Summing up different clippers to control harmonic content

[Vivek]

Trying to add up output from different clippers, all active at same time, to fine tune final harmonic content

Here is one of my attempts

(One or more of the diode legs can be asymmetric. Some output can be taken from a Pot thats in parallel to the clipping diodes)



Any possible benefit hidden here ?

[rankot]

I don't think it'll work like that.

[iainpunk]

i wouldn't do it like that, it would probably have a ''staircase effect'' when mixed in are different thresholds at different levels, but i haven't tried it yet, so i don't know. i think that filtering clipping with descending thresholds is a better idea...

the resistor and capacitor values should be changed around a bit, since the first LPF has to low a frequecny, 220nF should be more suitable. the resistor should stay low due to loading by the 2nd filter.

cheers, Iain

[Vivek]

But Iain, I want to clip signal in 3 or 4 different ways

and give the user the ability to mix in how much of signal he wants from each clipper.

I want the signal to be simultaneously clipped by 3 or 4 different sets of shunt clippers, and then summed up in user controllable percentages.

The low pass would already have been done in an earlier stage before the shunt diodes.

[11-90-an]

Breadboard it and see how it goes...

[Mark Hammer]

"Control harmonic content".  What does that mean?  Amount of clipping or quality of clipping?

Much of what distinguishes one clipping device from another is not so much how much harmonic content, but rather which harmonics and their relative proportion.  If one simply wants "more" then vary the gain of the stage feeding whatever does the clipping, or stick a booster in front.  If you want to alter the quality and pattern of the added harmonics, you need to play with the filtering and tone shaping, not just just change the diodes.

[Vivek]

Thanks for your comments, Mark Hammer

I understand that the ratio of 2nd,3rd,4th5th harmonics in a signal depend upon

The radians where the signal is clipped

Symmetric or Asymmetric clipping

So the idea I am trying to study is

If we clip a signal in 3 or 4 different ways

And then sum up in user controllable ratio

Do we get something useful ?

Suppose we take 95% Unclipped signal and add 5 % severely clipped signal , do we get great dynamics, but some stable amount of dirt even when the the signal drops ?


Suppose we take some signal that is only clipped at high levels, and mix it with some clear and some signal that is clipped at lower levels, do we get clipping that is more expressive since it has clipping sections that kick in at different signal levels ?


There are so many schematics that have switches to switch in different diodes. Each diode has different ratio of harmonics and different clipping radians.

I am trying to study the feasibility of having all diode legs working simultaneously, with pots deciding the mix that goes to the next stage.

[garcho]

It sounds to me like what you want is a mixer being fed by different distortion types and levels. I'm not sure what you mean by "work" but I do that kind of thing all the time (with a mixing board or in the box) and yes, many cool sounds are to be had. I don't think there's a Q&D passive way to do it, input buffer => parallel distortion circuitry => sum them with an op amp

[Vivek]

I'm not trying to repeat Andertons Quadrafuzz

[garcho]

Craig Anderton invented the summing amplifier configuration? I guess you're still on this "discover new distortion technique" kick. Do you have a DAW? I bet you'd have fun with Reaper or even Audacity and some of the countless free vst plug ins out there, lots of ways to get creative with sound routing and mangling, and a great way to try out some ideas for pedals (sonically anyway), especially EQ and clipping stages.

[antonis]

I don't think it'll work like that.

What he said..!! 

P.S.
Vivek, I've noticed in various threads than you selectively take into account only "friendly" to your wishes answers...
(don't get me wrong but, IMHO, this isn't a legitimate way of improvement..)

[rankot]

Try to simulate, it's quite simple circuit. But for what you intend, the only approach is to create few different distortion units and mix them.

[bluebunny]

I've looked at this thread a couple of times today, but haven't really wanted to add anything that may be misconstrued in any way as what actually may happen (it might just be the egg-nog talking, and we don't want that...).  But my gut reaction is that your passive scheme won't do what you intend.  I would be minded to keep things separated (and active) with op-amps, so you'd essentially arrive at an arrangement such as Gary and Ranko are suggesting.

[Rob Strand]

Whether the idea of mixing works gives you what you want is up for experiment.

As far as the circuit implementing the idea it should work however there's a few quirks which might not be so obvious:

In order for the clipper to clip the input level must be quite high one consequence of that is the opamp (say) driving the circuit needs to have a high gain.   Since the diodes are out of the feedback loop that opamp will clip and contribute a clipped sound - perhaps more than the diodes, depending on the gain.    The clean path signal for the mix ends-up not being clean!   The level control pots will have a a very strong signal on the feed side and a weaker level on the other, so the blend will be a little unnatural.  Perhaps a better way to do the mixing is to put level dividers to ground to only mix in the level off each clipper - these go to zero instead of "clean" when set to min.  Have one overall clean level.  The clean not actually being clean is easy to fix with a lower gain tap.   The sound of the clipped opamp feeding the clippers is harder to solve, you need to rejig something.

[amptramp]

I notice in the original drawing, the output does not go to any resistance to ground.  It would be best to have a virtual ground there, which takes the output to the inverting input of an op amp set up as a buffer or amplifier.  If the non-inverting input is set to a fixed voltage, the feedback will tend to keep the inverting input at the same voltage and the output of one clipping channel will not be able to backdrive the other channels.  That will at least enable you to set the contribution of each channel without any channels backdriving the others.  This will make the output controls independent of each other.

I am in favour of capacitive coupling or series L-C coupling of the clipping elements to avoid clipping lows (which would make the signal muddy) or highs (since you are already adding harmonic content and don't need to create additional dog-whistle frequencies).  Imagine a signal which is a low E at 82 Hz with a smaller high A signal at 880 Hz riding on top of it and it goes to a clipper that is all resistor and diode and therefore has no frequency selectivity.  If you clip the 82 Hz signal, there is a period in the waveform where the signal flatlines at the clipping level and you tend to square up the E but the variations from the A are lost since it is a smaller signal and during the flatline portion of the E, you lose the A entirely.  This gives you an A that is modulated by the E so you have the A gated at the low E rate so it exists in the linear region but vanishes once the signal goes beyond the clipping level.  Tis gives you a muddy sound since the A disappears during the extremes of the combined waveform.  With capacitive coupling to the clipper, you need progressively more signal amplitude before the lows start to clip and this is what you want.

[teemuk]

As far as the schematic example goes...
- all distortion sidechains are peak clipping
- all distortion sidechains are symmetric
As is, the sliders will only adjust the amplitude of harmonics resulting from symmetric peak clipping added to clean signal w/o said harmonics. Therefore I doubt that effects of these various controls have much difference from each other. All sidechains are doing pretty much the same thing, just at slightly different thresholds. And you already control the threshold with those controls.
But.. Go ahead. Build it and experience it yourself.