Fuzz Pedal

[Arch]

Hi, I am trying to build a simple fuzz guitar pedal as shown here:


I am using BC109 transistors, I need to check them to make sure I haven't destroyed them but so far things aren't working out for me. The waveform is expected to look more like this (or at least just the top half:


I am looking at it through an oscilloscope with a function generator outputting a sine wave and I am getting an input/output like this:



Now, it's not like the square-ish wave I was expecting. I am wondering if anyone has any hints based on the output waveform for what might be wrong, or expected wave forms at certain points in the circuit?

I will have an original one to compare against soon but I don't have it here at the moment
When I put my guitar through it - it makes a sound when I hit the string hard and it tends to fade right after. Also the fuzz pot doesn't seem to change the wave much
Tips? Tricks?

[AlfBaz]

That doesn't seem like a very good circuit. An electric guitars pick up has a very high output impedance, which means any current you try and take out of it will kill the signal. Secondly a passive pickup doesn't produce a very large signal certainly not enough to forward bias the base emitter junction of that input transistor without smashing the crap out of the strings.

One thing you could try to further your investigations is pre-amplifying the signal before putting it through that cct

[Owen]

Now, it's not like the square-ish wave I was expecting. I am wondering if anyone has any hints based on the output waveform for what might be wrong, or expected wave forms at certain points in the circuit?

That looks OK so far. Take a look at your input signal it's 400 mV peak-peak. That's way beyond an electric (passiv) guitar pickup. Typically a passiv pickup is round about 10 mV. If you reduce your input signal and select a 1x Probe you should see that squarewave signal.

[Shock]

I simulated your circuit and got the same.

I simulated this circuit which the author says is based on the original (I see the same circuit listed on the page where you got yours).  It uses positive ground and PNP germanium transistors,  I used two BC177 transistors and got an output similar to what you were expecting.  When you get that pedal you can easily check.  Make sure you don't daisy chain a negative ground pedal to positive ground pedal with the same external DC supply though.

[wiss]

Now, it's not like the square-ish wave I was expecting. I am wondering if anyone has any hints based on the output waveform for what might be wrong, or expected wave forms at certain points in the circuit?

That looks OK so far. Take a look at your input signal it's 400 mV peak-peak. That's way beyond an electric (passiv) guitar pickup. Typically a passiv pickup is round about 10 mV. If you reduce your input signal and select a 1x Probe you should see that squarewave signal.

Peak output from a passive humbucker can be several volts, even 10 V, into 10 Meg. Source impedance is something like 1 to 10 k at DC, lots of inductance  and a resonance at a few kHz.

[old gregg]

you waveform looks actually good.

The fuzz  pot doing nothing is normal, the original does the same. Put a 1K resistor instead and put a 100K pot in variable resistor at the input (way better).

for the fartty decay sound, check the bias at Q2 collector, it should be Vcc/2 (putting a trimpot is a better option)

[Owen]

Peak output from a passive humbucker can be several volts, even 10 V, into 10 Meg. Source impedance is something like 1 to 10 k at DC, lots of inductance  and a resonance at a few kHz.

Yes but peak doesn't mean continues like he's putting into the circuit via his signal generator. The avarage is just a few mV.

[deth502]

Peak output from a passive humbucker can be several volts, even 10 V, into 10 Meg. Source impedance is something like 1 to 10 k at DC, lots of inductance  and a resonance at a few kHz.

Yes but peak doesn't mean continues like he's putting into the circuit via his signal generator. The avarage is just a few mV.

yes, hence the statement in the op:

When I put my guitar through it - it makes a sound when I hit the string hard and it tends to fade right after.

[John Coloccia]

FWIW, everyone does positive ground fuzzes because negative ground fuzzes tend to have terrible oscillation problems.  One solution is to use a charge pump to get -9.  Then you don't have to worry about having isolated power supplies so long as you're careful to isolate the circuit.

[old gregg]

putting a charge pump is overkill, a negative ground version works perfect (I built several variation) with proper decoupling cap and feedback cap in the B/C path of the transistors (avoiding High HFE transistor too).

[Arch]

A little bit more info, my dad has an old arbiter fuzz face, one of the ones which seems like from right after they switched to the silicon NPN transistors and I am trying to replicate it. Was supposed to have it here today but I forgot to take it with me when I left  so I will have to wait two more weeks

That doesn't seem like a very good circuit. An electric guitars pick up has a very high output impedance, which means any current you try and take out of it will kill the signal. Secondly a passive pickup doesn't produce a very large signal certainly not enough to forward bias the base emitter junction of that input transistor without smashing the crap out of the strings.

