Diode clipping in Bazz Fuss question

[fishfude]

Hey Folks,

I have a question about the diode clipping in the Bazz Fuss...

Diagrams I have seen of diode clipping always have the diode(s) going to ground, with the signal travelling past getting diverted (correct way of looking at it?) once the diode's threshold voltage is exceeded. In the Bazz Fuss however you have the diode between the transistor's collector and base. In this case, what is happening here to create clipping?

Many thanks,

ff

[LucifersTrip]

here's one explanation from mac

http://www.diystompboxes.com/smfforum/index.php?topic=83203.msg691096#msg691096

[fishfude]

Nice one, thanks!

ff

[mac]

This little mod makes a better sounding BF.



The base resistor feeds the transistor, and when Vb = Vc the thing gets focus. If not it sounds like crap!!!
I realized about this when I was experimenting with Arsenio Novo's design,



This circuit sounds good only when Vc = Vb. And in this case it is exactly a BF with the PNP silicon acting as a diode.

That's why a darlington version of the BF using a 10k at the collector sounds better than a single silicon, a darlington needs very little base current at Rc=10k keeping the voltage across the diode almost zero, ie, Vb almos equal to Vb. A single silicon needs more current at Rc=10k and Vc is a little higher than Vb. In this case you need to increase Rc a lot to decrease the base current and so the voltage drop across the diode.

With the addition of a base resistor this problem is overcome.
For example, let's say you have a BC549C that have hfe: 470.
If Vc has to be equal to Vb to get a focused sound then it is easy to see that

ic*Rc = Ib*Rb

since ic = hfe*ib then

hfe*Rc = Rb

If you use a 10k at the collector then

Rb = 470*10K = 4700k = 4M7

This condition does not depend on Vcc nor of any emiter resistor. BUT it depends on the variations of hfe with temperature. Making the collector resistor an external pot you can adjust for small deviations.

mac

[Earthscum]

It's been awhile, but I've tried  that Q1/Q2 combo before, in a sense... and I remember excellent results (or at least promising). I started with an NPN as "diode", and tried using a pot (resistance) from the C of the "diode" to the C of the transistor. That, if I remember right, was boost to mud, no BF at all. I tried a PNP, as in yours, but with a pot (C to E) and I remember being able to get some interesting (but still BF sounds) out of it, down to a nearly clean boost. Gonna have to try it out again.

For more "clipping" sound you can try output from between 2 diodes (I tried this with schottky's, but might work with a pre gain stage and Si) and a 100R.
From Base to Collector: 100R - Diode - (output) - Diode
Without digging out my notes, I believe it works without the 100R, but the resistance seemed to smooth out the sound a bit, not so crossover distortion sounding.

[mac]

For more "clipping" sound you can try output from between 2 diodes (I tried this with schottky's, but might work with a pre gain stage and Si) and a 100R.
From Base to Collector: 100R - Diode - (output) - Diode
Without digging out my notes, I believe it works without the 100R, but the resistance seemed to smooth out the sound a bit, not so crossover distortion sounding.

I'll try this.
It remembers me a kind of octave up from a bazz fuss someone posted long ago. IIRC it was like this: B->10k->(output)->diode->C.

mac

[Mac Walker]

When the input voltage goes low, the diode begins to conduct, and clips at whatever forward voltage is typical for the diode (Ge, Si, LED, etc.).
When the input signal goes high, the transistor Vce goes to (nearly) zero volts, as the transistor becomes saturated.
So the diode is only working for half the cycle.
Here's a simulation:

