Sustain Problem Fuzz Pedal Design

[camzilla]

Hello,
I have been attempting to design my own fuzz pedal schematic based on the topology of the traditional transistor Big Muff Pi but using op amps, with the idea to be to land with a sound somewhere between the pumpkin big muff and the normal one. I have made a schematic and breadboarded it and I am happy with the sound but high notes do not sustain barely at all. The open low e string will ring out for about 15 seconds, but high notes like the 12th fret on the D,G,B,E strings only ring out for about 2 seconds before it sounds like they are swallowed by noise and it stops sustaining. If I put the breadboarded pedal after my compressor pedal it sounds great and all notes can sustain almost indefinitely fading smoothly into feedback and its great, but I would like the fuzz pedal to stand on its own. I am attaching the LT spice model of my circuit to this post. Would love any help to figure out how to improve this circuit and the sustain on the pedal.

I have tried increasing the gain of the input amplifier and of the clipping stages and these changes only seemed to make the circuit noisier and did not have significant affect on the sustain.

Please let me know if there is any more info I can give.

[Rob Strand]

Try a 1uF cap across R3.   

You could even drop the values of R2 and R3 down to 10k and use say a 47uF to 100uF cap.

FYI, R7 and R20 are not required as the inverting inputs get the DC bias from the feedback resistors.

[antonis]

Hi & Welcome..

Neither Sustain nor Volume pots work as they should to (voltage dividers..)
Both affect the gain of respective stage..

First stage exhibit a max gain of 22X (not 100X)..
I'd set R5 to a proper value (gain = 1 + R5/R6) and place a a Sustain pot on U1 out..
(like original Big Muff configuration, with R12 between lug1 and GND acting as CCW limiting resistor)

[Dormammu]

Redesign the bias network as suggested by Rob Strand. R7, R20 is not needed. The output amplifier also can be thrown away, leaving the just volume pot.
The first stage can be left as is, it does enough for its position.
You can choose a lower R6 for a slight increase in gain, between 1-3 kΩ.
It is also be good to add a series resistor to the input of the first stage op-amp, 1-5 kΩ, which will work better for interference suppression.

[antonis]

The output amplifier also can be thrown away, leaving the just volume pot.

And all the Tonestack loss..

[Dormammu]

The output amplifier also can be thrown away, leaving the just volume pot.

And all the Tonestack loss..

Nope. It will be sound different, but who say — bad ?
I built near 30 overdrives without overstaffing them with buffers and they all sound good.

[antonis]

I built near 30 overdrives without overstaffing them with buffers and they all sound good.

Good for you..!!!

P.S.
Last stage is a booster (for Tonestack loss recovery) rather than a buffer but let it be..

[FiveseveN]

but who say — bad ?

Mathematics.

[camzilla]

Thank you guys all for the quick and helpful responses!
I have been working on it this morning according to your suggestions.
I don't know if multiple post replies or individual is the proper etiquette here so I am just going to do one massive reply post to all suggestions.

Try a 1uF cap across R3.   

You could even drop the values of R2 and R3 down to 10k and use say a 47uF to 100uF cap.

FYI, R7 and R20 are not required as the inverting inputs get the DC bias from the feedback resistors.

The 1uF capacitor across the 2nd biasing resistor, R3, drastically decreased the noise. I was looking at the power supply ripple in my simulation and wondering how to negate it, I can't believe I didn't think of a filtering capacitor, super helpful! While breadboarding this I discovered an embarrassing breadboarding error that made my bias voltage around 5.5V instead of 4.5V, resolving this also helped the circuit sound better. Lastly, I removed the unnecessary biasing resistors R7 and R20. I was treating the feedback resistors and resistors to ground as a voltage divider so thought I needed to bias there, but was somehow missing that the capacitor blocks the DC so no current flows. It makes sense to me now how the biasing works there. Thank you so much for the suggestions!
If you don't mind me asking I do have a question about the filtering arrangement. You say if I reduced R3 down to 10k I should raise the value of the filtering capacitor significantly. R3 and the capacitor don't seem to form a frequency filter, what reasoning is there to increase the value of the filtering capacitor if I reduce R3?

