Skip to the content
2 layers35 × 47 mmRythm badge tudelft, the printed circuit board in 3D

Hardware design files for the SPC lab MIDGE wearable devices for in-the-wild social data collection.

Size
35 × 47 mm
Copper layers
2
Thickness
1.6 mm
Parts
42
Nets
62
Pads
226
Tracks, vias
276, 18
Pours
5
Routed
Yes
KiCad
KiCad 6 (file 20211014)
Created
28 Jun 2019
Updated
15 Oct 2026

Overview

Rythm badge tudelft is an open-source printed circuit board by TUDelft-SPC-Lab, from the TUDelft-SPC-Lab/spcl_midge_hardware repository on GitHub, published under the MIT licence. It is a 2-layer board of 35 × 47 mm with 42 parts on 62 nets, routed with 276 tracks and 18 vias.

It is built around the BMD-300-A-R, ICM-20948 and NTS0104BQ, with the TP4057 and 10118194-0001LF among its other parts. Its parts include 18 capacitors, 7 ICs, 5 resistors, 5 connectors and 3 diodes. It belongs with the lEDs and displays boards of this library.

Main components

5

README

README.md

SPCL Midge badge

Welcome to the Hardware repository of the Social Perceptive Lab for the MIDGE badge.

Test

The Midge badge is a 55x35mm wearable PCB for in-the-wild social data collection. It features

  • 2 microphones, with adjustable low and high cutoff frequencies
  • 9 axis IMU
  • SD card storage
  • 600mAh battery
  • BMD 340 processor with a BLE (bluetooth low energy) antenna

Learn more about the project itself here.

Learn more about the firmware and software here.

Versions

  • V1: first production ready version (2022)
  • V2: hardware upgrade (2026)
    • Bigger battery 500mAh -> 600mAh
    • Added battery port (battery can be plugged in instead of soldered in)
    • Newer processor (BMD 300 -> BMD 340)
    • Newer IMU (ICM-20948 -> LSM6DSOXTR)
    • Newer Magnetometer (ICM-20948 -> AK09918C)
    • Added an extension port

Motivation

Why are conference-badge-type wearable sensors important for us?

Wearable badges are a cheap, convenient way to record high-fidelity signals from individuals in social interaction. We have the need for such a device in the SPC lab because we study the social dynamics of small crowds (~25-100 people) interacting during an event. Conference badges allow us to equip every participant to record individual audio, body movement signals and proximity information, all at a relatively low cost per device.

Why not use cell phones?

Cell phones can be useful in sensing crowds but differences in the sensors, fidelities and recording options / software available to different cell phone models are an issue for analysis. If cell phones are to be provided by the scientist, then cost becomes an issue in comparison with wearable badges.

Why not use a wrist wearable?

Wrist wearables can be effective at capturing hand movements. While we could have designed our device with a wrist form factor, we think the badge form is more convenient to use and its position on the chest allows it to more reliably capture proximity and overall body movement through acceleration.

Why did we decide to modify the Rhythm Badge?

The open-source Rhythm Badge was designed with office environments in mind, where the goal was the collection of longitudinal data about a team over days or weeks. Proximity information and a low-frequency microphone allowed it to reconstruct the social network of team members. On the other hand our goal was to computationally analyse interactions both at the group and individual level from social signals (speech, movement, proximity) of subjects. This required higher fidelity sensors and a new device design.

How did we modify the Rhythm Badge?
  • Enabled full audio recording, with a frequency up to 48KHz with an on-board switch to allow physical selection between high and low frequency.
  • Added a 9-axis IMU to record pose. IMUs combine three tri-axial sensors: an accelerometer, a gyroscope, and a magnetometer. These measure acceleration, orientation, and angular rates respectively. The sensor information is combined on-chip by a Digital Motion Processor.
  • Added an on-board SD card to directly store raw data, avoiding typical issues related to packet loss during wireless data transfer.

What makes the project stand out?

This project is unique because it improves upon the state-of-the-art wearable sensors measuring face-to-face human social interactions. The improvement points are catered towards more fine-grained and flexible capturing of data. For instance, the choice between high and low frequency audio capturing mode is provided at data acquisition and the experimenters can freely choose the frequency best suitable for them (for example, whether they have high-frequency audio for transcription or privacy-preserving low-frequency audio for detecting speech activities). Additionally, the inclusion of the on-board digital motion processor directly provides orientation estimation with the newly included full 9-DOF inertial motion unit (IMU). This allows future applications (e.g., real-time F-formation detection) to be built directly on top of the outputs of the DMP. The above design considerations and implementation exemplify the uniqueness and new advantages that Midges offer.

Status of the project

Project outcomes

The design of the Midge is complete and 100 devices have been created. The device was successfully used for the ConfLab dataset. There it was used to record audio (1250Hz), IMU data (56Hz) and proximity data (1Hz) for 48 data subjects in the same space in a computer science conference. It has also been used to make smart drinking glasses and in several master and bachelor thesis.

Parties involved in the development

The development of the device was contracted to Ioannis Protonotarios for the Socially Perceptive Computing Lab at Delft University of Technology. Ioannis took as base the latest release of the MIT Media Lab's Rhythm badge and implemented both hardware and software changes and updates. Testing of the device was done in conjunction with Socially Perceptive Computing Lab members. The whole process took 3 years and two iterations of the device's hardware. Additional hardware, firmware and software improvements have been developed by the SIPLab-EdgeAI at the Costa Rica Institute of Technology and the Embedded Systems Group at Delft University of Technology.

Would you like to contribute to the project?

