The Tinkery

Chapter 9

The discovery happened on the third morning of the trimming process, while Silas was working on the generator’s output conductor with a flat file.

The trimming had been proceeding methodically. They had developed a sequence over the first two days: Silas would measure the current cross-section of a conductor at multiple points along its length, calculate the minimum cross-section required to carry the actual operating current with appropriate temperature rise plus a safety buffer, and determine how much material could be removed. June would then work the file while Silas measured, checking dimensions at regular intervals until the conductor reached its new specification. They marked each finished conductor with a strip of red tape and recorded the before and after weight in a column in the notebook.

The weight recovery was real but slower than either of them wanted. The bus bars had given back the most as the heavy rectangular copper stock, trimmed to a narrower profile on three faces, had shed nearly four pounds across the full distribution system. The motor feed cables had given back less, their conductor cross-sections already closer to the minimum workable gauge. The generator output conductor was the last major item on the list, and it was a substantial piece equaling the heaviest single conductor in the system, carrying the full generator output before the voltage regulator divided and distributed it.

Silas had been working the file across the conductor’s flat face for perhaps ten minutes when the color changed.

He stopped. He lowered the file and looked at the surface he had been cutting. The fresh copper he had been exposing with each stroke was the expected color. It was the bright salmon-pink of clean copper, catching the light from the windows. But beneath it, in the material the file had just reached, the color was different. Not pink but silver-white, the color of aluminum.

He set the file down on the workbench and picked up the conductor, turning it in the light from the south windows. The cross-section he had been working was now clearly visible: a copper outer layer, perhaps two millimeters thick on each face, surrounding a core of a different metal entirely.

He put the conductor down and stood looking at it for a moment.

Copper-clad aluminum. The conductor was, as it turned out, not solid copper. It had never apparently been solid copper. The aluminum core was why it had been oversized because aluminum’s electrical conductivity was lower than copper’s, requiring a larger cross-section to carry the same current, and the copper cladding had been added to improve the surface conductivity at the connection points. He had known this technology existed. He had used it, it seemed, and had then forgotten that he had used it, and had spent the intervening years remembering a solid copper conductor that had never existed.

The more immediate problem was the melting point. Copper melted at just over 1,980 degrees Fahrenheit. Aluminum melted at 1,220 degrees. A copper-clad aluminum conductor in a high-current application near a flash steam generator n less, running at sustained output, in an enclosed chassis was not a conductor he was willing to put on the Skiffer.

He picked up the file and set it on the other side of the workbench, the side where finished tools went, and looked at the conductor for another moment.

June was at the secondary bench, working on the trim documentation. He heard her pencil stop.

“What is it?” she said, without looking up, reading something in the quality of the silence.

“Come and look at this.”

She came and looked. He showed her the cross-section, the copper surface and the silver core beneath it, and explained what he was looking at and what it meant. He kept his explanation to the facts: the material, the melting point differential, the reason it had been oversized, the reason it could not go on the Skiffer.

June looked at the conductor for a long moment as she took in this latest setback. He watched her take a breath in through her nose, hold it, and let it out slowly before she said anything.

“What do we do about it?” she said. Her voice was level.

Silas looked around the Tinkery’s main floor , at the shelving rows receding into the building’s interior, at the catwalks on the second level, at the accumulated inventory of sixty years of building things from copper wire and copper strip and copper tube and copper sheet.

“We cast a new one,” he said.

“From what?”

He picked up his walking stick and gestured at the building around them. “There are easily ten thousand pounds of copper in this building. The trolley cable on the east wall alone is two hundred pounds of it. The door handles on the exhibition cases are brass, which is copper and zinc - we can separate those. The winding wire from the old motors I haven’t stripped yet.” He looked at her. “We’re not short of raw material. We’re short of time, so let’s not waste it standing here.”

~*~

Another day passed while Silas made arrangements.

The workbench had been cleared and reset. On it sat a two-piece open-ended iron box, its halves separated and lying flat, each piece approximately fourteen inches long and eight inches wide, with alignment pins on the parting faces and vent holes drilled through the walls at regular intervals.

