I'm developing a new approach to circuit analysis using AI, and I've created an open source tool that might interest the EasyEDA community.
Current PCB analysis falls into two categories:
- Design Rule Checking (like EasyEDA Pro's DRC)
- Manufacturing checks (like JLCPCB's DFM)
These are essential tools, but I'm targeting a different need: circuit intent analysis.
The goal: Describe what a circuit should do, and have AI assess if the design matches that intent. Think questions like:
- "What is this circuit designed to do?"
- "Does this Arduino shield have adequate isolation?"
- "Is this level shifter suitable to connect a 3.3V I2C device to a 5V Arduino?"
- "Are the decoupling capacitors properly placed for this FPGA?"
Current AI approaches require uploading schematics, Gerbers, or images—formats that are bulky and not optimized for semantic analysis.
My solution extracts three key files from EasyEDA:
- BOM (Bill of Materials) - What components are used
- Pick & Place - Where they're located
- Netlist - How they're connected
The tool easyeda_parser.py converts these three files into clean YAML — a lightweight, structured format perfect for AI analysis.
This gives LLMs the complete functional picture without the bloat of graphical formats.
The parser is open source, and available at: https://github.com/koendv/easyeda_parser/
My opinion is that "Export for AI analysis" should be native EDA functionality, built into the EDA tools themselves.
At the moment, the tool is limited to the information available in BOM, Pick and Place, and Netlist. A future native export might add stackup, trace widths, and trace length to the YAML, sufficient to do transmission line analysis, and allow queries like "Are the QSPI flash traces length matched?"
I'm sharing this early to gather feedback. Would this approach be useful for your workflow? What kinds of intent analysis do you want to perform?"
#Others#
Sign In Or Register Comment after
No comments yet. Be the first to comment!