http://falstad.com/circuit/#%24+1+5.0E-6+6.450009306485578+79+5.0+43%0Ar+336+272+336+352+0+1000000.0%0Ac+336+272+256+272+0+1.0E-7+0.8176313379746731%0Av+256+352+256+272+0+1+100.0+0.25+0.0+0.0+0.5%0Aw+256+352+336+352+0%0Aw+336+272+384+272+0%0Ar+432+176+432+256+0+10000.0%0Aw+384+272+384+320+0%0At+384+320+432+320+0+1+-0.4870619512236576+0.569804430220965+4000.0%0Aw+432+256+432+304+0%0Aw+336+352+432+352+0%0Aw+432+336+432+352+0%0Ag+256+352+256+368+0%0Aw+144+352+256+352+0%0Aw+144+352+64+352+0%0Av+64+352+64+160+0+0+40.0+9.0+0.0+0.0+0.5%0Ac+432+256+544+256+0+1.0E-7+9.999999980221386E-4%0AO+544+256+592+256+0%0Ad+288+192+224+192+1+0.805904783%0Aw+224+192+224+224+0%0Aw+224+224+384+224+0%0Aw+384+224+384+272+0%0Aw+400+192+400+256+0%0Aw+400+256+432+256+0%0Aw+432+176+432+80+0%0Aw+208+80+432+80+0%0As+288+192+352+192+0+0+false%0Aw+400+192+352+192+0%0Aw+208+80+64+80+0%0Aw+64+80+64+160+0%0Ao+16+32+0+42+1.6742321987285427+3.995838144404471E-155+0+-1%0A

You can subsitute LED's or zener diodes for a higher voltage output signal if desired, or add variable resistance as described above to change the response, also play around with the collector resistor, which is essentially setting the bias / gain for the circuit....

[thehallofshields]

This circuit sounds good only when Vc =Vb...

Okay Mac, I know this thread is old but I've read about every thread where you've commented on the Bazz Fuss and I haven't found my answers.

1. If Vc=Vb then aren't we left with 0 headroom? Is the added connection from V+ to Vb making a 0.7v difference to cover the Si Diodes voltage-drop?

2. When I add a resistor between V+ and Vb I get a very effective and cleanish Gate. In my buffered/boosted build its very sensitive, I like a 4.7m resistor. The result is no background noise, increased treble (like 15% more scuzz), a sharper attack and less sustain. Great for rock-rhythm playing. These don't seem to be the changes you are describing, are they?

3. What on earth is that Arsenico v2 "mode" control? Is that your contribution? I can't find it anywhere else.
It works great as a gate-control in my Bazz Fuss! Much, much better than say, a bias control in a Fuzz Face.
But its like voodoo to me, when we turn the pot what is that capacitor doing?
And isn't the 50k pot to ground dumping voltage (or current?) from the circuit?

[mac]

1. If Vc=Vb then aren't we left with 0 headroom? Is the added connection from V+ to Vb making a 0.7v difference to cover the Si Diodes voltage-drop?

That's DC bias. It gives a fixed and better DC reference for the collector. The transistor is fully on.
In a single transistor standard BF, 2n3904 or 2n5088, the diode is almost off so there is little current into the base, and the transistor is almost off. It's the signal that turns both on. A darlington works better because a tiny current from the diode turns it on.

Adding to Mac Walker, when a AC signal enters the base, the collector voltage goes up and down around the DC bias condition Vc=Vb.
When Vc goes down the diode does not turn on since it is reverse biased.
When Vc goes up, there is a point where the diode begins to conduct until it reaches its maximun forward voltage, limiting the collector voltage swing to a diode drop above the base, ie, soft clipping the signal.
If you were feeding the base with a sine voltage, the output should look like a square wave down, and a sine-ish wave up.

2. When I add a resistor between V+ and Vb I get a very effective and cleanish Gate. In my buffered/boosted build its very sensitive, I like a 4.7m resistor. The result is no background noise, increased treble (like 15% more scuzz), a sharper attack and less sustain. Great for rock-rhythm playing. These don't seem to be the changes you are describing, are they?

A single 2n3904 should sound like a standard darlington Bazz Fuss, but sharper.

3. What on earth is that Arsenico v2 "mode" control? Is that your contribution? I can't find it anywhere else.
It works great as a gate-control in my Bazz Fuss! Much, much better than say, a bias control in a Fuzz Face.
But its like voodoo to me, when we turn the pot what is that capacitor doing?
And isn't the 50k pot to ground dumping voltage (or current?) from the circuit?

