40.

The Unmodeled Bolt

Darsen Response Laboratory, Darsen, on Talon's Reach's world. AGH 163.

The two bolts lay on Hesna Bell's drawing because it was the only clear patch of bench left.

She looked at them, then at Marek.

“I need that.”

“I need these.”

“Then take them.”

He took one. She waited with her brown fingers on the corner of the sheet until he took the other.

The channel occupied the middle of the test room, between the workbench under the windows and the control stand. At the bench it had seemed a long, awkward object; mounted in its bed, with the water connections made and the guards waiting to be closed, it had acquired a compactness Marek liked. Nothing projected farther than it needed to. He could walk round it without turning sideways. That had cost him two afternoons and a revision to a brace which looked, now, as if it could never have been made any other way.

Marek Ulve, the orc fitter, knelt beside the moving section, his soft belly against his raised thigh, and put the first bolt through its joint. The short arm traveled with the channel, but these two plates were meant to stay fastened together. The movement belonged farther up, at the pivot. He seated the bolt, checked the fit, and took hold of the arm to test for unwanted play.

It held.

“Moving joint,” Hesna said from behind him.

He moved his broad brown-green hand so she could see the meeting edges.

“Again.”

He did it again. She entered the result before turning the page towards the fixed support. There the second bolt held the brace to the bed. It was from the same fabrication batch, taken from the same wrapped set; Marek had checked it with the first. It seated properly too. He tried the held pieces, found no play, and told her to enter it.

“I was going to.”

He had already reached for the marking pencil.

“You were looking at something else.”

“The line you told me to enter.”

He gave her a little more time.

Across the held edges at each position he drew a fine line. He also marked the bolt heads, M for the moving joint and F for the fixed support. The letters would be useful when the machine was taken apart. Hesna put the positions beside the letters in the fit record, and Marek signed the two checks.

She had tied her long dark hair up before coming into the room. Now she tightened it again, trapping a loose wave that kept falling towards the page, and fetched the supplied model record from the bench. Hesna had supplied the model of the channel's shape, supports and permitted conditions. The laboratory's statefold calculation had given an acceptable response. Marek had seen the result and had gone home pleased that evening, though he had waited until the following morning to say so. He liked a method that might let a customer reject a bad design before paying him to build the whole of it. He also liked getting paid to build this one, which had looked worth building.

Hesna ran her finger along the conditions for the physical trial. The operator checked them against the settings at the control stand.

“And the stop?” she asked.

The operator answered her, then looked at Marek. He gave the physical conditions he would be watching. They had agreed on them already. It was easier to repeat them now, with the channel quiet between them, than to discover halfway through that each had been waiting for the other to act.

Kiran Holloway arrived while Marek was putting the pencil away.

“Where can I stand?” they asked.

He showed them the observation side, where the moving joint would remain visible when the barred guard was closed. Holloway was visiting from the mutual-care town where they taught medical instrument work. They had come to examine the comparison between the calculated response and the physical one, and had brought something to write on. Marek was grateful that they had asked before finding their own place. There was a useful patch of floor beside the control stand which was useful because nobody stood in it.

Hesna gave Holloway the declared conditions. Marek let them see the two seated marks before the guard closed.

“You can follow that one from here,” he said, pointing to the moving joint. “The arm travels. This join should stay tight. You can see the line when it settles.”

Holloway changed their position slightly.

“There?”

“Yes.”

While the operator finished the last checks, Hesna moved a box of loose fittings from one end of the bench to the other.

“When this goes,” Marek said, “I'm going to stand here and look all the way down it.”

“I've got two trays to lay out.”

“After.”

“After what?”

“After I've looked.”

She considered the length of the bench. “You could look now. From that end.”

“I can see three boxes and your elbow.”

She moved her elbow. He laughed and took the empty fittings wrapper to the bin. It had been weeks since either of them had been able to put something down without first moving something else.

The operator started the wet-cycle run.

