Help with buffer followed by Electra

[Gargaman]

Hey, all
I'm trying to make a fuzz for a bassist friend. I really like Mark Hammer's Gruntbox but seems like I'm not ready yet for that fine tuning. I searched about buffing and blending between clean signal  and a simple fuzz path using op-amp like ROG Splitter and Blend; I got trouble to make the whole thing right but I'm glad the buffer worked fine, since I'm dealing with 072 for the first time. So I got this on breadboard, sounds ok for me although I don't know the benefits from a buffer before the Electra. From the beggining I wanted to put a simple tone control (SWTC or BMP tone stage) and don't want to waste the other side of the chip. How would you finalize this? Take the output of Electra back to the pin 6 (inverting) and tailor the tone around it? Tone control before the distortion? After all stages? Between stages? I appreciate any suggestions, do's and don'ts because I got to a point tired messing with this idea and will be nice to learn a bit about opamp inverting, non-inverting, phase, frequency response.
Thanks

[Gargaman]

Ok, I have a more concrete ask for help on this.
I can't make the other side of the chip to work.
I took the Electra output to the inverting pin 6 (thought that would be a good idea to get the signal in phase again);
calculated a gain of 4 then go to the BMP tone stage.
No output, although the tone pot is aparently doing his job on the noise floor.
What I am missing?
I took pin 5 both to ground (0V) and virtual ground (4,5V) but nothing change.

PS.: The schematic now shows the power section.

[antonis]

Place a resistor (10k or so)* between Q1 Collector and 104 cap..
(it is needed when hard clipping diode pair is ON)..

Disconnect TL072 +5 pin (non-iverting input) from GND and connect it to Vref..

Connect a big electro cap (10μF or bigger) between 2^nd op-amp output (pin 7) and 27k/472 junction..

(*) It should have minor impact on Electra stage gain due to 4M7/10k ratio primary role on gain..

P.S.
It should also be a good idea to move 100nF cap (104) between diode pair uppper side and 4k7 resistor..
(just to roll-off DC gain to unity)

[GibsonGM]

Should the gain pot, 100kB, be wired up as a variable resistor this way, rather than a divider?  I would think that the lower end would go to ground without the wiper connecting to it.   

[Gargaman]

Thks Antonis!
All right it's working through the second opamp, the tone works fine, but when I stick the diodes to ground..no sound.

Place a resistor (10k or so)* between Q1 Collector and 104 cap..
(it is needed when hard clipping diode pair is ON)..

Indeed, no change, but I can only tell because I bypassed the opamp-tone when following yours instruction step by step.

Disconnect TL072 +5 pin (non-iverting input) from GND and connect it to Vref..

Connect a big electro cap (10μF or bigger) between 2nd op-amp output (pin 7) and 27k/472 junction..

Did it.

It should also be a good idea to move 100nF cap (104) between diode pair uppper side and 4k7 resistor..

So, I moved the cap after the diodes and before 4k7 resistor. No sound when diodes engaged

[Gargaman]

Should the gain pot, 100kB, be wired up as a variable resistor this way, rather than a divider?  I would think that the lower end would go to ground without the wiper connecting to it.

Ok, gonna try, won't hurt.

[Gargaman]

Should the gain pot, 100kB, be wired up as a variable resistor this way, rather than a divider?  I would think that the lower end would go to ground without the wiper connecting to it.   

Ok, Mike. If I connect this way you suggested, when the gain is min, the sound is gone. So, I'll left the lug 1 floating instead.

[antonis]

It should have "weird" gain responce, Sir..

Q1 stage gain is considered the 4M7/(10k + Gain pot) ratio..
(roughly like NFB inverting op-amp with far less than infinite open loop gain..)

Pot wiper is connected to - almost- AC GND (TL072 pin 1) so by connecting pot lower lug to GND counteracts on gain by splitting the Base incoming signal..

e.g.
With pot set at 50%, as it is, gain is 4M7/(10k +50k)..
            //               , with lug 1 grounded, gain is the same but for a signal of 50k * (50k + 10k + h_FE*re) / 50k + (50k + 10k + h_FE*re) amplitude..

[antonis]

No sound when diodes engaged

Shorted diode pair or false switch placement..??

[Gargaman]

It should have "weird" gain responce, Sir..

Q1 stage gain is considered the 4M7/(10k + Gain pot) ratio..
(roughly like NFB inverting op-amp with far less than infinite open loop gain..)

Pot wiper is connected to - almost- AC GND (TL072 pin 1) so by connecting pot lower lug to GND counteracts on gain by splitting the Base incoming signal..

e.g.
With pot set at 50%, as it is, gain is 4M7/(10k +50k)..
            //               , with lug 1 grounded, gain is the same but for a signal of 50k * (50k + 10k + h_FE*re) / 50k + (50k + 10k + h_FE*re) amplitude..

