Open question #3 in docs/architecture.md §11. Requires the full cable drum (BOM 3.1) — the test is only valid at real distance.
Problem. The start pulse travels 400+ m of copper next to high-energy ignition systems. Nodes accept only a fixed-width 5 ms pulse and reject sub-millisecond spikes — but the width-validation scheme has not been tested against real interference.
Test plan
- Full-length trunk on the drum (coiled cable is a fine worst case for a go/no-go).
- Drive the start pulse through the RS-485 transceiver onto its pair; receive at the far end.
- Inject noise: a running ignition system / spark gap near the cable. Verify spikes are rejected and genuine pulses always pass.
- Capture edge quality and propagation with an oscilloscope.
Pass criteria. Zero false starts and zero missed pulses over the test session, with margins documented.
Validates decisions D01 (wired link) and D04 (broadcast pulse) as implemented.
Open question #3 in
docs/architecture.md§11. Requires the full cable drum (BOM 3.1) — the test is only valid at real distance.Problem. The start pulse travels 400+ m of copper next to high-energy ignition systems. Nodes accept only a fixed-width 5 ms pulse and reject sub-millisecond spikes — but the width-validation scheme has not been tested against real interference.
Test plan
Pass criteria. Zero false starts and zero missed pulses over the test session, with margins documented.
Validates decisions D01 (wired link) and D04 (broadcast pulse) as implemented.