Water spread through the channel, brightening the surface Marek had finished. The working section moved through its short travel and settled. When it settled, the fine line across the held plates remained whole. He watched the fixed support, then the moving section on its next return. The sound was a low, even rush with the small change he expected when the flow altered. There was room to hear it; the bed did not rattle, and the brace he had remade kept clear of the guard.

Hesna stood beside him for a while without asking anything. When she went back to the record, she was smiling.

They took the scheduled observations through the following shifts. Marek did other work within hearing of the run and found himself making room on the bench whenever he carried a tool back to it. He would clear a patch, look for a place to set down the object already occupying it, and put everything where it had been. Once, Hesna caught him doing it and moved her two empty trays under the bench without comment.

*   *   *

The changed sound came into the room while Marek was dressing the edge of a cover.

He stopped the file before he knew what he was listening for. The run continued through its familiar movement, but something under the rush now answered a fraction late. He laid the cover down and went to the observation side.

The moving section settled. The line at the held joint did not quite meet.

Marek bent until he had the same sightline he had used before the run. He could see a small step between the marks. At the fixed support the line remained whole.

“Look at the return,” he told the operator.

The operator came far enough to see, then went back to the controls. Hesna put down her pen.

“Which part?”

“The held plates. Here.”

He pointed beyond the bars. She crouched beside him, looked, and rose to check the operating record.

The operator eased the flow within their agreed settings. The rough sound thinned, then came back. The moving joint failed to meet its mark again, with more room between the edges.

“Stop it,” Marek said.

The operator was already at the control. The flow diminished, the moving section settled and the room became quiet enough for Marek to hear water draining beyond the bed. He kept looking at the mark. It did not come right when the machine stopped.

“These stayed inside the declared range,” Hesna said, showing him the last operating sheet.

“Write the adjustment down as well.”

“It is.”

They waited until the operator had isolated the run and opened the guard. Marek put a hand on the moving arm. Its new freedom was slight and unmistakable. He could feel it before seeing the broken line shift. The bolt had ceased to hold the plates as it had held them at the start.

He fetched a support and took the weight before loosening anything. His first thought was to fit another fastening and check the alignment. A part could fail. He had replaced enough of them. Then he saw the edge of the channel beyond the joint, where the shape he had spent an afternoon finishing no longer followed its support.

“Bring the straightedge,” he said.

Hesna brought it and held the far end while he laid it along the working section. There was daylight under it where there should not have been. He shifted it, checked from the other side, and made her hold it again.

“We'd have to take that whole length out,” she said.

He tried the adjacent seat. It had drawn out of line too. The damage ran into the part they would need sound in order to replace the first one.

He sat back on his heels. Water cooled the knee of his trousers where he had knelt. The cover he had been filing lay on the bench, ready for an assembly that would no longer take it.

“This one's gone,” he said.

Hesna kept hold of the straightedge until he took his end off the channel.

“The bed?”

“I'm talking about the channel.”

“I know. Can we use the bed again?”

He looked at it. She waited, with the straightedge tucked under her arm, while he checked the mounting. It was a question he would have to answer whether he liked the time she had chosen for it or not.

“Yes,” he said. “That looks sound. I'll check it properly.”

The operator asked where he wanted the recovered parts. Marek cleared a shallow tray and divided it with a strip of wood. He wrote moving joint on one side and fixed support on the other. The damaged bolt, still marked M, went into the first. He removed its sound mate from the fixed support and put it beside the second label.

Hesna brought the initial fit sheet and the run observations. The early entries were in her hand; others belonged to the operators. Marek found the first recorded failure of the moving joint to settle to its mark, the adjustment, and the wider displacement that had followed it. Then came the stop. He had seen each of those things. In the record they occupied less room than the two successful fitting checks.

The operator began cleaning the water from around the bed. Marek got up and went to help move the disconnected line out of the way. When he returned to the bench, Hesna had put the favorable calculation beside the trial record.

“Leave it there,” he said.

She looked up, as if he had told her to do something she was already doing.

