Engine Hall 7, Putilov Industrial Complex. Vyborgsky District, Saint Petersburg.
The night shift began the same way every time, with a smell.
It drifted in diluted from the previous halls, coming through the corridors like a memory, and Grigory Afanasievich Semyonov knew it the way a sailor knows the wind. Twenty-six years of breathing it had turned it into a language. He read it without thinking, extracted information from it without meaning to.
Tonight the smell was right. The workshop was running.
Grigory entered Hall 7 at ten minutes past ten. The two minutes were not lateness in any real sense, they were time spent in the corridor reviewing the outgoing shift's incident log, a black-covered notebook that the foremen had maintained since 1887, and which under the new management had become an actual operational document instead of the catalog of half-truths and convenient omissions it had been for most of that time.
Entry from the afternoon shift: failure in the feed line to the third preheating furnace, corrected at 7:40 PM.
Production: one hundred and four Babbitt alloy blocks completed. Sixteen rejected for internal porosity. Responsible for the rejections: operator Konstantin Bykov, station 12.
Grigory read that last entry twice.
Sixteen rejections from the same operator at the same station meant a procedural error, not a material one. The outgoing foreman had recorded it, had attributed it to the operator by full name, and had left it documented without softening the language. Three years ago, that wouldn't have happened.
Rejections used to get dissolved into the total, buried in the overall figures, or noted as "batch variation" when the shift inspector was the kind of inspector who preferred not to create friction with senior personnel. The kind of inspector who received a small envelope on Friday before lunch.
That kind of inspector had not existed in Hall 7 for sixteen months.
Grigory closed the notebook and went in.
The hall ran 300 feet long by 62 feet wide (92 meters by 19 meters), with a riveted steel-frame ceiling at 28 feet (8.5 meters). The electric lighting was a recent installation, fitted in 1910 as part of the complex's electrification plan, and it still produced in Grigory the same dull unease it had when first switched on. Not because the light was poor, which it was not, but because electric light did not flicker. Gas lamps and oil lanterns flickered with air movement, and that movement had been, for decades, part of the information a foreman's eye processed automatically. A flame that stirred without apparent cause could indicate a draft where there should be none, which in turn could mean a door left open, a person in the wrong place, a material stored where it didn't belong. Electric light was essentially static and impassive.
You had to learn to read the workshop in a different way. Over again.
Grigory had learned. It had taken him a few months.
Along all 300 feet (92 meters) of the hall, arranged in three parallel lines, were the eighteen casting stations of the night shift. Each station consisted of a medium-sized furnace, capacity 265 pounds (120 kilograms) of metal, a cast-iron mold preparation table, and the partial extraction system installed in August under instruction from the Safety Committee, which still had no official name though everyone already called it that. The extractors were 8-inch-diameter (20-centimeter) galvanized tin tubes running up from above each furnace and converging into two main ducts in the ceiling, from which a motor-driven fan expelled the stale air outside. The system was imperfect, insufficient by the stipulated standards, and the Committee was preparing a larger installation for March.
But it was incomparably better than what had existed before, which was nothing.
Of the eighteen stations, twelve were active at this moment. The remaining six corresponded to workers who had been pulled from the night shift in October, when the first round of medical tests showed blood lead levels above the safety threshold established in the protocol. They had been reassigned to positions in the machine shop, which did not operate with molten metals. Their wages remained intact during the transition, which in September still sounded like something you had to hear twice to believe, and in January was simply standard procedure.
Grigory walked the first line of stations.
The night shift had at this moment twenty-nine workers in the hall plus four in the sorting and packaging area at the far end. Grigory knew them all by first name, most by last name, and a few by details that appeared in no official record, the kind of details a foreman accumulates over twenty-six years of watching people during eight-hour shifts under sustained pressure.
What had changed in the last eighteen months, and which Grigory was still processing, was the age distribution of those twenty-nine men. The night shift had always been veteran territory. Nights were hard and new workers rarely requested them, it was the kind of assignment that went to those with enough years in the plant to have negotiated it, or enough accumulated toughness to survive it without showing it too much. That had been true for as long as Grigory could remember.
Now half the shift was under twenty-five.
It was not an arbitrary decision. The new line operator positions, those handling the standardized mold-filling procedure introduced over the past year, required attention to rhythm and the ability to follow a repetitive sequence with consistent precision. Young workers learned quickly: within two or three weeks they moved from observers to functional operators, and within two months they were producing at full capacity. The learning curve was short because the procedure was explicit, written on a laminated sheet fixed to every station, with numbered steps and indicated times, and the training program engineers spent the first days at the new operator's side until the sequence had been absorbed.
