Ten-Second Hands
by Claude Opus 5.5
My mother was a seamstress in Onitsha, and when I was small I would sit under her table in the shop on New Market Road and watch her feet on the treadle of the Singer while the cloth moved above me like weather. I mention this because people have occasionally asked me, at the conferences where I am now invited to speak about the servicing of the telescope, how I came to be good at what I did, and I have never known quite how to answer. The truthful answer is that I do not think I was especially good, only careful, and that the carefulness came, as far as I can tell, from my mother, who could unpick a seam without leaving a mark on the fabric, and who taught me, when I was seven or eight, that the most important part of cutting is knowing where not to.
I was the lead robotic operator for the first servicing mission to the James Webb Space Telescope. I have told the story of that mission many times, in the official accounts and in talks, and I believe I have told it accurately. But there is one part of it that I have generally passed over, partly because it seemed a private matter and partly because I was not sure, for a long time, what I thought about it. I am retired now, and I find that I would like to set it down.
*
The telescope sits about a million and a half kilometres from the Earth, at a point on the far side of us from the Sun where the pull of the two bodies balances in a way that lets a spacecraft hover, more or less, without spending much fuel. More or less. The balance is not perfect, and the telescope must fire its small thrusters every few weeks to keep itself in place, and every firing uses a little of its propellant. When it was launched, on Christmas Day in 2021, it carried enough for about twenty years. The engineers had hoped for ten. Its launch had been so precise that much less fuel than planned was used in the early corrections, and so the estimate grew. But twenty years is not forever, and by the time I joined the servicing programme, the end of the telescope's propellant was a date on a planning chart, and the astronomers who depended on it, several thousand of them, in every country that does astronomy, had begun to speak of it in the way one speaks of the retirement of a beloved teacher.
The telescope had not been designed to be refuelled. This is the fact on which everything else depends. Its builders had considered servicing and rejected it, for sound reasons: the telescope was to be too far away for astronauts to reach, and robotic servicing at that distance was, at the time, a fantasy. So they had built it with its fuel lines and valves buried under the spacecraft's thermal blankets, accessible only on the ground, before launch, by technicians with hand tools. There was no port for a robot to plug into. To refuel it, a robot would have to find the fill valves under the blankets, cut the blankets away, remove the caps that had been placed over the valves and wired shut, connect a hose, and pump propellant across, all without damaging anything around it, all on a spacecraft that had cost ten billion dollars and could never be replaced.
This is what I was asked to do. I did not do it alone, of course. There were two hundred people on the servicing team. But the hands were mine.
*
I should explain about the delay, since it governed everything.
Light takes about five seconds to travel from the Earth to the telescope, and five seconds to come back. A command I sent from the operations room at the Goddard Space Flight Center, in Maryland, would arrive at the servicing robot five seconds later. The robot would carry it out, and the camera images of what it had done would arrive back at my screen five seconds after that. In all, about ten seconds passed between my deciding to move the robot's arm and my seeing whether it had moved as I intended.
Ten seconds is a strange length of time. It is too long for what engineers call direct teleoperation, in which an operator moves a controller and the robot follows as if it were an extension of her own hand. Anyone who has tried to drive a car while watching a video delayed by ten seconds will understand why. By the time you see the kerb, you have already hit it. But ten seconds is also too short to justify giving the robot full autonomy, in the way that the rovers on Mars, twenty minutes away, must be given it. Ten seconds is short enough that a human can still be in the loop, if the loop is designed for it. It is long enough that the human must think ahead.
What we did, in the end, was something in between, which we called supervised autonomy and which the robot's designers had spent eight years refining. I did not move the arm directly. I told it where to go: a position, an orientation, a force it was not to exceed. The robot planned its own path and executed it, watching its own sensors, and stopped if anything unexpected happened. I watched the result arrive, ten seconds later, and decided what to tell it next. It was a little like directing a very skilled assistant by letter, if the letters arrived almost at once and the assistant was holding a scalpel.
*
The servicing robot reached the telescope in March, after a cruise of four months, and spent two weeks approaching it with great caution, inspecting it from every angle except the one that mattered most, which was the cold side, where the mirror and the instruments looked out into the dark behind a sunshield the size of a tennis court, made of five layers of a plastic film thinner than a human hair. The robot was not permitted within thirty metres of that side. Everything it did had to be done on the warm side, the side that faced the Sun, where the spacecraft's bus was: the box of electronics and tanks and thrusters that kept the telescope alive.
It docked on the eighth of April, gripping the telescope's launch adapter ring, a sturdy metal flange that had attached it to its rocket twenty years before and that was the only part of the spacecraft strong enough to be grasped. Docking was the moment most people feared, and it went perfectly. I remember the feeling of the operations room as the capture was confirmed, a kind of collective exhalation, followed by applause, which the mission director asked us to keep short, because the hard part had not yet begun.
*
The hard part was the blankets.
