Chương 186: CHAPTER ONE HUNDRED AND EIGHTY-SIX: THE MECHANISM FAILS

The interrupted-thread breech failed on the eleventh of November, on the forty-third test cycle, at the worst possible moment, which was in front of everybody. Briggs had scheduled the demonstration deliberately. Forty-two clean cycles over five weeks had convinced him the design was sound enough to show the council, and showing the council was how a development programme kept its authorisation. The demonstration was in the assembly hall with fourteen people present.

Marcus, Hargrove, Hollis, Danny, Vance, Price, and eight from the forge complex. Lena ran it, because Briggs had insisted. She solved it, she demonstrates it. Cycle forty-three: the breech block seized at the quarter turn. It did not fire. It did not fail dangerously. It simply stopped — the block a third of the way into its rotation, immovable in either direction. Lena tried to reverse it. It did not reverse.

She set the piece down on the bench very carefully and stepped back and said, into a room of fourteen people, "It's seized." There was a silence of about four seconds. Then Briggs said, "Good." Everyone looked at him.

"Forty-three cycles," Briggs said.

"It failed on forty-three and it failed closed and inert, which is the failure mode we chose this design for." He came forward to the bench.

"If we'd built the falling block, we would have found out about this in year three with somebody's hand on it. Lena — talk me through what you felt." She was pale but her voice was steady.

"It bound at about thirty degrees," she said.

"Gradually — not a catch. It got heavier through the last ten degrees and then it stopped. It didn't jump or grab."

"Progressive resistance," Briggs said.

"That's thermal. Not a defect — expansion." He turned the piece under the lamp.

"How many cycles since the last cooling interval?"

"Nineteen," Lena said.

"And the previous forty-two were run how?" She stopped.

"In batches of ten," she said.

"With cooling between." If you spot this tale on Amazon, know that it has been stolen. Report the violation.

"There it is," Briggs said.

"The tolerance is correct cold and it is not correct at sustained operating temperature. Nineteen consecutive cycles put enough heat into the block that it grew past the thread clearance." He set the piece down.

"Which means the design isn't wrong. The clearance is." Hargrove said: "How much of a problem is that?"

"Depends entirely on the answer to a question I don't have yet," Briggs said.

"If nineteen cycles is the limit, this thing is useless — an engagement runs longer than nineteen cycles. If nineteen is what you get with this clearance, and a different clearance gives you a hundred, then it's a number we adjust." He looked at Lena.

"What do you think?"

"Clearance," she said immediately.

"The seizure was progressive. A design fault would be abrupt. Progressive means it's a dimension, and dimensions are adjustable."

"Agreed," Briggs said.

"How do we find the right one?"

"Test to failure at four different clearances," Lena said.

"Same thermal loading, same cycle rate. Find the clearance that survives a hundred consecutive." She paused.

"Which means four barrels and four blocks and probably six weeks."

"Which means the prototype timeline moves," Marcus said.

"By about two months," Briggs said. Marcus looked at the seized mechanism on the bench.

"Move it," he said. Hargrove shifted in his chair.

"Can I ask a question that will be unwelcome?"

"Always," Marcus said.

"The falling block would not have had this problem," Hargrove said.

"The wedge geometry doesn't bind on thermal expansion the same way. We chose the slower, safer design and we have just spent five weeks and are about to spend six more on a problem the other design would not have presented." He paused.

"I am not arguing to change. I want it on the record that the safe choice has a cost and we are paying it."

"Noted and correct," Briggs said.

"It also has a benefit, which is that Lena is standing there with all her fingers."

"The falling block wouldn't have taken her fingers on cycle forty-three," Hargrove said.

"It would have taken someone's in year four when nobody was watching for it."

"Yes," Briggs said.

"That's the trade. Visible cost now against invisible cost later." He paused.

"I have been an engineer for forty-five years and I have never once regretted paying the visible one." Danny, who had not spoken, said: "Can I say something about the record?" Marcus nodded.

"This failure should go in the journal," Danny said. The room turned.

"We publish the technical guides," Danny said.

"We publish what works. We have never once published a failure." He paused.

"Every craftsman in three counties reads those guides and takes them as authority. And what they learn from them is that the march's engineering works the first time, which is a lie, and which means when their thing fails on cycle forty-three they conclude they are bad at it."

"It's a weapons programme," Hargrove said.

"We don't publish weapons work."

"Publish the thermal expansion problem," Danny said.

"Not what it's for. The problem is universal — any threaded closure under sustained heat has it. Mill gearing has it. Furnace doors have it. Publish: we built a threaded closure, we ran it forty-three times, it seized, here is why and here is how we are finding the answer." He paused.

"And put Lena's name on it, because she called it as a dimension rather than a defect inside of ten seconds, and every apprentice in this region should know that a nineteen-year-old did that." The room was quiet. Lena said, from the bench, "I don't want—""You don't have to want it," Briggs said.

"It's true and it should be written down." He looked at Marcus.

"He's right about the guides. I've been publishing successes for six years and I never thought about what that teaches."

"Write it," Marcus said.

"Both parts. The failure and the method."