No one had to say yes
Getting a test instrument onto a submarine is not mainly a technical problem. It is a problem of getting a lot of people, none of whom work for you, to say yes. Then you go to sea and run the trial anyway.
Three chains of command, none of them mine
Every trial started on land, with paperwork. Per vessel, end to end, that meant shipyard-level authorization and fleet-level authorization before anyone went near the boat. Then, aboard, I briefed the crew, worked their local approval process, and got the department head to sign. Only after that could we notify the start, do the physical install, and close the loop back through ship and shipyard signature.
Not one of those approvers reported to the program. Not one of them had to say yes.
The install itself was hands-on, done by two or three of us: a temporary work surface clamped to shipboard handrails, the acquisition computer racked, and data and communications cabling run through the engine room, the control room, and maneuvering.
The schedule breaks on arrival
I was the designated trial lead at least twice. On a trial the team was seven engineers on two shifts, a 24/7 watch stationed around the boat, some of them depot-level and in-service engineers from outside my organization. The watch schedule gets built in advance against a hard constraint: certain positions cannot be filled by junior engineers, and the senior ones cannot go on every trip.
Then you arrive and the schedule breaks. Safety briefs, radiation dosimeters, rack assignments, and crew briefings hit everyone at once. I re-planned coverage live, under a running trial, and made sure people ate.
Underway, there is no second take
Instrumentation failures underway have no undo. The boat does not stop, and the trial cannot be re-run. Every second an instrument is offline is data that will never exist. So when something failed, I fixed it on the spot, and I documented exactly what was lost, when, and for how long. Post-test analysis had to know which windows were valid and which were not.
What I took from it
- Authority is not the same as getting a yes. Separate chains of command had to approve every trial, and none of them worked for me. The yes came from making the case clear, every time.
- The plan is a starting point, not a script. The watch schedule was built in advance and broke on arrival every time. The job was re-planning it live without dropping coverage.
- Underway, there is no second take. A failure you do not fix is data gone for good. So you fix it, and you write down exactly what the gap was.