Improve Options For a Piezoelectric Preamp

Here, I've tried to design a Acoustic Guitar Piezoelectric preamplifier. I knew piezo electric preamps are not like conventional non-inverting differential amplifier because piezoelectric pickups are of very high impedance.

Although, I don't have any experience in designing a customized preamp, but I'll have to give it a try.

The circuit is basically a instrumentation amplifier. Gain is set by R6, which is: G = 1+(2R/R6), here R is 4.7K, so the gain here is G ≈ 14.8. in final design, I'll replace it by a 4.7K preset.

Next thing, I think it needs a DC blocking capacitor at input to pass only AC signals from the piezo transducer. Will I use a high pass filter there? like:
**broken link removed**
What should be the optimum C and R value so that the frequency response is not become too affected (specially the lower ones), what if, C = 1µF and R = 4.7KΩ, ie. signals below 1/(2*Pi*R*C) = 33.86Hz will cut off. Is there any wrong? Any other idea?

Will I need to add any capacitor in output? I think it's not needed because I'm using a split power supply, so I don't need a decoupling capacitor. Right?

Now in LTSpice, I've done it using LT1001, but in real world I'll do (I've done it in breadboard) it using TL072. Later I'll be using TL074 to save space a bit. Will NE5532 outperform TL07x in this application? (oh! again the classic TL072 vs NE5532 battle)

Please tell me some other improvement options.

How well are the resistors matched for the instrumentation amplifier? Even very slight variances (<1%) create large problems.
Most piezo devices that I have worked with use an FET fed from a constant current right at the device to condition their signal or they are used in "charge mode".

How well are the resistors matched for the instrumentation amplifier? Even very slight variances (<1%) create large problems.

Click to expand...

From wikipedia:
The ideal common-mode gain of an instrumentation amplifier is zero. In the circuit shown, common-mode gain is caused by mismatches in the values of the equally-numbered resistors and by the mis-match in common mode gains of the two input op-amps. Obtaining very closely matched resistors is a significant difficulty in fabricating these circuits, as is optimizing the common mode performance of the input op-amps.[3]

Most probably, you are pointing at this issue. I didn't think of it very much at first. I used standard 4.7KΩ ¼ Watt carbon resistors. I've measured them using digital meter (UNI-T UT10A). They are:

4.70KΩ Δ=0Ω
4.65KΩ Δ=50Ω
4.69KΩ Δ=10Ω
4.62KΩ Δ=80Ω
4.71KΩ Δ=10Ω
4.63KΩ Δ=70Ω

1% of 4.7KΩ is 47Ω. So simply these resistors are NOT too closely matched. I'll try to get a new set of resistors with less tolerance.

Most piezo devices that I have worked with use an FET fed from a constant current right at the device to condition their signal or they are used in "charge mode".

Click to expand...

What was the application of your piezo transducers?
Can you show me an example schematic of "FET fed from a constant current".

Here I'm attaching a low quality mp3 sample sound of my guitar recorded with this preamp without any filter.

Most of what I did Piezo-wise was help develope electronics for them. I have used them to measure pressure inside combustion chambers of autos and also inside M59 gattling guns. I measured oscillations of continuous caster molds and acceleration/velocity/displacement of shaker tables.
Here is info about signal conditioning:

**broken link removed**