The Modules

core_lights — Filament Control

The heart of CozZie Glow. Two PWM sliders — one for the 5W bulb, one for the 21W — each
running from 0 to 65,535. Precise numerical entry available for exact values. Live percentage
display for each filament. This is the module that started everything.

mod_alarm — Alarm & Auto Cycle

Scheduled sunrise and sunset simulation. Configurable wake time and sleep time. The module
slowly ramps PWM from minimum to maximum over a configurable duration — default 20
minutes for sunrise, 15 for sunset. The ramp uses a power curve rather than a straight line,
mimicking the non-linear rate of natural light change. Manual trigger buttons for on-demand
sunrise or sunset cycle. Stop button halts any active cycle instantly.

mod_thermal — Sensors & Output

Four temperature zones in real time — room ambient, tent interior, bulb temperature, and
heated blanket. Pi CPU temperature monitored separately. Live load bars show PWM
percentage for each filament and calculated total amp draw against PSU maximum. Load bar
turns yellow at 75% and red at 90% — visual warning before trip threshold.

mod_audio — Audio DVR

Records audio via USB microphone connected to the Pi using ALSA. Configurable sample rate
— default 48,000Hz. Event threshold triggers automatic recording on sound events. Tag Now
button manually marks a timestamp in the log. Recordings stored to USB thumb drive. Storage
limit configurable. Event log shows all recorded clips with timestamps.

mod_camera — Sleep Cam v2

Motion detection running against the IP camera RTSP stream. Pixel difference analysis at
160x90 resolution — lightweight on the Pi but sensitive enough to detect movement.
Configurable threshold, cooldown period, and clip length. Pi clock timestamp burned directly
onto every clip — immune to the camera's unreliable internal clock. Motion mask allows
exclusion zones to reduce false triggers. Spike log records intensity of each detection event.
Timelapse mode available. Audio from Pi USB microphone included in clips. Detection events
rated by movement intensity and logged with full timestamps.

This module detected 127 motion events on night one. Every one legitimate.