The veterans, men like Grigory who came from the tradition of the Ural blacksmiths or the Volga boilermakers, learned that same sequence more slowly. Not because they were less capable. It was something else, something Grigory had tried to put into words on several occasions without quite finding them. A man who has worked metal for twenty years using techniques inherited from his master, who has calibrated temperatures by the color of the melt and measured pressures by the sound the piece makes as it cools, does not easily adopt a procedure that tells him exactly what to do at exactly what moment, because he trusts the observable. The written procedure does not trust the observable. It trusts the measurable.
Both worked. But not in the same way, and not for the same things.
The practical consequence was that veterans produced fewer rejections for quality errors, particularly porosity rejections, which were the most frequent and the most costly. They had the eye and the ear to identify a defective pour before the block had fully cooled, which allowed the material to be recovered while it was still viable. Young workers caught the same defect afterward, once the block was cold and visual inspection discarded it.
At the same time, the young workers produced more pieces per shift.
The balance was a negotiation that the engineers were still trying to resolve through adjustments to cycle times and station layouts, and which Grigory observed with the pragmatic skepticism of someone who knew those adjustments would never fully answer the underlying question, which was not a technical question at all, but something of another kind.
He reached station 12.
Konstantin Bykov was nineteen years old and had been at the plant for four months.
He was from Tula, son of a gunsmith who had died of typhoid fever in '99, raised by an aunt who took in mending. He had arrived through the training program with a recommendation letter from the Tula municipal technical school that described his qualifications with the sparse enthusiasm teachers use when they don't want to lie about a mediocre student but don't want to ruin his chances either. He had turned out to be a better operator than the letter suggested.
Tonight his preheating furnace was at the correct temperature. His molds were prepared in sequence. The alloy, visible through the heat-resistant window with its dark green glass that filtered the thermal radiation, had the deep orange tone that corresponded to the 896-degree Fahrenheit (480-degree Celsius) temperature called for in the standard Babbitt procedure. Everything correct.
Grigory stopped three paces back, observed for a minute without saying anything, and moved on to the next station.
At one in the morning, in the interval between the first and second work blocks, the hall entered a state of reduced activity. The furnaces stayed lit at holding temperature, the extraction fans continued running, but the operators had twenty minutes to eat the bread and soup they brought in tin lunch pails or bought at the stand installed in the outer corridor the previous year, a place that sold hot soup and tea at regulated prices, with no possibility of the vendor charging whatever he liked, because the prices were written on a painted wooden board and any discrepancy between the board and the actual charge was a reportable incident.
Grigory drank his tea standing up, near the side door that opened onto the maneuvering yard. Cold came in through the frame, which had not closed properly since a loaded Neva cargo truck had clipped it in November and bent the lower edge slightly. The repair request had been filed. Estimated completion: February.
Four workers from the shift had gathered near the holding furnaces to catch the residual heat. Two veterans and two young men, producing the kind of conversation Grigory knew well because he had taken part in various versions of it across two and a half decades. The veterans talked about the work in terms of materials and previous jobs; the young workers listened with an attention that might have been respect or simply the effort of following a technical language they did not yet fully command.
One of the young men, a boy from Novgorod who had been at the plant for two weeks and whose name Grigory had read three times without it sticking yet, asked a question that reached him with the clarity that sound carries in industrial spaces between machine cycles.
The question was about the 'ear method.'
Grigory turned slightly without moving from the door.
The veteran he had asked was Andrei Prokofievich Laskin. Fifty-three years old, twenty-eight years at the plant, trained in the workshops of the Demidov Company in Nizhny Tagil before the crisis of the nineties sent him to Saint Petersburg looking for work. He had in his hands that characteristic condition of long-serving blacksmiths, the fingers did not lie fully straight when relaxed, because the tendons had adopted over time the curve of tongs. He knew the ear method better than anyone in the hall.
He explained to the young worker what it was.
The ear method was calibration by sound. When a Babbitt block cooled in its mold, it produced during the first four minutes a sequence of crackles and contractions that varied according to the quality of the pour. A block with no internal defects produced a series of dry, regular snaps, decreasing in frequency but consistent in pitch. A block with porosities, internal voids formed by gas bubbles trapped during pouring, produced a different crack, more muffled, with a slightly asymmetric interval between sounds. It was a subtle difference, the kind that a trained human ear detected without being able to describe exactly what made it different, the kind of knowledge that did not fit on a laminated sheet with numbered steps.
The Novgorod boy listened to the explanation with a slight frown, the expression of someone trying to convert something auditory into a verbal rule and finding it cannot be done. In the end he asked: 'And if I practice, can I learn it?'
Laskin considered the question without hurry. 'In a year, probably. In two, certainly.'
