Data Catalog
Discover and request access to industrial published data products by participants in the BACK-X data space. Metadata is based on DCAT-AP with backend-domain extensions.
BACKX.LineMachineEvents.v1
Line and backend-machine events with plant context
FAIR description
Unified stream of production and machine events in backend processes (dispensado, curado, test, AOI), enriched with line, machine, recipe and product context. It is the basis for the plant digital thread, traceability by line/machine/product, status dashboards, OEE and alarms, and data preparation for analytics and AI.
Industrial Context
Data Product Structure
Event identification
event_idGUID · EventBase.idevent_typeProcessEvent | AlarmEvent | InspectionEvent | CommandEvent…created_timestampDateTimeupdated_timestampDateTimePlant context
line_id→ lines.idline_namestringmachine_line_idMachine instance on linemachine_namestringmachine_typedispenser | oven | AOI | loader…Product context
customer_product_idstring (nullable)part_numberstring (nullable)Technical metadata
source_system"ProcessTracker vX.Y"ingestion_timestampDateTime en BACK-Xversion_schema"BACKX.LineMachineEvents.v1"Usage Policy (ODRL)
UC3M and ONIZEA may use the data for development and validation of AI/analytics models and for building BACK-X internal dashboards and tools.
Redistribution of detailed events to third parties outside the consortium is not permitted without explicit consent from MSTECH.
Aggregated views only: KPIs by product/day/line. Derived data products will retain lineage reference to the source data product.
These conditions are automatically enforced by the Eclipse EDC Connectorof BACK-X in ODRL format during data-contract negotiation.
Access Interface
Cross-platform access through the participant EDC connector. Automatic contract negotiation.
EDC_BACKX endpoint
Download of Parquet/JSON files with daily history for offline analytics.
v1 — batch
Streaming access for near-real-time monitoring. Planned for v2.
v2 — coming soon
BACKX.PCBManufacturingEvents.InternalRaw.v1
Internal raw PCB manufacturing events with complete plant, product, machine, recipe, material, defect and quality context
FAIR description
The most granular dataset in the BACK-X space. It contains internal operational events from SMT/PCB lines captured from AOI, ICT, ERP, MES, FCT and ProcessTracker. Each record represents a manufacturing or inspection event enriched with line, machine, product, PCB design, component, material, recipe-parameter, test/defect, quality KPI and full traceability information. Primary source for diagnostics, root-cause analysis and generation of derived products.
Industrial Context
Data Product Structure · Main groups
Identification + source system
event_idGUID · unique event identifiersource_system.nameAOI | ICT | MES | FCT | ERP…source_system.connector_idIngestion EDC connectorsource_system.tenant_idLogical data tenantPlant, Line and Product
plant_context.plant_idPlant identifierplant_context.line_nameSMT line nameplant_context.shift_idProduction shiftproduct_context.work_order_idWork orderproduct_context.serial_numberUnit serial numberPCB Design + Component
pcb_design_context.board_typerigid | flex | rigid-flexpcb_design_context.layer_countNumber of layerspcb_design_context.surface_finishENIG | HASL | OSP…component_context.package_typeQFN-32 | BGA-144 | SOT-23…component_context.msl_levelMoisture sensitivity levelTest, Defect and KPIs
test_and_defect_classification.test_resultPASS | FAILtest_and_defect_classification.defect_codeOPEN_JOINT | SHORT…test_and_defect_classification.root_cause_categorymachine | material | process…quality_and_kpi.first_pass_yield_flagbooleanquality_and_kpi.cycle_time_sCycle time in secondsUsage Policy (ODRL)
Sensitive dataset for internal use only. Not visible in the public catalog. Access only through direct communication with MSTECH and a pre-established contract.
Internal process diagnostics, root-cause analysis, quality monitoring and generation of derived data products (Diagnostics.Restricted, Patterns.Aggregated) within the consortium.
