Total tonal control in a classic fuzz? An attempt at powerful elegance...

[Brossman]

Hey Folks,

I've always enjoyed the snarling spit of classic fuzz but always wished I had the same control over tonal shaping that I find in some more conventional OD's, or even in an amplifier. I also despise needless complication (though being a novice, that can be easily accomplished ).

I was inspired to make a FF clone a la Joe Gagan's EasyFace. The stability and gain control are there, but not the EQ'ing I'm looking for. So... How to maximize tonal shaping while minimizing controls? Yes, there are many one-knob configs, but none of those quite does it for me...

Over and over again, I've found the James to be remarkably intuitive and powerful for tone shaping, and it does not introduce as much loading as most BMT EQ's (eg, Fender/Marshall/Vox). Here's what I have so far:



I believe this should work well enough without the 3rd transistor stage. However, that stage should act to 1) buffer for consistent output impedance, and 2) recover potential volume losses in the passive tone stack.

Here's a link to a quick demo I made of the partial circuit. No gain or volume, and no 3rd stage - just tone controls:

https://youtu.be/jS_vCpiIAsg

------------------
A couple notes:
i) In changing values for R11 in Duncan TSC - with "Bass" and "Treble" controls maxed - I found that varying from 50k to 300k varies the notched frequency from around 600Hz to 1.6kHz [a HUGE change in fuzz character]. Stock value in the James is 180k, so I've used a 50k in series with a 250kA pot to vary across this range. I call this the "Sculpt" feature... it's intuitive, but complex to describe.
ii) That same change also expands or contracts the Bass control's range; Its midrange content is heavily affected, and therefore it is quite reactive with the "Sculpt" control
iii) With BMT at neutral settings, the James give a fairly flat EQ; By halving the value of C7 with an identical cap in series, DTSC shows a +3dB boost right at 1k.


So... thoughts? Suggestions?

[aron]

Sounds fine to me. Cleans up nicely! Nice range of tones.

[antonis]

Nice..!! 

(but, IMHO, for a FF you could save 2 caps by deleting C4/5/6/7 and shorting Bass pot with a single cap while placing another cap in series with Treble pot wiper..)

[PRR]

> James ....does not introduce as much loading

That may be relatively true (must be some reason it was THE most popular 2-knob in mono systems). But loading does matter significantly.

I can't read the values on your plan? I took Duncan's defaults scaled-down 10:1 for transistor impedances. Then I toggled source 1k and 10k, load 10k and 1Meg.


5dB and 10dB changes just by interface (before you touch the knobs!) is not a disaster but is inconsistent, and you can't "see" why the sound changed.

Also: I don't see how your Q3 works unless it is exceptionally leaky.

[antonis]

I don't see how your Q3 works unless it is exceptionally leaky.

Eagle eye, Paul..!!
(the rest of us were focused of Tone control..)

@Brossman: Move R16 right leg from Q3 Emitter to Collector..

[Brossman]

@Brossman: Move R16 right leg from Q3 Emitter to Collector..

Thanks! I sorta just copied it from the first part, but something told me it wasn't right to copy half of a Darlington... I figure the base probably needs a bias 

...But loading does matter significantly.

I can't read the values on your plan? I took Duncan's defaults scaled-down 10:1 for transistor impedances. Then I toggled source 1k and 10k, load 10k and 1Meg.

Ah, yes.  *Ahem* I... did not consider that, frankly. I think I'm so used to working with tube circuits I didn't think much about it. I did model changing the signal impedance to 1k, , but compared to the scale of the components therein, of course it wouldn't show much effect...

R14 is a 1M to ground that will feed the base of Q3 through C8... it hadn't occurred to me to consider these effects:
1) Whether R14 is necessary, or just part of the model for output impedance;
2) That ALL control schema need to be scaled to accommodate different ranges of impedance found in transistor circuitry.

