FPVSoundLogger

A DIY RP2040-Zero based onboard audio recorder for FPV drones. It uses a MEMS mic and SD card module to capture sound, automatically triggered by Betaflight arming via PinioBox. Recordings match DJI goggles video length for seamless editing. Includes hardware design, MicroPython code, and examples.

Enjoy this project or find it useful? Feel free to buy me a coffee ☕ to support its development.

Project Background

I first started experimenting with the excellent and promising OpenIPC project, driven by my love for tinkering, hacking, and coding. The idea of building an open FPV video system was incredibly appealing. But after many tests, the freeze issues and limited range eventually convinced me to switch to DJI for reliable everyday flying. (But I promise — I’ll come back to OpenIPC!)

The DJI O4 Air Unit provides fantastic onboard video recordings. But (because there's always a “but”), it still doesn’t record audio. And audio truly adds something special — whether for aggressive FPV flying or even some cinematic flights.

So I wanted a solution that would be simple/lightweight/inexpensive and capable of recording high‑quality audio and perfectly* synchronized with DJI recordings. This project is the result of that idea.

Démo

Features

• Autonomous module
• Work with Betaflight/INAV
• Enable/disable recording
• Auto record using arm/disarm trigger
• Audio/video share the same duration *

Prerequisites

• A flight controller running Betaflight/INAV with at least two available Pinio outputs
• A transmitter (TX) capable of assigning AUX channels (EdgeTX, OpenTX, etc.)
• A stable 3.3V output on the flight controller and a common ground
• A free switch or similar to toggle recording modes
• Basic soldering tools for connecting power, ground, and signal wires and plenty of flux — <u>flux is life</u>

How It Works

• Power up the drone — In Disable mode module's LED Red blinking 🔴
• Use one of your TX switches to Enable recording mode — the LED now blinks Green 🟢
• Arm: the module immediately starts writing audio data to the SD card, and the LED now blinks Blue 🔵
• Once disarmed — whether after a normal end of flight or a dramatic crash — the module stops recording and finalizes the audio file with the proper header
• Remove the SD card and retrieve the WAV file that matches the exact flight duration
• Import the audio into your favorite video editor and drop it on the timeline — it syncs effortlessly with the DJI footage !

Duration consideration

⚠️ Due to DJI's camera record initialization delay after arming, a typical 0.5-2 seconds seconds delta can be observed and you may need to offset the audio track in your editor. ⚠️

Format: rec_XXXX_MM-SS.wav

• 🔵 Index (XXXX): Auto-incremented counter (0001, 0002, 0003...)
• 🟡 Duration (MM-SS): Recording length in minutes and seconds

Architecture / Hardware Overview

Components

| | INMP441 | RP2040-Zero | SD Card SPI Module | | - | :-------: | :-----------: | :-------------: | | |<img src="resources/images/INMP441.png" height="150"> | <img src="resources/images/rp2040-Zero.png" height="150"> | <img src="resources/images/SDCardModule.png" height="150"> | | Description | The INMP441 is a compact digital MEMS microphone that outputs audio through an I²S interface.It delivers low‑noise, high‑quality audio capture with a built‑in amplifier and ADC | The Raspberry Pi Pico uses the RP2040 dual‑core ARM Cortex‑M0+ microcontroller running at up to 133 MHz with 264 KB RAM and 2 MB flash.It offers rich connectivity like I²C, SPI, UART, PWM, ADC, and powerful PIO blocks for custom protocols.Compact, low‑cost, and easy to program in MicroPython or C/C++.| Micro SD TF Card Memory Shield Module provides an easy way to add microSD storage to microcontrollers.It uses an onboard card socket and level shifting for safe 3.3 V–5 V operation.Ideal for large dataqqsets via SPI. | | Size | 12x14mm | 18x24.5mm | 18x21mm | | Weight | 0.36gr | 1.66gr | 1.46gr | | AliExpress | INMP441 | RP20240-Zero | SD TF Module |

Wiring

Alternatives

An alternative is to use the FC’s 5 V rail to power the Pi Pico. The SD Card SPI module can convert data signals to 3.3 V levels, allowing it to use the Pi Pico’s 5 V rail. However, according to the INMP441 datasheet, the maximum supply voltage must not exceed 3.63 V, so you must power it from the Pi Pico’s 3.3 V rail.

