XycLOps II: Automated Circuit Optimization Platform

Version: 2.0 (Fall 2025)
Backend Engine: Xyce (Sandia National Labs)
Framework: Python 3.10+ / Tkinter / SciPy

---

📖 Overview

XycLOps (Xyce Loop Optimizer system) is a desktop application that automates the tuning of analog circuit parameters. By interfacing with the Xyce parallel electronic simulator, it iteratively adjusts component values (resistors, capacitors, inductors) until the circuit's output matches a user-defined target curve.

XycLOps II represents a complete architectural overhaul, introducing support for Multi-Physics Optimization (Transient, AC, and Noise), interactive Visual Target Editors, and a modern, responsive user interface.

---

✨ Key Features in v2.0

• Multi-Physics Engine
  Expand beyond traditional transient simulations: optimize circuits across Time-Domain, AC Frequency Response, and Noise Spectral Density — all within a unified workflow.

• Asynchronous Threading Architecture
  The solver now runs on a dedicated background thread, fully decoupled from the UI. The application remains smooth and interactable even during heavy multi-physics optimization sweeps.

• Interactive Visual Target Editors
  Design target behaviors by drawing directly on the plot — step inputs, piecewise-linear shapes, and custom responses. A fast “Apply to Target” button instantly converts visuals into optimization goals.

• Data Stream Optimization
  Smart plot downsampling (100k+ point support) prevents GUI lag while maintaining visually accurate waveforms — ideal for long transient runs and wide-band AC sweeps.

• Safe Process Management
  A robust, thread-safe Abort system empowers users to immediately cancel simulations without crashes or orphaned processes.

• Modern UI Framework
  A centralized theming engine replaces default Tkinter visuals with a polished, professional interface utilizing real-time plotting and flat UI design.

• Workspace-Aware Logging
  Every optimization session auto-generates timestamped, versioned directories under
  netlist-results/<netlist>/<session> — ensuring clean organization and preserving previous results.

• Robust SPICE Parser
  Enhanced compatibility with LTSpice-style .PARAM usage.

---

🛠️ System Requirements

1. Operating System: Windows 10/11, macOS, or Linux.
2. Python: Version 3.10 or higher.
3. Simulator: Xyce must be installed and added to your system's PATH.
   • Verification: Open a terminal and type Xyce -v.

---

📦 Installation

It is recommended to run XycLOps II in a virtual environment to manage dependencies (numpy, scipy, matplotlib, pillow, etc.).

1. Setup Virtual Environment

Windows:

# Create the environment
python -m venv xyclopsvenv

# Activate it
.\xyclopsvenv\Scripts\activate

macOS / Linux:

# Create the environment
python3 -m venv xyclopsvenv

# Activate it
source xyclopsvenv/bin/activate

2. Install Dependencies

pip install -r requirements.txt

---

🚀 Running the Application

To launch the XycLOps II user interface:

# Ensure your virtual environment is active!
python main_app.py

Note for Windows Users: The application includes multiprocessing.freeze_support() handling, making it safe to package with PyInstaller if needed.

---

📖 User Workflows

1. Upload Netlist

• Click "Upload Netlist" and select a valid SPICE file (.cir, .net, .txt).
• The system parses components and automatically clears any stale state from previous sessions.

2. Parameter Selection

• The tool lists available components.
• Select the components (e.g., R1, C3) you want the optimizer to tune.

3. Optimization Settings

Choose your analysis type and configure targets:

📈 Transient Analysis

• Settings: Configure Stop Time, Time Step, Start Time, and Max Step.
• UIC: Toggle "Use Initial Conditions" to bypass DC operating point calculation.
• Target: Use the Heaviside Editor to visually draw a step response target (e.g., 0V to 5V rise).

🔊 AC Analysis

• Settings: Select Sweep Type (DEC/LIN/OCT) and frequency range (Start/Stop Hz).
• Target: Upload a CSV defining the target Bode plot (Gain vs. Frequency).
• Response: Optimize for Magnitude (dB) or Phase.

📉 Noise Analysis

• Settings: Select Output Node (e.g., V(out)) and Input Source.
• Target: Define a target noise floor (e.g., 10 nV/√Hz).
• Quantity: Optimize for Output Noise (onoise) or Input Referred Noise (inoise).

4. Run & Monitor

• Click "Begin Optimization".
• The Dashboard shows a real-time plot of the simulation vs. your target.
• A Sidebar streams logs from the backend.
• A floating Convergence Window displays the real-time percentage improvement.

---

🔧 Developer Tools & Testing

XycLOps II includes utilities for profiling and regression testing.

Startup Profiler

Diagnose slow launch times by measuring library import overhead:

python analyze_startup_time.py

Stress Test Harnesses

Headless scripts located in backend/manual_tests/ verify the solver engine without the UI.

Transient Stress Test:

python -m backend.manual_tests.long_instr_amp_test

Optimizes a 7-variable Instrumentation Amplifier with strict tolerances (10^{-14}).

AC Stress Test:

python -m backend.manual_tests.long_instr_amp_ac_test

Performs frequency domain optimization from 10Hz to 1MHz.

---

📂 Project Structure

• main_app.py: Application entry point.
• frontend/: UI code (Tkinter), including ui_theme.py and visual_curve_editors.py.
• backend/: Core logic. curvefit_optimization.py (Solver) and netlist_parse.py (SPICE Parser).
• runs/: Automatically generated session logs and result artifacts.

---

⚠️ Troubleshooting

• "Xyce not found": Ensure the path to the Xyce executable is in your system Environment Variables.
• Graph Flatlines: Try enabling "Default Bounds" in the settings menu to give the optimizer a wider search space.
• Slow Startup: Run analyze_startup_time.py to identify if a specific library is causing delays.