He picked up the cover and put it with the other parts they would have to examine.

*   *   *

By the next working morning he wanted it to be the batch.

He had slept, come back, checked the damaged assembly in better light, and still wanted the same thing. A faulty piece of metal was an unpleasant but familiar proposition. He could reject a batch. He could tell the maker what had happened and find another supply. The channel had been built to a description both he and Hesna had accepted, with fitted relations he had checked himself. He would rather discover a bad piece in it than begin again with that description.

He had taken hold of the tray when Hesna pushed the earlier operating sheets towards him. Her hair was secured more tightly today, and she had put the ink far from the edge, where he could not knock it over while reaching for the pieces.

“Look at these first.”

“We checked the settings.”

“The last part. If an earlier shift went outside the range, this wasn't the trial we specified.”

“Would that let us keep the design?”

“We'd need a new build to test it.”

He left the tray where it was and read the entries nearest him. Hesna followed the rest, flattening the turn of a sheet before moving on. They reached the ordinary observations before the changed vibration.

“Nothing,” she said.

“All inside?”

“Yes.”

She left the sheets spread in their order.

“Then we'll need them,” he said.

He set a small comparison joint at the clear end of the bench. Its lower piece clamped to the wood. The upper piece had a short handle, with stops around it to contain any movement and keep hands clear of the meeting edges. He had chosen it because its fastening fit matched the one they needed to examine. It could show whether the recovered pieces held without putting water through the wrecked channel again.

Marek took the bolt from the moving joint side of the tray. The M was still visible on its head. He fitted it, checked the seating, and put his hand on the handle.

The upper piece rocked. The small clack was louder on the bench than he had expected.

Hesna came round the end. He moved his hand aside.

“Here.”

She took the handle and moved it within the stops. Her first movement was too cautious to show much. She tried again, felt the loose relation, and looked at the base clamped to the bench.

“The lower one's still?”

He put his hand beside it so she could watch both together. She repeated the movement.

Then he took the damaged fastening out, returned it to its labeled compartment and fitted the one from the fixed support. He used the same seating check. Hesna kept her place beside him.

This time the handle held. He tried it, let her try it, and then put the damaged piece back. The unwanted rocking returned.

“Same batch,” she said.

“Yes.”

“And both held at the beginning.”

“I checked them.”

“I watched you.”

He took the damaged fastening out again and returned it to its label before reaching for the sheet. Her two entries and his signature were there. Under them lay the observations from the run, the moving joint's clean returns continuing well beyond the first start.

Marek rubbed the side of his jaw with the back of his hand. The blanket explanation he had brought in that morning had become less comfortable. It did not explain why this particular fastening had changed in this particular place while its mate had held, or why the change had developed through the run.

“I still want the batch examined,” he said.

“So do I. Look at this one.”

Hesna moved the record until he could see the entry she meant. It was the last ordinary observation before the changed vibration. The supplied operating conditions were still permitted. She laid out the missed return, the adjustment, the larger displacement and the stop in the order they had happened.

Marek followed the little differences in the words describing the joint. He could remember the motion more readily than the words. The first small missed return had looked capable of settling. The next had not.

He went to the disconnected channel. The arm rested on the temporary support he had put under it. Marek left it there and showed Hesna the pivot, then the plates below it.

“Every return came through this,” he said, pointing to the empty fastening hole. “The brace at the other end stayed on the bed.”

Hesna looked from one position to the other. Both fastenings had belonged to the same permitted run. Only one had been worked repeatedly by the traveling section.

“Where was that held before it shifted?” she asked.

He showed her the old seated mark. She compared it with the two fit entries she was carrying, then went to the model record.

The supplied representation included the travel he had shown her. It included the range of the water conditions and the places where the channel was held. Hesna laid that description beside the observation sequence and found the entry for the fastening relation.

Marek waited while she checked the following part.

“That stays fixed,” she said.

“It's supposed to.”

