Wiznet makers

Building a High-Performance PoE Network Speaker with the W55RP20

This article shares the development process of a real-time audio streaming network speaker built using the W55RP20 System-in-Package (SiP), which combines the RP2040 microcontroller and the W5500 Ethernet controller into a single chip.

The W55RP20 is a cost-effective yet powerful networking solution, making it an excellent choice for PoE (Power over Ethernet)-based network audio devices. This project successfully validates both the hardware and software architecture required to build a PoE-powered network speaker using the W55RP20 and a high-performance audio amplifier.

1. Project Overview and Hardware Configuration

The W55RP20 provides a powerful networking platform at a very competitive cost. To facilitate rapid integration into real-world applications, a standard pin-compatible development board named W55RP20-4032 was designed and developed.

Hardware Specifications

• MCU: W55RP20 (RP2040 + W5500 + Flash Memory)
• Audio Amplifier: TAS5825P
  • Class-D architecture
  • Hybrid-Pro algorithm for high efficiency and low heat generation
• Interfaces:
  • I2S for audio data transmission
  • I2C for amplifier configuration and control
• Debug & Display:
  • Dual NeoPixel (SK6812) LEDs
  • SSD1306 OLED display for status monitoring

2. Development Environment: Arduino vs. Pico C/C++ SDK

Initial testing was performed using the Arduino environment with the lwIP networking stack due to its ease of development and rapid prototyping capabilities. However, because lwIP relies on a software TCP/IP stack, network throughput was limited to approximately 2 Mbps.

While this bandwidth was sufficient for CD-quality audio streaming (44.1 kHz, 16-bit), future expansion to high-resolution audio formats (96 kHz, 24-bit and above) required greater network performance.

To overcome this limitation, development was migrated to the Raspberry Pi Pico C/C++ SDK, enabling direct control of the W5500 hardware TCP/IP stack and significantly improving network efficiency.

Later, the Arduino environment was enhanced by integrating and optimizing the W55RP20_Ethernet3 library, allowing utilization of the hardware TCP/IP offload engine (TOE). As a result:

• Arduino + lwIP throughput: ~2 Mbps
• Arduino + W55RP20_Ethernet3 throughput: ~6 Mbps
• Pico SDK + W5500 Hardware Stack (UDP): ~17 Mbps

Related Articles

• Arduino lwIP Environment
  https://nexp.tistory.com/4295
• Arduino lwIP UDP Throughput Test
  https://nexp.tistory.com/4306
• Arduino W55RP20_Ethernet3 UDP Throughput Test
  https://nexp.tistory.com/4308
• W55RP20 Pico SDK UDP Throughput Test
  https://nexp.tistory.com/4309

3. Technical Challenges and Troubleshooting

One of the most significant challenges during development was eliminating audio noise and ensuring stable real-time audio streaming.

Eliminating Byte-Split Noise

Reference:
https://nexp.tistory.com/4303

Previously, 32-bit audio samples were transmitted as four separate 8-bit segments. Timing mismatches occasionally caused bit shifts, resulting in audible distortion and waveform corruption.

Solution

Instead of splitting the audio data into individual bytes, complete 32-bit (int32_t) samples were transmitted atomically. This ensured that the DAC always received intact audio samples, completely eliminating waveform corruption.

Improving W55RP20 Ethernet Performance

Reference:
https://nexp.tistory.com/4309

The Arduino implementation was upgraded to use the W5500 hardware TCP/IP offload engine through the W55RP20_Ethernet3 library.

Key improvements included:

• Enhanced UDP throughput using iPerf
• Optimization specifically for network speaker applications
• Successful transmission of 96 kHz high-resolution audio
• Verification of audio quality at higher sampling rates

Solving UDP Transmission Issues on W55RP20

Reference:
https://nexp.tistory.com/4311

Since real-time audio applications require low latency, UDP was selected as the primary transport protocol.

Key achievements:

• UDP throughput optimization
• Stable real-time audio streaming
• Reliable 96 kHz high-quality audio playback
• Reduced packet loss and transmission jitter

4. Integrated Dashboard and Streaming Features

Beyond basic audio playback, a Python-based audio streaming dashboard was developed to provide real-time control and monitoring capabilities.

Real-Time Audio Streaming

Users can select audio sources directly from a PC and stream them instantly to the network speaker.

Text-to-Speech (TTS)

Entered text is automatically converted into WAV audio and broadcast over the network speaker system.

YouTube Audio Streaming

By entering a YouTube URL, the system can extract audio in real time and stream it directly to the speaker.

Audio Visualization

A real-time waveform viewer displays the output WAV signal, allowing users to monitor audio activity visually.

5. Conclusion and Future Development (PoE)

This project successfully demonstrated that the W55RP20 provides a stable and high-performance platform for real-time network audio applications.

Through extensive hardware and software validation, reliable transmission of high-quality audio over Ethernet was achieved while maintaining low latency and excellent network efficiency.

The next phase of development will focus on completing a fully integrated PoE-powered network speaker solution, including:

• IEEE 802.3af/at PoE support
• Enhanced audio processing features
• Multi-speaker synchronization
• Network audio broadcasting and paging functions
• Commercial-ready hardware design

The results confirm that the W55RP20 is a highly capable and cost-effective platform for next-generation network audio and PoE speaker applications.