Also... it doesn't much help that I indeed had *no* component values in my first LTspice layout  (cuz I couldn't figure out how through fuzzy-brained late-night insomnia... another day)

Here's the revised layout:


Thanks so much Paul and Antonis!
- Brienne

[Brossman]

I've been doing a little bit of modelling so I can show how the sculpt control and boost control will work.
I just couldn't help myself.
Folks I'm having WAAAAY too much fun with this

Given my better understanding (I think? I hope!  Thanks Paul), I have revised the tone stack to be more appropriate to a transistor circuit. I tried to tune the circuit to see where nominal settings produce the "flattest" EQ that a James can get.

* Every graph shows me sweeping through the range of the mid sculpt control. Key shows where in the resistance range of that control each plot lies.


Bass + Treble max - creates a mid scoop.


Bass max, treble at noon.


Bass min, treble at noon.


Same as above, but mid boost engaged (C from Treble to ground is reduced from 33n to 10n)


Flat EQ with mid boost engaged


Bass + Treble at min, with mid boost.


** In the last 3 plots, the green trace is the same as magenta, but no mid-boost engaged

------
OK babble brains, so what?

I think where I've put the cap I change as a mid boost hits a particularly useful range of frequencies. From some of these plots we can see that the mid-boost has a greater effect as B+T are lower. The mid boost also expands the mid-sculpt's range into lower frequencies while giving some pretty whopping boost. [I was wrong in my first post... I meant to say >6dB of boost. Here, you can see what I mean]

I also wanted to show how subtle that mid range control looks on the plot compared to IRL conditions. Even with B+T maxed, the notch ranges from 600-1.6kHz - an astounding range for controlling the character of meat in a fuzz. With B+T at minimum, and with the boost engaged, the peak ranges from 550-2kHz.

I have to say, I'm pretty impressed. Feels stupidly simple to me, and that's what I love!
-------

Oh, and before I forget, here's the schematic of the updated tone stack from DTSC:

[Brossman]

Antonis -

IMHO, for a FF you could save 2 caps by deleting C4/5/6/7 and shorting Bass pot with a single cap while placing another cap in series with Treble pot wiper..)

I don't know I understand how to accomplish this. Are you saying to make it more like the old Bassman tone stacks? I wonder if changing the arrangement would still allow the flexible mid-shifting control...

[nocentelli]

Antonis -

IMHO, for a FF you could save 2 caps by deleting C4/5/6/7 and shorting Bass pot with a single cap while placing another cap in series with Treble pot wiper..)

I don't know I understand how to accomplish this. Are you saying to make it more like the old Bassman tone stacks? I wonder if changing the arrangement would still allow the flexible mid-shifting control...

I think this is the sort of thing antonis was suggesting. Ignore the component values, I just took them from the Midfi demotape fuzz which has this kind of simplified James/Baxandall arrangement.

[antonis]

What Leo said..
(without shift necessarily been trimpot..)

R14 is a 1M to ground that will feed the base of Q3 through C8... it hadn't occurred to me to consider these effects:
1) Whether R14 is necessary, or just part of the model for output impedance;
2) That ALL control schema need to be scaled to accommodate different ranges of impedance found in transistor circuitry.

1)R14 isn't needed for a real world (complete) circuit..
(it's just there for estimation of succeeding stage impedance and loading calculation purpose only..)
BTW, we (both me and EQ) don't like Q3 ultra low input impedance..
Of course, you can simulate it in DTSC but your example of 220k is far away from reality..!!
(without R15/16/17 resistor values it can't be calculated precisely but for a grounded CE amp w feedback resistor bias its value shouldn't be higher than 2k5 - 5k, say..)
IMHO, Q3 stage should be a CE amp of high gain BJT with Emitter degeneration (without bypassing Emiter resistor..) biased with a resistive voltage divider(*) (like Big Muff output stage, say..) with a gain of about 5 (precise gain value should be considered taking into account EQ signal loss & OUTPUT impedance)..
(*) oprionally bootstrapped..
(Of cource, a JFET CS amp should get rid all the about issues but a self-respecting Fuzz ought to be strictly bipolar..)

2) see above..