“Here. In what we supplied.” She drew the sheet closer and showed him. “It holds like the original fit. Through the run.”

He looked from the entry to the two bolts in their tray. Their present difference occupied nothing in that description. The calculation could vary the conditions they had supplied; this held relation remained what it had been when he first checked it.

“Where does the later play go?”

She started to point farther down the page, then stopped.

“It doesn't.”

They worked back through the supplied representation. Hesna checked the parts he could not read unaided and explained the relation when he asked. Marek put the damaged fastening beside the earlier fit check, then beside the observations that followed it. In the small fixture it could still be made to show the play; the other, returned to the same fixture, held.

The ordinary fastening had fatigued as the moving joint worked it. They had supplied the calculation with a relation that never acquired that changing condition. Marek had checked its first fit correctly; Hesna had checked the response the supplied model could give. Neither check had put the omitted interaction there.

He sat down on the stool and read the observation sequence once more. Knowing what had happened made each entry less puzzling. It did nothing for the channel laid open in the middle of the room.

Hesna pulled out a fresh sheet.

“I'll need the position of both pieces and the starting fits with this.”

“They're in front of you.”

“To keep with it.”

He found another pen. While she recorded the model's unchanged relation, he copied the fit and position entries, keeping the originals beside him. His first attempt at the moving joint's little outline was too small to show where the plates had shifted. He drew it again.

*   *   *

The contract representative came with a folded plan and left it folded.

Marek had cleared a place for it, but the representative put it under one arm while Hesna explained the result. They listened to the account of the lost prototype, asked what could be reused and then asked for the delivery date.

“We have to revise the description before we commit another build,” Hesna said.

“How long?”

She gave the work they could name. She did not give a number of days for what they had not yet done.

The representative looked towards the channel.

“I had people ready to receive this.”

Marek began to explain that the bed was sound, then stopped. The representative was still looking at the place where the channel had drawn out of shape.

“They can't receive it,” he said.

“No. What can I tell them about the next one?”

“I can't give you another date yet,” Hesna said. “I'll send you an estimate when we have enough to make one.”

“As soon as you do.”

“Yes.”

There was no quarrel. Marek would have found a quarrel easier to return to work after.

He carried the unused chair away and set the teaching drawing on the bench.

The laboratory had arranged a subsequent comparison for learners. He and Hesna could show the supplied model, its accepted response and the physical result, with the omitted interaction identified in the accompanying account. Marek had redrawn the channel neatly enough that the moving arm could be distinguished from its support at a glance. Beside it, Hesna had kept the original model description and response. The paragraph about the prototype loss lay at the bottom.

It was an accurate paragraph. He had made sure of that.

Holloway read the proposed material and asked where the recovered fastenings would be.

“We can mention the comparison here,” Marek said, pointing.

“I read it. Where will the pieces be?”

He nodded towards the tray at the other end of the bench.

“Then we should put them with this.”

“They'll hear what happened.”

Holloway looked at the clean drawing again. “Let them try the failed result.”

Marek moved a box to reach the tray. The wooden divider came loose, and he held it upright with one finger while he lifted the whole thing. He could put it in front of every learner who came through. He could spend the next part of his working life showing people the joint he had fitted and the channel it had cost them. It would be a remarkably durable piece of work after all.

“It isn't the only thing we've made here,” he said.

“I haven't asked you to make it the only thing.”

Holloway waited while he set the tray down.

Hesna clamped the guarded comparison joint beside the tray and fitted the damaged fastening. She held its handle steady in the position the drawing showed.

“If I keep it like this,” she said, “what can they see?”

“A held joint.”

She took her hand away.

Marek put his own on the handle and rocked it. The movement was small enough to disappear in his drawing, however accurately he drew the pieces. He had given the learner the state they expected and a paragraph telling them that something else had occurred.

He put the fixed-support mate beside the fixture.

“Both,” he said.

Holloway nodded. “And where each came from.”

“The labels stay.”

“The record too,” Hesna said. “They need to see that both held before we started.”