Neither Sustain nor Volume pots work as they should to (voltage dividers..)
Both affect the gain of respective stage..

First stage exhibit a max gain of 22X (not 100X)..
I'd set R5 to a proper value (gain = 1 + R5/R6) and place a a Sustain pot on U1 out..
(like original Big Muff configuration, with R12 between lug1 and GND acting as CCW limiting resistor)

I apologize for the labeling error, that was a relic from an earlier draft of the circuit, but I do now have the gain at 22x intentionally, 100 times was very very noisy and seemed impractical. I am aware that I made the controls for volume and sustain controlling gain of the amplifiers, I thought that it would have the same effective function while giving a wider range of possible values than a passive setup, and I am using a TL074 chip so I have the extra amplifiers there already. I do know that the output amplifier has the downside of potentially adding its own clipping if the gain is set too high. Besides being different from the big muff, should this setup work in theory? Or does having the load of a passive potentiometer on the circuit improve performance in some way?
You also mention that the tonestack is very lossy, and I was also concerned with that in my LTspice simulation, but I was not clever enough to find a way to improve it. My tonestack is directly copied from the traditional Big Muff Pi, just changed the filter values to equivalent filters so I could use capacitors I had on hand. I suppose I could lower the value of the potentiometer to reduce the load on the signal, though I think this would also decrease the range of the tone control. How would you suggest improving upon this?


While troubleshooting I plugged the output jack into the output of the input amplifier to see if the sustain issue arises there or in the clipping stages and the problem was present there too, so my problem is probably in the input stage. When I isolated this stage it was barely louder than the raw guitar signal, despite being theoretically 20 times gained. I assume that there must be some problem in me being dumb and breadboarding the circuit wrong or a connection coming loose in the breadboard since the connections have become quite finnicky after all of my fiddling. I will re-breadboard the circuit again and see where that takes me.

Once again thanks all for the suggestions! This has been very helpful and if anyone knows any other improvements or catches any errors please let me know.
Here is an updated schematic after the changes I have made today:

[antonis]

You also mention that the tonestack is very lossy, and I was also concerned with that in my LTspice simulation, but I was not clever enough to find a way to improve it. My tonestack is directly copied from the traditional Big Muff Pi, just changed the filter values to equivalent filters so I could use capacitors I had on hand. I suppose I could lower the value of the potentiometer to reduce the load on the signal, though I think this would also decrease the range of the tone control. How would you suggest improving upon this?

It's a passive Tonestack..
Make make R18 47k and don't bother anymore..
(15dB boost should be fine..)

[Rob Strand]

The 1uF capacitor across the 2nd biasing resistor, R3, drastically decreased the noise. I was looking at the power supply

With the previous connection it's not only noise from the PSU rail but also the potential for oscillations since any connections to Vref could modulate the Vref and feed into other parts of the circuit.

Was there any improvement in the sustain on the higher notes.

If you don't mind me asking I do have a question about the filtering arrangement. You say if I reduced R3 down to 10k I should raise the value of the filtering capacitor significantly. R3 and the capacitor don't seem to form a frequency filter, what reasoning is there to increase the value of the filtering capacitor if I reduce R3?

For your circuit as shown the filtering/low impedance Vref circuit is the most important aspect.  You can use 1M+1M+1uF and all will be fine.  The main parameter is the RC product which determines the Vref filter cut-off frequency.   

In general however, when there are more connections back to Vref the 1M+1M+1uF doesn't represent low enough impedance.  That's where the lower 10k version can help.  For example, in some cases you might connect R6+C2 back to Vref instead of ground,  so now the Vref needs to have a low impedance compared to R6+C2.   There's many other instances where things can connect back to Vref.   Another form of connection is where the circuit feeds a DC current into Vref, that will upset the Vref voltage.  Lower Vref resistors will help reduce the change in Vref and in extreme cases you need to buffer Vref with an opamp.

[antonis]

R3 and the capacitor don't seem to form a frequency filter, what reasoning is there to increase the value of the filtering capacitor if I reduce R3?