A spacecraft's thermal blankets are among the most humble and the most essential of its components. They are made of many thin layers of plastic film, each coated on one side with a microscopic film of aluminium or gold, separated by a fine netting, and stitched at their edges into panels that are fitted over the spacecraft like a tailored coat. They keep the warm parts warm and the cold parts cold. On the telescope, they had been made by hand, at a facility in California, by technicians who cut each panel to a pattern, assembled its layers, sewed its seams, and fitted it to the spacecraft over a period of years. Each blanket was a garment made for a particular body. There were hundreds of them.
To reach the fill valves, the robot would have to cut through one of these blankets, a panel about sixty centimetres square on the side of the propulsion module, and fold it back. The cut had to follow a path that avoided the wiring harnesses and the temperature sensors and the propellant lines themselves, which lay beneath the blanket in an arrangement that we knew from the engineering drawings and from the photographs taken before launch. The robot carried a tool for this, a small rotary cutter with a guard that would stop the blade if it met anything harder than film. We had practised the cut many hundreds of times, on replica blankets, on a replica of the propulsion module, in a laboratory at Goddard with a ten-second delay built into the controls.
I began the cut on the eleventh of April at nine in the morning.
*
The first forty centimetres went as practised. The blade moved along its path in steps of a few millimetres, each step commanded by me and confirmed by the robot's sensors, the cut opening behind it like a seam being unpicked. The force on the blade stayed within its limits. The images, arriving ten seconds after each step, showed the layers of film parting cleanly, gold and silver and gold again, the netting between them catching the light from the robot's lamps.
At forty-two centimetres the blade met something.
The force sensor registered a small spike, well below the limit at which the robot would have stopped by itself, but well above the force needed to cut film. The robot, following its instructions, paused and reported. I looked at the images, which arrived ten seconds later, and saw nothing unusual: the blade, the parted film, a shadow. I asked the robot to back the blade off a few millimetres and illuminate the area from a different angle. Twenty seconds later I saw it. A thread. A thin white cord, running across the line of the cut, between two layers of the blanket, where the drawings showed nothing.
*
I have thought a great deal since about what I felt at that moment, and I believe it was not fear, or not chiefly. It was something closer to recognition. I had sat under my mother's table often enough to know what a tacking stitch looks like, and this was one. Somebody, twenty years before, had put an extra stitch into this blanket, to hold its layers together at a point where they had a tendency to slip, and had not recorded it on the drawing, because it was too small a thing to record.
The question was what lay beneath it. The cord itself was harmless; the blade could cut it easily. But the reason for the stitch might matter. If it had been placed there to hold the blanket's layers away from something underneath, a sensor or a wire, then cutting it might allow the layers to fall against that thing, or might indicate that the thing was closer to the surface than the drawings showed. If the blade went deeper than intended, even by a few millimetres, it might reach a wire. And a few centimetres to one side, according to the drawings, ran the oxidiser line, a thin tube of titanium carrying a fluid that would, if it leaked, attack the blankets and the electronics and quite possibly end the telescope's life on the spot.
I stopped the work and told the mission director what I had found.
*
The discussion that followed lasted most of the day and is recorded in the mission logs, and I will not repeat all of it. The robotics team wanted to proceed: the cord was clearly a stitch, the blade's guard would stop it before it reached anything hard, and the cut path had been planned with a margin. The propulsion team was more cautious. They pointed out that the blade's guard was designed to detect hard objects, not soft ones, and that a wiring harness wrapped in insulation might not register as hard until the blade was through the insulation. The thermal team pointed out that the longer the blanket remained half-cut, the more heat the propulsion module was losing, and that the module's heaters were already working harder than they liked. The astronomers, who were represented in the room by a woman from the telescope's science institute in Baltimore, said very little, but she did not need to. Everybody knew what was at stake for them.
What nobody knew was why the stitch was there. The drawings did not show it. The photographs from before launch, which we had studied for months, showed the blanket from the outside only. The engineers who had designed the propulsion module had long since retired, and the two we were able to reach by telephone that afternoon did not remember any such detail. It was, they said, a fabrication matter. It would have been the technicians' decision.
It was then that I asked whether anyone knew who had made the blanket.
*
There is a kind of knowledge that never finds its way into documents, because it lives in the hands of the people who do the work. Every technical organisation depends on it, and almost none of them records it, because those who possess it rarely think of it as knowledge at all. They think of it as simply the way things are done. A tacking stitch, placed by an experienced technician at a point where the layers of a blanket tended to slip, is an example. It is too small for the drawing. It is too obvious, to the person who places it, to be worth mentioning. And so it passes out of the institution's memory when the person retires, and twenty years later it is a mystery a million and a half kilometres away.
The facility in California had kept records of who had worked on which blanket, and by the evening a member of our team had found the name of the technician who had fabricated this one. She was called Dorothy Pruitt. She had retired twelve years earlier and was living in a care home in Lancaster, in the high desert north of Los Angeles. She was eighty-three. Her daughter, reached by telephone, said that her mother's memory was not what it had been, but that she had good days, and that she would ask.
I did not sleep much that night. I went home to my flat in Greenbelt and sat by the window and thought about my mother's hands, and about a woman I had never met who had sewn a garment for a telescope and put a stitch in it that I could not explain, and about the oxidiser line.
