Riccardo eMMe

Observatory Journal

Lightning Jun 18, 2026 1 min read Draft

Handling Noise and False Positives on AS3935

A practical note about disturbers, false positives, wiring distance, local interference and why lightning detection is not just plug and play.

This note is marked as draft. It is part of the public project log, but it may evolve as the observatory changes.

Lightning detection is noisy by nature

The AS3935 can detect lightning-related radio signals, but it can also react to electrical noise and local interference.

During early tests, the sensor reported disturbers and unknown events. That was expected: the first milestone was not perfect detection, but understanding the signal quality.

Possible noise sources

Local electronics, wiring, power supply quality, sensor distance from the ESP32 and electromagnetic interference can all affect readings.

Even a piezo lighter can generate events that are useful for testing interrupts, but not representative of real atmospheric lightning.

Next improvements

The next step is to improve physical placement, reduce interference, tune the AS3935 parameters and clearly separate lightning events from noise and disturbers.

The project should not hide false positives. It should classify them and make them visible.