“This is the flask,” Silas said, picking up the bottom half. “The top half is the cope. The bottom is the drag.” He set it back down and picked up an object from the bench beside it - a wooden form in the shape of the letter H, cut from a piece of close-grained mahogany. The H-shape was the cross-section of the bus bar they needed: two rectangular flanges connected by a central web, the geometry that distributed current across the full width of the connection while minimizing material and therefore weight. The wooden pattern was smooth, its edges slightly tapered so it could be drawn from the sand without disturbing the cavity walls. “This is the pattern. It took me most of yesterday evening to get the taper right.”

June picked up the pattern and turned it over, running her thumb along the taper on one of the flange edges. The angle was slight but consistent around the perimeter. “And the sand holds the shape of this after you pull the wood out?”

“Green-sand,” Silas said. He patted the dark, damp mound in the bucket beside the bench. “Sand mixed with clay and enough water to give it cohesion. The clay is what holds the grain together when you compact it around the pattern - the water activates the clay’s binding properties.” He picked up a handful and squeezed it in his fist, then opened his hand. The sand held the impression of his fingers cleanly, the surface slightly shiny where the clay had been compressed. “Too dry and it crumbles when you draw the pattern. Too wet - ” He set the handful back. He had noticed, ramming the first batch of sand into the drag the previous evening, that his hands had felt less certain than they once had at this task - the rhythmic, graduated pressure of the ramming rod requiring a consistency of force that his arms had taken longer to find than he remembered. He had told himself it was sufficient. He was no longer certain it had been. “Too wet causes its own problems. We’ll get to that.”

He reached for the ramming rod - a turned hardwood handle fitted to a flat iron foot - and began packing the green-sand into the drag around the pattern, working from the edges inward, the rod moving in short, overlapping strokes. “You want the sand dense enough that it won’t shift when the copper hits it. Not so dense that you’ve closed off the porosity the steam needs to escape through.” He picked up a fine wire and began pushing it through the compacted sand at intervals - thin vent holes, spaced two inches apart across the full face of the mold. “The steam has to go somewhere when the hot metal hits the moisture in the sand. These give it a path.”

~*~

He found the furnace under three layers of canvas tarpaulin in the northwest corner of the main floor, behind the display of early electrical equipment and beside a stack of iron pipe that had been destined for a pressurized water experiment that had never moved past the planning stage. The tarpaulins came off in order, releasing the particular smell of stored canvas and old iron into the air of the Tinkery. Underneath them sat a drum of heavy iron construction, approximately three feet in diameter and four feet tall, banded with iron hoops at six-inch intervals and sitting on three cast iron feet that had pressed permanent impressions into the floorboards over the years of its storage. The lid was firebrick set in an iron frame, fitted with two ports through which a pair of carbon-graphite electrodes descended into the furnace interior - each electrode the diameter of a man’s forearm, black and dense, the graphite crystalline structure faintly visible in the afternoon light.

Silas put his hand on the iron flank of the furnace.

“Héroult-type arc furnace,” he said. “I built it in 1875 for the Atlantic telegraph project. We needed to smelt copper alloys to specific compositions for the cable conductors, and a coal forge introduces carbon contamination into the melt through the combustion gases.” He cranked the inspection gear on the side of the furnace, his knuckles stiff on the iron handle, working the mechanism that tilted the lid for interior access. The gear was stiff with disuse but moved. “In a coal forge, the copper absorbs oxygen from the fire during melting. Oxygen in solution makes the finished metal brittle, with reduced conductivity. Here, the heat source is the electric arc between the carbon rods and the metal charge. No combustion gases. No oxygen introduction from the heat source.”

June was looking at the two carbon electrodes with an expression that was not quite apprehension but was adjacent to it. “And it won’t explode?”

“The arc itself is contained inside the furnace body.” Silas straightened from the inspection gear and picked up the bundle of old trolley cable sections he had cut and coiled that morning - heavy stranded copper, the outer insulation stripped back to bare metal, the bundles tied with wire to keep them compact. He packed them into the clay-graphite crucible inside the furnace, working methodically, filling the available volume densely. A loose charge melted unevenly; a tight charge gave the arc a consistent target. He poured a layer of crushed charcoal over the top of the copper charge, covering it completely. “The charcoal consumes the residual oxygen in the furnace atmosphere above the melt. It burns off before the copper reaches temperature.”

He closed the lid and moved to the switchboard.

