TIDUFG2 December   2025

 

  1.   1
  2.   Description
  3.   Resources
  4.   Features
  5.   Applications
  6.   6
  7. 1System Description
    1. 1.1 Terminology
    2. 1.2 Key System Specifications
  8. 2System Overview
    1. 2.1 Block Diagram
    2. 2.2 Design Considerations
      1. 2.2.1 Sensor Selection
    3. 2.3 Highlighted Products
      1. 2.3.1 TLV387
      2. 2.3.2 TLV9054
      3. 2.3.3 MSPM0G5187-LP
      4. 2.3.4 LOG300
      5. 2.3.5 UCC28881
      6. 2.3.6 TPS709
  9. 3System Design Theory
    1. 3.1 Current Sensor
    2. 3.2 Hybrid Integrator
    3. 3.3 Band-Pass Filter
      1. 3.3.1 Log Amplifier
      2. 3.3.2 Current Low-Pass Filter
      3. 3.3.3 Non-isolated Voltage Sensing
      4. 3.3.4 Auto Labeling Circuit
        1. 3.3.4.1 Line Voltage Sensing
        2. 3.3.4.2 Arc Gap Voltage Sensing
        3. 3.3.4.3 Differential to Single-Ended Conversion
      5. 3.3.5 Power Supply
  10. 4Hardware, Software, Testing Requirements, and Test Results
    1. 4.1 Hardware Requirements
    2. 4.2 Software
    3. 4.3 Test Setup
      1. 4.3.1 Arc Testing Setup
    4. 4.4 Test Results
  11. 5Design and Documentation Support
    1. 5.1 Design Files
      1. 5.1.1 Schematics
      2. 5.1.2 BOM
    2. 5.2 Tools and Software
    3. 5.3 Documentation Support
    4. 5.4 Support Resources
    5. 5.5 Trademarks
  12. 6About the Author

3.3.4 Auto Labeling Circuit

The arc-labeling circuit consists of two isolated amplifiers, and an isolated comparator. The arc-labeling circuit gathers and labels arc signatures in a controlled lab environment. The design uses the fact that in a lab environment, the AC line voltage and the arc gap voltage are available, unlike in the field. Figure 4-2 shows the typical arc test setup in the lab. An arc generator generates reproducible arcs at different operating conditions. The relation between the arc gap voltage, measured across the arc generator and line voltage, serves as an indicator for an arc. This information then labels the data sampled by standard arc detection signal chain. The labeled data then trains an embedded AI model. This portion of the board works with a compatible TI LaunchPad development kit and draws power from the 3V3 LaunchPad output. The MSPM0 SDK includes a software example to transfer data from the LaunchPad to the connected computer.