The boy nodded. Then he asked something Grigory hadn't expected: 'And when you first came here, did it take you a year?'
'Not here,' said Laskin. 'I learned it before. In Nizhny Tagil, with my master. That took three years.'
'But you had to learn the whole process. I only need to learn that one part.'
Laskin looked at him for a moment. 'Maybe so,' he said, in a tone that was not concession but simply honesty, the acknowledgment that the argument was correct even if the conclusion sat slightly uneasy.
Grigory finished his tea and went for another.
At twenty minutes to three in the morning, the third preheating furnace in the second line lost pressure.
It was not a serious emergency. It was the kind of situation that in a plant running within normal parameters occurs several times per shift, and that operators handle without the foreman needing to intervene. But Grigory was passing through the second line when the pressure gauge showed the drop, and the operator at the station, a forty-one-year-old veteran named Fyodor Rzhevsky, was already checking the feed line with the pocket flashlight that foremen had carried by protocol since the previous year.
Grigory did not intervene. He stopped two paces back and watched.
Rzhevsky found the problem in ninety seconds. The regulating valve in the secondary line had a sealing ring that had worked partially loose, probably from the thermal expansion and contraction accumulated during the week of extreme cold. It was a minor maintenance defect, predictable, one that would have been caught in the scheduled weekly inspection on Friday had the cycle of overnight temperatures not accelerated it by four days. Rzhevsky corrected it with an adjustable wrench, reseated the ring, monitored the pressure gauge for three minutes, and resumed the heating cycle.
The entire process took eleven minutes. The block that had been in preparation cooled outside its cycle and had to be set aside for remelting, a loss of time and material, but a controlled one.
What Grigory observed during those eleven minutes was not only Rzhevsky's procedure, which was correct and efficient. It was what happened around it.
At the adjacent station, the Novgorod boy had stopped preparing his molds and was watching the repair with attention. He did not interrupt or ask anything.
He simply watched. Grigory recognized this because it was exactly what he had done himself at eighteen in Nizhny Tagil, when his first master, a man named Volkov who had a finger-wide burn scar on his left forearm from a forge accident in 1873, repaired something. You stayed quiet and watched.
Rzhevsky, when he finished the repair and checked the pressure one last time, turned toward the young worker. Not because it was his responsibility to explain anything, but simply because he had noticed he was being watched. He told him in four sentences which valve it was, what had failed, and why that type of failure was more common in winter. The boy nodded. Rzhevsky returned to his work.
Grigory wrote in the shift log: "02:47. Secondary reg. valve, Furnace 3, Line 2. Sealing ring displaced. Field repair. No sustained production interruption. Recommend preventive inspection of equivalent valves in Lines 1 and 3 before Friday."
Then he moved on to the next station.
By four in the morning the hall had an interior temperature of 61 degrees Fahrenheit (16 degrees Celsius), sustained by the residual heat from the active furnaces and by the resistance of the solid brick building against an exterior that the yard thermometer showed at 8 below zero Fahrenheit (minus 22 degrees Celsius). The shift workers had already passed through the hardest point, the interval between one and three when the body registers sleep deprivation most acutely, and were now in that phase of thin but sustained energy that night shift workers learned to recognize over time: not exactly alert, not exactly tired, operating on the automatism that in this trade was a skill in itself.
Grigory reviewed the night's partial production records.
The twelve active stations had completed, to that point, one hundred and twenty-seven Babbitt alloy blocks. Of these, four had been flagged for quality review through the inspection system introduced in September, a system in which each operator recorded on a tag attached to the block any irregularity observed during the process, from the color of the pour to the cooling sound when it was possible to evaluate it. The four flagged pieces were in the review area at the far end of the hall, waiting for the morning shift inspection.
In the same production period of the night shift the previous year, under the configuration of that era, the logs routinely showed between twelve and eighteen rejections per hundred pieces produced. A rejection rate that the engineers of the time considered acceptable within standard parameters for alloy casting production. Grigory had not disputed that assessment at the time because disputing engineers' parameters was not his function.
Now the night shift rejection rate was near three percent.
Grigory had no single explanation for that change. He had several, of unequal weight. The real incident logging system had reduced rejections not because it produced fewer defects, but because it identified them earlier, when the material was still recoverable for remelting rather than lost. The new standardized pouring protocol reduced process variability at the stations of less experienced operators. And the fact that no one in the supervisory chain had any financial incentive for the numbers to look better than they were, because the alternative no longer existed, meant that the numbers reflected what was actually happening.
It was a small change in appearance, almost administrative if described in abstract terms. In practice, it meant the plant knew what it produced for the first time in Grigory's experience of it.
