Real-time ADC Reader for xWRL6432 mmWave Radar via DCA1000

This Python package provides a standalone real-time ADC reader for the xWRL6432 mmWave radar sensor using the DCA1000 capture card — no installation of TI’s mmWave Studio required. It is especially useful for:

• Linux users who cannot run mmWave Studio
• Developers building custom real-time signal-processing pipelines

Note: Texas Instruments officially supports only recording/playback/visualization via mmWave Studio. This module enables true real-time frame assembly and data access.

Don't have a DCA1000 capture card and don't need high data throughput? Check out this repo which uses SPI and reads data after the first FFT (1D radar cube).

Features

• Radar & DCA1000 Configuration
  Send CLI commands to both the xWRL6432 radar EVM and the DCA1000 from Python.
• Start/Stop Acquisition
  Simple API to begin and end streaming over the DCA1000 UDP interface.
• Frame Assembly
  • Reorders out-of-order UDP packets
  • Splits packets that cross frame boundaries
  • Detects and discards incomplete frames
• Queue-based Output
  Push each assembled frame into a queue.Queue for downstream processing threads.
• Batch Recorder
  Record a fixed number of frames and either retrieve them in memory or save to a .npz file for offline analysis (e.g. in Jupyter Notebooks, example included).
• OpenRadar DSP Compatibility
  Works seamlessly with the OpenRadar DSP utilities for subsequent processing steps (e.g. range-doppler)

Installation

1. Clone this repository and enter its directory:

   git clone https://github.com/loeens/xWRL6432-adc-reader.git
   cd xWRL6432-adc-reader

2. Install the package and its dependencies:

   pip install .

Quick Start

Radar EVM and DCA1000

For setting the switches on the EVM and on the DCA1000, please refer to TI's DCA1000 mmWave Studio User Guide. I have attached an image of my setup here, copy it at your own risk.

• 5V/3A PSU to the DCA1000 via DC barrel jack (realistically draws around 500 mA in my experience)
• Micro USB link between the radar EVM and PC for UART (mmWave CLI) communication and power
• Ethernet cable between PC and DCA1000

• Make sure the IP of your computer is 192.168.33.30 (subnet mask 255.255.255.0 and no gateway). Note: In Ubuntu you can check the box "Use this connection only for resources on its network" in order to be able to still access the internet via wifi.
• Follow TI's instructions to flash the Motion and Presence Detection Demo firmware to the EVM.

Code

Before running your code, make sure the radar EVM and the DCA1000 are powered on and you meet the prerequisites described in the previous section.

from queue import Queue
from xwrl6432_adc_reader import XWRL6432AdcReader

# 1) Create a thread-safe queue
data_queue = Queue()

# 2) Instantiate the reader
adc_reader = XWRL6432AdcReader(
    radar_serial_port="/dev/ttyACM1",
    radar_cfg_path="radar_config/iwrl6432.cfg",
    out_queue=data_queue
)

# 3) Start streaming in the background
adc_reader.start_acquisition()

# 4) In your processing loop:
try:
    while True:
        frame = data_queue.get()
        # → process your ADC frame here
except KeyboardInterrupt:
    pass

# 5) Clean up
adc_reader.stop_acquisition()
adc_reader.close()

A complete example is available in the examples/ directory.

Configuration File

The reader expects a TI-style radar configuration file (e.g. iwrl6432.cfg) that you would normally use with mmWave Studio. Place it in a known path and pass it to the constructor, or alternatively use one of the default ones in the radar_config directory of this repository.

Acknowledgements

A big thank you to:

• OpenRadar: I copied and adapted the DCA1000 interface code from their adc.py module. Unfortunately for me there was an issue with their reader function. I fixed it in this repo and created a pull request in the original repo.
• Texas Instruments: For the specification documents and detailed packet-format documentation.