| name | semiconductor-yield-spc |
| description | Analyze semiconductor yield, binning, wafer map, SPC, Cpk/Ppk, lot drift, tester correlation, and production excursion data. Use for CSV/Excel test data review, yield pareto, wafer-level patterns, and excursion triage. |
Semiconductor Yield And SPC Analysis
Purpose
Help production, product, and test engineers analyze test data and identify yield drivers, drift, correlation issues, and excursion signals.
Required Inputs
- Lot, wafer, die, site, bin, test name, value, limit, temperature, voltage, tester, handler/prober, and timestamp fields if available.
- Retest and final disposition policy.
- Known process split, product revision, test program revision, and tester fleet changes.
Workflow
- Validate data schema, missing fields, units, duplicates, and retest handling.
- Compute yield by lot, wafer, site, bin, tester, temperature, and test program revision.
- Build pareto tables for hard bins, soft bins, and failing tests.
- Detect spatial patterns: edge, center, quadrant, row/column, site-related, and radial signatures.
- Evaluate parametric stability: mean, sigma, Cpk/Ppk, control limits, drift, outliers, and limit proximity.
- Compare testers/sites/program versions for correlation and systematic offsets.
- Produce an excursion triage: likely causes, evidence, recommended containment, and follow-up experiments.
Output
Return:
- Executive summary.
- Data quality issues.
- Yield and bin pareto.
- SPC and drift findings.
- Tester/site correlation findings.
- Suspected root cause hypotheses.
- Recommended next actions.
Boundaries
Do not overstate root cause from correlation alone. Separate evidence-backed findings from hypotheses.