Circuit Simulation MCP Server

A Model Context Protocol (MCP) server that provides circuit simulation capabilities using PySpice. This server allows you to create, simulate, and analyze electronic circuits through a simple tool-based interface.

✅ System Status

Fully Tested & Production Ready - PySpice 1.5 + ngspice 44 compatibility verified on ARM64/Apple Silicon macOS.

• 100% Test Coverage: All simulation functions validated and working

• Platform Verified: macOS ARM64 with Homebrew ngspice 44 installation

• Component Support: Resistors, capacitors, inductors, voltage/current sources, diodes, LEDs, transistors

• Analysis Types: DC, AC, and Transient analysis all fully functional

Related MCP server: ltspice-mcp

Features

• Circuit Creation: Define circuits with various components (resistors, capacitors, inductors, voltage/current sources, diodes, transistors)

• Multiple Analysis Types:

  • DC Analysis (steady-state)

  • AC Analysis (frequency response)

  • Transient Analysis (time-domain)

• Circuit Management: List circuits, get detailed information, export data

• Visualization: Generate plots of simulation results

• SPICE Integration: Uses PySpice for accurate circuit simulation

• Built-in Debugging: Comprehensive error reporting and circuit validation

Installation

Prerequisites

1. Python 3.8+

2. Ngspice: The underlying SPICE simulator

   • macOS (Recommended - Tested Configuration):

     # Install ngspice with ARM64 compatibility  
     /opt/homebrew/bin/brew install ngspice libngspice

   • Ubuntu/Debian: sudo apt-get install ngspice

   • Windows: Download from ngspice website

Install the Package

# Clone the repository
git clone <repository-url>
cd circuit-sim-mcp

# Install in development mode
pip install -e .

Environment Setup (macOS ARM64)

For optimal compatibility on Apple Silicon Macs, set up the library path:

export DYLD_LIBRARY_PATH="/opt/homebrew/Cellar/libngspice/44.2/lib:$DYLD_LIBRARY_PATH"

Add this to your shell profile (.zshrc, .bash_profile) for persistence.

Dependencies

The package will automatically install:

• mcp - Model Context Protocol Python SDK

• pydantic - Data validation

• numpy - Numerical computations

• matplotlib - Plotting (optional)

Usage

Running the MCP Server

The server can be run as a standalone MCP server:

python -m circuit_sim_mcp

Using with MCP Clients

Configure your MCP client to use this server:

{
  "mcpServers": {
    "circuit-sim": {
      "command": "python3",
      "args": ["-m", "circuit_sim_mcp"],
      "env": {
        "DYLD_LIBRARY_PATH": "/opt/homebrew/Cellar/libngspice/44.2/lib"
      }
    }
  }
}

Note: The DYLD_LIBRARY_PATH environment variable is crucial for ARM64/Apple Silicon compatibility with ngspice 44.

Available Tools

1. create_circuit - Create a new circuit with components and validation

2. validate_circuit - Check circuit for common issues and get recommendations

3. simulate_dc - Perform DC analysis (steady-state voltages)

4. simulate_ac - Perform AC analysis (frequency response)

5. simulate_transient - Perform transient analysis (time-domain behavior)

6. plot_results - Generate professional plots and visualizations

7. export_data - Export simulation data (JSON, CSV, TXT formats)

8. list_circuits - List all created circuits

9. get_circuit_info - Get detailed circuit information and netlists

10. create_example_circuit - Create guaranteed working example circuits for learning

11. debug_simulation - Debug simulation failures with detailed SPICE output

12. list_available_models - Show available component models and parameters

Example

Here's a simple voltage divider example:

import asyncio
from circuit_sim_mcp.server_basic import CircuitSimServer

async def main():
    sim_server = CircuitSimServer()
    server = sim_server.server

    # Create a voltage divider circuit
    components = [
        {
            "component_type": "voltage_source",
            "name": "V1",
            "voltage": 10.0,
            "nodes": ["vin", "gnd"],
            "source_type": "DC"
        },
        {
            "component_type": "resistor", 
            "name": "R1",
            "resistance": 1000.0,
            "nodes": ["vin", "vout"]
        },
        {
            "component_type": "resistor",
            "name": "R2", 
            "resistance": 1000.0,
            "nodes": ["vout", "gnd"]
        }
    ]
    
    # Create the circuit
    result = await server.call_tool("create_circuit", {
        "name": "voltage_divider", 
        "components": components
    })
    
    # Perform DC analysis
    dc_result = await server.call_tool("simulate_dc", {
        "circuit_name": "voltage_divider",
        "output_nodes": ["vin", "vout", "gnd"]
    })
    
    print(f"DC Analysis: {dc_result}")

if __name__ == "__main__":
    asyncio.run(main())

See examples/simple_voltage_divider.py for a complete working example.

