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

19 February 2017

How to make a King-Z Piezo Buffer Pedal

I've been busy on a quest to tame the piezo pickup. The up-shot is a new pedal. In this post I'm going to share the circuit with you and probably all sorts of other things too...

Disclaimer: I'm no expert. If I want to find out whether a fence is electric... I'll touch it. But just because I did... doesn't mean you need to too. What I've gathered here below is the end result of hours of experimentation. In all truth, I've had far more failures than successes. And it goes without saying... if you spot errors below... please let me know!

Let's start with a reminder of what this project is all about. About 2 years ago I blogged about how it can be hard to get a good sound out of piezo pickups and that I had a hunch that this problem could be solved with some clever circuitry.

The nub of the issue stems from something called "Impedance". Electric guitars with your standard wound pickups are typically what you might call high impedance. Piezo transducers tend towards what I've been calling ultra-high impedance.

Most amps and recording equipment is tailored to handle high impedance signal, but not necessarily ultra-high. As such you can find that piezo pickups sound quiet and flat through them. Worse still, as you try to compensate for the poor signal by upping gain, you can end up adding noise. It's often a no-win situation.

Enter the "Piezo Buffer" circuit, which attempts to do some impedance matching and also to provide a little bit of gain before the signal hits your amp or recording studio. You can buy these circuits or you can make your own. I've chosen to make one. Bloody hell it's been a journey of discovery!

Let me introduce you to the King-Z Piezo Buffer...

Here's the circuit diagram. I'll explain some of what goes where and why soon. First of all, here is a list of the parts you will need for the circuit as drawn above:

N-JFET - 2N5457
LED - 2.8V
R1 = 10k
R2 = 10k
R3 = 3.3k
R4 = 5k
R5 = 1k
R6 = 1k
R7 = 2.2k
R8 = 300k
R9 = 680k
R10 = 1M
R11 = 2.7M
R12 = 4.7M
R13 = 10M
C1 = 330pf ("331")
C2 = 0.1uf ("104")
C3 = 220uf
C4 = 34nf

An explanation for R4 and R5...

I haven't found this being done in any of the buffer circuits I've found on the internet. This may be because most are targeting a 9V DC battery to power the circuit. The King-Z circuit also requires 9V DC, but I haven't used a battery. Instead, I'm using a "Wall wart" power supply. These things can be pretty variable in how much voltage they actually provide. The one I'm using for this project is rated at 9V, but I've measured it at closer to 13.8V.

I don't want the power supply to become a problem, so I've used a "voltage divider" technique to drop the voltage of the current that gets passed to the drain ("D") of the JFET.

The resistors I'm using will drop the voltage to about 2.3V which is going to suit the 2N5457 JFET far better.

 An explanation for C3 and C4...

This is an attempt at AC filtering. The idea is to use the inherent AC signal-blocking capabilities of capacitors to reduce noise on the circuit. This is noise that might being leaking in through the power supply. I may be using a DC power supply, but remember that this is being powered by AC mains. Some of this AC signal may make its way through to my circuit and result in unwanted noise.

The general principle here is to use capacitors that block the AC frequencies that you don't want. Typically, people are using a big capacitor for low frequencies and a small capacitor for high frequencies. I've even seen people considering a third mid-range filter, but this seems like overkill to me.

By big, I'm talking in the region 100-220uf. I settled for 220uf in the end. I'm using a high voltage electrolytic capacitor for C3.

By small, we're talking somewhere in 20-50nf. I'm using a polyester film capacitor for C4, but I reckon a ceramic capacitor would work just as well.

 An explanation for C1 and C1...

These capacitors are there to isolate the signal flowing from IN to OUT from any DC interference coming from the JFET. Again, they're predominantly filtering.

Both are tiny. I'm using ceramic capacitors here.

An explanation for R3 and R8 to R13...

First of all let me call out that R8-R13 should NOT be all connected at the same time! My intention is to only connect one of these using a pedal switch. So at any point in time, one of them will be connected to the circuit and the other 5 will not.