External distribution. Re-export without anonymization. Any use outside the consortium without explicit consent from MSTECH. external_distribution_allowed: false
Privacy level: raw_sensitive· Retention: 1825 days · Visibility: private_internal
Access Interface
Exclusive cross-platform access through an EDC contract. Automatic ODRL negotiation.
EDC_MSTECH → EDC_BACKX
Download of Parquet/JSON files with daily history for internal offline analytics.
v1 — batch
Acceso en streaming casi en tiempo real. Planned for v2.
v2 — coming soon
BACKX.PCBFailureDiagnostics.Restricted.v1
Restricted PCB failure diagnostics dataset with anonymized traceability, process measurements and root-cause analysis
FAIR description
PCB failure diagnostics dataset at unit, batch or inspection level with hashed sensitive identifiers (pseudonymized). Includes PCB, BOM, component, material context, detailed printing, placement and reflow measurements, recipe parameters, inspection/test results and failure analysis with estimated economic impact. Suitable for predictive quality models, rework prioritization and process-condition assessment.
Industrial Context
Data Product Structure · Main groups
Identification + Traceability
record_idRST-XXXXXXXX · Unique diagnostic IDtraceability.production_batch_id_hashAnonymized batchtraceability.machine_typeprinter | placer | reflow_oven…traceability.production_timestampDateTime UTCPCB + BOM
pcb_context.board_typerigid | flex | rigid-flexpcb_context.surface_finishHASL | ENIG | OSP | ImmersionTinbom_context.total_component_countintegerbom_context.high_risk_component_countintegerProcess measurements
process_measurements.printing.paste_deposit_volume_pctnumberprocess_measurements.placement.x_offset_ummicronsprocess_measurements.reflow.peak_temp_c°Cprocess_measurements.reflow.time_above_liquidus_ssecondsprocess_measurements.reflow.nitrogen_usedbooleanFailure analysis
failure_analysis.defect_codeMISSING_COMPONENT | OPEN_JOINT | COLD_JOINT…failure_analysis.defect_severitycritical | high | medium | lowfailure_analysis.root_cause_confidence0.0 – 1.0failure_analysis.estimated_cost_impact_eurnumberfailure_analysis.corrective_actiondescriptive stringUsage Policy (ODRL)
Visible in the catalog for discoviewy. Data access and transfer are subject to prior request and must be explicitly accepted by MSTECH as Data Provider.
Defect analysis, predictive quality models, rework prioritization, economic impact estimation and process-condition assessment. Only Level 2+ participants with an active EDC contract.
Redistribution to third parties outside the consortium. Re-identification of hashes. Use without an active EDC contract. contract_required: true · redistribution_allowed: false
Privacy level: pseudonymized· Visibility: restricted · Mandatory Eclipse EDC contract
Access Interface
ODRL contract negotiation through DSP. Requires trust level 2+.
Contrato EDC activo
Parquet/JSON files with diagnostic records for the agreed period.
v1 — batch
Near-real-time diagnostic alerts. Planned for v2.
v2 — coming soon
BACKX.PCBFailurePatterns.Aggregated.v1
Aggregated and anonymized dataset of PCB failure patterns by design, process, component, defect and production period
FAIR description
Patterns aggregateds y anonymized de PCB failures calculados sobre ventanas temporales (semanal / mensual) por agrupaciones como patrón de defect, design-process, paquete-process o familia de component-process. Sintetiza conditions technicals de PCB, component, materiales y process junto con métricas de quality: tasa de failure, yield, rework, scrap, tamaño muestral, multiplicador de riesgo y recomendaciones accionables. Útil para benchmarking y prevención de defects sin exponer unit-level data sensibles.
