What Data Looks Like for This Use Case
Data for semiconductor packaging and testing scenarios comes from wafer test maps, MES systems, bonding and molding equipment logs, ATE tester data, X-ray/SAM images, and customer specification requirements. Daily produced documents include packaging process cards, bonding and molding parameter sheets, wafer incoming material maps, test program version files, yield reports, and failure analysis reports. Data is updated per batch, per test piece, and daily. It covers the full process from bare die assembly to finished product testing.
Constraints Imposed on Workflow Orchestration by These Characteristics
Multi-source heterogeneous data sources require configuring data validation steps between workflow nodes to ensure uniform format for parameters and test results exported by different devices. High-frequency update business rhythm requires binding real-time triggering mechanisms to workflows, to avoid using outdated wafer incoming material maps or test program versions. Diverse document types require splitting workflow nodes by document category, to separately handle parameter configuration, test data, failure analysis, and other content. Severe consequences of errors require setting exception branches in workflows, to automatically trigger the exception report generation process when risks such as mixed batches or program mismatches occur.
Configuration Settings
| Configuration Item | Suggested Setting | Rationale |
|---|---|---|
| Trigger Node | Triggered by batch completion event | Matches the per-batch update business rhythm, avoids repeated workflow execution |
| Context Variables | Bind variables to wafer incoming material ID and test program version | Ensures the workflow calls exclusive data for the current batch, avoids cross-batch confusion |
| Node Execution Timeout | 600 seconds | Covers typical time consumption of ATE test data pulling and X-ray/SAM image analysis |
| Failure Branch Trigger | Triggered by execution failure of the node with nodeId as the test node | Accurately locates exceptions in the test link, automatically generates corresponding exception reports |
| Historical Context Count | 3 entries | Matches short process communication for a single batch, avoids redundant data occupying context |
| Maximum Input File Size | 1000 MB | Accommodates multiple documents such as wafer test maps and yield reports for a single batch |
Three Common Misconfigurations
- Workflow does not complete all nodes and shows no error: The workflow status displays as completed, but no exception report is generated. The cause is that branch validation for
nodeIdis not configured, skipping the test result verification step. - Workflow associates outdated historical records: The generated process documents reference old test program versions. The cause is that the
variablescontext for the current batch is not bound, and global historical data is called instead. - API calls to the workflow cannot upload files: The API request returns a parameter error. The cause is that the allowed formats for the file receiving node are not configured in the workflow, and the document types exported by packaging and testing equipment are not matched.
How to Verify Proper Configuration
- Manually trigger the workflow for the corresponding batch, check whether the wafer incoming material map data bound to the current batch is automatically pulled.
- Simulate execution failure of the test node, check whether an exception report in the corresponding format is automatically generated.
- Call the FastGPT v4.17.0 API to upload documents for a single batch, check whether the workflow normally receives and starts execution.
- View the workflow running logs, confirm that the
variablesparameter is bound to the test program version of the current batch, and there is no redundant historical data.
The values provided here are common starting points, and should be evaluated against the reader's own samples.