Introduction
Noise is the enemy of reliable PIR detection. This guide identifies noise sources and provides practical mitigation strategies.
Types of Noise in PIR Systems
1. Thermal Noise (Johnson-Nyquist)
Fundamental noise from thermal agitation of charge carriers. Present in all resistors and the pyroelectric element itself.
Mitigation: Limit bandwidth, choose low-noise components, operate at lower temperature if possible.
2. 1/f Noise (Flicker Noise)
Noise that increases at lower frequencies. Significant in PIR because signal of interest is also low frequency (0.1-10 Hz).
Mitigation: Use chopper-stabilized amplifiers, AC coupling, and avoid DC amplification.
3. EMI/RFI
External electromagnetic fields coupling into sensor or wiring. Sources: motors, relays, radio transmitters, switching power supplies.
4. Power Supply Noise
Ripple and spikes on power lines affecting sensor operation.
5. Environmental Noise
Thermal fluctuations from HVAC, sunlight, moving air.
6. Microphonics
Vibration causing mechanical movement of the pyroelectric element.
Noise Mitigation by Category
Circuit-Level Mitigation
- Bandwidth limiting: Filter out noise outside signal band (0.1-10 Hz)
- Low-pass filtering: Reduce high-frequency noise (EMI)
- High-pass filtering: Remove DC drift and very low-frequency noise
- Shielding: Enclose sensor in grounded metal shield
- Twisted pair wiring: Reduce magnetic coupling
- Ferrite beads: Suppress high-frequency noise on cables
- Decoupling capacitors: Stabilize power supply (0.1µF + 10µF near sensor)
PCB Layout Guidelines
- Keep sensor connections short
- Guard ring around high-impedance nodes
- Separate analog and digital ground planes
- Avoid routing sensor traces near switching signals
- Use ground plane under sensor
Software Noise Reduction
// Moving average filter
const int samples = 5;
int buffer[samples];
int index = 0;
int filter(int newSample) {
buffer[index] = newSample;
index = (index + 1) % samples;
int sum = 0;
for (int i = 0; i < samples; i++) sum += buffer[i];
return sum / samples;
}
Measuring Noise
Use an oscilloscope to measure sensor output with no motion. Characterize:
- Peak-to-peak noise amplitude
- Frequency content (FFT)
- Stability over time
- Response to environmental changes
Case Study: Noisy Industrial Environment
A PIR sensor near a large motor triggered randomly. Solutions implemented:
- Added ferrite beads on sensor cable
- Installed sensor in shielded enclosure
- Added 0.1µF and 10µF capacitors at sensor
- Used twisted pair for power/signal
- Moved sensor 2 meters from motor
Conclusion
Noise in PIR systems comes from many sources, but systematic mitigation at circuit, PCB, and system levels can achieve reliable operation even in challenging environments.