0
\$\begingroup\$

I want to record sound with a piezo transducer. I am connecting it to my sound card's input which has a good mic pre-amp. The levels are quite good and the sound seems quite clear but there's a lack of low end.

I want to build an impedance-matching circuit to improve the sound, to make it full range.

My priorities are:

  • Full frequency range and good quality sound.
  • Simple to build, because I have very little experience in electronics.
  • Must not cost a lot.

I've seen some options on the internet, but they are old and most include the functionality I don't need, introducing additional undesirable complexity.

I was wondering if there are now better and possibly simpler, modern solutions to this problem, possibly making use of the newer elements of better quality.

What would be a good approach to solving this, considering all of the above?

\$\endgroup\$
  • 1
    \$\begingroup\$ what is the frequency response of the piezo transducer? \$\endgroup\$ – jsotola Nov 3 '18 at 3:10
  • 1
    \$\begingroup\$ impedance matching in not going to help, there's a lack of low end beacuse piezos basically don't go there. \$\endgroup\$ – Jasen Nov 3 '18 at 3:17
  • \$\begingroup\$ Thus you need LOW END BOOST. \$\endgroup\$ – analogsystemsrf Nov 3 '18 at 4:30
  • \$\begingroup\$ @jsotola I have no Idea about the FR, bought them cheap off of aliexpress, don't think there were any specs. Is the input FR of the piezo the same as it's output FR? What should I look for in the spec if I want to know the INPUT FR? \$\endgroup\$ – John Doh Nov 3 '18 at 11:39
  • \$\begingroup\$ The frequency response should be in the form output/input. You get attenuation and resonance at certain points depending on the physical characteristics of the material. \$\endgroup\$ – Jaywalk Dec 17 '19 at 20:23
0
\$\begingroup\$

The information representing signal 'coming out' of the piezo transducer is neither voltage nor current. Impedance is the quotient of voltage and current but the signal representing the information, that comes out of the transducer, is actually a charge.

The best type of amplifier, i.e. which adds the least of noise, is therefore a charge amplifier. Mainly a low-noise op-amp with the output of the piezo transducer to the inverting input and a capacitor from the op-amp's output back to the inverting input as feedback element.

This will cause the voltage over the transducer to stay zero so that there is no charge going into the parasitic capacitance of the transducer itself, which would have led to attenuation of the signal, i.e. the charge going into the input terminal of the amplifier which, because the input pin of the op-amp is zero, will now all flow into the feedback capacitor.

Because the same charge goes into the feedback capacitor as is generated by the piezo transducer, the smaller this feedback capacitor, the larger the output voltage, what we normally would call amplification.

If you make a drawing, you will see what I mean. Especially when you put a sign '0 V' near the inverting input pin of the op-amp.

A charge amplifier will give you 'full range', but as @Jasen commented, the amplifier isn't the real problem in this respect. You'll need to find out the frequency response of the piezo transducer, and for the frequency range where it's not 'straight' you'll have to build a compensating circuit.

\$\endgroup\$
  • \$\begingroup\$ Thanks for the answer Joe. Could you please draw such a circuit for me so that I could build it and try it? Can you recommend a good op-amp? \$\endgroup\$ – John Doh Nov 3 '18 at 11:52
  • \$\begingroup\$ You could have googled by now for 'charge amplifier circuit' and then it would have come up with amongst others this [description](https:metrology.hut.fi/courses/s108-180/Luento3/varvah.pdf). Then you could have started designing the circuit that's suitable for your transducer, and come back here with questions if that wouldn't work. In other words, show some effort, then people are more than willing to help out in case of problems. \$\endgroup\$ – joe electro Nov 3 '18 at 12:27

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.