One thing you could try to further your investigations is pre-amplifying the signal before putting it through that cct

The thing is, it works. That's how they were built 

Now, it's not like the square-ish wave I was expecting. I am wondering if anyone has any hints based on the output waveform for what might be wrong, or expected wave forms at certain points in the circuit?

That looks OK so far. Take a look at your input signal it's 400 mV peak-peak. That's way beyond an electric (passiv) guitar pickup. Typically a passiv pickup is round about 10 mV. If you reduce your input signal and select a 1x Probe you should see that squarewave signal.

I am trying to understand my measurements. If I connect my multimeter to the other end of the guitar cable I can even get readings up to 1.4v (2.8vpp) hitting all the strings. Is this the incorrect way to measure something like this? Is this like when I connect my function generator directly to the oscilloscope and the volts are way off? My intentions were to measure the guitar, often saw values from 200-400mv so I set my function generator to 400mvpp and put it as the input of my circuit

I simulated your circuit and got the same.

I simulated this circuit which the author says is based on the original (I see the same circuit listed on the page where you got yours).  It uses positive ground and PNP germanium transistors,  I used two BC177 transistors and got an output similar to what you were expecting.  When you get that pedal you can easily check.  Make sure you don't daisy chain a negative ground pedal to positive ground pedal with the same external DC supply though.

I am just using 9v batteries for now  I haven't got all the risky connection combinations figured out yet so batteries will do for now.
So by simulating what I have, you think this is working correctly? I will definitely compare to the real one but there is some sort of fading happening where I hit a string hard and it makes a sound and then almost immediately cuts out in a soft way. I wonder if the output is much higher than the real one and the amp is doing something special. All just speculation until I can compare I guess

you waveform looks actually good.

The fuzz  pot doing nothing is normal, the original does the same. Put a 1K resistor instead and put a 100K pot in variable resistor at the input (way better).

for the fartty decay sound, check the bias at Q2 collector, it should be Vcc/2 (putting a trimpot is a better option)

I can take a look at Q2. What is the reasoning for Vcc/2? Is there some theory behind that? Or is this from experience?

[wiss]

A little bit more info, my dad has an old arbiter fuzz face, one of the ones which seems like from right after they switched to the silicon NPN transistors and I am trying to replicate it. Was supposed to have it here today but I forgot to take it with me when I left  so I will have to wait two more weeks

That doesn't seem like a very good circuit. An electric guitars pick up has a very high output impedance, which means any current you try and take out of it will kill the signal. Secondly a passive pickup doesn't produce a very large signal certainly not enough to forward bias the base emitter junction of that input transistor without smashing the crap out of the strings.

One thing you could try to further your investigations is pre-amplifying the signal before putting it through that cct

The thing is, it works. That's how they were built 

Input impedance for guitars are in the range of 100 k to 10 Meg.
Distortion pedals can have quite low input impedance, this will result in a loss of high frequency contents, which may very well be part of the "sound" of the pedal.

Now, it's not like the square-ish wave I was expecting. I am wondering if anyone has any hints based on the output waveform for what might be wrong, or expected wave forms at certain points in the circuit?

That looks OK so far. Take a look at your input signal it's 400 mV peak-peak. That's way beyond an electric (passiv) guitar pickup. Typically a passiv pickup is round about 10 mV. If you reduce your input signal and select a 1x Probe you should see that squarewave signal.

I am trying to understand my measurements. If I connect my multimeter to the other end of the guitar cable I can even get readings up to 1.4v (2.8vpp) hitting all the strings. Is this the incorrect way to measure something like this? Is this like when I connect my function generator directly to the oscilloscope and the volts are way off? My intentions were to measure the guitar, often saw values from 200-400mv so I set my function generator to 400mvpp and put it as the input of my circuit

The volts as measured by your scope (dmm) is correct, it is the generator that is unaware of the particular load present.

I simulated your circuit and got the same.

I simulated this circuit which the author says is based on the original (I see the same circuit listed on the page where you got yours).  It uses positive ground and PNP germanium transistors,  I used two BC177 transistors and got an output similar to what you were expecting.  When you get that pedal you can easily check.  Make sure you don't daisy chain a negative ground pedal to positive ground pedal with the same external DC supply though.

