# Robotic wire harness assembly trial log: fields and calculations

Version: 2026-09-23. Companion files: robot-trial-template.csv and robot-trial-demo.csv.
The CSV files use UTF-8 with BOM and comma delimiters. Open them with Excel’s “From Text/CSV” function or LibreOffice. Field names remain in English so teams working in different languages can combine records. The table below explains each field.

Each row represents one complete-process attempt on one harness from an agreed starting point. Define unit_definition before the trial. In the example, the boundary runs from the start of feeding until acceptance/rejection and the cell being ready for the next unit. Retries remain in the same row and do not count as new units. Create another batch_id when the workpiece, fixture, program, condition or timing boundary changes. Use a separate batch for trials that cover only the insertion stage, even for the same workpiece; do not include them in the whole-process denominator. Add actual acceptance methods to this suggested log according to the product requirements of the project.

| Field | Meaning and entry instructions |
| --- | --- |
| data_kind | Use measured for actual measurements; the demonstration file uses synthetic_example and must not be included in measured-data summaries |
| batch_id / trial_id | Trial batch with consistent conditions / unique attempt ID within that batch |
| unit_definition | Workpiece unit, starting point and endpoint; whether feeding, inspection and recovery are included |
| drawing_revision / component_lot | Harness drawing revision / references for terminal, housing and cable lots |
| fixture_program_revision / condition_id | Fixture and program revisions / reference to the trial-condition record |
| first_pass_ok | 1 if all batch acceptance checks are completed without a retry; 0 if failure to pass first time is confirmed; blank if unknown |
| final_ok | Acceptance outcome after permitted retries: 1 or 0; leave blank if not determined |
| retry_count | Number of additional automatic retries within this attempt; explicitly enter 0 for none; leave blank if unknown |
| human_recovery | 1 if manual recovery occurred, 0 if it did not, blank if unknown; this is not a count of people |
| occupancy_seconds | Seconds occupied, including failures and recovery within this attempt; leave missing values blank rather than filling them with 0 |
| first_failed_stage | First failed stage, such as feed, alignment, insertion, locking or inspection; enter none if no stage failed |
| lock_check / electrical_check | Check status under the project method: pass, fail, not_measured or not_applicable; reference methods and limits in evidence_ref |
| evidence_ref | Location of raw logs, inspection methods and criteria, and images/video under the same ID; demonstration data has no measurement evidence |
| notes | Records of damage, terminal backout, wire sequence, abnormal recovery or additional checks |

## Recalculate the demonstration records

All 6 rows in robot-trial-demo.csv are fictional data illustrating the recording method. No physical test took place. They share the same unit_definition, conditions and revisions, and have complete first-pass outcomes, final outcomes and occupied times:

- Started attempts N = 6.
- 4 rows have first_pass_ok = 1; first-pass yield = 4 / 6 = 66.7%.
- 5 rows have final_ok = 1; final acceptance proportion = 5 / 6 = 83.3%.
- Total occupancy_seconds = 30 + 30 + 50 + 70 + 30 + 30 = 240 seconds; mean occupied time is 40 seconds per attempt.
- 1 row has human_recovery = 1; retry_count totals 2 retries.
- T004 fails during insertion, so subsequent locking and electrical checks are not_measured. This does not prevent a final failure decision for the whole unit.

Before summarizing actual data, check that batch conditions are consistent, trial_id values are unique and the records for each metric are complete. If 1 final outcome is missing among 6 started attempts, report “known successes, known failures, 1 missing record, 6 started attempts” and do not yet claim a final acceptance proportion for all attempts. Also report missing time values; do not divide the recorded portion by the total number of attempts. An electrical check that was not performed cannot be marked pass. Whether electrical checks are required depends on the trial purpose and acceptance requirements for the product.

The log has no automatic calculation macros and supplies no universal contact force, retry limit, sample size or equipment safety parameters. Enter the actual limits before execution. The proportions in the log are descriptive sample results and do not establish production capability.


## Original CSV values and their meanings

The original files retain stable English values. The following table explains their meaning; demonstration identifiers do not identify actual parts or tests.

| Original value | Meaning |
| --- | --- |
| `measured` | Actual measurements from an executed trial. |
| `synthetic_example` | Fictional demonstration data; exclude from measured summaries. |
| `DEMO_ONLY` | A demonstration-only batch identifier. |
| `T001–T006` | The six individual demonstration attempt identifiers. |
| `one harness; feed start to accept/reject and cell ready` | One harness, timed from the start of feeding through acceptance or rejection and restoration of the cell to readiness. |
| `DEMO-D1` | Fictional drawing revision. |
| `DEMO-L1` | Fictional component lot. |
| `DEMO-F1-P1` | Fictional fixture and program revision. |
| `DEMO-C1` | Fictional trial condition. |
| `feed` | Feeding stage. |
| `alignment` | Alignment stage. |
| `insertion` | Insertion stage. |
| `locking` | Locking stage. |
| `inspection` | Inspection stage. |
| `none` | No failed stage, when used in first_failed_stage. |
| `pass` | The specified check passed. |
| `fail` | The specified check failed. |
| `not_measured` | The check was not performed; this is not a pass. |
| `not_applicable` | The check does not apply to this trial. |
| `none; synthetic data` | There is no physical-test evidence; the data is fictional. |
| `Illustration only; no physical test` | The row illustrates recording and does not describe a physical test. |