LG1 Gain block for "Unit Process / Basic Block" AIAB

[mdcmdcmdc]

Any existing circuit that uses differential distortion?

A few forum members helped me out with this design:

https://www.diystompboxes.com/smfforum/index.php?topic=127201.0

[bowanderror]

Any existing circuit that uses differential distortion?

There are a bunch of differential fuzzes & distortions out there, but they tend to sound pretty harsh & unmusical IMO. An open-loop differential amplifier is a different beast, with a much softer clipping mode and quite unique transfer curve.

Some of the most pleasant sounding soft clipping follows the hyperbolic tangent (tanhx) transfer function, which is also desirable for sine waveshapers. There are a variety of methods for sine waveshaping, including the overdriven differential pair, voltage-controlled amplifier (often diff pair), OTA, Middlebrook & Richer, and piecewise linear approximation.

Although quite common in synth circuits for both waveshaping & distortion, I haven't seen many overdrive pedals that utilize these circuits. The Hagerman Distort is based on this old dynamic gamma correction patent, which uses a differential pair w/o feedback to mimic various nonlinear transfer curves. The device that originally got me interested in this is the Zvukofor Tanhx (user manual with more details), a stereo soft-clipping studio preamp. Based on some posts I saw of his Colour Module, I believe Zvukofor's design uses a JFET differential pair, but I have yet to trace it.

I played around with Thomas Henry Tri-to-Sine Waveshaper from his book:



This sounded really nice on the breadboard, but still needs to have the level dialed in as the diff pair can go from no clipping to too much with a very small change in signal level:



I'd like to keep working on a tanhx-style design like this, so let me know if anyone else is interested!

[ElectricDruid]

I'd like to keep working on a tanhx-style design like this, so let me know if anyone else is interested!

Yeah, absolutely interested! I've used a design like this to give a bit of "tube grit" (marketing bullsh!t - take with a big pinch of salt) to an electronic organ I built. It sounded great, giving a bit of compression and "squash" to the sound without completely overpowering the sound or losing the chords. I think it's a great design, if you're not after a really hard distortion. For rich fruity blues/jazz grit, it's perfect. Metal? Maybe not so much!

[bowanderror]

I think it's a great design, if you're not after a really hard distortion. For rich fruity blues/jazz grit, it's perfect. Metal? Maybe not so much!
[/quote]

I'm with you on a good lower gain OD. I'm going to make a new thread with my previous post so we don't clog up this one.

[Vivek]

Yeah, absolutely interested! I've used a design like this to give a bit of "tube grit" (marketing bullsh!t - take with a big pinch of salt) to an electronic organ I built. It sounded great, giving a bit of compression and "squash" to the sound without completely overpowering the sound or losing the chords. I think it's a great design, if you're not after a really hard distortion. For rich fruity blues/jazz grit, it's perfect. Metal? Maybe not so much!

Doesn't amount of harmonics depend much more on how hard you drive the clipper rather than shape of transfer function of clipper ?


I mean, suppose we have a bluesy Tanhx clipper

And then we decide to feed it with a signal that is 50 times bigger

Won't you get metal worthy distortion ?



Fooling with gain of Tanhx function :
https://www.desmos.com/calculator/2857hjenmf

It can go from "hard clip" to mild based on gain more than based on transfer function shape

[ElectricDruid]

Doesn't amount of harmonics depend much more on how hard you drive the clipper rather than shape of transfer function of clipper ?

Yes, but if you wanted a hard clip, why would you use a very soft-clipping function and then drive it incredibly hard (with all the extra noise you'd get from all that gain) instead of using a hard clip and much less gain and therefore much less noise?

The *point* of the differential pair (tanh) function is that you get subtle distortion at very low gain levels, so you can add "colour" without adding obvious distortion, until you turn it up a little bit and start to get some grit coming in, still without hard distortion.

There are better options for hard distortion which will give you more harmonics with less noise.

[bowanderror]

To keep from sidetracking this thread, here is a dedicated new one.

[Vivek]

Doesn't amount of harmonics depend much more on how hard you drive the clipper rather than shape of transfer function of clipper ?

Yes, but if you wanted a hard clip, why would you use a very soft-clipping function and then drive it incredibly hard (with all the extra noise you'd get from all that gain) instead of using a hard clip and much less gain and therefore much less noise?

The *point* of the differential pair (tanh) function is that you get subtle distortion at very low gain levels, so you can add "colour" without adding obvious distortion, until you turn it up a little bit and start to get some grit coming in, still without hard distortion.

There are better options for hard distortion which will give you more harmonics with less noise.

I feel as follows :

Noise is most noticeable when signal is low.

Small signal Gain is the slope of input output graph very close to the zero crossing.

All transfer functions with same slope at zero crossing have same gain at zero crossing, irrespective of the function that generated that slope at zero crossing.

In fact, if you look at transfer curve only at zero crossing, you cannot predict which transfer function created that slope at zero crossing.

I'm trying to say : all methods to create a particular transfer slope at zero crossing will lead lead to same noise for signals close to zero crossing.

[Rob Strand]

Here's conceptually different way of setting up clipping.

The idea is to set the clipping point then use the exponential shape of the BJT (or whatever) to round-off the knee just under that point.

The idea for the circuit starts from a current sensing detector, like you see on a simple constant current source.  See fig 5.1B.

https://sound-au.com/ism.htm

A scheme to do this is show in the first circuit on this schematic,



However we can see that a "VBE Multiplier" type clipper is actually doing the same thing except it senses the current in the feedback resistor.   The advantage of the first method is we can separate the gain setting feedback resistor "R2" and the clipping.
The clipping point is programmable in each case and isn't related to how close the opamp is to the rails.

The output waveform is as follows,



Here's the difference between the new and conventional, not much difference in the waveform,



The reason they are the same is they both using the same idea and both are stretching the exponential characteristic of a diode/BE junction.

So we can keep going with the idea and prevent following the exponential characteristic down to zero voltage.   The idea is to add an offset to the reference voltage used to set the thresholds.



As expected, the characteristic is much sharper, and the difference is now visible,



It might be possible to use a single resistor between the bases instead of connecting to ground.   As shown the positive and negative clips points can be set independently.

If we wanted to soften the characteristic we would need to do the opposite and set the transistors just under the on point.   Perhaps adding diode here an there will help do that.


It occurred to me if we removed the current sense resistor altogether we arrive at something very similar to the early Hotcake,

https://www.diystompboxes.com/smfforum/index.php?topic=125157.0

The difference being the version I've posted controls both polarities and feeds the transistor back to the -ve input instead of the offset pin of the opamp.   Which ends-up being like the tri-sine converter at the end of the 1st page of that thread!.