I am just using 9v batteries for now  I haven't got all the risky connection combinations figured out yet so batteries will do for now.
So by simulating what I have, you think this is working correctly? I will definitely compare to the real one but there is some sort of fading happening where I hit a string hard and it makes a sound and then almost immediately cuts out in a soft way. I wonder if the output is much higher than the real one and the amp is doing something special. All just speculation until I can compare I guess

you waveform looks actually good.

The fuzz  pot doing nothing is normal, the original does the same. Put a 1K resistor instead and put a 100K pot in variable resistor at the input (way better).

for the fartty decay sound, check the bias at Q2 collector, it should be Vcc/2 (putting a trimpot is a better option)

I can take a look at Q2. What is the reasoning for Vcc/2? Is there some theory behind that? Or is this from experience?

Having that node voltage halfway maximizes the available output swing, if you have a pot there you could adjust for that parameter instead (viewing on a scope).
The crappy sustain might very well be due to wrong biasing somewhere, to low voltage on a collector or a base for example, probe around the circuit with a scope and a low level input, 1 to 10 mV. There should be a fair amount of gain.
The "square" will not look like a square, it is high-pass filtered!

[John Coloccia]

Crackling followed by loss of volume/sustain is almost always a bias issue.

[Paul Price]

The fuzz pot has to have some big effect on the sound  or you have a wiring problem around the pot, a short somewhere.
Make sure you have the polarity correct of the electrolytic capacitors in your circuit.

Check the battery. The overload cutting out of sound might just be due to a bad (weak) 9v battery.

For the circuit to work properly, the base of the input transistor must be biased at approx + .55 to .65V. This requires about .6mA to flow through the fuzz pot (1k) to develop this bias voltage. So, .6ma through the 8.2k resistor creates approx.  5V drop and then the second  transistor will have an optimal (Vbat/2)  +- voltage swing  possible with the 9V battery and give you the best and symmetrical squared waveform fuzz sound. The output cannot be anything but a distorted squarewave except at the highest freq. tones from the guitar. This is due to the small valued capacitor (.1uf) feeding the output.

The idea here is to over amplify the input signal from the guitar to create an output signal to your guitar amp that is somewhat square and full of harmonics(fuzz), but still retains some symmetry of the input signal in the output +- waveform excursions, and this symmetry gives the smoothest fuzz sound, a dirty clipped distorted squarewave.

Guitar pickups can put out several volts, the amplitude is directly proportional to the applied violence upon the strings and also the tonal frequencies. Feedback from your monitor speakers to your guitar strings/pickup sustains the fuzz effect once initiated by a finger or a plectrum.

[Arch]

Been pretty busy, but I think I sorted all my problems out. The output of the guitar was too high and caused the fuzz pot to not make a difference. Lowered my volume on the guitar and it seems to work right. The waveform also seems to match the original one for the most part but can match with a little bias adjustment on the first transistor

[vvangelovski]

That doesn't seem like a very good circuit. An electric guitars pick up has a very high output impedance, which means any current you try and take out of it will kill the signal. Secondly a passive pickup doesn't produce a very large signal certainly not enough to forward bias the base emitter junction of that input transistor without smashing the crap out of the strings.

One thing you could try to further your investigations is pre-amplifying the signal before putting it through that cct

That's actually part of the design and the specific sound of the fuzz face - the low input impedance. I don't know if it was done by accident or on purpose, but it is a "design feature" of the circuit. You can read about it here: http://www.geofex.com/article_folders/fuzzface/fffram.htm

@Arch the circuit is sensitive to the hfe - current gain factors of the transistors you're using. Specifically it want's Q1 to be of lower hfe than Q2. From experience building different fuzz circuits, the BC108/BC109 silicon transistors can vary in hfe quite a bit, although they will always have adequate minimum hfe and no leakage as opposed to GE transistors. What is recommended for GE FF is Q1- hfe~70-90 Q2 - hfe~110-130. I've found that it helps hand pick the transistors for hfe to get a good fuzz sound. The way I measure them is with a Peak Atlas DCA55, but you can also build your own circuit for measuring hfe and leakage, just google for it.

[old gregg]

lot of old guitar pedals have a low input impedance. Don't know if it was made on purpose but what people designated as "Warm" is that high frequency cut off for those pedal. Put a buffer in front of it and all this goes away.

Never thought I'll be talking stompboxes on the eev blog. Seems that artistic/music isn't so far away from engineering, cool !!