Component Types

Supported Components

• Resistor: {"component_type": "resistor", "name": "R1", "resistance": 1000.0, "nodes": ["n1", "n2"]}

• Capacitor: {"component_type": "capacitor", "name": "C1", "capacitance": 1e-6, "nodes": ["n1", "n2"]}

• Inductor: {"component_type": "inductor", "name": "L1", "inductance": 1e-3, "nodes": ["n1", "n2"]}

• Voltage Source: {"component_type": "voltage_source", "name": "V1", "voltage": 5.0, "nodes": ["n1", "n2"], "source_type": "DC"}

• Current Source: {"component_type": "current_source", "name": "I1", "current": 1.0, "nodes": ["n1", "n2"], "source_type": "DC"}

• Diode: {"component_type": "diode", "name": "D1", "nodes": ["n1", "n2"], "model": "D"}

• LED: {"component_type": "diode", "name": "LED1", "nodes": ["anode", "cathode"], "model": "LED"}

• Transistor: {"component_type": "transistor", "name": "Q1", "transistor_type": "npn", "nodes": ["collector", "base", "emitter"], "model": "2N2222"}

Alternative Value Fields

Components also accept "value" as an alternative to specific field names:

• Resistor: "value": 1000 (same as "resistance": 1000)

• Capacitor: "value": 1e-6 (same as "capacitance": 1e-6)

• Sources: "value": 5.0 (same as "voltage": 5.0 or "current": 1.0)

Node Names

⚠️ Important: Node names cannot be Python keywords. Use descriptive names like:

• vin, vout (instead of in, out)

• vcc, vdd, gnd (instead of class, def, if)

Branch Structure

This repository has two main branches with different feature sets:

🌟 main branch (Current - Basic Version)

• Purpose: Core circuit simulation functionality

• Target: Users who need basic circuit analysis

• Features: Essential circuit creation, simulation (DC/AC/Transient), and data export

• Complexity: Simple, focused, easy to understand and extend

🚀 advanced-features branch

• Purpose: Full-featured version with AI-powered analysis

• Target: Professional users and complex circuit design

• Features: Everything in main + intelligent datasheet prompting, complexity analysis, datasheet-based components

• Complexity: Advanced features for sophisticated circuit analysis

To access advanced features:

git checkout advanced-features

Architecture

• server_basic.py: Basic MCP server implementation using FastMCP

• circuit.py: Circuit representation and component definitions

• simulator.py: PySpice integration and simulation engine

• __main__.py: Entry point for running the server

Development

Running Tests

python -m pytest tests/

Adding New Components

1. Add the component class to circuit.py

2. Update the Component.from_dict() method

3. Add netlist generation in Circuit._component_to_netlist()

Troubleshooting

PySpice Version Compatibility

Fixed in this version: PySpice 1.5 is patched to work with ngspice 44. If you encounter version compatibility issues:

1. Verify ngspice version: ngspice --version should show version 44.x

2. Check PySpice patching: Our system automatically handles PySpice 1.5 compatibility with ngspice 44

3. For fresh installations: Follow the macOS ARM64 setup above

Library Path Issues (macOS ARM64)

If you see library loading errors:

# Set the correct library path for ARM64 Homebrew ngspice
export DYLD_LIBRARY_PATH="/opt/homebrew/Cellar/libngspice/44.2/lib:$DYLD_LIBRARY_PATH"

PySpice Import Errors

If you get PySpice import errors:

1. Ensure ngspice is installed and in your PATH

2. Check architecture compatibility (x86_64 vs arm64)

3. For ARM64 Macs, use the Homebrew paths shown above

4. Reinstall dependencies if needed:

   pip uninstall numpy PySpice
   pip install numpy PySpice

Simulation Failures

If simulations fail:

1. Use the debug tools: Run debug_simulation tool to get detailed SPICE output

2. Check circuit validation: Use validate_circuit tool to identify common issues

3. Verify ground connections: Ensure your circuit has a 'gnd' or '0' node

4. Check component values: Extreme values can cause convergence issues

Advanced Features

For advanced features like intelligent datasheet prompting, circuit complexity analysis, and datasheet-based components, check out the advanced-features branch:

git checkout advanced-features

License

MIT License - see LICENSE file for details.

Contributing

1. Fork the repository

2. Create a feature branch

3. Make your changes

4. Add tests

5. Submit a pull request