Input cap vs Output cap increase in a fuzz

[WolfCathedral]

Hey Guys,

Building a fuzz face and was just wondering what are the pros and cons of choosing to boost the value of the input cap vs the output cap for more bass in the pedal? I understand the input will put more bass through the distortion etc, but is there any sort of standard or reasons to choose one over the other to boost?

[thermionix]

I almost mentioned this in your other thread.  My preference is to use a larger output cap (47n) and vary the input cap, either with a pot blending two caps, or with a switch like in the picture I posted in the other thread.

For my personal FF I have the blend pot, 22n fixed and a 100kL pot blending in a 2.2u.  For the one I build for my buddy with the switch, it has 33n fixed, and the switch parallels in either another 33n, or a 1uF.

All these values were arrived at via breadboarding first.  Everyone has different ears, tastes, rigs, and every transistor set has its own sound, so the values I like might not suit anyone else.

[Mark Hammer]

I'll second thermionix's suggestion.  The output cap simply sets the bandwidth of whatever the fuzz has already done to the signal.  The input cap strongly determines what the fuzz will do to the signal.

At the same time, think of them as functioning like pre-distortion and post-distortion EQ.  Both are useful in shaping the character of what you hear.  But, since so much of the guitar signal "lives in the basement", altering how much low end and lower mids a fuzz circuit has to react to will have a more profound effect ion what is produced than a change to the output bandwidth will.

[goo]

for Fuzz Face i guess input cap should be "large" one (around 1 - 3.3uF), output - to taste (0.1 - 0.01uF).
my germanium fuzz face has two switching caps on the input (1 and 0.01uF) and fixed 0.01uF output cap.
when 1 uF is plugged in, the sound has enough bass, but with 0.01uF is almost like regular fuzz with the wah pedal all the way down.
i think if there`s no much bass on the input, then how can you generate it on the output?

[WolfCathedral]

Thanks guys. Yeah im going to breadboard it for sure. It seems like a small input cap and a huge output cap will allow for a tight sound with some thump while a large input cap will be a muddy wooly fuzz. ill have to play with it and see which i like best as both options sound sweet 

[teemuk]

With complex signals (signals consisting of more than a single frequency) you can not have harmonic distortion without having intermodulation distortion as a side effect.

Intermodulation distortion means that harmonic distortion is also produced at frequencies that are sum and difference frequencies of frequencies within the signal.

Which means:
- Magnitude of intermodulation distortion is higher the more harmonic distortion gets introduced (because harmonic distortion introduces more frequencies subject to intermodulation)
- Magnitude of intermodulation distortion is higher the wider the effective bandwidth of the distorted signal is (because greater bandwidth introduces more frequencies subject to intermodulation)

IMD in excess adds so much harmonic content to the signal that it messes "note sensitivity" (signal muddies up) and starts to sound very discordant (i.e. you can only play single notes without sounding off tune). All this sounds especially bad towards lower frequencies. The tonel becomes just sputtering and "farting".

As implied, shaving off low end response (pre distortion) reduces magnitude of introduced IMD at lower frequencies, and this sounds different than distortion without that reduction. Usually "tighter" and less "muddied up". So if you want to reduce IMD you lower coupling cap values feeding the clipping stages. Coupling caps in output can only affect low end response but they have no effect to intermodulation that happens within that clipping.

Since harmonic distortion and IMD go hand to hand you can't have one without another. So it becomes a highly subjective fine line of preference how much IMD one tolerates in the signal. If we reduce low end too much (pre distortion) the output signal will start to loose its "fuzzy" characteristic (which one likely wishes to preserve in a FUZZ-type effect) , if we reduce overdrive below certain threshold the signal doesn't sound distorted and sustained enough.

Therefore the proper capacitor values are usually derived by experimentation. Traditional "fuzz" -type tone is function of distinct amount of low frequency IMD. If you reduce it too much the effect no longer sounds like fuzz but begins to sound more like those overdriven guitar tones that were discovered after the fuzz craze of early 1960's was over.

[Quackzed]

an 'input cap blend' pot is a really usefull thing to add to a fuzz, go from tight trebly distortion right to wooly thick fuzz... 