The switchboard was bolted to the brick wall of the northwest corner, positioned when he had built it to be within reach of the furnace controls while remaining outside any arc flash radius from the furnace itself. It was a slate panel, four feet wide and three feet tall, mounted in a frame of polished brass angle iron. The knife switches were heavy copper-jawed units, each handle wrapped in turned hardwood for insulation, arranged in a logical sequence across the panel face - mains first, then the arc circuit, then the electrode feed control. Two glass-faced dial ammeters occupied the top of the panel, their faces six inches across, the needles currently resting at zero against their left-side stops. The connecting cables from the switchboard to the furnace were thick, uninsulated copper bar stock mounted on ceramic standoffs along the wall.

“That wheel,” Silas said, pointing to a ceramic-handled rheostat mounted to the wall beside the switchboard - a large wheel, eighteen inches in diameter, its rim wrapped in strips of ceramic tile for insulation. A mechanical linkage ran from the rheostat’s shaft up the wall and across to the electrode feed mechanism on the furnace lid. “When I close the main switch, you turn that wheel clockwise. The linkage lowers the carbon rods toward the copper charge. Turn it slowly - the gap between the rod tips and the metal needs to close gradually. When the arc strikes, you’ll know. Hold the wheel position when it does. Don’t let the rods advance further.”

“How will I know when it strikes?”

Silas looked at the furnace. “You’ll know.”

He gripped the handle of the main knife switch - heavy wood, the copper jaw behind it massive enough that closing it required a deliberate downward pull rather than a snap. He pulled it down.

The CLACK of the contacts closing was loud in the Tinkery, a hard metallic report that bounced off the brick walls. The cables on the wall beside the switchboard thickened slightly - a faint vibration in the copper bar stock, visible if you were looking at it - and the air near the connection points carried a sharp smell of heated varnish from the insulating lacquer on the ceramic standoffs warming under the current.

“Now,” Silas said. “Lower the rods. Slowly.”

June turned the wheel clockwise. The linkage moved, and inside the furnace the carbon rod assembly descended toward the copper charge.

The arc struck without additional warning. The sound was a sustained CRACK that was not a single event but a continuous tearing, crackling roar, as though the air inside the furnace had become a different substance entirely. Violet-white light erupted from the seams of the furnace lid with an intensity that lit the northwest corner of the Tinkery and threw moving shadows back through the shelving rows toward the center of the building. The ammeter needles on the switchboard swung hard to the right and held there, quivering against their upper stops.

June stopped the wheel and held it.

“Hold that position,” Silas said, his voice raised over the arc noise. “If the rods advance further they’ll contact the melt directly. That shorts the arc circuit and sends the full current through the crucible at once.”

Through the tinted viewing glass set into the furnace lid - dark green glass, thick enough to reduce the arc’s output to something the eye could tolerate - the copper charge was changing. The stranded wire bundles were not melting in the gradual way of metal approaching its temperature in a coal forge. The arc was depositing energy into the charge at a rate that bypassed the intermediate stages. The copper was liquefying from the point of arc contact outward, the solid metal at the edges of the crucible slumping into the growing pool at the center. Four minutes after arc strike, the crucible held a pool of molten copper that reflected the arc light back in pulses, its surface moving with the gas evolution from the charcoal layer.

Silas reached into his waistcoat pocket and produced a small cylinder of copper foil, twisted closed at both ends. Inside it was calcium boride powder - the boron source. Boron’s function in copper refining was specific: it sought dissolved oxygen in the melt and combined with it chemically, forming boron trioxide, which was insoluble in copper and rose to the surface as slag. The alternative was phosphorus, but phosphorus left residue in solution that reduced conductivity. Boron left nothing behind.

He used a long iron rod to slide open the charging port on the furnace lid - a small rectangular opening, just wide enough for the foil cylinder - and dropped the packet into the arc zone.

The color coming from the furnace seams changed immediately and completely. The violent violet of the copper arc was replaced by a spectral green the color of graveyard mist - sharp and cold, the green of a flame burning elemental boron, visible for the three or four seconds it took the calcium boride to disperse through the melt. Then the green faded.

“Power off,” Silas said, and reached for the main switch.

The CLACK of the contacts opening was as definitive as the one that had closed them. The arc sound stopped in the same instant.

