Simple Variable hardness of clipping idea

[Steben]

Hi Guys.

In the constant search for idea in order to go analog when "modeling" sound I thought of the hardness of clipping.
Think of a Blackface or Plexi compared to a classic AC Vox. The first have rather hard clipping, the latter rather soft onset without the feedback.
By using a classic idea of parallel placed clipping pairs of diodes yet making it variable we can model in a very simple way the clipping character.
When looking at I/V curves of diodes I noticed diodes are very complex. We all know they dont have a fixed knee voltage, nor do they have a fixed resistance beyond the knee voltage. And there is differnce between different kinds of diodes. This means if we put them in parallel, they are parallel very complex resistors.  Parallel means the addition of currents flowing, which is the sum of current curves. With a variable resistance in series with on or another diode pair to ground we can shift the curves. Doing so we "round the knee" even more of the highest treshold pair.
Parallel diodes (same polarity of course) are often quoted as "dangerous". That is because they are in high current circuits, think of LED lighting. But with low level signals this is irrelevant.

[Kipper4]

Hey mate what is it like in practice?
I've sent something like it before over on AMZ but I've not really seen it used much.
Have you an accompanying circuit?

[Steben]

Hey mate what is it like in practice?
I've sent something like it before over on AMZ but I've not really seen it used much.
Have you an accompanying circuit?

All classic "pair of diodes" clipping circuits can use this. With the pot all up, you have diodes in parallel which according to the type of low treshold diode (in this case schottkies) give very smooth clipping onset, but with less clarity and firm clean headroom ("voxy"). Sounds a bit like diode ladders. With the pot down, it is almost only the pair of high treshold diodes (more marshallesque).
it doesn't work with for example silicon diodes and LED's. The diodes need an overlap in the current curve in order to influence the total picture. Germanium and silicon might work.

[Mark Hammer]

I've installed pots in series with clipping diode pairs in a few pedals, to reduce the intensity of clipping.  Works fine.  The only thing to be concerned with is that it not only changes the clipping quality, but also the volume.  So if you adjust the clip you have to tweak the volume control.

[Nitefly182]

Every time I try a configuration like this I feel like the result is not worth the pot is occupies. The effect in the examples I have used is subtle and I'd almost always use the pot space for something else and maybe offer a diode lift setting as a compromise.

[teemuk]

Indeed. (Listening of sine waves excluded) the difference between soft and hard clipping is astonishingly subtle in practice. The perceived differences (in distortion tone) are more often results of differences in clipping thresholds than differences in 'knees' of the transfer function.

[Steben]

I've installed pots in series with clipping diode pairs in a few pedals, to reduce the intensity of clipping.  Works fine.  The only thing to be concerned with is that it not only changes the clipping quality, but also the volume.  So if you adjust the clip you have to tweak the volume control.

Great reMark. Though I feel an idea rising of morphing between a 30W Vox and a 50W Marshall



EDIT: did you add a pot to one pair with no extra pair added? Because having two pairs is the base of my story. One pair with a resistor gets you towards TS style sounds (if the opamp is not clipping).

[Steben]

Indeed. (Listening of sine waves excluded) the difference between soft and hard clipping is astonishingly subtle in practice. The perceived differences (in distortion tone) are more often results of differences in clipping thresholds than differences in 'knees' of the transfer function.

I agree. But the non-subtle ideas are spread enough.
There is in the "wet" clean tone of amps (low gain) definitely a different character. Usually this is linked to the addition of a feedback loop. This feedback loop reduces the non-linear response of an amp. In other words: it tightens the "knee". Typical Voxes have no feedback loop and are described as "lush". This does not apply to high gain sounds. Once in high gain, the differences between a 50 000$ vintage amp and a 50$ pedal are subtle indeed. That's why "plexi" pedals used to multiply like rabbits: just stick the correct midtrebly EQ in front of a clipper and voila, you have your "stack" sound.
One of the things I've learned in recent times are facts about even order harmonics for example.
They are praised for their "musical" character, although they are omnipresent in nasty fuzzes and tone suffers more in intermodulation distortion. Opposite of this is very soft symmetric clipping as in a TS style circuit. This is reflected in opinions about tube preamp (even order) vs poweramp (uneven with crossover compression) clipping: in theory, the uneven harmonics and crossover is put down as nasty, yet in practice power amp drive is found to be musical and full of dynamic expression.

[amz-fx]

I played with this idea a few years ago and ended up with this arrangement:



There is not much to be gained by varying the resistance before the Schottky diodes once you have it dialed in. The silicon resistance is where most of the action is.

Best regards, Jack

[Digital Larry]

I played with this idea a few years ago and ended up with this arrangement:



There is not much to be gained by varying the resistance before the Schottky diodes once you have it dialed in. The silicon resistance is where most of the action is.

Best regards, Jack

Hi Jack!  So what does a sinewave look like through that?  Seems like it would require fairly delicate balancing so that the clamping of the lowest vf diodes would still allow enough voltage to the high vf diodes for them to do anything.

[Steben]

I played with this idea a few years ago and ended up with this arrangement:



There is not much to be gained by varying the resistance before the Schottky diodes once you have it dialed in. The silicon resistance is where most of the action is.

Best regards, Jack

Wow. well, looks a big leap from silicon to LED at first sight

[bool]

An obvious piecewise approach to clipping; and imho the correct unwrapping of Jack's comment wrt where the "action is" would be to decide whether LEDs are going to be the "main" clippers or merely a sort of a "final limiter", thus leaving the bulk of the tone-shaping action to Si diodes. Schottkys look like they are there to provide some low-level "bloom" and nothing more.

[aron]

Yes, Jack had something like this long ago. It DID work, but it was subtle and (to me) not worth pot.
My favorite diode strings have long been documented in the schematics page.

[Rob Strand]

An alternative to paralleling is blending.

For example put two silicons in each direction (total 4x silicons) to get the Si level closer to the LEDs.  Use a separate R1 for the silicons then use a largish valued blend pot say 100K between the Si and LED paths and take the output off the wiper. Both  R1's needs to be lowish compared to the blend pot say 1k to 10k.

Like Aron mentioned it seems hard to justify a pot but as an experiment the pot convinces you of that.     The simplest approximation is a switch:  All 4xSi, All 2xLED, 2xSi up + LED down.