Industrial Context
Data Product Structure · Main groups
Identification + Periodo
record_idAGG-XXXXXXXX · Unique aggregated pattern IDaggregation_leveldefect_pattern | design-process | package-process…production_period.period_typeweekly | monthlyproduction_period.period_startdate homeproduction_period.period_enddate finPCB + Component + Process
pcb_context.board_typerigid | flex | rigid-flexcomponent_context.package_typeTQFP-64 | BGA | QFN…process_context.process_stageprinting | pick_and_place | reflow…process_context.peak_temp_cnumbermaterial_context.solder_alloySnPb | SAC305 | BiSn…Defect
defect_context.defect_categorysoldering | placement | material…defect_context.defect_codeCOLD_JOINT | OPEN_JOINT | MISSING_COMPONENT…defect_context.defect_severitycritical | high | medium | lowdefect_context.test_stageAOI | ICT | FCT…Metrics + Recommendation
quality_metrics.failure_rate_pctnumberquality_metrics.risk_multipliervs. reference baselinequality_metrics.units_processedinteger · min. 50recommendation.risk_score0–100recommendation.avoid_combination_flagbooleanUsage Policy (ODRL)
Quality benchmarking, risk-combination detection, prevention of recurring defects and prioritization of process improvements. Compatible with Level 1+ participants with an active contract.
Commercial use allowed only under explicit license. Minimum threshold of 50 units per group guarantees differential privacy. license_type: commercial_use_only
Attempted re-identification of manufacturers or lines from aggregates. Unauthorized redistribution. redistribution_allowed: false
Privacy level: aggregated_anonymized· Visibility: public_catalog_restricted_access· Minimum threshold: 50 units
Access Interface
Contract negotiation via DSP. Compatible with Level 1+ participants.
Contrato EDC
CSV/Parquet/JSON files with patterns for the weekly or monthly period.
v1 — batch semanal
Queries filtered by period, defect type, process or component family.
API BACK-X · coming soon
PCB SMT Defect Risk Prediction & Scenario Optimization Service
AI inference service to estimate PCB/SMT defect risk, compare scenarios and prioritize operational inspection
FAIR description
AI inference service to estimate defect risk in PCB/SMT components, run single or batch predictions, compare before/after process scenarios and return a risk ranking, explanatory drivers, recommended actions and enriched similar historical patterns.
Industrial Context
Service Structure
API operations
POST /predictSingle prediction · JSON in/outPOST /predict/batchBatch prediction · JSON or CSVscenarioComparisonCompare variants against the baseline scenario through batch submissionMain inputs
package_typeComponent package typepitch_mmPin / pad pitch in millimeterstime_above_liquidus_sTime above liquidus during reflowpeak_temp_cProcess peak temperaturepaste_deposit_volume_pctRelative solder-paste deposit volumex_offset_um / y_offset_umComponent placement offsetInference outputs
risk_scoreNormalized risk scorerisk_levelLow, medium, high or critical levelfailure_probabilityEstimated failure probabilitymost_likely_defectMost likely defectrecommended_actionsRecommended operational actionssimilar_historical_patternsSimilar anonymized patternsEDC
asset_idpcb-smt-defect-risk-ai-service-v1asset_typeHttpDatatransfer_typeHttpData-PULLfrontend_edr_exposureforbiddenUsage Policy (ODRL)
Inference through the Data Plane Public API and EDR access after contract negotiation. Use is allowed to estimate risk, prioritize inspections, compare scenarios and generate operational rankings.
The EDR must not be exposed in the frontend, and results or historical patterns must not be redistributed outside the contract. frontend_edr_exposure: forbidden
Historical patterns must remain anonymized and the plant region must be anonymized before returning enriched results.
Data protection: personal_data: false · production_sensitive_data: true · historical_patterns_anonymized: true
Access Interface
REST endpoint to evaluate one board, component or lot with JSON input.
POST /predict
JSON or CSV input to generate risk ranking and compare multiple process variants.
POST /predict/batch
HttpData asset published as an invocable service through the Data Plane and EDC contract.
contractdef-pcb-smt-defect-risk-ai-service-v1
More data products on the way
Additional derived data products, AI services and operational KPIs will continue to be added. External electronics manufacturers (OEMs) will be able to access public views of BACKX.PCBFailurePatterns.Aggregated.v1 under contract.
Join to publish data