Absolutely. Some of the contributions that we currently find most important in our application setting (sensing of small crowds) are: Improvements to audio sensing. Although our Midge was an improvement over the Rhythm badge in quality of audio sensing, audio quality is still an issue. Especially, the front-facing microphones of the device introduce significant cross-contamination. Potential solutions include exploring beam-forming, directional microphones or addition of headset support. Software improvements. There is significant room for improving the software of the device. The hub, which serves as monitoring and control centre for all connected Midges is especially relevant. The current command-line interface would benefit from being replaced by a user-friendly web-based interface that displays the status of connected midges and allows for quick bulk actions like eg. disabling sensors, starting/stopping recordings, etc.

Additional hardware details

  1. We've tested the midge with Gigastone 8GB, microSDHC, class 10, V10 sdcards
  2. We've used standard 3.7V, 500 mAh Lipo batteries

Issues on the PCB

Currently we are not aware of specific problems with the PCB, all functions work correctly (battery charging, cut-off switches, SD card, audio recording, IMU recording, status LEDs). If you do find a problem, please open an issue in the github repository.

Parts

42
QtyValueFootprintReferences
7100nFCapacitor_SMD:C_0603_1608MetricC1, C9, C10, C14, C15, C16, C18
71uFCapacitor_SMD:C_0603_1608MetricC6, C7, C8, C11, C12, C13, C17
310uFCapacitor_SMD:C_0603_1608MetricC2, C3, C5
3MP34DT05-Acustom_libs:MP34DT06U2, U3, U7
210MResistor_SMD:R_0603_1608MetricR1, R2
22kResistor_SMD:R_0603_1608MetricR5, R6
110118194-0001LFdigikey-footprints:USB_Micro_B_Female_10118194-0001LFJ1
11k?Resistor_SMD:R_0603_1608MetricR4
1503182-1852503182-1852:MOLEX_503182-1852J3
1?Capacitor_SMD:C_0603_1608MetricC4
1AP7345DPackage_DFN_QFN:X2-DFN1612-8U4
1BAT+Connector_Wire:SolderWirePad_1x01_Drill1.2mmJ4
1BAT-Connector_Wire:SolderWirePad_1x01_Drill1.2mmJ5
1BMD-300-A-Rdigikey-footprints:Bluetooth_Module_BMD-300MOD1
1CHRG_LEDLED_SMD:LED_0603_1608MetricD1
1ICM-20948ICM-20948:QFN40P300X300X105-24NU5
1LED_SmallLED_SMD:LED_0603_1608MetricD3
1LOGOrythm_badge_tudelft:flyG***
1NTS0104QFN50PX250X300X100-15N:QFN50PX250X300X100-15NU6
1OFF-LOW-HIButton_Switch_SMD:SW_SP3T_PCM13SW2
1OFF-ONButton_Switch_SMD:SW_SPDT_PCM12SW1
1STDBY_LEDLED_SMD:LED_0603_1608MetricD2
1SWDConnector_PinHeader_1.27mm:PinHeader_2x05_P1.27mm_VerticalJ2
1TP4057Package_TO_SOT_SMD:SOT-23-6U1

Similar boards

  • 2 layers108 × 95.3 mmBlue Wizard, the printed circuit board in 3D

    Blue Wizard

    wizarddata/Blue-Wizard-Mod

    Bluetooth Controller for the Kinesis Advantage line of keyboards using the BlueMicro firmware

    47 parts38
  • 2 layers126 × 100.4 mmDragon joke, the printed circuit board in 3D

    Dragon joke

    DC801/DC26PartyBadge

    Design and software for the DC26 DC801 badge

    91 parts8
  • 4 layers56 × 38 mmCanary, the printed circuit board in 3D

    Canary

    jetkov/canary-pcb

    98 parts
  • 4 layers29 × 31.5 mmCelsian dev pcb, the printed circuit board in 3D

    Celsian dev pcb

    malcircuit/Celsian-PCB

    PCB layout for the Celsian BLE environment sensor

    39 parts
  • 4 layers34 × 28 mmDevice, the printed circuit board in 3D

    Device

    jayis1/hardware-design-lab

    Full-cycle hardware device designs — each device in its own subfolder with architecture, schematics, PCB layout, and software stack

    30 parts
  • 10 layers180 × 100 mmJetson agx thor baseboard, the printed circuit board in 3D

    Jetson agx thor baseboard

    antmicro/jetson-agx-thor-baseboard

    Open hardware baseboard for Nvidia Jetson Thor AGX T5000 System on Module

    1,213 parts48Apache-2.0
  • 6 layers107 × 68 mmCm4-baseboard, the printed circuit board in 3D

    Cm4-baseboard

    antmicro/cm4-baseboard

    Experimental baseboard supporting System on Modules with Raspberry Pi CM4/CM5 pinout

    516 parts36Apache-2.0

Questions

How big is the Rythm badge tudelft PCB?

It measures 35 × 47 mm and has 2 copper layers, 1.6 mm thick.

What parts does the Rythm badge tudelft use?

42 parts, built around the BMD-300-A-R, ICM-20948, NTS0104BQ and TP4057. The full list of its parts is on this page, with links to each part's symbol and footprint where PCBKit's library has them.

Where can I download the Rythm badge tudelft KiCad files?

From the TUDelft-SPC-Lab/spcl_midge_hardware repository on GitHub, at the commit this page was made from: the board is Hardware_Design_v1/PCB/KiCad/rythm_badge_tudelft.kicad_pcb. PCBKit opens the board in the browser, with nothing to install.

Can I use the Rythm badge tudelft design in my own project?

Its repository publishes it under the MIT licence: you may reuse it on that licence's terms (keep the author's notice, and for a copyleft licence share your changes alike).