From Python to Silicon
This project goes beyond building a guitar pedal, it engineers the supply chain, power infrastructure, and instrumentation required to validate it.
This repository documents a complete systems engineering project, spanning software-driven logistics, custom power regulation, embedded instrumentation, and empirical analysis of non-linear analog circuits.
A technical report covering the design process, thermal analysis, and spectral validation.
While the hardware, firmware, and organic signal analysis are fully validated, the Transfer Function analysis is currently under active development.
- Completed: Instrument analysis confirming the "Soft Knee" topology and Harmonic Asymmetry (Notebooks
01&02). - In Progress: I am currently implementing Angelo Farina's (2000) Exponential Sine Sweep (ESS) Deconvolution method in Notebook
04. This will allow for the mathematical separation of the Linear Impulse Response from the Harmonic Distortion products, providing a Bode plot of the system with linear and harmonic components explicitly separated.
The primary objective was to validate the claim that CMOS Hex Inverters (digital logic chips), when biased into a linear class-A region, exhibit soft clipping behavior and harmonic asymmetry comparable to vacuum tube triodes.
The Evidence: Using the custom RP2040 DAQ built for this project, I captured the saturation characteristics of the Red Llama drive (Chapter 6 of the report).
- Time Domain (Left): Note the "Soft Knee" compression at the peaks. Unlike silicon diodes which shear waveforms off flatly, the CMOS chips round off the transients, preserving dynamic feel.
- Frequency Domain (Right): The spectral fingerprint reveals a dominant 2nd Harmonic (Octave). This even-order harmonic content characterizes what audio engineers call 'warmth', confirming the tube emulation hypothesis.
This project is divided into four interdependent subsystems, each enabling the next.
- The Problem: Manual BOM management is nondeterministic and error-prone.
- The Solution: A Python-based logistics engine that parses PDF documentation, subtracts local inventory, and utilizes "Nerd Economics" to calculate heuristic safety stock.
- Status: Production / External Repo
- The Problem: Audio circuits require a low-noise floor, but standard wall-warts are noisy Switch Mode Power Supplies (SMPS).
-
The Solution: A custom fabricated 12V
$\to$ 9V Linear Regulator with thermal management to support high-current loads. - Key Tech: L7809CV, Schottky Protection, Thermal Dissipation Analysis.
- The Problem: Validating the procurement and power systems requires a sensitive analog load.
- The Solution: A clone of the Way Huge Red Llama, utilizing CD4049 CMOS Hex Inverters for soft-saturation.
- Modification: Replaced D2 (1N4001) with 1N5817 (Schottky) to recover 0.4V of headroom.
- The Problem: I needed to verify the harmonic content of the Red Llama, but didn't own an oscilloscope.
- The Solution: A custom-built RP2040-based oscilloscope.
- Architecture: "Store-and-Forward" capture engine using MicroPython
@nativeemitters. - Performance: 97.8 kSps (Calibrated), 12-bit depth, 1.3mV Noise Floor.
.
├── docs/ # The Engineering Monograph (LaTeX/PDF) & Analysis Figures
├── oscilloscope-rp2040/ # Firmware (MicroPython) & Analysis Pipeline
│ ├── firmware/ # RP2040 Logic (MicroPython)
│ ├── notebooks/ # Jupyter Lab Analysis
│ │ ├── 01_instrument_acquisition.ipynb # Organic Signal Capture 🟢
│ │ ├── 02_instrument_analysis.ipynb # Topology Validation 🟢
│ │ ├── 03_transfer_acquisition.ipynb # Log Sine Sweep Gen 🟢
│ │ └── 04_transfer_analysis.ipynb # Farina Deconvolution 🟡
│ ├── src/ # Shared Analysis Library (Metrics, Plots, DSP)
│ └── schematics/ # KiCad/Python Signal Conditioning Schematics
├── red-llama-build/ # The Device Under Test (DUT)
│ └── procurement/ # Star Ground Artifacts (BOMs, Manuals)
└── power-regulator-12v-to-9v/ # The Linear Power Supply Design