So what you end up with is a single resistance on the JFET Gate ("G") and a single resistance on the JFET Source ("S"). These are there to bias the JFET which in the simplest terms, is how you get this component to work.

I'm using an 2N5457 N-JFET. There are different types of JFET available that should work just as well. They've all got different tolerances and operating requirements. Go read the datasheet on the one you want to use and adjust the input voltage and gate biasing resistors to get the best sound you can.

What I've discovered through my own investigations and experimentation is that actual JFET operation isn't always in line with datasheets. There can be a margin of error in how they're designed to operate and how they actually perform in the wild. I guess it's difficult to make them exactly to spec. As such, you'd do well to experiment with your settings here. What you are trying to achieve is the cleanest sound you can... with no clipping!

R8 to R13...

These are simply a range of different size resistors that can be switched in and out to try and get the best sound out of whatever you have attached to the input. This is the magic... the whole point of the King-Z pedal! I've found a huge variability in piezo performance with a swathe of different impedance and noise issues. The bigger the resistor the higher the impedance the circuit will handle. Well, that's the theory ;-)

Oh... one last thing here on JFET biasing: Always have some resistance on both the Gate and Source pins or you risk damaging the JFET (especially if you're driving it towards the higher end of its operating voltage scale). If your circuit stops making sound... power off... ground everything... check the resistor biasing... power on. You might get lucky and it will start working again.

 An explanation for R1 and R2...

R1 and R2 are intended to provide protection to the circuit from unwanted noise coming from the input and output stages. They complement the C1 and C2 capacitors, acting to dampen rogue signal as it bounces back and forth on the line in the Input and Output stages. This can be a problem where there are differences in impedance - which is exactly what we're dealing with here!

Finally, the last bit to explain is the LED. I have two resistors here because I couldn't find a single 3.2k resistor in my box of resistors. You obviously don't need to have an LED in your circuit, but a word of caution here... I've found that it does have an impact on how my circuit performs. I ran a test where I didn't hook up the LED and it drastically affected the sound I was getting (in a bad way). If you want to get similar results to me, I'd recommend that you keep the LED... OR you bias your circuit without the LED attached.

It's one thing to design a circuit, but a whole different thing to actually build it. Here's an overview of how I hooked everything together for this build. I like to be able to take things to bits easily so you'll see that I'm using 6 terminal blocks that I can use to connect the pedal up.

IN = the positive line from the Audio IN
OUT = the positive line to the Audio OUT
V+ = the positive line in from the DC Power Supply
LED = the positive line to the LED. Connect the negative LED to Ground.
Ground = the line in/out to Ground. Ensure that this gets connected to the pedal case to ensure optimum shielding.
H1 - H6 = the switchable resistors to calibrate the pedal. To make H1 active, connect H1 to Ground; To make H2 active, connect H2 to Ground; etc. I've connected these all up to a switch to allow me to easily change this. Always have one active in the circuit at any one time.

When it all goes wrong...

Let's be honest and admit that things are going to go wrong. Two years ago I rage-quit this project and shoved everything in a box. I had started out with high hopes, but quickly discovered that a winning smile just wasn't going to cut it.

I hit all the same sorts of issues the second time around... only this time I managed to get through them! In the interests of learning from my mistakes, I'll share some of them here for you.

The biggest issue I've had all along has been noise in the circuit. Every time that I've bread-boarded a circuit I've never got the clean sound I've been looking for. It's a common problem. I've tried all sorts of things. Let's discuss a couple:

Power Supply Noise: You'll hear about the potential for AC noise to make its way into your circuit from the power supply. It's easy to check for this by swapping the wall wart for a battery. I don't think I had this problem. Even so I still think the C3/C4 capacitor filters are a good idea.

Grounding: Grounding is important. I think I subscribe to the school of thought that all grounds within your circuit should converge at one point, but to be honest, I have this sneaking suspicion that a lot of people make too much of this a cause of circuit noise. In my experience it seems that you'll get one of 3 outcomes irrespective of how you do your grounding: 1. No sound at all... the circuit is broken; 2. A godawful buzzing like a laser... things are touching that shouldn't (i.e. a short-circuit); 3. It works. If it works, any noise you hear is most likely coming from somewhere else.