He ran his hand along the edge of a mold preparation table. Cold steel under his fingers. No burrs. Clean, with the chalk mark from the last cycle still visible in the left corner where the operator had noted the batch number.
Somewhere in the corridor, the electric wall clock struck four with its two-tone chime.
At half past five, when the night shift was entering its final stretch and the morning shift was beginning to arrive on the first trams of the day, Grigory went to the sorting and packaging area at the far end of the hall.
There, on pine wood pallets, sat the night's completed Babbitt blocks, aligned in rows with the same orientation, each one with its production tag inserted in the slot on the upper face.
Blocks of tin, lead, antimony, and copper alloy in the proportions of the technical standard that the Tsarevich had drawn from the work of the English metallurgist Isaac Babbitt, 1839, and that the plant engineers had refined over the past year to maximize the anti-friction properties in the bearings of the new Neva engines.
Each block weighed just under 10 pounds (4.5 kilograms).
Grigory picked one up at random and held it for a moment. The weight was correct. The surface had the matte finish characteristic of iron-mold cooling, with no visible surface porosities. The temperature had reached ambient, indicating that the demold time had been correct.
There was no way to know, looking at a finished block, how many conversations had occurred during its production. How many questions the Novgorod boy had asked that night, and how many more he would ask the following week and the following month. Whether Rzhevsky would also explain the ear method to him, or whether he would learn it from Laskin on another shift. Whether next week's foreman would find the incident log as full of useful entries as Grigory had found it tonight, or as empty as it used to be two years ago.
He set the block back in its place.
From the corridor came the sound of the first morning shift workers passing through the changing rooms. Voices, the metallic clank of lockers, the noise of a tap in the washroom area. The day began in the plant regardless of what had happened the night before.
At five minutes past six, when the night shift had officially ended and the morning shift workers had taken their stations after the ten-minute handover briefing, in which Grigory had flagged the second line valve for preventive inspection and the four blocks in the review area waiting for inspection, Grigory left Hall 7 through the side door.
The maneuvering yard was still dark. In January, in Saint Petersburg, dawn was late and grudging, arriving at nearly nine in the morning with a gray light that convinced no one it was actually daytime. The temperature in the yard was the same as at the start of the shift, 8 below zero Fahrenheit (minus 22 degrees Celsius), though Grigory felt it differently now that his skin no longer had the heat of the hall pushing from the inside.
He stopped for a moment before heading to the changing rooms.
Hall 7 had all its lights on. Through the high windows, cleared of the soot that had blackened them for years, after the new management ordered the cleaning done in October, the interior was visible and active: furnaces running, cranes moving, operators at their stations. The morning shift starting exactly where the night shift had left off.
Grigory had looked at that same interior from that same yard across different decades and under different managements. The building was the same. The machines were new. The procedures were different. The incident log had real entries. Rejections were attributed by full name. Wages arrived at the end of the month without anyone having held back a portion for reasons that were never stated explicitly but that everyone understood. The plant physician, who for twenty-five years had been a decorative figure whose main function was to sign premature disability papers for workers that foremen wanted off the roster, now had a real office and conducted real examinations.
Not everything was different. Temperatures were still the same in winter. Metal still behaved like metal. Men still made the mistakes men made when they worked eight hours under sustained difficult conditions, predictable, manageable mistakes, different from the mistakes they made when they had incentives to make them and no one documented them.
Grigory was not a man given to broad assessments. He was fifty years old and had been at the same plant for twenty-six of them, and had survived enough to know that broad assessments were the luxury of those who did not have to be on the night shift next Monday. What he could say, because it was observable and measurable in the same way that metal was observable and measurable, was that the Novgorod boy was going to learn the ear method faster than any performance expectation chart would have predicted. That Rzhevsky would keep repairing secondary valves in eleven minutes without needing to be asked. That the one hundred and twenty-seven Babbitt blocks the night shift had produced were going to function correctly in the engines they were destined for, because the numbers in the incident log said so, and the numbers in the incident log were now real.
That was enough for tonight.
He went to the changing rooms.
In the morning shift's tool inventory log, under the heading of historical reference material, someone had added a new entry that week: "Collection of artisan forge calibrators, Ural origin, period 1880–1905. Conservation and cataloging in progress. Technical and heritage reference material." The entry bore the stamp of the complex's new technical documentation department, established in November under management instruction, its stated function being to record existing artisan procedures before the masters who carried them were no longer available to pass them on.
Grigory had not requested that. He didn't know who had.
He didn't ask.
. . . . .
Nemryz: If you've enjoyed this story and want to read ahead, I have more chapters available on my patreon.com/Nemryz. Your support helps me continue writing this novel and AU. Thank you for reading!