Remember to click G+1 if you like a post... It will make me happy :-)

14 February 2015

Dreaming of building a Piezo Buffer

Today's post has taken a bloody age to come to fruition. It's a tiny step in a giant journey. I'm working on a new electronics project to build a Piezo Buffer. The problem I have is that I'm starting from knowing diddly-squat about Electronics. There has been so much to learn that it's been quite overwhelming at times. I’m going to try and give you as much info as I can as I work through this. Let's hope that I can string it all together. Today, things are mainly theoretical. Come on… let's get stuck in...

First of all, I'll set the scene by revealing what a Piezo Buffer is, and why I think I might possibly need one.

You've seen a handful of posts from me over the years where I've discussed piezo pickups. I've fitted them into some of my instrument builds and always wrestled with getting a good sound out of them afterwards. I've had varying degrees of success, but it's never felt like I've ever properly cracked it.

The piezo pickups that I've been using are the ceramic under-saddle variety; They sit under the bridge and operate by turning the vibration generated by the moving strings into AC signal.

All types of piezo pickups (aka piezoelectric transducers) have a similar trait: They are what is termed as "ultra-high impedance". I’m struggling to come up with an easy way to explain Impedance, but for now, lets just say that the higher the impedance the lower the current that flows. Devices such as Amps, Recording Studios and Mixing Decks all want to be fed by "low impedance" devices - i.e. (relatively) Lots of current. For these devices "High" impedance is bad; "Ultra-high" is really bad.

You might be in the habit of using a "Hi-z" input when recording electric guitars direct. Whether you realise it or not, by doing this you are "matching" the high impedance of the guitar to the low impedance input requirements of the studio. You’re accounting for the difference in output and input impedance in an attempt to get a better signal and hence a better sound. Not all Hi-z inputs are created equal, but they're all trying to do the same thing.

Whilst a good number of devices come equipped with Hi-z inputs ("z" here referring to "Impedance"), none I've worked with come with built-in Ultra-Hi-z inputs. The answer to this problem is what I’m calling a "Piezo Buffer" which is a device whose primary job is to convert a signal flowing from an ultra-high output impedance device to suit the needs of a low impedance input.

Got it?

I've researched all manner of potential ways of doing the buffering, but invariably, I keep coming back to one deceptively simple design that has been posted by Scott Helmke.

The solution I talk about in this post owes a lot to Scott’s inspired work. I'm calling it the Mark 1 because I am pretty certain that once I start breadboarding that this will evolve into something else.

As is the case with Scott's, the heart of my buffer is a component called a "2N5457" transistor. It's what's referred to as an "N-Channel General Purpose Amplifier" or more generically an "N-Channel Junction Field Effect Transistor" (JFET). The strength of this particular transistor lies in its ability to handle extremely high impedance input from devices that are unable to supply much current. If you've been paying attention then you'll be pointing at the screen right now saying, "but surely, that's exactly what piezo pickups are!"


Pretty much the whole Piezo Buffer circuit is as it is because of the needs of the 2N5457.

I stumbled across this great little app the other day called iCircuit. The picture above was created in it. It's been a bit of a revelation for me that I can experiment with circuits without actually having to build them for real. Just as well really, because I'm still waiting for parts to arrive in the post. iCircuit is a circuit-building app with a cool simulator mode. You can switch the circuit on and see what happens!

Now, I'm not saying that this App is perfect or even if it is an accurate simulator, but it has allowed me to explore ideas and that's keeping me happy at the moment. I've spent hours and hours playing with it.

Let me give you a quick run-down of the highlights of the Mark 1 Piezo Buffer circuit. I hope to be experimenting with it soon:

The signal generator top left is purely to simulate current generated by a piezo. It won't appear on the finished board.

Rp is a pull-down resistor. This sets the input state. I've read that this should be 10 times the input impedance. I'd like to experiment with a pot here to see what effect different values has on the quality of the sound. In Scott Helmke's design he is using this pull-down resistor to effect a sort of "gain". This is definitely worth trying out.

Rd, Rg, Rs, Cin and Cout are a set of resistors and capacitors used to calibrate the 2N5457. I won't explain the values I am intending to start with just yet (though you can see them on the schematic), Suffice to say that there is some science behind them that I will explain in a later post, once I've had a chance to validate the circuit.

The two capacitors marked 800MHz and 1900MHz is a part of the circuit that I'm not totally sold on yet. I may drop it altogether. What it is meant to do is to provide a degree of protection from RF interference by shunting part of the signal to ground. The issue I have with this is that I don't think that I'll need it and I don't like that it is removing from the signal. Before I commit, I'll experiment with it and see if my ears can tell the difference. Maybe I could make it switchable? This is the one thing in this circuit that I don't think iCircuit is modelling properly, but what do I know?

The only other thing of note is the 9V power supply to run the 2N5457.

I'll call out the grounding issues I've been having with iCircuit. I seem to remember reading that you need to be specific with ground connections to give iCircuit the best chance of simulating the circuit properly. What I've found is what seems like trivial differences (to me at least) in ground connections can have significant difference in how the circuit simulation plays out. What I have above was achieved through some pretty delicate trial and error. I can't wait to find out whether my real circuit needs to be put together with such precision.

Okay, there's the groundwork laid for this project. I'm still a little confused as to how to go about turning the circuit diagram above into a functioning pedal. Let's hope I can figure all that out! Rest assured, you'll get all the highs and lows here on the blog. Until next time...


  1. The two shunt capacitors in parallel will add together, giving you 43pF.

    1. Thanks for the insight MarkyH. I have the circuit pretty much as good as I'm going to get it now and I never did add in the RF shunt. I don't know if you caught any of my later updates on this project, but I hope to post the circuit I've ended up with maybe even this weekend.

    2. MarkyH - I've been doing some reading into capacitors in parallel and what I've been learning might warrant a post in itself. Put simply, the suggestion is that although you can work out an overall capacitance for capacitors in parallel just as you've explained, there is a physical difference to how separate capacitors will perform over a single "equivalent" capacitor. Although equivalent in capacitance, they're not equivalent in terms of how they respond to different frequencies.

      This article explains it far better than I can:

      So there may be real value in having capacitors in parallel... though I've read about some pitfalls in doing this. Ha ha... swings and roundabouts!