[AlfBaz]

I can't help but feel we're drifting into the world of audiophoolery here

The same "fuzz" sound can be easily achieved by simply overdriving a 741, it's simply a matter of driving the thing into saturation giving what is essentially a square wave output

My old fender twin reverb and most marshalls have an input impedance of 1M, Boss overdrive pedals, 1M. Copious effects gear, 1M

To change the "tone" of the guitar's signal by "overloading" the pickup is just stupid. The vibrating string is the source of energy and by loading the pick up you are essentially dampening the string and essentially pissing away any tonal qualities you may have gained from things like paying all that money for pickups wound by nude Californian virgins, trying all the different brand strings, paying top notch for guitars made from dense wood's, brass nuts, hundred dollar guitar leads etc, etc...

I remember guitar seminars with some locally famous guitarists years ago. What a lot of them tried was to match impedance for max power transfer. No idea what effect this had on the sound as their rigs where huge (back in the day where it was hip to have 10 unit high rack mount effect units)

And if having low input impedance load to your guitar is so "cool" then what's the obsession with tubes? Isn't it supposed to be all about their high input impedance?

[Nerull]

If you're using a fuzz pedal, you aren't going for audio quality. If you have amps, I would expect you would grasp that distortion pedals are not used to produce pristine, clean audio.

[AlfBaz]

If you're using a fuzz pedal, you aren't going for audio quality. If you have amps, I would expect you would grasp that distortion pedals are not used to produce pristine, clean audio.

No but if you want punch, sustain and ear splitting harmonics...

I guess like all things to do with sound, it's always subjective

[dfmischler]

If you're using a fuzz pedal, you aren't going for audio quality.

Audio quality is in the ear of the musician, in this case.  Granted that he is not looking for what an audiophile wants (flat response, no distortion or compression).  But anyone who has played with a guitar and a nice tube amp at levels that push the power tubes into saturation knows that there is a desired quality that is pretty hard to achieve with a few transistors (the harmonics don't seem to come out right, etc.).

[AlfBaz]

knows that there is a desired quality

Argh, I clicked it fully expecting to hear a nice full sounding power chord, instead I got words... I think I've just been mooned

[dfmischler]

There are audio links further down on that page.  Read down and click 'em.

[AlfBaz]

There are audio links further down on that page.  Read down and click 'em.

To late, had to scratch that itch now got satriani blaring away 

[Nerull]

If you're using a fuzz pedal, you aren't going for audio quality.

Audio quality is in the ear of the musician, in this case.  Granted that he is not looking for what an audiophile wants (flat response, no distortion or compression).  But anyone who has played with a guitar and a nice tube amp at levels that push the power tubes into saturation knows that there is a desired quality that is pretty hard to achieve with a few transistors (the harmonics don't seem to come out right, etc.).

While that may be the case, a fuzz face is not designed to produce anything like a nice warm tube overdrive. The silicon version even more so, think 'Spanish Castle Magic'. Loads of gain and a very harsh sound. Complaining that the pedal design isn't optimum for sound quality is kind of missing the point.

[vvangelovski]

What's specific about the fuzz face (especially the germanium version) is it's ability to clean up with the guitar volume knob and playing dynamics especially on single coil pickups. Yes a full on fuzz sustaining for days can be achieved with an opamp, and there are lots of them on the market. Believe me I've tried to blind test different circuits with 2-3 experienced guitar players, they know the difference, at least they know what they like more. So besides the harmonics content when it comes to fuzz people are either after soaring endless sustain or sensitivity to playing dynamics, or somewhere in between. That's why you have a bazillion fuzz circuits and really they do sound different, at least when you're playing through them.

[old gregg]

The same "fuzz" sound can be easily achieved by simply overdriving a 741, it's simply a matter of driving the thing into saturation giving what is essentially a square wave output

There's still differencies in sounds. I've never liked op-amp distortion whereas I really appreciate fuzz face and fuzz (certainly to do with the internal Op-Amp construction, push pull and crossover distortion and so on). Tube vs transistors is the same thing, it's the way they distort and messed up the sound.

 
about loads, it's just the way it way made in that time. The famous Clapton treble booster works the same. Rise the input impedance (or put a buffer in front of it) and the sound's horribly ear-piercing. Don't know why they didn't use higher input impedance, it seems absurd today, maybe it was a cost thing or it was intentional 

What's more, some guitaristes (Warren Haynes from Gov't mule for instance) accumulate meters and meters of cable to tame high-end on his huge rig. Ok, this is a bit silly, a RC filter will suffice but what your ears like isn't a math equation.

In the same field, a funny thing to do : go to a guitarist forum, start a thread "what's the best buffer I could get" and watch the show