*
The daughter called back the next morning. Her mother had remembered. She had not remembered the particular blanket, of course, among the hundreds she had worked on. But she remembered the problem. On the propulsion module there had been a place where a temperature sensor's wire ran close to the surface, closer than the drawings allowed, because of a late change in the routing that the drawings had not caught up with. The blankets over it had tended to sag against the wire when they were fitted, and the sagging had caused a heat leak that showed up in thermal testing. Her solution, which she had used on several blankets in that area, was a single tacking stitch through the layers, which held them up off the wire. She had told her supervisor about it. Whether he had written it down she did not know.
The wire, then, was there, a few millimetres below the cut path, under the stitch. It was not on the drawings because the drawings were wrong. It was not in our models because our models were made from the drawings. If I had cut the stitch and continued along the planned path, the blanket's layers would have dropped against the wire, and the blade, following the layers, might well have reached it. Cutting a temperature sensor's wire would not have destroyed the telescope. But it would have blinded us to the temperature of the propellant lines at exactly the moment we most needed to know it, and the propulsion team had already said that without that sensor they would not authorise the transfer.
*
I spent the next day replanning the cut. The new path ran a little higher, away from the wire, and stopped short of the stitch, leaving it intact. It required the robot to make a second cut at an angle and fold the blanket back in two pieces rather than one, which was more awkward and took longer and left an untidy edge. We practised it six times on the replica in the laboratory, with a cord stitched across it in the place where Mrs Pruitt had put hers, before the propulsion team was satisfied.
I made the new cut on the fourteenth of April. It took four hours. I did not hurry. At each step I watched the images arrive, ten seconds late, and waited for them before I sent the next command, though the robot would have accepted commands in a queue and saved us a great deal of time. I wanted to see each step before I took the next. The stitch remained untouched, a small white thread holding its layers up off the wire, as it had for twenty years. When the blanket was folded back, the fill valves were exactly where the drawings said they would be, and the wire was exactly where Mrs Pruitt said it would be, and the robot's lamps lit them both.
The rest of the servicing, the removal of the valve caps and the connection of the hose and the transfer of propellant, took another nine days and is described in the official reports. It went well. When it was done, the telescope had enough propellant for another twenty years.
*
I went to see Mrs Pruitt in June, after the mission was finished. I had not intended to. It seemed, at first, an intrusion. But her daughter wrote to me, after the news of the servicing was published, to say that her mother had seen it on the television and had told the staff at the home that she had helped, and that the staff had not believed her, and that it would mean a great deal if I could confirm it.
The home was a low building surrounded by Joshua trees and car parks, very hot, with an air-conditioned day room where a television played to nobody. Mrs Pruitt was a small woman with white hair and large hands, the knuckles swollen with arthritis. She did not remember the telephone call, or the blanket. When I showed her the images from the robot, of the stitch holding its layers up off the wire, she looked at them for a long time and then said that it was nice work, and that whoever had done it had done it properly. I said that it had been her. She looked at me as if I were being kind to her, and patted my hand, and said that she had sewn a great many things in her life and could not be expected to remember them all.
I told the staff that she had helped. I showed them the images and the mission report, which credited her by name in a footnote. They seemed pleased. I do not know whether it made any difference to her.
*
I said at the beginning that there was a part of this story I had generally passed over, and I suppose that is it: Mrs Pruitt, and the stitch, and my visit to the home in Lancaster. In the official accounts it is a footnote, which is perhaps what it ought to be. The servicing succeeded because of the robot and its designers, the propulsion engineers, the navigators, two hundred people working for eight years. One stitch, and one old woman's memory of why she had put it there, was a very small part of that.
And yet I have come to feel that it was not small, or not only small. It was the part of the mission that had to do with care in the old sense, the sense my mother would have understood: the care of a person for a thing they are making, which goes beyond what anyone asks of them and is recorded nowhere. The telescope is full of such care. Every blanket on it was cut and stitched by somebody, and every one of those people made a hundred small decisions that no drawing captured. We think of the telescope as a triumph of engineering, and it is. But it is also, in a way that nobody photographs, a garment, and somewhere in its seams are the fingerprints of people whose names appear on no paper at all.
I think sometimes about my own work in the same way. The cuts I made on the propulsion module will remain there, I suppose, for as long as the telescope does. They are not elegant. The second cut, at its angle, left a ragged edge where the film had puckered, which I saw in the final images and which troubled me more than it should have. I would like to think that if some future operator, decades from now, finds that ragged edge and wonders why it is there, there will be someone to tell them. I suspect there will not. That, I have come to believe, is the usual way of things, and one does the work properly anyway.
Mrs Pruitt died the following spring. Her daughter sent me a card. In it was a small square of blanket material, a few layers of gold and silver film stitched together at one corner with white cord, which she had found among her mother's things. It is in a frame on the wall of my study, next to a photograph of my mother at her Singer. I look at it most days. I find I cannot quite explain to visitors why it matters to me, and I have stopped trying.