What this category of data looks like
Data sources for production ledger data in semiconductor wafer fabrication include MES systems, equipment FDC collection modules, defect detection and metrology equipment, wafer map files, process recipe libraries, and experimental data. Daily documents cover process flow cards, batch operation records, equipment recipes, SPC control charts, exception handling sheets, defect maps, yield reports, and more. Data updates follow three rhythms: single-batch wafer runs, real-time parameter collection, and daily summaries. Fields include nanoscale process parameters, batch IDs, equipment operating durations, defect position coordinates, and more, with subdivided units such as nanometers, seconds, and wafers.
Constraints on HTTP Interfaces and External Systems Posed by These Characteristics
The multi-source data collection feature requires interfaces to support batch transfer of multiple fields, and needs to adapt to the data formats of different systems such as MES and equipment FDC modules. The update rhythms of batch-based, real-time, and daily summaries require interfaces to support on-demand or scheduled invocation modes to avoid invalid requests. The nanoscale parameters and subdivided measurement units require fields transmitted by the interface to retain high-precision formats to prevent data distortion. The large amount of process data from ultra-long process chains requires the interface to have reasonable request body size limits and timeout thresholds to avoid transmission interruptions or timeouts.
Configuration Settings
| Configuration Item | Recommended Approach | Rationale |
|---|---|---|
Authorization | Bearer <API key assigned by FastGPT> | Complies with the authentication specifications for FastGPT external interfaces, and adapts to the interface invocation requirements of wafer fab MES systems |
variables | Map wafer batch process parameters and yield data to global variables, with no more than 800 records passed per single request | Matches the batch-based update characteristics of wafer fabrication ledgers, and avoids overloading interface requests |
responseData | Structured JSON containing batch ID, defect map index, and yield report link | Meets the core needs of external systems for parsing ledgers, and facilitates rapid data integration |
| Interface timeout duration | 600 seconds | Covers the time required for ledger summary of ultra-long process chains, and avoids request interruptions caused by complex processes |
| Request body size limit | No more than 1000 MB | Adapts to the transmission of ledger data including large files such as wafer maps and SPC charts |
Three Common Misconfigurations
- Phenomenon: When an external platform calls a published interface, global variables are not correctly included in the request body, and core ledger fields are missing from the returned results. Cause: The global variable mapping switch is not enabled in the interface configuration, and fields such as wafer batch process parameters and yield data are not bound as variables recognizable by the interface.
- Phenomenon: "Cannot read property"-type errors are prompted during interface testing, and returned content cannot be parsed normally. Cause: The
responseDataparsing rule is not configured according to the field structure of wafer fabrication ledgers, causing external systems to fail to identify special fields such as defect maps and yield report links. - Phenomenon: Interface calls frequently time out, and summary requests for large-volume ledgers cannot be completed. Cause: The timeout threshold is not adjusted according to the ultra-long process chain characteristics of wafer fabrication, and the default timeout duration is insufficient to cover the summary time of multi-process data.
How to Verify Proper Configuration
- Call the interface with simulated wafer batch test data, check whether the request body contains preset global variable fields, and verify that the
variablesparameter is passed correctly. - Check whether the
responseDatareturned by the interface contains core ledger fields such as batch ID, defect map index, and yield report link, and confirm that the data format meets the configuration requirements. - After configuring the
Authorizationauthentication parameter, call the interface with an invalid key, check whether a correct authentication failure prompt is returned, and verify that the authentication logic is effective. - Adjust the interface timeout duration to a reasonable interval adapted to the process chain, call a test request containing multi-process data, and confirm that the request is not interrupted. This verification process applies to FastGPT v4.17.0.
The configuration values provided above are common starting points, and should be evaluated against the reader's own specific samples.