Build

In the interest of keeping things lightweight and considering the very low current required for the system to operate, 28‑AWG silicone cable is more than sufficient, and you shouldn’t add more than 1 g to the whole setup.I cut and glued a small foam circle — the kind you find at the bottom of FC/ESC boxes, as dense as possible — to make a windscreen/pop filter.

Be careful not to get any glue into the microphone’s input opening.To keep everything in place, I used a ridiculously heavy piece of heat‑shrink tubing to “thermal‑zipper” the whole thing, along with a bit of hot glue (B7000 is obviously the best option, though hard to find) to secure the small 4‑pin JST connector against the assembly.

In my rough version installed in my Jeno 5", the module weighs about 10 g, but it’s definitely possible to do much better.

A voltage spike can cause parasitic artifacts in the recording. I haven’t tested this yet, but as mentioned in the INMP441 datasheet it is recommended to use a small 0.1 µF ceramic capacitor between VDD and GND.

Examples

<img src="resources/images/build_example.png" width="829"> <img src="resources/images/build_example_2.png" width="829">

Installation & Configuration

Pi Pico: Flashing MicroPython on an RP2040 with Thonny

1. Install Thonny

• Download and install Thonny from the official website.
• Open Thonny.

2. Put the RP2040 board into bootloader mode

• Plug in the board while holding the BOOTSEL button.
• Release the button once connected: the board should appear as a USB drive named RPI-RP2.

3. Install MicroPython from Thonny

• In Thonny, go to:
  Tools → Options… → Interpreter
• Select:
  MicroPython (Raspberry Pi Pico / RP2040)
• Click Install or update MicroPython.
• Select your RPI-RP2 device.
• Start the installation.

<img src="resources/images/MicroPythonInstall.png" width="829">

4. Automatic restart

The board will reboot with MicroPython installed.
Thonny should now display a >>> prompt.

5. Quick test

Type:

print(f"WE ARE {'fpv'.upper()}")

<img src="resources/images/MicroPythonInstalled.png">

6. Project files to upload

├── main.py                    # Main application (auto-runs on boot)
   └── lib/
       ├── sdcard.py              # SD card driver
       └── fpv_sound_logger.py    # Core logger class

• In Thonny, go to View > Files
• In your computer's file panel (left side), navigate to your project folder
• Select the lib folder
• Right-click → "Upload to /"
• Then select main.py
• Right-click → "Upload to /"
• Thonny will upload the entire structure automatically

Note: This method preserves the folder structure. The lib folder and its contents will be uploaded correctly.

<p align="left"> <img src="resources/images/lib_upload_to.png"> <img src="resources/images/lib_main_uploaded.png"> </p>

TX && Betaflight configuration

All you need to know about Betaflight's Pinio && PinioBox can be found here, I also got some excellent advice from Betaflight’s official discord

• You'll need to configure AUX channel on your TX to Enable/Disable recording

  I configured it to use the center position of an three-position switch mapped to AUX6 on my Radiomaster Pocket running on EdgeTX.

• Provide a common ground from the flight controller to the module, along with a clean 3.3V or 5V supply

  On my JHEMCU GF20F722, I used the 3.3V rail that was originally intended for a DSM receiver.

• Configure your flight controller’s Pinio outputs using CLI to deliver the required logic levels:

  • One Pinio controls record enable/disable

  • The second Pinio follows the arm/disarm sequence to automatically start and stop recording.

  [...]

  resource pinio 1 B10 # remap RX3 pad

  resource pinio 2 B11 # remap TX3 pad