How does a fuzz pedal make fuzz?

[Meehah]

I've been building a few fuzz pedals lately and am pretty confident building them. I understand the basics and a bit more of the parts work. However I can't understand how the fuzz pedals work. I understand how the diodes clip but not the all transistor ones like the Fuzz Face, Fuzzrite, Tone Bender and all the others work. I understand that it comes from the transistors over driving but I don't understand why the transistors are overdriving in the first place. I've done experiments with the circuits and read and watched everything I can find about it but none of if it explains this. To sum up, how are the transistors overdriven to distort the sound?

[bluebunny]

Read R.G.'s "Technology of the Fuzz Face".  It explains how the circuit works and should answer your questions about how the transistors are being used.

[R.G.]

Maybe this will help.

1. Fuzz is distortion, generally clipping of the tops and/or bottoms of a signal.
2. Distortion is heard by the human ear as a blast of additional frequencies that were not in the original signal. How and why this happens is a long answer by itself.
3. To cause clipping, you drive a circuit with a signal bigger than it can respond linearly to.
4. For diode clipping, you drive a diode with a signal bigger than the diode turn-on voltage, about 0.6V for silicon, about 0.2 to 0.3V for germanium.
5. For transistor-only distortion, you drive the circuit with a signal that, when multiplied by the transistor circuit gain, is bigger than the transistor-circuit's output can swing.

For instance: if you have some transistor circuit which is powered from 9Vdc, it cannot provide an output bigger than 0V to 9V because that's all the power supply will let it swing. In practice, it will be to some degree smaller than the power supply, depending on the circuit's design. If this circuit has a gain of 10, and it can't put out more than 9v total of output signal, then any input signal bigger than 9V/10 = 0.9v will cause it to distort. There will be some degree of fuzz.

This is why you hear people talk about "high gain" distortion circuits. The bigger the gain, the smaller the signal that can cause distortion, so the more distortion you get from the same small signal.

[Plexi]

Don't believe what R.G. says...
Fuzz is made with magic! 

[teemuk]

Add to R.G.'s answer, distortion can have different "timbres" depending on amplitude of added harmonic frequencies. e.g. asymmetric clipping sounds different than symmetric clipping, soft clipping sounds different than hard clipping, etc.

Intermodulation distortion of complex signals has great effect to this timbre because it introduces majority of distortion's harmonics. To shape timbre it has been customary to "pre-emphasize" the signal before clipping distortion is introduced so that intermodulation distortion has less effect at certain frequencies, and perhaps more on others.

Modern "hi-gain" tones are usually created by radically hi-pass filtering the signal around 1kHz. This reduces effect of IMD introduced by frequencies below the cut off point and produces "tight" and "pronounced" tone, typically with lots of sustain.

"Fuzz" effects are similar in that they clipping distort the signal but pre-emphasizing is less radical. More "lows" are clipped. This produces the classic "fuzzing" sound, which at very high gain levels tends to blurb out into mushy fartiness as intermodulation distortion turns the signal into incomprehensible mess.

[Mark Hammer]

I like to distinguish between overdrive, distortion, and "fuzz", based on how long the  harmonic content lasts.

A picked/plucked/strummed guitar string generates its most harmonic content and greatest amplitude in the first fraction of a second, and quickly declines thereafter.  Depending on a bunch of factors, the speed with which either the harmonic content, and/or the amplitude, decays will vary.  For instance, part of what makes big jazz boxes sound "fat" is not only their size and hollow form, but the "floating" wooden bridge that many have.  This alters both the decay in volume, and the drop in harmonic content.  A strummed chord may sustain in such guitars for a long time, but at a MUCH lower level, and with much less harmonic content, than at the moment immediately after picking.  That sharp drop makes such guitars sound and feel more percussive.

So what does this have to do with fuzz?  Simple, the overdrive/fuzz circuit merely responds to what you feed it, and the duration of any added harmonic content will depend on the properties of that input signal.

If I have a room with a 2-metre ceiling, some people walking in that room will bang their heads, just not all that many.  And those that do will quickly learn to bend a little to avoid it.  Lower the ceiling to 1.7 metres and a lot more people will be banging their heads, but not everyone.  The very tall people will likely not be able to avoid banging their head, but the people around average height will only need to bend a little to avoid it, and all the children and many women will have no problem.  Lower the ceiling still to 1.5 metres, and the majority of people will be banging their head against it.

Let's flip that around.  Make the ceiling 2 metres, but insist that everyone in the room wear shoes that raise their height by 10cm.  Headbanging ensues, but not for everyone.  Okay, now make those platform shoes 20cm, and the headaches will start for many.

Ultimately, one of the things that differentiates the sound of a pedal, and whether one labels it an overdrive, distortion, or fuzz, is going to be whether the threshold at which clipping and noticeable added harmonic content is generated, remains well above the average input signal level, or whether the gain that the circuit applies to the input signal pushes it up high enough such that, even after the string has settled down after the initial pluck, it still remains in the vicinity of that threshold.

For me, what counts as "fuzz" is something that manages to keep the string sizzling well after picking.  Overdrives will produce lots of bite in response to picking, but only add modest amounts of lower-order harmonics as the string settles down at a now lower amplitude.  That is, the manner in which we users tend to differentiate between overdrive, distortion, and fuzz seems to depend on whether the circuit's maximum impact seems to be the same a second or more after picking, or is only shortlived.

While not the whole story, the "fuzziness" of even some modest overdrives can be increased by simply boosting the level of the input signal so that it remains well above the clipping threshold for as long as one holds the note and allows the string to vibrate.  Boxes like the TS-9 only provide a maximum gain of 118x, while others (e.g., ProCo Rat) will apply gains well into the thousands.

[bluebunny]

Let's flip that around.  Make the ceiling 2 metres, but insist that everyone in the room wear shoes that raise their height by 10cm.  Headbanging ensues, but not for everyone.  Okay, now make those platform shoes 20cm, and the headaches will start for many.

This analogy is making my head hurt.      Perhaps Paul can pitch in with one that involves relatively benign tractors or fences?   

[robthequiet]

Big hair helps with the banging, witness the historic MTV Headbanger's Ball. /s

Some really great pedals do their distortion by putting a couple of FET gain stages together. So the distortion consists of amplifying in the first stage so the second stage has to cope with an input signal too big for it to render clean. According to legend, the FETs distort in a more pleasing manner to some people.

Same thing basically if you run a fully-cranked clean boost into an amp clean channel, overwhelming (the technical term is "slamming") the amp input preamp stage. The Boss FA-1 (opamp-based) will do this nicely. Check out the Brown Sound in a Box for an interesting self-contained configuration.