Then rebuilt it again.
Every time, the same result emerged.
At the projected polar orbital cycle, repeated thermal contraction shifted the regulator housing into resonance with its operational actuation frequency.
The individual stresses remained within qualification limits.
The combined effect did not.
Not immediately.
That was the danger.
The system would look healthy.
For months.
Possibly years.
Then degradation would accelerate.
Isaac documented everything.
Forty-seven slides.
Six weeks.
Every assumption.
Every tolerance.
Every uncertainty.
He did not walk into the boardroom claiming the satellite would explode.
He said the exact opposite.
“This is a slow failure.”
That was what made it dangerous.
Cameron Ashford sat at the far end of the table.
Cameron was Nexora’s vice president of technical programs, though internally people simply called him the man who could make a launch date survive anything.
He was brilliant in his own way.
Political.
Organized.
Smooth under pressure.
He understood customer expectations.
Budget cycles.
Congressional committees.
Program milestones.
He also hated uncertainty when uncertainty threatened a schedule.
Giselle Hartmann, Nexora’s chief executive, sat beside him.
She had led the company through six years of rapid growth.
Under her, Nexora had doubled revenue and become one of the most aggressive competitors in national-security space contracting.
The company needed Prometheus.
A contract renewal depended on it.
A delay of four months would cost tens of millions.
Possibly more.
Isaac finished slide forty-seven.
Silence.
Cameron folded his hands.
“Your model parameters fall outside the validated heritage range.”
“Yes.”
“Then the model isn’t validated.”
“The environment is outside the test profile.”
“The test profile was approved.”
“By the customer.”
“Yes.”
“Based on the data we gave them.”
Cameron’s expression cooled.
“We gave them qualified test data.”
“You gave them data generated under separated stress regimes.”
Giselle spoke for the first time.
“Isaac, are you saying the qualification process was fraudulent?”
“No.”
“Then choose your words carefully.”
“I’m saying it’s incomplete.”
The legal representative wrote something.
Isaac noticed.
He continued anyway.
“The thermal cycle and vibration load are interacting.”
“We need combined-environment testing.”
“How long?”
Cameron asked.
“Four months if we redesign the valve housing in parallel.”
Cameron looked toward Giselle.
Four months.
The number sat in the room like a bill no one wanted to pay.
Cameron returned his attention to Isaac.
“The contract requires hardware in eleven months.”
“I know.”
“The customer has accepted the current design.”
“They don’t know what I know.”
Cameron’s jaw tightened.
“You’re suggesting we reopen a signed technical baseline based on an unvalidated cross-domain model.”
“I’m suggesting we tell them.”
“Tell them what?”
“That the qualification envelope doesn’t represent the mission environment.”
“And if your projection is wrong?”
Isaac answered immediately.
“Then we lose four months.”
“And if we delay?”
“We may lose the contract.”
“Yes.”
“And if we don’t?”
Isaac pointed toward the slide.
“You may lose the vehicle.”
The room went quiet again.
Cameron leaned back.
“You’re very comfortable spending money that isn’t yours.”
Isaac looked at him.
“I’m comfortable admitting uncertainty before hardware is in orbit.”
Giselle finally lifted her eyes.
“Can you sign the design approval with a technical reservation?”
“No.”
“Can you sign the current qualification status as acceptable while continuing secondary analysis?”
“No.”
“Why?”