PIP Stereo Piezo Preamplifier

(This page is part of a series. See the rest of my contact microphone pages here.)

The piezo sensors in contact mics and hydrophones need a buffer preamp to sound their best. There are many practical options listed in my page Choosing a Contact Mic For Field Recording. I have tested most of them, and still yearn for a lower noise floor. In 2020 I was in search of a cheap, simple preamp that could run on the plug-in-power (“PIP”) provided by small recorders with 1/8″ jacks. So I optimized my existing experiments and came up with this. For my students, I paired this preamp with the sensor design described here.

Download the schematic, PCB layout files, and 3D-printing files: ZLP Stereo Piezo Preamp v1

The JFET Desert

The lowest noise is usually found with discrete JFET transistors rather than opamps. Unfortunately, the well-regarded options (2SK170, BF862, PF5102 …) are being discontinued at an alarming pace, so I’ve been on a quest to find an available FET that can become the heart of an ultra low noise preamp. The verdict was, you can’t do it with traditional “through-hole” components anymore. But there are some options left in “surface mount” (SMD) packages.

The Circuit

I’m not an engineer and there is nothing novel about my design. It’s just a stereo pair of common-source JFET impedance converters. I tried to optimize it for the low voltage and current available from plug-in-power. I tested many SMD FETS and discovered that the 2SK209 (IDSS grade Y) performed the best for me.

The signal passes through back-to-back zener diodes (D1 & D2) to help protect the FET from excessive voltages. (Piezos can produce massive signals when struck or dropped!) C1 & C2 are the optional pad capacitors switched by S1. R1 & R3 are the gate resistors for the JFET. I used 1M but you could extend the bass response lower using a larger value here (at the expense of more risk of clipping and saturation). Q1 & Q2 are the 2SK209 N-JFET transistors. You can substitute others if you re-bias the circuit, but not many will work well. R2 & R4 are the drain resistors. They form half of the bias network that sets the operating point of the JFET. The source resistors are inside the recorder (usually 2.2k). You can re-bias the circuit by adjusting the drain resistors.

Beware that JFETs have massive “spec spread” so there’s no guarantee that 2 will behave the same. For best results, you should buy a lot and sort them by IDSS. They’re too tiny to write on, so you need little pill boxes or some plain old masking tape as seen here:

Bill of Materials

Ref

Qty

Value

Package

Notes

Manu.

Part #

Mouser Link

C1, C2

2

.33uF

Capacitor – SMD 1206 / 3216 size (X7R)

pad capacitor

Yageo

CC1206KRX7R9BB334

https://www.mouser.com/ProductDetail/YAGEO/CC1206KRX7R9BB334?qs=7s%252B3O6pAiyC1aFicMfaP8g%3D%3D

D1, D2

2

Zener 5.1v

Dual Diode Array – SMD SOT-23 size

FET protection

Diodes Inc

DZ23C5V1-7-F

https://www.mouser.com/ProductDetail/Diodes-Incorporated/DZ23C5V1-7-F?qs=qzbtemnCuJBGR93phcTpQw%3D%3D

Q1, Q2

2

2SK209

JFET transistor – SMD SOT-23 size

IDSS grade Y

Toshiba

2SK209-Y

https://www.mouser.com/ProductDetail/757-2SK209YTE85LF

R1, R3

2

1M

Resistor – SMD 1206 / 3216 size

gate resistor

Yageo

RT1206FRE071ML

https://www.mouser.com/ProductDetail/YAGEO/RT1206FRE071ML?qs=8cPjvKtxWv6lYTF8Ipr29w%3D%3D

R2, R4

2

150R

Resistor – SMD 1206 / 3216 size

source resistor

Yageo

RT1206FRE07150RL

https://www.mouser.com/ProductDetail/YAGEO/RT1206FRE07150RL?qs=Fz%2FrpjPuTcGfiK6dkE8z3g%3D%3D

S1

1

PAD SWITCH

DPDT Slide Switch SMD

0.1” pin spacing

C&K

JS202011SCQN

https://www.mouser.com/ProductDetail/CK/JS202011SCQN?qs=LgMIjt8LuD95JYWxZ7NvZA%3D%3D

PCB Layout

All components are SMD and you don’t need to drill any holes. I used the largest common SMD size so it’s solderable by hand. Of course it would be easier in a DIY reflow oven.

The schematic and PCB layout were done with the free open source KiCad software. I set up the board for single layer home etching with fairly thick traces. The board is tiny so I made a panel of 20 that fits on a 6″ x 4″ copper-clad board. I scored and snapped the boards to separate them. (Hard work, not especially recommended.)

Enclosure and Finishing

Solder 2 piezo sensors to the IN pads and a shielded 1/8″ cable to the OUT pads (both have shared grounds on their middle terminals).

In the download I included an STL for a basic 3D-printed plastic enclosure with a cutout for the pad switch. OpenSCAD source file is included. I used small zip-ties to hold the enclosure together.

I didn’t show it in the photo, but you should wrap the board in grounded foil shielding tape (over an insulating layer of kapton or similar). The circuit is very sensitive to RF interference so we need all the help we can get!

Mods and Improvements

  • The preamp has plenty of low-noise gain for listening to quiet sources. Even with the pad engaged, it’s still not the best preamp for percussive situations. (I think you could alter the bias point to lower the gain and improve dynamic range at the expense of noise.)
  • You can get even more gain if you put a 10 uF capacitor in parallel with the source resistors (R2 & R4).
  • There is no power supply filtering here. It wouldn’t be possible since the audio output is also the power input. Badly shielded cables and/or dodgy recorders will be audible.