Silas cranked the lid open. He took a long-handled iron skimming rod - flat iron plate on the end of a four-foot handle - and moved it across the surface of the copper pool, collecting the boron trioxide slag that had risen to the top: a glassy, dark scum that came away cleanly and left the copper surface beneath it with a stillness and a reflectivity that made it look less like metal and more like a pool of liquid set on fire.

He looked at it for a moment.

“Beautiful,” he said quietly. The heat coming up from the furnace was pushing against his face and fogging the lower edge of his goggles. He did not step back.

He picked up the long-handled iron tongs, reached into the furnace, seated the tong jaws around the crucible, and lifted. The crucible came clear of the furnace with slow, deliberate movements, the molten metal in it shifting with the change in angle, its surface tilting and then settling.

He carried it to the flask on the floor, which was locked closed with its iron pins and lying flat, the cope on top of the drag, the pour hole - a tapered channel cut through the cope - positioned over the central web of the H-shaped cavity. He straddled the flask, the crucible in both hands, and tilted it over the pour hole.

The copper entered the mold in a bright, continuous stream, orange-white at the pour hole, darkening to orange as it disappeared into the sand. It ran smoothly, filling the pour hole and dropping into the cavity below.

Then the flask lifted.

The bottom half - the drag - rose a quarter inch off the floor on one side as the steam pressure from the moisture in the sand found the path of least resistance upward through the parting line. The THUMP of it shook the floor planking under Silas’s feet. A geyser of steam and ejected copper droplets erupted from the vent holes on the near face of the flask, the copper droplets trailing light as they hit the floor and immediately dulled to black-red. The cope rose off its alignment pins, venting pressurized steam from its parting face with a sustained hiss, then dropped back down.

Silas set the crucible down on the floor brick he had positioned for it and stepped back.

“Back,” he said, and June, who had already moved, moved further.

The hissing continued for several seconds, then reduced. Copper had leaked from the parting line and was cooling on the floor in flat, irregular puddles, their surfaces going through red to dark red to black as the heat dissipated into the boards. The smell of scorched wood and steam filled the space.

Silas removed his goggles. He looked at the flask. He looked at his hands, which had a slight tremor in them that he was aware of and chose not to address directly.

“Too much water in the sand,” he said. “And not enough vent holes. The copper hit the moisture in the sand and the water converted to steam faster than the vents could pass it. The pressure had nowhere to go except up through the parting line.” He looked at the cooled copper puddles on the floor. “I rammed the sand yesterday evening. I was not as precise as I should have been with the moisture content.”

June looked at the flask. She looked at the cooled copper on the floor. “Are we done for tonight?”

Silas looked at her.

“Strip the telegraph cable off the east wall,” he said. “All of it. And those trolley cable bundles in the corner - chop them into six-inch lengths. We reset.”

~*~

He spent the first hour of the reset doing nothing that involved copper. He spread the new sand batch on a sheet of iron plate and positioned it near the furnace, where the residual heat from the first melt could begin drawing moisture from it. He checked it at twenty-minute intervals, pressing a handful between his palms and observing how it held the compression. When it reached the consistency he wanted - cohesive under pressure, no free moisture visible at the surface when compressed, the clay doing its binding work without the water doing damage - he began ramming the drag.

He used the ramming rod with deliberate, measured force, working in layers of approximately one inch, each layer compacted before the next went in. He counted his strokes per layer and kept the count consistent. When the drag was rammed to the parting face, he drew the pattern - a slow, straight pull, straight out of the sand along the draft axis, the mahogany coming free cleanly and leaving the H-shaped cavity intact in the sand behind it.

He used the fine wire for the vent holes, spacing them at one-inch intervals this time, working across the full face of the cope and the drag both, each hole poked through to the exterior of the flask so the steam generated at the mold face would have an unobstructed path outward. He counted the holes as he made them. Forty-seven in the cope, forty-one in the drag.

He closed the flask, seated the alignment pins, locked the iron clamps. He looked at it for a moment, then picked up the fine wire again and added eight more vent holes along the parting line edge of the cope.

The second melt took the same time as the first. The arc struck at the same point in the electrode descent, the ammeter needles swung to the same position, the copper charge liquefied on the same schedule. The boron packet produced the same spectral green flash and left the same clean mirror surface behind it.

