Open question #1 in docs/architecture.md §11 — the highest-risk item in the project. Per decision D15, no batch purchases, PCB orders, or public timelines until this passes.
Problem. A datasheet "≤ 1 ms response time" specifies delay, not repeatability. Timing accuracy lives in the jitter of the full path: beam → sensor → optocoupler → capture.
Test plan
- Slotted disk on a motor at stable RPM breaking the beam; hundreds of cycles.
- Measure timestamp spread with a logic analyzer (BOM 2.4).
- Run the same rig on the brand-name sensor and on low-cost industrial clones (BOM basket 1, validation-first option).
Pass criteria. < 1 ms total path jitter (beam → timestamp).
The outcome decides
- the sensor BOM — brand vs clone is a several-fold cost difference for the open BOM;
- whether D12 (no line-scan photo finish) stands — see its "would change it" clause.
Independent reproduction is a top-tier contribution — report via the Bench validation result issue form or right here with raw data attached.
Open question #1 in
docs/architecture.md§11 — the highest-risk item in the project. Per decision D15, no batch purchases, PCB orders, or public timelines until this passes.Problem. A datasheet "≤ 1 ms response time" specifies delay, not repeatability. Timing accuracy lives in the jitter of the full path: beam → sensor → optocoupler → capture.
Test plan
Pass criteria. < 1 ms total path jitter (beam → timestamp).
The outcome decides
Independent reproduction is a top-tier contribution — report via the Bench validation result issue form or right here with raw data attached.