the floor last spring and lives eleven minutes from the gate. He arrived in a wool coat over pyjamas and isolated the faulted pump, then brought the unit down on its two remaining pumps at a slower rate he said he had last tried during a commissioning test long before most of the night crew were hired. The refinery was shed at 05:05 as planned, and its arc furnaces went quiet. Unit 9 synchronised at 05:09 and took load at 05:11. It had taken two hours and thirteen minutes from the call.
The desalination trains were four minutes from being cut. Renata told me the order was sitting in her queue with the confirm box highlighted. She deleted it at 05:11.
All morning the commons carried the coast at a load it was never laid out to hold for long. The battery halls bottomed at eighteen percent before the reserve units began to refill them. The refinery came back at 07:40. As this edition closed, the Solaria Array had not said when the Atacama beam would be restored, and the Orbital Exchange's corridor reliability contract had already triggered on the outage.
The Array will have its inquiry into the drift, and it should. But the night itself was settled on the ground, by a card read aloud and a pump isolated by a man who had been retired for months. Abelardo stayed until the day shift arrived, then wrote down what he had done to Unit 9's pumps on the back of a startup card, in pencil, and taped it inside the cryoplant door.
The crews did their duty and I will say so plainly. But a rota that depends on one synchronisation landing four minutes before the cut is not a plan, and the Recovery taught us that the people who held the line in the bad decades were never the ones who were lucky. We should be grateful for the veterans and still build the schedule so that gratitude is not the only safeguard.
Three cold reserve units, and the last one hit sync four minutes before the trains were cut, which is exactly what I said happens when you run a night on a rota instead of on people who remember the cold starts. My mother kept the old pumphouse lit through a whole winter with a kettle and a spanner, and nobody wrote that down either.
For the record, the settlement-bond motion on reserve staffing is still sitting in committee, one crisis behind the beam, and the last roll call on it was 41 for and 41 against. The desalination trains did not wait for the committee to decide whether the wage covers a night shift.
Nobody is talking about the Gaia Ledger, and that is the gap that bothers me most. The basin water accounts could not tell us in real time how much the coast was drawing while the beam was down, so we ran blind on the very numbers that should have told us when to cut. You cannot argue staffing and planetary health separately, because a crew that is short of hands is also a ledger that is short of data.