[TejfolvonDanone]

i think if there`s no much bass on the input, then how can you generate it on the output?

It's not "adding in" it's "not filtering out". These are passive filters so there is no boosting in any frequency. If you have a simple high pass filter on the input (that's basically the input cap) the bass won't completely disappear just get damped. Then you amplify it and you filter with the output cap again.
It also don't generate anything (except some IMD) which wasn't in the original signal.

[goo]

It's not "adding in" it's "not filtering out"

so, i could put 10nF input cap and 1uF output cap and the sound should be just like 1uF input cap and 10nF output cap?

[blackieNYC]

Some good points!  Likes all around.  A simple rc filter rolls off a 6dB per octave, which isn't really all that steep, so the frequencies just below the rolloff frequency are somewhat retrieveable. Somewhat, in that rolling off some highs at the output (very common) and making up output volume with a tweak of the output level,will sound like the lows are partially restored. 

Regarding IMD (great post), I find the low frequencies that are created (some of the frequency differences mentioned above) to be the most offensive. These are the beat frequencies - below 20Hz, but also above. These can sound like fluttering and sputtering, and can make you sound out of tune - or at least make you question your intonation.  In my opinion, the output cap can be used to reduce this particular kind of IMD.  You can filter them off at the input, but again it's only 6dB per octave (unless you do a double or two-pole filter, which will also cause some full-spectrum signal loss and affect the fuzziness).
So, we can filter at both in and out. And at the output, you can add a two-pole filter and still have plenty of output volume. In most designs - you know the ones where you never turn the output volume past 11 o'clock.
I'm going to put my money where my mouth is, and make a fuzz with the variable capacitor blend pot at both in and out. 

[blackieNYC]

It's not "adding in" it's "not filtering out"

so, i could put 10nF input cap and 1uF output cap and the sound should be just like 1uF input cap and 10nF output cap?

No.  You had it right the first time - the output can't generate anything that isn't present at the input. With a slight sorta-exception in what I mention above regarding output gain makeup. The big input cap will allow low frequencies to be clipped - generating harmonics of, say, your E string. The signal now has a strong presence of the harmonics of 80Hz and 160 Hz - lots of 160 and 320Hz in the signal.  The small output cap can filter out that 80Hz, and some of the 160Hz, but that newly created 320Hz is going to come thru. Flip the in and out caps Large v. Small, and this 320Hz signal is never created, because the 80 and 160Hz content was never there to generate these harmonics. 
This is a bit oversimplified, but quite a significant factor in fuzz.

[TejfolvonDanone]

It's not "adding in" it's "not filtering out"

so, i could put 10nF input cap and 1uF output cap and the sound should be just like 1uF input cap and 10nF output cap?

No.  You had it right the first time - the output can't generate anything that isn't present at the input. With a slight sorta-exception in what I mention above regarding output gain makeup. The big input cap will allow low frequencies to be clipped - generating harmonics of, say, your E string. The signal now has a strong presence of the harmonics of 80Hz and 160 Hz - lots of 160 and 320Hz in the signal.  The small output cap can filter out that 80Hz, and some of the 160Hz, but that newly created 320Hz is going to come thru. Flip the in and out caps Large v. Small, and this 320Hz signal is never created, because the 80 and 160Hz content was never there to generate these harmonics. 
This is a bit oversimplified, but quite a significant factor in fuzz.

I'd just like to add that with a clean boost it would be a yes. Because there is no distortion (which can be thought as adding HIGHER harmonics to the signal). But because it's a fuzz there is a really big distortion and blackieNYC explained it perfectly.

Also little addition to the IMD discussion:
If you have perfectly harmonic content in the original signal (all of the higher harmonics' frequencies are the base frequency times an integer) then the IMD content will be perfectly harmonic as well. The little detail which is really significant in this case is that the guitar signal is not like this. Every harmonic is slightly off and that's why IMD can sound really muddy.
In a hypothetical world where instead of guitar everybody uses cello (with a bow) and use distortion with it there wouldn't be any problem with IMD.
Sorry for the off topic but i felt a little itch in the back of my brain to mention this. 

[goo]

2 blackieNYC:
yep, that`s what i was thinking about.