In the next couple of posts, I’ll build a guitar fuzz pedal. More specifically – a Fuzz Face clone. It is a simple, yet, iconic pedal. It’s been out there since ’60s. Used most famously by Jimi Hendrix, but essentially used by many others like Gilmour, SRV etc.
Even more specifically 🙂, I’ll be doing a clone of silicon version of it, something similar that Gilmour used on The Dark Side of the Moon album.
If you look it up, there are several versions of the Fuzz Face. In the above schematic, I kind of used the simplest one. It uses NPN transistors, BC109 in the schematic. The pedal used BC109C and BC108C transistors among others, but I’m pretty confident nearly any NPN transistor (with enough gain) will work (the question is only how good will they sound and behave when played).
Getting to Know the Effect
The effect is super simple – 2 transistors, 2 pots, 4 resistors and 3 capacitors. The simplest effect I did had just a couple of components less. It is a deceptively complex to analyze though. Here’s my attempt to “decompose” it:
Volume control is just a simple audio/log tapered potentiometer. It controls level of output of Fuzz Face. The effect will produce about 0.3Vpp output at wherever the fuzz control is set. This is a way to control the loudness of that signal.
Fuzz potentiometer controls the gain of the circuit. The gain controls how hard clipping and thus distortion will be. More gain, harder the clipping. C2 capacitor forms low pass filter with the pot and at highest gain setting cut-off frequency is about 7Hz, this means there’s no real effect on sound. As the gain is lowered the cut-off frequency increases so at lower settings it could affect lower frequencies a bit.
Coupling capacitors as in all circuits I looked at so far are there to block any DC component coming into the effect and preventing any leaving it and thus affecting the next effect in chain. C1 and C3 can affect frequency response though.
C3 forms a low-pass filter with volume potentiometer. At highest volume the cut-off frequency is around 32Hz, but with lower volume it starts attenuating low frequencies more. At half volume, for example, the cut-of frequency is about 64Hz.
C1 forms a low-pass filter with input impedance of the effect essentially. It’s a bit trickier to determine this, but roughly, input impedance is going to be between 500 ohm and 1K ohm so the cut-off frequency is roughly 70-150 Hz.
Yes, the input impedance is not ideal and will load guitar pick-ups affecting higher frequencies, but at the end of the day, it is the sound that matters.
The Amplifier Part
The core of the effect is two stage amplifier. This one in particular is a bit strange (or rather unfamiliar to me) – it uses shunt-series negative feedback (through R3 and RV1 pot). Apparently – this is great for current amplifiers since it drives input impedance down and output impedance up. This goes contrary to the rule of the thumb that input impedance should be as high as possible and output impedance as low as possible for a guitar effect.
Two stage amplifier is normally used to generate more gain. BC109C and similar have lots of gain already (between 420 and 800 according to the datasheet). The 2nd amplifier would multiply the gain of the first one, but negative feedback drives this down.
Having said all of the above, that’s not really how the circuit is used, the first stage more or less is used to generate soft clipping (or no clipping at all for lower signal/low fuzz setting) and drive the 2nd stage into very hard clipping that translates into our lovely fuzz sound 🙂.
Amplifier Stage 1
The first stage of the amplifier is in common emitter configuration, despite not looking like it that much. Since there is no degenerative resistance at emitter, the gain would be (based on biasing) about 300 (if my calculation is correct). Negative feedback, introduced through R3, lowers this gain to around 200.
Fuzz control does not have that much of an impact on the gain of the amplifier. The gain is between roughly 150 and 200 (43.5-46dB), so relatively modest impact of the control on the gain. The biggest purpose (if we look very simplistically I suppose) is to act as a sort of a reference voltage to the input signal to keep the transistor on.
The transistor will be working as a linear amplifier for small input signals – single coil pickup, single note softly played for example; but, it will start clipping as soon as chords or stronger playing is involved.
The biasing is such that collector voltage of the transistor is at around 1.3-1.4V. Since this is not close to the half of the 9V power source, the clipping will be asymmetrical.
Amplifier Stage 2
The 2nd stage is a bit more conventional common emitter configuration. The gain goes from the minimum of about 8.5 to the maximum of about 230 (18.6-47dB).
What is not that conventional is that it is directly coupled with the previous stage. There’s no coupling capacitor, which is fine, but since base is biased at around 1.3V there is very little upward swing possible because ultimately the voltage will be limited by VBE of the first and VBE of the second transistor – and that is nearly a constant value.
The collector voltage of the second transistor is biased at around 3.4V but output is not taken from there. The output is taken from between R1 and R4 resistors, and that point is biased at around 8.7V which gives very little room before it starts clipping.
The second stage is driven harder than the first and it causes asymmetric hard clipping.
The actual maximum gain of the both stages combined is (due to feedback network) about 57dB (700). Due to the biasing and clipping, the actual output is limited to around 0.3Vpp regardless of how hard the guitar is played (there’s just harsher distortion for harder playing).
Notes and Other Reads
The above is super simple overview how the circuit works. For way better and more in-depth analysis see ElectroSmash and geofex web sites. Just keep in mind that they are mostly focused on germanium transistor versions, but the analysis is nearly the same.
Due to differences in gain and slightly different operating characteristics between silicon and germanium transistors, the various values might be different if you look at the other analysis. For example, for a sweet spot for gain for germanium transistors 70-130 is mentioned. Silicon transistors in my analysis have way higher gain (at least 400).
Any silicon transistor (at least based on my SPICE analysis) with gains of over 100 will work, and give a hard clipping asymmetric distortion. However, whether they are going to sound the same is a different matter. Especially since transistors work in saturation/cut-off for large portions of the input signal.
As mentioned above, nearly every component in the circuit has impact on the behaviour of all other components, so experimenting with it could be fun, but also down-the-rabbit-hole experience.
Let’s simulate the effect in SPICE, here’s the schematic:
And here’s the input signal, power and simulation commands:
Here’s the SPICE file if you want to give it a try:
Frequency response is looking like this:
As mentioned above, there’s a cut-off frequency roughly between 70Hz and 150Hz. Then there’s another cut-off again well over 20KHz. This somewhat depends on level of fuzz control. Volume control just lowers the output and has very little effect on frequency response.
The main thing to look at is actual transient analysis diagram that will give us the looks of the output signal and how distorted it is:
Here’s another transient diagram that shows side by side input, output signal and also signal before going through output coupling capacitor (in red):
That signal in red shows the hard clipping, green diagram shows that there is some slight filtering on the output. The comparison shows difference between high fuzz setting (left) and low fuzz setting (right).
The above is actually with input level of 10mV peak (20mVpp) which is a bit low – say single coil pickup, single note payed softly. If I plug in something hotter – like a humbucker (single note, normal attack, or single coil, a chord strongly played), here’s how the output looks like:
High fuzz setting on the left and low fuzz setting on the right again.
Characterizing the Effects
Before I dive into breadboarding stuff, let’s get some basic characteristics for the effect:
|Parameter||Fuzz Face (Silicon BC109C)|
|500-900 ohm (depends on fuzz control)|
|0-120 Kohm (depends on volume control)|
|Current Draw||1 mA|
All values I got through simulation and all are approximate.
For way better and more in-depth analysis see ElectroSmash and geofex web sites. And more info on Fuzz Central. And even more info on Analog.Man – with details on various transistor options and their effect on sound.
Nice getting started with LTSpice tutorial from Sparkfun.