Why We Photograph Every Ground Return

Return paths are easy to dismiss until a UART starts speaking in riddles. In our bench-first cohorts, we ask participants to photograph the ground braid, annotate the shot with net names, and pair it with a single scope capture. The ritual feels slow on day one and indispensable by week three.

We keep the bar practical: one overhead shot, one close-up of the star point, and a short note on what changed since the last image. That trio is enough for async mentor review without bloating your drive.

The second paragraph of discipline is about honesty. If a student skips the photo because the bench looks messy, we treat that as signal — not shame, but a cue to reset cable dress before chasing bugs. Neat returns correlate with faster bisects, even when correlation isn’t causation.

Finally, we archive the photos beside firmware tags. When a regression lands, the image diff often explains more than the git diff. It is operational hygiene, not theater — and it keeps junior engineers from narrating ghosts.