This project reads the supply voltage (Vcc) and internal temperature sensor of an ATtiny85 using only its built-in ADC channels.
No external components or sensors are needed — just the microcontroller and Serial output.
- 🧠 Exponential Moving Average (EMA) filtering for smooth readings
- ⚡ Vcc measurement using the internal 1.1V bandgap reference
- 🌡️ Temperature sensing via internal sensor
- 💡 LED blink on Pin 4 as heartbeat indicator
- 🪶 Minimal RAM and CPU usage
- 🔧 Easy calibration via top-of-sketch constants
- ATtiny85 (tested with 8 MHz internal clock)
- µC must be flashed with the Spence Konde Optiboot bootloader to assign correct ADC-channels
- Serial connection (e.g., via USB-Serial adapter, TinyUSB board, or similar)
- LED on pin 4 (optional, for visual heartbeat)
- ATTiny showing Vcc in Volts and Temp in °C
- disregard the first 20 or so values of the output
- Temp: 25.4 °C | Vcc: 3.274 V
- Temp: 25.3 °C | Vcc: 3.269 V
- Temp: 25.3 °C | Vcc: 3.270 V
At the top of the sketch, you’ll find two important constants:
// This holds the Temp-correction-coefficient
// common values: 300 - 318. The lower the value the higher the output
#define TEMP_REF 308 // Temp coefficient
// adjust ADC_REF. The higher the value the higher the output
#define ADC_REF 988 // 1.1V internal reference in mVUse these to calibrate against a trusted thermometer and known power supply.
🧪 How It Works
The bandgap channel (ADC 0x0E) allows calculation of Vcc by comparing the internal 1.1V reference to the actual supply.
The temperature sensor (ADC 0x0F) gives a raw ADC value that is approximately linear but varies per chip — a manual offset is used to bring it close to ambient.
The EMA filter helps smooth both values while keeping RAM usage low and speed high.
🧵 License: MIT License — free to use, modify, and share. Attribution appreciated.
💡 Author: Created by ToS with technical input from ChatGPT.