Copper Mountain Vna

CI License Python 3.10+ MCP eng-mcp-suite

Drive Copper Mountain Vector Network Analyzers from any MCP-compatible AI client. TCP/IP SCPI over the wire — 45 tools spanning sweep, calibration, measurement, and export.

Quick start · Tools · Workflows · Documentation

What is copper-mountain-vna-mcp?

copper-mountain-vna-mcp is a Model Context Protocol server that exposes a Copper Mountain VNA — RVNA, TRVNA, S2VNA, or S4VNA family — to any MCP-capable AI assistant. The server speaks raw TCP/IP SCPI to the Copper Mountain software socket server (default port 5025), then surfaces sweep configuration, calibration, S-parameter capture, marker math, and Touchstone export as plain MCP tools.

Drive it from Claude Desktop, Claude Code, VS Code Copilot, or any other MCP client. SCPI is generated server-side, so the AI agent never needs to know the underlying command set — it just calls vna_configure_sweep, vna_trigger_sweep, vna_get_s_parameter and gets back structured data.

What copper-mountain-vna-mcp does well:

• 🤖 AI-native via MCP. First-class Model Context Protocol server with 45 tools. Any Claude / LLM agent can drive it.
• 🐍 Python + MCP surfaces. Import coppermountain_vna_mcp for direct driver access, or run as an MCP server for AI-agent automation.
• ⚡ Direct TCP/IP SCPI. No VISA dependency, no proprietary middleware — just asyncio sockets straight to the Copper Mountain software socket server.
• ✅ Validated against 4 VNA families. RVNA (S11), TRVNA (S11/S21), S2VNA (full 2-port), S4VNA (full 4×4).
• 🔒 AGPL-3.0-or-later. Network-copyleft open-source. Independent of Copper Mountain Technologies (trademark disclaimer below).

Disclaimer: This project is not affiliated with, endorsed by, or sponsored by Copper Mountain Technologies. "Copper Mountain" is a trademark of Copper Mountain Technologies. This is an independent open-source project.

Quick start

Install

pip install coppermountain-vna-mcp

Or from source:

git clone https://github.com/RFingAdam/copper-mountain-vna-mcp.git
cd copper-mountain-vna-mcp
pip install -e ".[dev]"

Supported instruments

Two surfaces, same answer

import asyncio
from coppermountain_vna_mcp.driver import CopperMountainVNADriver
from coppermountain_vna_mcp.models import SParameter, SweepConfig

async def main():
    async with CopperMountainVNADriver("127.0.0.1", 5025) as vna:
        await vna.configure_sweep(SweepConfig(
            start_freq_hz=1e9, stop_freq_hz=2e9,
            num_points=201, if_bandwidth_hz=1000, power_dbm=0,
        ))
        await vna.trigger_sweep()
        data = await vna.get_s_parameter_data(SParameter.S11)
        marker = await vna.marker_search_min()
        print(f"Best match: {marker.frequency_hz/1e9:.3f} GHz @ {marker.magnitude_db:.1f} dB")

asyncio.run(main())

{
  "mcpServers": {
    "copper-mountain-vna": {
      "command": "python",
      "args": ["-m", "coppermountain_vna_mcp"],
      "env": {
        "VNA_DEFAULT_HOST": "127.0.0.1",
        "VNA_DEFAULT_PORT": "5025"
      }
    }
  }
}

Start the VNA software socket server

The VNA software runs on Windows. Enable the socket server via command line:

:: S2VNA / S4VNA
S2VNA.exe /SocketServer:on /SocketPort:5025 /visible:on

:: RVNA / TRVNA
RVNA.exe EnableSocket:5025

Or in the GUI: System > Socket Server > Enable.

Configuration

Configure via environment variables or .env file:

Tools

45 MCP tools, grouped:

Full tool reference in docs/tools.md.

Calibration

Standard cal flows are exposed as discrete tools so an agent can walk them step by step.

SOL (1-port reflection):

1. vna_calibration_start(method="SOL")
2. vna_calibration_measure_standard(standard="SHORT", port_num=1)
3. vna_calibration_measure_standard(standard="OPEN", port_num=1)
4. vna_calibration_measure_standard(standard="LOAD", port_num=1)
5. vna_calibration_apply()
6. vna_correction_on()

SOLT (2-port full): SHORT/OPEN/LOAD on both ports, then THRU between them, then apply.

Calibration is frequency-specific — recalibrate when the sweep range changes substantially.

Measurement templates

Pre-configured sweep + marker + trace setups for common DUT classes. Load by name with vna_load_template, then vna_apply_template to push to the VNA.

await vna.load_template("antenna_wifi_5ghz")
await vna.apply_template()

# Custom filter — provide parameters at load time
await vna.load_template("filter_bandpass_custom",
                        center_freq_hz=1.5e9, bandwidth_hz=100e6)
await vna.apply_template()

Workflows

copper-mountain-vna-mcp fits in the following eng-mcp-suite workflow bundles:

• lab-automation — pair with mcp-rs-spectrum-analyzer, mcp-rs-siggen, and mcp-rs-cmw500 for end-to-end RF bench-test workflows driven from a single agent session.
• antenna-tuning — feed return-loss / VSWR traces into mcp-nec2-antenna or mcp-openems for iterative match optimization.

Install the bundle with:

eng-mcp-suite install --workflow lab-automation

Documentation

• 📘 Quick Start — install through first call.
• 🛠️ Tool reference — every MCP tool, every argument.
• 📐 Usage examples — a practical end-to-end S11 walkthrough.
• 🏗️ Architecture — how this MCP fits in eng-mcp-suite.
• 📝 Changelog

Part of eng-mcp-suite

_{This MCP server is part of}

eng-mcp-suite

_{Part of eng-mcp-suite — an open umbrella of MCP servers for RF / EMC / PCB / signal-integrity engineering. Drop into the lab-automation workflow bundle with eng-mcp-suite install --workflow lab-automation.}

Contributing

Contributions welcome.

1. Pick a GitHub issue.
2. Fork + branch (feature/your-thing or fix/your-bug).
3. Run the local check suite:

   ruff check . && mypy src/coppermountain_vna_mcp
   pytest --cov=coppermountain_vna_mcp
   

4. Open a PR — link the issue, request review.

License

AGPL-3.0-or-later. Copyright 2025-2026 Adam Engelbrecht (RFingAdam). Relicensed from Apache-2.0 in v0.3.0 to align with the eng-mcp-suite toolkit-wide AGPL move.

Acknowledgments

• Copper Mountain Technologies — for documenting the SCPI command set used by RVNA / TRVNA / S2VNA / S4VNA software.
• The MCP working group — for the Model Context Protocol specification.