Ok, gonna keep it for later studies. 
I thought that it was a simple pre-gain..but seems to complicate all around. The gain pot itself makes more sense with diodes on, so it may be placed on a better location, but I liked how it acts when Electra is boosting clean and dirty too.

Shorted diode pair or false switch placement..??

I don't think so because, I'm switching manually on the breadboard; and if I take the output from the Electra to volume out, the switching goes ok. (volume drop, sound distorts).

[GibsonGM]

I see Antonis, it is 'unusual' in that sense.   I would move the gain pot to the emitter of Q1 - although it may be just fine as it is!   
Generally I see this without a buffer in front of it, which helps with issues of weak input signal.  In this case, that may well not cause any issues.

[antonis]

I would move the gain pot to the emitter of Q1

You're free to do so, Sir but it shouldn't be a "Gain" pot anymore..
(Q1 Emitter is AC grounded so pot should act as bias trimmer..)

[GibsonGM]

https://i.stack.imgur.com/8hKP8.png

I would move the gain pot to the emitter of Q1

You're free to do so, Sir but it shouldn't be a "Gain" pot anymore..
(Q1 Emitter is AC grounded so pot should act as bias trimmer..)

No??  It works when I do it with triodes    And does not shift the bias.  A single pot would of course change bias!

[antonis]

 

You didn't mention "is series with Emitter bypass cap" Sir, did you..??

P.S.
It should work fine for a CE amplifier with single resistor or voltage divider bias..
Here, stage gain is a combination of open loop gain (R_Collector / R_Emitter) and closed loop one (R_Feedback / R_IN)
i.e. A_closed = A_open / (1 + A_open * Feedback) results into A_closed = 1 / Feedback only for A_open very high..
Q1 A_open definately can't be considered as very hign so the A_closed is partially dependent on (and not totally idependent of) A_open, hence "Gain pot" setting, as per your suggestion, can't "directly" respond in gain setting..

[Gargaman]

Place a resistor (10k or so)* between Q1 Collector and 104 cap..
(it is needed when hard clipping diode pair is ON)..

It should also be a good idea to move 100nF cap (104) between diode pair uppper side and 4k7 resistor..
(just to roll-off DC gain to unity)

Ok, by unity you mean 4,68 (22K/4.7K), right?
Also, what makes 10k (between Q1 Collector and 104 cap) necessary here, but not un-necessary on Electra by itself?

[Gargaman]

Shorted diode pair or false switch placement..??

What if I tell you that the diode switch works only if I take the output directly from the 10K resistor (bypassing the cap)??
I was about to give up and use only one side, but with the tone control and distortion, the volume must be at maximun.
I tried again to use other side.
So here is where I am:
INPUT > INVERTING (6) > ELECTRA > NON-INVERTING (3) > TONE > OUT
I tried switching this order every manner you imagine.
Gonna update the schematic.

[antonis]

Ok, by unity you mean 4,68 (22K/4.7K), right?

Not actually..

By unity I mean unity (X1) 'cause a cap in the gain branch of NFB loop (in series with 4k7 resistor) acts as "open" circuit for DC, hence TL072 inverting configuration works as a DC unity gain amp..
No DC current flows through 4k7 resistor, hence no voltage dividing effect from 22k/4.7k, hence DC Vout = DC Vin, hence Vout/Vin = 1..

Also, what makes 10k (between Q1 Collector and 104 cap) necessary here, but not un-necessary on Electra by itself?

Just different points of view..
Diodes to GND severely load Q1 Collector by dominating its current sourcing capability.. This might be welcome from distortion viewpoint but it shouldn't be "incorporated" into hard clipping effect..

[Gargaman]

Oh, I see.
How would you come with an arragement to make a gain of 4?
My intention was to compensate the tone control loss before it.
Well, maybe switching order have messed up things, I ended with some very strong synth fuzz putting EQ before opamp non-inverting; although the switching diodes shut the sound off in every arragement.
I got lost here, nevermind

[antonis]

 

Obviously, you didn't get me..
(my teaching skills suck..)

Signal (AC) gain indeed is 4.68 (22k/4.7k)..
(AC finds its way to GND through cap where DC doesn't)..!!
DC gain is rolled-off to unity for various reasons, the most significnat here been not saturating output due to DC gain..
For bi-polar supply ideal op-amp, biased at GND (0 Volts), high DC gain shouldn't be a problem 'cause many times zero equals zero..
(for real world op-amp, DC offset should multiplied by gain but could easily further "isolated" via an output cap..)

Here, a DC gain of 4.68 means that amp output has to produce 4.68 X 4.5V (non-inverting bias level) hence permanently hitting power supply rails (saturated)..

[Gargaman]

Ok, sorry, didn't get that AC/DC thing.
So, if the diodes were between opamp and capacitor, the DC will take a way through the diode and the DC gain will be differente than unit?
So, for whatever AC gain (buffer, 4x, 10x, etc) I'm setting an opamp, I must set it for unity DC gain?