But they DO form a LPF..!!
(of 1/2πRC cut-off frequency, where R = R2//R3)

Even if R3 didn't involve in LPF calculation, by decreasing its value you should also decrease R2 value (for retaining Vcc/2 bias voltage) hence increase cap value..

P.S.
I think you'll find below application note interesting enough..
https://www.analog.com/media/en/technical-documentation/application-notes/AN-581.pdf?doc=AD7699.pdf

[Dormammu]

I built near 30 overdrives without overstaffing them with buffers and they all sound good.

Good for you..!!!

P.S.
Last stage is a booster (for Tonestack loss recovery) rather than a buffer but let it be..

How then does this circuit work, without any boosters?  And yes, I add a tonestack before the volume.

[Dormammu]

Once again thanks all for the suggestions! This has been very helpful and if anyone knows any other improvements or catches any errors please let me know.
Here is an updated schematic after the changes I have made today:

I completely lost out of sight — I would increase R9 to 200k and R11 to 200-500k.
It also seems an interesting idea to remove the diodes from the 2nd clipping stage.

[antonis]

How then does this circuit work, without any boosters?  And yes, I add a tonestack before the volume.

Did you, by any chance, also see Elvis next day..??

P.S.
@Dormammu:
I clearly told you that without any recovery stage (booster or anything else) after a passive Tonestack, any loss will pass to output and further dominated by next stage potential low input impedance..
You clearly stated that a passive Tonestack right after a hard clipping configuration and before a low value Volume pots works just fine..
Do you find any reason to keep arguing..??

[Dormammu]

@Dormammu:
1.  I clearly told you that without any recovery stage (booster or anything else) after a passive Tonestack, any loss will pass to output and further dominated by next stage potential low input impedance..

2. You clearly stated that a passive Tonestack right after a hard clipping configuration and before a low value Volume pots works just fine..

Do you find any reason to keep arguing..??

1. I don't understand at all what it is about, because i did not observe negative effects in relation to the construction of stompboxes. Are your opamps completely unable to handle with the load ?

2. Of course, I take a volume pot a little bigger, but not too much. Didn't mention it, because who knows — he understands.

In the context of the presented circuit — we can no longer argue, because always 4 units.   

[camzilla]

Quite a delay since my last message, it was the final week of some Summer classes, but I have completely re-breadboarded the circuit and made some updates and it seems to be working as expected now!

The 1uF capacitor across the 2nd biasing resistor, R3, drastically decreased the noise. I was looking at the power supply

With the previous connection it's not only noise from the PSU rail but also the potential for oscillations since any connections to Vref could modulate the Vref and feed into other parts of the circuit.

Was there any improvement in the sustain on the higher notes.

All notes now sustain about equal or greater to the audible acoustic sustain which is as much as I could expect. Somewhere along the troubleshooting process I reached a stage where notes would sustain indefinitely like one of the Op amps was self oscillating, but unfortunately I didn't pay much attention to what was happening in the circuit at that time since I was focusing on the  many other intolerable noises coming out of it at the time, but something I might look into for a future project. Thanks for all the insight on filtering. The pedal is quite noisy so I may experiment moving down to a 10k biasing arrangement, but I think the noise might just be from having a high gain pedal with so many exposed conductors on the breadboard. I also currently lack the high frequency attenuating capacitors across the output and inverting input of all the amplifiers since I do not have any capacitors on the scale of picofarads at the moment. I likely make an order to add these before I build it to make sure it is as quiet as possible, but at the end of the day noise isn't the end of the world. Into the clean channel on my katana it can get some smooth-ish synthy fuzz sounds and into the lead channels it gets a decent high gain crunch so I'm not disappointed for being my first circuit.

R3 and the capacitor don't seem to form a frequency filter, what reasoning is there to increase the value of the filtering capacitor if I reduce R3?

But they DO form a LPF..!!
(of 1/2πRC cut-off frequency, where R = R2//R3)

Even if R3 didn't involve in LPF calculation, by decreasing its value you should also decrease R2 value (for retaining Vcc/2 bias voltage) hence increase cap value..

