Piezo pickup mod - wood Stompbox Piezo

[brett]

RE: Piezo buzzer/Piezo pickup
Hi.  I've got a nice stompbox working with 7mm plywood top and bottom (overall 300x300x90mm).
The pickup is a piezo wafer, taken put of a plastic buzzer, and hot-melt glued to the inside wall.
It picks up EVERYTHING, including some high-frequency stuff, making it sound like the small drum that it is.  I wanted more bass.  Like a floor tom.  In frequency, my guess is that I want more frequencies down where the low E is on a guitar.
So... I thought a low pass filter would work.
I measured the output impedance at around 660k ohms (+/-80k).  This seems to be rarely measured.  On the internet, the values are mostly 1 or 2 megohms.  (I measured the dynamuc impedance, while using it as a drum).
So.. a capacitor between the 'hot' signal wire and the ground wire will make a LOW-PASS FILTER, hopefully making the box sound lower in tone.
A quick calc and a 2.2 nF (0.0022 uF) looked like a good roll-off frequency.
Experimenting with capacitor values, I found it fairly mild.  A 4.7nF cap worked really well. 
So that's my report:
Piezo buzzer pickups are good.
They have Zout of 660k (lower than generally quoted)
A low-pass filter of 1nF to 10nF rolls off treble.
For strong bass, mild midrange and no treble, a capacitor value of 4.7 nF works well.  Connect it from signal to ground.
Thanks for reading.
ps These piezo transducers have many potential uses, from pickups to reverb send and receive to signal processing.

[brett]

[Rob Strand]

The situation is quite a bit different to that!

The 660k parallel resistance should be in *parallel* with the piezo capacitance (C0 = 1nF to 40nF).  Perhaps even better is 660k to ground but it's still not correct.  In practice it doesn't do much to the response.

The series resistance is much lower than that;  pulling a number out of the air maybe in the broad ball park of 1k.   The series resistance works against the capacitance of the piezo element, and limits the highs in charge mode.

For voltage mode, as you have drawn it, the thing that limits the lows is the resistance (low frequency impedance) of the *next stage* (not shown you pic) and the capacitance of the piezo itself (C0=1nF to 40nF).  The low frequency cut-off is f3 = 1/(2*pi*Rin*C0)

When you add a parallel capacitance it does two things:
- reduces the output due to a capacitive divider formed with C0
- increases the Thevenin equivalent capacitance of the source,
  Ceq = C0 + Cp ; Cp = added parallel capacitance.

What you might be seeing is some boost in low frequencies due to the new lower cut-off of,
f3 = 1/(2*pi*Rin*Ceq) = 1/(2*pi*Rin*(C0+Cp))
if your C0 is about 4.7nF then you will half the f3 frequency.
If your C0 is 1nF then you will substantially lower the cutoff.

The normal way to decrease the low-end is to use a preamp with a higher impedance.

[Mark Hammer]

One sees "cajons" for sale a lot these days, as simple replacements for larger percussion setups for street musicians or smaller ensembles.  I wonder if anyone uses piezo discs to mic, and ultimately process, them.  I guess that would make them "whompboxes" rather than stompboxes.  Tongue-in-cheek remarks aside, it's a serious question.

[Matthew Sanford]

You could go the active route https://www.richardmudhar.com/blog/using-piezo-contact-mics-right/
He gives an excellent explanation for why piezos lose bass

[drdn0]

FWIW, slap a couple of piezos in parallel. I found best results using 3 on the noise box things I was making. Actually made it sound like a useful noise generator rather than just tinny screeching.

[Mark Hammer]

Piezo discs vary in size, and size is correlated with resonances.  One wants something that is at least an inch in diameter.  The other thing is that, as we know from stompbox design, attenuating treble can make bass seem to be more prominent.  So how one affixes a piezo disc to a vibrating surface can shape what parts of the frequency response can seem dominant.  Something that can damp mids and highs, like many double-sided tapes, or even hot melt glue, can create the sense of bass being more prominent.

[Rob Strand]

This guy does a good of demonstrating the mounting (and damping) problem:

EDIT:

Another one which demonstrates the damping in a different way,

[brett]

Thanks, everyone. 
I've made SO many mistakes. 
Of course the intrinsic capacitance is to ground.  Doh!
And as explained the low-fass filter works in concert with the input impedance of the amp... which is 470k.  My calcs around the 660k output impedance of the piezo just happen to give similar results.  Pure coincidence.
Thanks all !
ps more boxes are on the way with different materials and piezo placement.  Heavier wood on the sidewalls + piezo on the sidewall was awful.  Trying piezo on the top next... so many possibilities!