0
\$\begingroup\$

I run a parking facility that has recently suffered a rash of thefts, accomplished by cutting through our chain link fence.

I am trying to design a simple circuit using a piezoelectric vibration sensor, which will allow us to detect fence cutting in real time and dispatch security.

Because the piezo film can generate high voltages, I need to scale the output to 0-3.3V so as not to damage the MCU I'm using to read the values (ESP32). I would also like to boost the signal, so that small vibrations are easily detectable.

I have tried a number of different circuit designs, including the control board provided by the manufacturer of the piezo film. None come close to the sensitivity I can observe by hooking up the piezo to my oscilloscope. As a crude comparison, with the piezo film taped to my desk, the oscilloscope can detect light finger taps on the desk several inches away, whereas my circuits require explicitly bending the film before any signal registers.

I am not an electrical engineer, and have a very limited understanding of signal conditioning, however, I'd very much like to understand where I've gone wrong, and not just have the answer dropped in my lap. Much appreciated.

What I think is my most correct attempt: Falstad Circuit Simulation

The op-amp I'm using is LM2904. Selected because this is what DFRobot is using in their board.

\$\endgroup\$
10
  • \$\begingroup\$ Your real circuit will have hundreds of feet of wire, you can't ignore that. I'd suggest go looking for a pre-engineered solution personally. The reason your circuit isn't working but an O-scope is, could be the difference in input impedance. An O-scope is about 1Meg or maybe much higher. Your circuit as shown is 500k in the simulator, what is it in real life???? \$\endgroup\$
    – Kyle B
    Commented Oct 4, 2022 at 0:26
  • \$\begingroup\$ Can you differentiate between a fence cutting event and someone just shaking or rattling the fence? False positives could be the real obstacle. \$\endgroup\$
    – JYelton
    Commented Oct 4, 2022 at 0:29
  • \$\begingroup\$ The piezeo sensor would probably go off constantly in wind, or when local wildlife bumps into it. \$\endgroup\$
    – Polynomial
    Commented Oct 4, 2022 at 0:35
  • \$\begingroup\$ @KyleB The full plan is to scale this up to several dozen devices transmitting wirelessly over LoRaWAN, so the wire length should be negligible. The version being discussed here is a proof of concept. My resistor in the simulator is actually a potentiometer, 500k-1M ohms. Very low sensitivity regardless of how I set the pot. \$\endgroup\$
    – Ryan
    Commented Oct 4, 2022 at 0:48
  • \$\begingroup\$ @JYelton, Polynomial -- indeed very possible. We need to get this out on a fence to test that. If I get this working, it will serve as a proof of concept. \$\endgroup\$
    – Ryan
    Commented Oct 4, 2022 at 0:52

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.