Marek looked at his clean drawing. It was still useful. He moved it along the bench to make room for the tray, then brought the short observation sequence over with the fit record and Hesna's account of the omitted relation.

The tray would need its divider fixed before people handled it. He found a strip that fit better and trimmed it while Holloway and Hesna discussed the order of the comparison. He could hear where they disagreed. Holloway wanted the pieces tried before the explanation; Hesna wanted the original fit available before a learner decided one had simply been loose at installation. They arranged it so both could happen.

Marek brushed the shavings into his hand. The M on the damaged head was still readable. He left it alone.

*   *   *

At the first teaching comparison, the learner nearest Marek traced the moving arm on the drawing without touching the paper.

“That travels,” they said. “These stay together.”

“Try them.”

The learner looked at the handle, then at him. Marek showed the outer grip and the stops. The fastening marked M was fitted in the comparison joint, with its moving joint label laid beside it.

The learner moved the handle. It rocked against the lower piece and came back with its small clack. They leaned closer and tried once more, watching the two pieces rather than the handle.

“It's loose.”

“Now this one.”

Marek removed the damaged fastening into its labeled place and fitted the sound mate from the fixed support. The learner kept watching. When he finished the seating check, they tried the same movement. The handle held.

They looked at the two labels.

“Were these both in it?”

“Yes.”

Holloway put the original fit record where the learner could read it. The learner followed the two entries and Marek's signature, then looked at the moving-joint specimen again.

“When did that one stop holding?”

Marek brought the observation sequence closer.

He let the learner find the early returns. At the first recorded missed mark, they paused and asked whether that had been outside the supplied conditions. Hesna showed the checked readings. The learner went on to the adjustment, the worsening play and the stop.

“It kept getting worse.”

“At that joint,” Marek said. He pointed to the fixed support's observations. “This one kept holding.”

The learner stood back to see the channel as well as the drawing. For a moment Marek thought they would ask him to explain the whole thing from the beginning. Instead they followed the arm's line in the air, then stopped their hand at the held plates.

“So this one was being worked with it.”

“Yes.”

Hesna gave them the supplied representation. She did not wait for them to find an entry they had not learned to read. She showed the fastening relation and followed its unchanged state through the part of the run it described.

“What would the comparison need to include?” she asked.

“That it might come loose.”

“What would tell us more than that?”

The learner looked back at the observations. They put one finger beside the first fit and another beside the later play.

“What happened between these.”

Hesna drew the record towards her far enough to point without covering their hands.

“That's work we still have to put into it.”

Marek gave the learner the original trial's result: the fastening-fatigue interaction the model had omitted, the lost prototype and the delayed contract.

“Would another kind of bolt have saved it?” the learner asked.

“We haven't tested one.”

“But something stronger?”

“Maybe. I'd want to see how it held through the run.”

He drew the observation sequence back beside the model description. The learner stayed looking at both.

The others came round to try the guarded comparison. Holloway made room at the end of the bench and asked one of them to say which piece they were taking before the exchange. Marek caught himself reaching to do the substitution for that learner, then left his hand beside the tray while they read the label and asked about the fit. He answered. They made the comparison.

Afterward he and Hesna checked the pieces back into their two places. The sound one still had its F. The damaged one went beside the moving joint label, with the short history folded where it could be taken out without lifting either specimen. Holloway gathered their notes and spoke with a learner who had been waiting to ask about another instrument. Marek heard only the beginning of the question before the operator called him to the test bed.

One of the reusable supports needed a decision before it could be put away. He examined it, marked the work he wanted done and returned to find Hesna measuring the small space left beside the teaching tray.

“Both trays?” he asked.

“One.”

“You've revised that.”

She slid the empty tray towards him. “There. Bare wood.”

He looked at the clean bottom of the tray, then along the crowded bench to the unfinished cover. He left the tray where it was and fetched the cover. Hesna moved her papers just far enough for him to lay it flat, then sat beside the model record with her pen.