b1500-powermeter-rollover 1.0.1

Keysight B1500A + Thorlabs PM100D synchronized LIV sweep with automatic optical rollover detection (CUSUM, EWMA, rolling average, regression)

b1500_powermeter_rollover

Synchronized Keysight B1500 + Thorlabs power-meter IV sweep with four-algorithm rollover detection.

© Veronica Gao Zhan – May 2026

---

Features

Feature	Detail
IV sweep	Point-by-point sourcing via B1500 SMU (IV or VI mode)
Optical power	Real-time Thorlabs PM100D / PM400 readout per point
Rollover detection	4 algorithms: CUSUM (default), EWMA, Rolling Average, Regression (sklearn)
GUI	PyQt5 scrollable control panel + live 2×2 matplotlib canvas
CLI	Full headless operation with argparse
CSV export	Timestamped file per sweep with rollover summary header

---

Package Structure

b1500_powermeter_rollover/
├── __init__.py            ← public API re-exports
├── __main__.py            ← python -m b1500_powermeter_rollover
├── config.py              ← SweepConfig, MeasurementPoint, RolloverResult
├── b1500_controller.py    ← thread-safe VISA driver for Keysight B1500
├── powermeter_controller.py ← thread-safe VISA driver for Thorlabs PM100D/PM400
├── rollover_detector.py   ← instrument-agnostic online rollover detector
├── engine.py              ← SynchronizedMeasurementEngine (no GUI dependency)
├── cli.py                 ← build_parser() + run_cli()
├── gui/
│   ├── __init__.py
│   ├── worker.py          ← MeasurementWorker (QThread)
│   └── main_window.py     ← SynchronizedMeasurementGUI (QMainWindow)
├── requirements.txt
└── pyproject.toml

---

Installation

# From the B1500/ folder (editable install)
pip install -e ".[all]"

# Or install only the core (no GUI, no ML)
pip install -e .

# Or manually
pip install pyvisa pyvisa-py pyusb numpy PyQt5 matplotlib scikit-learn

---

Quick Start

GUI (interactive)

python -m b1500_powermeter_rollover

CLI (headless / automated)

# List connected instruments
python -m b1500_powermeter_rollover --list

# Run an IV sweep with CUSUM rollover detection
python -m b1500_powermeter_rollover \
    --b1500 GPIB0::17::INSTR \
    --pm    USB0::0x1313::0x8078::INSTR \
    --start 0 --stop 2.5 --steps 26 \
    --compliance 0.5 --dwell 0.1 \
    --rollover --method cusum --threshold 90 \
    --output ./results --name MyLaser

Python API

from b1500_powermeter_rollover import (
    SweepConfig,
    B1500Controller,
    ThorlabsPowerMeterController,
    SynchronizedMeasurementEngine,
)

b  = B1500Controller()
pm = ThorlabsPowerMeterController()
b.connect("GPIB0::17::INSTR")
pm.connect("USB0::0x1313::0x8078::INSTR")

cfg = SweepConfig(
    mode="iv", start=0, stop=2.5, steps=26,
    enable_rollover=True, rollover_method="cusum",
)
eng = SynchronizedMeasurementEngine(b, pm, cfg)
eng.on_log = print
data = eng.run()

r = eng.rollover_result
print(f"Peak power {r.peak_power:.4e} W at {r.peak_voltage:.4f} V")

Standalone rollover detector (no instruments)

from b1500_powermeter_rollover import SweepConfig, RolloverDetector

cfg = SweepConfig(rollover_method="cusum", rollover_threshold=0.90)
det = RolloverDetector(cfg)

peak = 0.0
for power in my_power_readings:
    peak = max(peak, power)
    triggered, info = det.update(power, peak)
    if triggered:
        print("Rollover detected!", info)
        break

---

Detection Algorithms

Method	Latency	Notes
cusum (default)	1–3 pts	Lower-sided CUSUM (Page 1954). Scale-invariant. O(1)/sample
ewma	~window/2	Exponential moving average. Tunable with --alpha
rolling_avg	window pts	Classic windowed mean. Robust to impulse noise
regression	window pts	sklearn LinearRegression (batch) + SGDRegressor (online)

---

Acknowledgements

This project was developed using vibe coding — an AI-assisted development workflow powered by GitHub Copilot. The architecture, code structure, and implementation were generated through iterative natural-language prompting and human review.

License

© Veronica Gao ZHan