Yeap, it's my own.
The mode control R4 lets you find a Bazz Fuss sound, a clean sound around the middle, and octave up to the other side WITH drive/gain pot at zero.
It is a voltage divider that sets bias for the pair. C4 filters the power supply noise.

I think it is time to make a LTSpice simulation of a BF

mac

[PRR]

> If Vc=Vb then aren't we left with 0 headroom?

Collector can pull-down lower than Base.

Collector-Emitter voltage can approach 0.050V.

Here's a slightly simplified form, showing Base and Collector voltages. You see the Collector can idle at ~~0.6V and pull-down to 0.05V. Without the diode, it will let-up to over 7V; with the diode the up-swing is limited to 1.2V. The complete swing is nearly symmetrical: +/-0.6V.



Also note how "carefully" I had to trim Mac's "~~2.2Meg" to suit the specific simulated transistor that I used. It really wants to sit at 1.3V or 0.2V, has no special affinity for Vb. HIGHLY hFE critical, and will drift with temperature. OTOH, whan BIG signal is applied it *will* tend to this condition, self-adjusting; so maybe it only has to be "close" to boogie well.

[PRR]

> What on earth is that Arsenico v2 "mode" control?

It is a simplification of the "1.295Meg" resistor trim. Use a fixed 1Meg, then find the 1.295 or 2.2 ratio by dividing-down the +9V until the current is what you need. It uses a low-value trim and covers a more likely range than a rare 2meg or 5Meg trimmer.

> And isn't the 50k pot to ground dumping voltage (or current?) from the circuit?

It sucks directly against the 9V battery, not the audio. The suckage is a small part of the overall power consumption.

The cap is a frill to catch power crap from coming right into the input, and a Good Idea.

> It works great as a gate-control in my Bazz Fuss!

Mac gave you the condition for "good gain and limited output".

You have to bias carefully (or make the circuit bias itself) to get an output that is like the input only bigger.

If you go past that, gain goes to zero until you SMACK it. "Gated". Like a car with bent steering, which wants to run into the right guard-rail. If you YANK it, it comes over-center, will even smack the left guard-rail, but always relaxes back to riding the right guardrail. That's Bad Amplification. (But maybe good guitar effect.)

[thehallofshields]

If you go past that, gain goes to zero until you SMACK it. "Gated". Like a car with bent steering, which wants to run into the right guard-rail. If you YANK it, it comes over-center, will even smack the left guard-rail, but always relaxes back to riding the right guardrail. That's Bad Amplification. (But maybe good guitar effect.)

Okay steering alignment and bias make a good metaphor. In my buffered build 1meg and below add such gating that the gain control and guitar-volume knob become a little useless. I like the sound of a higher value, but its still acting more gated than just letting the diode do the biasing. This is why I'm asking.

I believe I understand but to clarify;
With  the standard BF circuit, Vb should be a tad too low, which has a drawback of necessitating a large input signal to turn the transistor on. -Correct?

So, in the BF circuit, what sonic differences do we expect to hear from Vb being too low or too high?

[mac]

The complete swing is nearly symmetrical: +/-0.6V.

Paul,
I ran simulations with single transistors and darlingtons.
Single transistors make a simetrical curve, but darlingtons have a square wave down.

http://www.diystompboxes.com/smfforum/index.php?topic=106176.msg967278#msg967278

OTOH, whan BIG signal is applied it *will* tend to this condition, self-adjusting; so maybe it only has to be "close" to boogie well.

When you use germaniums, single o darlington, and non-leaky germ diodes or just 4148s, it sounds even better.
A self biasing circuit might help because the sweet spot is temp dependent and too narrow, like a Dirac delta function 
I thought something like John Hollis' Rock Face. Ideas?

Mmhh... I never checked if the tuning condition is more stable when you have an emiter resistor.

That's Bad Amplification. (But maybe good guitar effect.)

That's a pretty commercial name!
It also describes my circuits

mac

[thehallofshields]

Now why would that be that Darlington's nearly Square the negative half while a single NPN is nearly symmetrical? Is this because of the slight difference in bias? I would have expected, if anything, for it to be the other way around?

I ordered some mpsa13's. I'll report back what difference I hear.