Breadboarding: Breadboards are pretty flaky; It's easy to have bad connections and wires crossing over each other. Housing a circuit properly isn't guaranteed to fix noise issues, but it might go a long way towards it. Give it a go.

Input Noise: There's not a great deal you can do here, but there are a couple of things to check. Long cables are bad. Try short ones while you're building. Try different piezos for comparison. Just like anything else, you could simply be dealing with a bad one, or one that is suffering from shielding problems. Rule this out before you throw in the towel.

Output Noise: Check that your amp isn't introducing the noise. Cut the gain. Make sure that no effects are being added. Cut all EQ. Switch to the clean channel. Use a short cable. Rule this out as a cause.

Shielding: In my opinion poor shielding is the biggest cause of noise in any circuit. You only need to pinch the positive line in from your audio signal with your thumb and finger and hear the noise intensify to know that this is a real problem. The good news is that boxing your circuit will go a long way to solving this problem for you. While you're bread-boarding, switch off plugs in the vicinity that you're not using. Hide your phone. Move your light further away. They'll all be contributing.

You've been very patient to get this far, so I'll reward you with a listen to the King-Z Buffer in action. It's set to about 680k and I've got my Sharkfin homemade uke plugged into it. The line from the pedal is fed into my Line6 Spider IV 15watt guitar amp. It's on the clean channel, a little gain, no EQ, with a little chorus and lots of echo.

The first thing you need to take note of is how quiet it all is while I show you around the set-up, before embarking on a trip of buffered ukulele psychedelia :-)

Noise = nada

Thanks where thanks are due...

I slip in a quick picture of Rod Naylor who wrote a wonderful book on woodcarving that I bought last weekend. I reckon Rod might just become a hero of mine in the not-too distant future ;-)

Back to circuits...

I want to make a few quick call outs before I forget. The first is to Scott Helmke who's Mint Box Buffer circuit has been truly inspirational. This is the design that made me think that I could actually do this. I've pondered it so many times trying to unravel the dark magic. I never did manage to get his circuit working the way he says it should, but that doesn't matter. Thank you Scott for sharing your work.

I also want to thank Petre Tzv Petrov whose circuit I stumbled across towards the end of my quest. It turns out that our two solutions share a lot of the same sort of ideas, and better still, Petre is able to explain his circuit far better than I am able to explain mine. Thank you Petre for sharing your work. It helped me to validate a lot of what I had done and taught me some of the reasons why I did it!

That's your lot. I never mentioned that I've published a new comic. Perhaps I'll do a post on this soon. In the meantime check this out.

6 November 2016

It's all a matter of perspective

A couple of weeks ago I was taken by a couple of a new ideas for making pictures inspired by some fantastic work I stumbled across on the internet. Being the uneducated fool that I am, I don't really know what to call this new style, but I can certainly explain what I got up to in my experimentation...

First of all, let's give credit where credit is due. Here is the picture that inspired my adventures. It's a lino-print by London-based artist, Paul Catherall called Trellik Blue. He's done a series of print runs from this particular design but this one is my favourite. I love the colours; I love the simplified design; I love the geometery; I love this picture!

The subject is a distinctive housing block in a pretty distinctive suburb of London. Here's what wikipedia has to say on Trellick Tower:
Trellick Tower is a 31-storey block of flats in Kensal Town, Royal Borough of Kensington and Chelsea, London, England. It was designed in the Brutalist style by architect Ernő Goldfinger after a commission from the Greater London Council in 1966, and completed in 1972. It is a Grade II* listed building and is 98 metres (322 ft) tall (120 metres (394 ft) including the communications mast).
I really do need to find out more about the wonderfully named Ernő Goldfinger! Althought the architecture might be described as Brutalist, I'm not sure that this picture warrants the same title.