1
\$\begingroup\$

I have been trying to detect a very low frequency, very low strain signal with different types of piezos (hobby type, hobby type enclosed in plastic casing, shielded contact mic type, and flexible piezo film type). The signal is generated from a small stream of water passing through a pipe 2" in diameter.

If I put these sensors on the pipe, and measure the AC signal directly with my oscilloscope, I cannot detect a signal. Although I have not been able to do a frequency analysis of the water pipe, I would guess that the water passing through the pipe at the frequency I'm trying to detect to be <10Hz. Additionally, the water passing through the pipe would create what I'd call "very low strain."

If I turn the water up to a "high" amount, the signal is easily detectable -- the frequency of the water is what I'd estimate 100 - 200 Hz and there is more strain on the pipe.

I have also created voltage amplifier to see if I can amplify this (undetectable) low strain signal I mentioned at the top of this post. The amp seems to only amplify noise or interference.

In short, I think that the piezos I have cannot detect a low strain <5Hz signal.

Can anyone recommend a way to detect <5Hz low strain signal? It appears ultrasonic sensors could be a solution for this problem.

\$\endgroup\$
  • 1
    \$\begingroup\$ Can you use a strain gauge? (Having said that, I'm not sure that I fully understand your setup. Some description of what you are ultimately trying to achieve with these measurement might help.) \$\endgroup\$ – Nick Alexeev Oct 28 '15 at 2:51
  • 1
    \$\begingroup\$ Ultrasound and 5Hz are not going together. Also piezo is high frequency transducer (from few kHz to Mhz). \$\endgroup\$ – Marko Buršič Oct 28 '15 at 6:48
  • 2
    \$\begingroup\$ Low frequencies + Piezo strain gauge = charge amplifier. A 'normal' op-amp configuration will discharge the Piezo device very quickly. \$\endgroup\$ – Chu Oct 28 '15 at 7:53
  • \$\begingroup\$ Are you actually detecting strain, that is the change of distance between two points on the pipe? Or are you detecting acceleration, that is the rate at which the speed of part of the pipe surface is moving is changing? The first is readable all the way down to DC, but needs two points of attachment. The second has a frequency slope in its sensitivity, so no response at DC, response rising with increasing frequency, and needs only one point of attachment. You can use a 'strain gauge' in either configuration. \$\endgroup\$ – Neil_UK Oct 28 '15 at 9:15
  • \$\begingroup\$ @NickAlexeev a strain gauge is a good idea. For now I have not been able to use one. Perhaps for the next version I will try this. Is there a strain gauge you'd recommend for detecting the strain that running water creates passing through a pipe? I was just checking out some links on Mouser for things like this: mouser.com/ProductDetail/Intersil/ISLEM-BDGSTKEV1Z/… \$\endgroup\$ – Thomas Wilk Oct 30 '15 at 17:10
2
\$\begingroup\$

To detect low frequencies with a piezo requires an extremely high impedance amplifier, for example if the element is 10nF you'd need something of the order of 3-10M ohms.

Try a low bias current CMOS op-amp with (say) 4.7M load resistor (parallel to the piezo). Choose one with bias current << 1nA ( easy- there are parts with << 1pA).

Your circuit may not be good enough for practical use if the resistance gets much higher than a few M - condensation from the cold water, for example would have to be avoided even at the suggested resistance level.

\$\endgroup\$
  • \$\begingroup\$ I wonder if a TIA opamp configuration would be best? (Physically the piezo is a current source with big shunt capacitance.) One could even put a cap in the feedback path and integrate the current. \$\endgroup\$ – George Herold Oct 28 '15 at 13:49
1
\$\begingroup\$

For wide frequency response of a piezo consider a Charge Amplifier instead of a voltage amplifier. The lower cutoff frequency is give by the relationship of the resistor and capacitor in the feedback loop. The opamp needs to be a high impedance JFET or CMOS type.

To test your input stage, I recommend putting the piezo element onto a speaker element and drive it at <10 Hz to verify the frequency reponse.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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