1
\$\begingroup\$

What is a suitable sensor for detecting very faint vibrations on metal e.g. on a railroad where the train is miles away? Is this even feasible? Just a hobby project for experimentation.

I thought about an accelerometer like ADXL345 but in the above use case it would be more like sound - super faint.

Someone said a piezo vibration sensor - but how could that possibly pick up vibrations miles ahead? I'm thinking more of something that would be suitable for detecting an earthquake or tsunami far away. Ideas welcome.

\$\endgroup\$
3
  • 2
    \$\begingroup\$ It's feasible, but what are you trying to detect? You need to know the spectrum of the sound ie its characteristics. \$\endgroup\$ Jun 20, 2022 at 9:31
  • \$\begingroup\$ Search for "seismometer." \$\endgroup\$ Jun 20, 2022 at 16:25
  • \$\begingroup\$ A MEMS microphone with direct contact has chances \$\endgroup\$
    – Jens
    Jun 20, 2022 at 16:34

3 Answers 3

2
\$\begingroup\$

Piezoelectric wafer sensors work well for this, but using them effectively requires more knowledge than a hobbyist typically has, in terms of the circuit design needed. You can't simply connect them to the input pins of an ADC. You need op-amps and other types of signal-conditioning circuitry.

Strain gauges would also probably work to measure vibration, and have their advantages, but the amount of voltage change that a strain-gauge would experience in response to vibration would be extremely small, so you'd need an extremely high-precision data-acquisition system to measure it in a useful way. I'd say that would be beyond what a hobbyist could conceivably build—although you could buy an off-the-shelf DAQ and then interface it with some strain-gauges mounted to the metal rails you want to monitor.

An extremely sensitive accelerometer would be your best bet as a hobbyist. I don't have a specific one to recommend, but check the specifications and compare. They can pick up more vibrations than you would expect. The ADXL345 is probably not the best choice. It is focused on being ultra-low power, not high sensitivity. It only has a resolution of 10 bits. For this application, I'd look for something with higher resolution. A single-axis accelerometer should be sufficient; you aren't trying to measure acceleration or determine tilt in all three axes, only vibration. Perhaps consider something like the ADXL103 or ADXL203 (https://www.analog.com/media/en/technical-documentation/data-sheets/adxl103_203.pdf). This gives you extremely high precision measurements in either one or two axes, and it's packaged on a single IC, including all the signal conditioning, which would be very convenient for integration. But there are lots of options in this space worth considering.

\$\endgroup\$
1
  • \$\begingroup\$ I really like the idea of just using an IC - nice and easy to package. I did not know the 10 bits of resolution from ADXL345 is considered low. That's good to know, thank you. I will check out the two axis ADXL103. \$\endgroup\$
    – T.K.
    Jun 27, 2022 at 18:52
2
\$\begingroup\$

Another suggestion - assuming this is a steel rail then glue a magnet to it and place a free standing coil over the magnet. Any vibrations will induce a current in the coil. Amplify/filter and digitize.

You can also do the reverse, with the coil fixed to the rail and the magnet free standing.

It's an R&D project at undergrad level

\$\endgroup\$
1
\$\begingroup\$

Gramaphone cartridge with the stylus resting on the rail and a suitable preamp? Capacitive displacement sensor? Laser interfeometer?

Plenty of ways to measure small vibrations.

\$\endgroup\$
3
  • \$\begingroup\$ In other words, a "contact microphone". \$\endgroup\$
    – Dave Tweed
    Jun 20, 2022 at 11:36
  • \$\begingroup\$ A laser interferometer, in the right conditions, can be used to measure incredibly faint gravity waves from billions of miles away. \$\endgroup\$
    – rdtsc
    Jun 20, 2022 at 12:08
  • \$\begingroup\$ Self-mixing Laser interferometer + paper in the link youtube.com/… \$\endgroup\$
    – DKNguyen
    Jun 20, 2022 at 13:57

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.