# Instructions for the RS-485 bench record sheet

`rs485-bench-template.csv` is a blank template. `rs485-bench-planned-example.csv` only illustrates a set of plans. The 10-meter trunk, 1-meter/0.1-meter stubs, 115200 bit/s and 10000 frames are arbitrary demonstration plan values, not a recommended configuration, paper data or measured results. Replace them for the actual equipment.

## How to use the three planned rows

R01 saves the original configuration. R02 changes only the stub from 1 meter to 0.1 meter. R03 restores the 1-meter stub. Keep the transceiver, supplies, trunk, termination, bias and messages unchanged. The example does not claim that shortening the stub will necessarily resolve the fault; first observe whether the change is repeatable.

Replace every `Same as R01` with the actual configuration or a reference to an immutable baseline record. Termination fields include built-in device switches and external resistors. Bias fields include actual resistors and the fail-safe conditions of the receiver in use. Measure the stub from the trunk branch point to the transceiver interface, rather than recording only the short external cable.

## Result fields and units

`attempted_at_utc` is the actual time. Record `received_frames`, `error_frames` and `timeouts` according to the error definitions written in advance. Do not assume that the three are necessarily mutually exclusive or can automatically be added together; explain the counting convention in `error_definition`. If calculating an error rate, state separately whether the denominator is transmitted or received frames.

Use `measured_rise_time_ns` for the actual 10% to 90% rise time in nanoseconds. Record `idle_differential_v` and `common_mode_v` in volts, and bind the measurement point, polarity, reference ground and observation conditions to the evidence. `evidence_ref` locates the waveform or log. `conclusion` contains only judgments supported by this observation; `next_decision` states the conditions requiring broader validation.

All measurement, evidence and conclusion fields are blank in the example. Blank means unknown; 0 means that zero was actually measured. They are not interchangeable. Use `record_type=measured_trial` for actual records.

## Recheck the article's demonstration calculations

Assume propagation velocity fraction v=0.67 and c=3×10^8 m/s.

- Stub screening: 100 ns × 0.67 × 3×10^8 m/s ÷ 10 = 2.01 m; 20 ns gives 0.402 m.
- Round-trip propagation over 100 m: 2 × 100 m ÷ (0.67 × 3×10^8 m/s) ≈ 0.995 microseconds.
- Single bit time: approximately 104.17 microseconds at 9600 bit/s, and approximately 8.68 microseconds at 115200 bit/s.

The assumptions for these hand calculations come from the article. They are not R01–R03 test results and do not automatically determine whether termination can be removed. The CSV contains no calculation formulas, macros or pass decisions. Determine comparison and stop conditions from the device documentation and project requirements.