An engineering archive documenting the design and implementation of a complete polyphonic digital synthesizer built around the classic 8051 architecture.
A complete digital synthesizer engineered from the ground up.
Hand-written 8051 Assembly. Custom Hardware. Professional Audio Design.
The RTAL WELLENKRAFT in its original working environment (together with the first Mutable Instruments SHRUTHI in a 19-inch rackmount chassis from the lab)
*Figure 1 – The RTAL WELLENKRAFT in its original 2009 working environment*
WELLENKRAFT_SOUND_DEMO_GitHub.mp4
Figure 2 – The RTAL WELLENKRAFT Frontview
How far can a classic 8051 microcontroller be pushed?
Most engineers would have chosen a modern ARM processor or a dedicated audio DSP.
I deliberately chose a different path.
The Silicon Labs C8051F120 had become my preferred embedded platform after years of professional development. I knew its architecture, timing behaviour and peripherals intimately.
The challenge was simple:
Could a professional polyphonic wavetable synthesizer be built entirely around that processor?
The answer became WELLENKRAFT 8051.
- 🎹 8 Voice Polyphony
- 🌊 Two DDS Oscillators per Voice
- 🎛 Digital Filter per Voice
- 🎚 Independent DCA / DCF Envelopes
- 🎵 Complete MIDI Implementation
- ⏱ MIDI Clock Synchronisation
- 💾 External SPI Flash Wave Memory
- ✨ SPIN FV‑1 Stereo DSP
- 🖥 Ctrlr Editor
- ⚙ 100% 8051 Assembly Firmware
This project was never about proving that old hardware is "better".
It was about demonstrating what is possible when hardware, software and system architecture are designed together with a deep understanding of the platform.
Every engineering decision was intentional.
The C8051F120 was the processor I trusted most.
Years of embedded development had given me confidence in its deterministic behaviour, interrupt latency and peripheral architecture.
My work at the Technical University of Dortmund also provided access to professional development tools and laboratory equipment.
For real-time synthesis, predictability is everything.
Assembly Language allowed absolute control over execution timing, RAM usage and hardware access.
No compiler overhead.
No hidden optimisations.
Every instruction exists for a reason.
📷 Insert assembly source screenshot.
Benchmarking showed that one processor could generate four complete voices while maintaining deterministic timing.
Instead of simplifying the synthesis engine, I duplicated it.
Processor A: - Voices 1--4
Processor B: - Voices 5--8
The result is true eight-voice polyphony without sacrificing synthesis quality.
The synthesis processors should create sound---not compute long reverbs.
The SPIN FV‑1 provides dedicated stereo effects while the 8051 processors remain focused on deterministic synthesis.
Several effect programs were written specifically for WELLENKRAFT 8051 and stored in external EEPROM devices.
- Ctrlr Editor
- MIDI
- Processor A
- Processor B
- Audio Mixer
- FV‑1 DSP
- Stereo Output
- Main editor
- Oscillator page
- Filter page
- Modulation page
- Effects page
- Preset management
Future documentation will explain:
- DDS implementation
- 24-bit phase accumulators
- Voice allocation
- Digital filters
- Envelope generators
- MIDI parser
- Interrupt scheduler
- Flash memory organisation
- Performance optimisation
Looking back, the most valuable lesson from WELLENKRAFT 8051 is that outstanding embedded systems are built through understanding---not simply by selecting faster hardware.
Knowledge of the architecture proved more valuable than processor performance alone.
Many concepts developed here later influenced subsequent RTAL synthesizers.
firmware/
hardware/
schematics/
ctrlr/
docs/
images/
audio/
RealTimeAudioLab (RTAL) documents complete engineering projects spanning embedded systems, digital audio, custom hardware and electronic musical instruments.
Every repository is intended to preserve both the finished instrument and the engineering journey that created it.