P.S.
I think you'll find below application note interesting enough..
https://www.analog.com/media/en/technical-documentation/application-notes/AN-581.pdf?doc=AD7699.pdf

Thanks for the insight. I thought that this filter would help for the noise originating from the 9v source but wouldn't apply to the ac oscillation from the guitar signal since it "enteres" through the 4.5V node so it experiences C1 and R2 in parallel?
Seems some good info in that article, I'll throw it in my overflowing bookmark folder of resources!
The 47k output amplifier gain resistor seemed to be a good spot with a nice volume range.

I completely lost out of sight — I would increase R9 to 200k and R11 to 200-500k.
It also seems an interesting idea to remove the diodes from the 2nd clipping stage.

I was experimenting with increasing these resistors, but although increasing them should theoretically not change much since so much signal goes through the clipping diodes raising them seemed to raise the noise level a lot. When I did this though I changed them both to 470k, maybe progressively gaining it as you suggest will work better. Thinking about adding higher value resistors on a switch for a higher gain mode if I have space on my board in the design phase.
I also thought it would be cool to have an op amp hard clipping mode like you suggest, I was thinking I might achieve it the same way as some of the existing muff pedals and add a switch to bypass the tone stage, which would cause my ouput amplifier to hard clip since the signal would be much larger without the lossy tonestack.



I am on to the next phase of trying to get this all on a perfboard and into a box. I was hoping to fit into a 1590B enclosure that I have already, but looking at the component count I might have to upsize. Looking for software online I found Fritzing and am starting to work on a layout now. Seems like it could be quite a puzzle since everything will have to be point to point instead of having access to a ground plane like on PCBs. As always appreciate all your responses and help so far been very instrumental. Let me know if there's any expert advice on perfboard layouts or any further rookie mistakes on the circuit!

Edit: I'm thinking I might go the PCB fabrication route just for the ease of design. The $20 may be worth saving space and time.

Here is version 4 of the circuit:

[Rob Strand]

All notes now sustain about equal or greater to the audible acoustic sustain which is as much as I could expect. Somewhere along the troubleshooting process I reached a stage where notes would sustain indefinitely like one of the Op amps was self oscillating, but unfortunately I didn't pay much attention to what was happening in the circuit at that time since I was focusing on the  many other intolerable noises coming out of it at the time, but something I might look into for a future project. Thanks for all the insight on filtering. The pedal is quite noisy so I may experiment moving down to a 10k biasing arrangement, but I think the noise might just be from having a high gain pedal with so many exposed conductors on the breadboard. I also currently lack the high frequency attenuating capacitors across the output and inverting input of all the amplifiers since I do not have any capacitors on the scale of picofarads at the moment. I likely make an order to add these before I build it to make sure it is as quiet as possible, but at the end of the day noise isn't the end of the world. Into the clean channel on my katana it can get some smooth-ish synthy fuzz sounds and into the lead channels it gets a decent high gain crunch so I'm not disappointed for being my first circuit.

Sound like you have got it sorted.   Yes, the small caps across the feedback resistors will certainly help prevent oscillation.   Larger values help as well but they are also used to fine tune the tone/fizz.

[duck_arse]

you could probably omit C3 without much effect on the filter cutoff. the "big muff" tonestack has a DC path to ground, so you might want an isolating cap between U3 output node and C4/R13. and R20 should return to V/2, bias, 4V5, instead of ground as it is attached to the biased opamp input.

[camzilla]

you could probably omit C3 without much effect on the filter cutoff. the "big muff" tonestack has a DC path to ground, so you might want an isolating cap between U3 output node and C4/R13. and R20 should return to V/2, bias, 4V5, instead of ground as it is attached to the biased opamp input.

C3 was a vestige since I was planning on having the larger capacitor C10 on a switch labeled "Smooth" but the sound with the larger capacitor was so much better I decided to have it hard wired. I will take that out to lower the component count since you're right its insignificant compared to the other cap.

On that isolating cap suggestion, my thinking was that no DC current could flow because of the capacitors in the tonestack and the coupling capacitor after the tonestack, so I thought an isolating cap after U3 would have no effect. Am I missing something there?

Thanks so much for catching that error on the filter with R20 to ground after the tone stage. That makes sense that it should rather be connected to the virtual ground of 4.5V than genuine ground. I will make that change.