The Independent Review Put Six Failures Back On The Table

The review began the following Monday in the company lab. Preston stayed out under a board instruction. I attended as a paid outside witness, with Elva's approval and a written agreement that gave me no control over the findings.

The fee was $85 an hour. I put the first twelve hours against our furnace bill.

Celia Rook, the independent safety engineer, took the chair at the end of the bench. She had asked everyone to leave their slide decks closed.

"Show me a raw event," she said.

Tamsin loaded the first camera record. A small slab of packed snow moved through our marked test lane. A bright flag on the reference rope crossed the painted line.

On the sensor trace, nothing crossed the alert threshold.

"Did the reference fail?" Celia asked.

"No," I said. "You can see both distance marks."

"Did the sensor fail?"

"It didn't detect that event. We hadn't proved why."

She nodded and wrote my answer down exactly.

We went through all twenty-seven runs. Four known movements had no valid alert. Two runs had gaps that made a result impossible to judge. Twenty-one had useful records and valid detections.

The investor summary called all twenty-seven successful.

Celia drew three columns on a board: PASS, MISS, NO RESULT.

"A missing result isn't a pass," she said.

It was such a small sentence. I had spent months trying to get it heard.

Ines sat beside a sealed box of old controller boards. She asked few questions. When she did, they took the room straight to the place everyone wanted to skip.

"Who changed this row?"

"Where is that approval?"

"Why does the repeat have the same event number?"

The system had allowed someone to replace the first run with a later attempt in the summary. The old record still existed underneath. Nobody reviewing only the final spreadsheet would know to look.

Preston's instruction explained the choice: use the best repeat.

That afternoon, Celia asked for the change request I had filed in February. Tamsin found it in the preserved ticket history.

I had requested $14,230 for a cold chamber rental and enough staff time to repeat the startup tests. Preston had denied it. His note said the buyer would cover further development after closing.

"Was that disclosed as unfinished work?" Celia asked Ines.

"No. It was presented as completed validation."

Celia hired a separate electronics specialist for the bench work. The board approved the extra cost. They cooled two controllers in a chamber, kept the reference input steady, and logged the battery voltage through startup.

The ten-day closing extension ran out while they worked. Aster Reach and the board signed another, allowing up to six weeks for the review. Ines sent me the notice because it affected the proposed access schedule. Nobody could claim that a sale deadline had silently become my consent.

No one needed mountain access for that test.

Three mornings later, Celia called me back to the lab.

"We reproduced a failure," she said.

At low temperature, the battery dipped during startup. A filter meant to block electrical noise could stay closed afterward. The radio still sent its regular check-in, so the screen showed a green connection while real event signals went unheard.

"Connected wasn't the same as listening," I said.

"Correct. We have more work before assigning every field miss to it. But the flaw is real."

Two cable faults explained the data gaps. They belonged in the report too. A single clean explanation would have been easier to sell, but we had six records, not one story to polish.

The pilot crews received a notice that afternoon. It described the missed detections, told them to stop relying on the units, and gave them a direct number for the review team.

One crew lead called me.

"We kept our spotters," he said. "You made enough noise about that."

"I'm glad."

"Wish we'd heard about the misses."

"So do I."

He gave Celia copies of the instructions his crew had received. Those instructions had called the green connection light a readiness check. The correction would have to reach the people using the screen, not just the people changing its code.

I sat in my truck afterward and unfolded the grocery list. Mara had added onions in her small handwriting beside my crust line.

I drew a second arrow under it. A weak layer could carry weight for days before it failed. Waiting did not make it stronger.

On Friday, Celia sent a draft finding to everyone who needed to answer factual questions. It supported the six unresolved failures and said the winter program had not met its own test rules.

I read it twice before reaching the attached approval sheet.

My name stood at the bottom of a document I had never seen.

Above it were the words: ALL TWENTY-SEVEN RUNS ACCEPTED FOR FIELD RELEASE.

Ines called before I could pick up my phone.

"Wilfredo," she said, "did you sign that page?"