LED Lamp Tester Automation (TTML)
Automated LED lamp testing for London Underground, reducing test time from ~10 minutes to 30 seconds. Designed the Raspberry Pi system, hardware, Python/OpenCV scripts, Node-RED workflow, and operator test screen for accurate, traceable testing.
💥 Impact
Replaced a slow, manual testing process with a fully automated system, improving throughput, accuracy, and repeatability. Enabled scalable testing of production batches with efficient, traceable results.
✨ Highlights
Extremely satisfying project: I designed the end-to-end solution, combining hardware, software, and process flow, successfully integrated into the manufacturing cell.
📝 Details
- Automated voltage, current, LED count, and colour verification tests using Raspberry Pi and camera.
- Python/OpenCV scripts validated full and partial illumination patterns, and correct LED counts per colour.
- UI provided operators with clear test instructions and real-time feedback.
- Node-RED orchestrated test sequences, executing Python and bash scripts for each step.
- Test results logged per lamp using RFID serial numbers, stored in CSV files for analysis and process optimisation.
🤝 Contributions
- Designed and implemented the full hardware and software system, including relay board control and test sequencing.
- Developed Python/OpenCV scripts to count LEDs, verify colour within wavelength tolerances, and check illumination patterns.
- Built a simple UI to inform operators of the test progress and results in real time.
- Used Node-RED to coordinate workflow, calling Python and bash scripts for each test step.
- Implemented data logging with RFID-linked CSV files for traceability and later analysis.
🗓️ Timeline
Completed within a few months.
🧑🤝🧑 Team
Sole developer, collaborating with production engineers on mechanical design and assembly; manager provided guidance as required.