Silas skimmed the slag. He lifted the crucible. He carried it to the flask.

He tilted it over the pour hole.

The copper went in and did not stop. It filled the pour hole in a smooth column, descended into the cavity, filled the H-shaped void in the sand, and rose up the riser tube - a secondary channel cut into the cope to serve as an overflow and a pressure relief - appearing at the riser opening as a bright bead of copper that meant the cavity below was full. Thin threads of steam came from the vent holes, straight and clean, each one a small straight column that rose to the ceiling and dispersed.

No pressure event. No flask movement. No ejected copper.

Silas set the crucible on the floor brick.

He looked at the flask. He looked at the riser, where the copper bead was cooling from bright orange to red.

“That,” he said, “is a clean pour.”

They let the flask sit. The riser bead went from red to dark red to black. The flask itself radiated heat into the space around it, the iron warming under the touch of a hand held near it without contact. At the forty-minute mark, Silas crouched beside the flask and pressed the back of his hand against the cope’s exterior. Warm, not hot.

He picked up the iron pry bar, seated it against the flask latch, and broke the latch open. He pushed the cope off the drag. Sand cascaded from both halves, dark and smoking at the mold face, lighter and cooler further from the cavity. In the drag, partially buried in sand, the casting was visible - the H-shape of the bus bar, its surface scaled with sand and oxide, a pour sprue projecting from the central web where the pour hole had been, and a riser sprue at the opposite end.

He gripped the casting with his heavy leather forging tongs and lifted it from the sand. It was still hot enough to be uncomfortable to hold barehanded. He carried it to the water barrel and lowered it in.

The sound filled the Tinkery completely - a sustained, high-pitched shriek of steam that whited out the space around the barrel for three seconds and left the air thick with water vapor when it cleared. Silas lifted the casting from the barrel, dark and wet, and carried it to the anvil.

He picked up the cold chisel and the two-pound mallet and sheared the pour sprue off at the casting face with two blows, then the riser sprue with two more. The copper was still warm enough to cut cleanly without being hot enough to deform under the mallet. He turned the casting over and knocked the flat faces against the anvil to remove the adhered sand.

Then he picked up the forging hammer.

“Why are you hitting it?” June said. She was watching from the other side of the anvil, her arms crossed.

“Casting leaves the crystal structure in a relaxed state.” Silas brought the hammer down on the flat face of the H’s upper flange. The sound in the Tinkery was a clear, ringing CLANG that came back from the brick walls a fraction of a second after the impact. “The copper solidified from the outside of the cavity inward, and the crystals grew in whatever direction the cooling gradient dictated. Random orientation, variable grain size.” Another blow, the same position, the same force. “Under the cyclorotors the Skiffer will vibrate continuously. Random grain structure in a conductor carrying high current, under sustained vibration, develops fatigue cracks at the grain boundaries.” Another blow. The surface of the casting was brightening under the hammer, the oxide scale breaking away and the fresh copper beneath coming up salmon-pink under the impact. “Cold working closes the microscopic voids the deoxidation process may have missed. It refines the grain size and gives the grains a consistent orientation relative to the direction of stress. The result resists fatigue cracking.”

He worked across the upper flange, then the lower, then the web, maintaining the same rhythm - each blow placed, not swung, the hammer coming down with controlled force rather than dropped weight. The casting was compressing visibly under the work, the dimensions changing by fractions of a millimeter, the surface finish going from rough and scaled to smooth and dense.

The final blow rang against the anvil. Silas held the hammer at his side and looked at the bus bar - the completed, forge-hardened, solid copper H-shaped conductor, ready for machining to its final connection dimensions.

He held it up by the tongs and turned it in the light from the east windows, which were beginning to show the first gray of approaching dawn.

He tossed it across the anvil to June. She caught it with both hands and grunted at the weight.

“Wire brush it down,” Silas said. He set the hammer on the anvil and straightened, his hands on his lower back. “Connection faces get a flat file after the brush. I want the contact surfaces flat to within half a thousandth.”

June was already turning the casting over in her hands, examining the surface, running her thumb along the web face. She picked up the wire brush from the bench.

“Tomorrow we wire the generator,” Silas said. He picked up his walking stick from where it was leaning against the anvil base. “And after that, you fly.”