show your creative clipping circuits

[Reinerterig]

So I have been experimenting with soft clipping, trying to figure out all the different ways diodes can react to an input signal. Like the difference between diodes in parallel and series.

If anyone is interested, here are some basic circuits I ended up with using mostly the same diodes on all the circuits --as to keep the experiment consistent and changing the values would only change the clipping threshold-- and trying to keep the resistance in the feedback loop at 500K to keep the gain constant --where there are two resistors in the loop you can change the values to be..say 100k and 400k this will ensure you have the same gain but might change the symmetry--.

LTSpice file:
https://www.dropbox.com/sh/i2wvm8fac1l8axz/AAC0WJ2KH-d5yCytbnzEq8E6a?dl=0

I was surprised to see what wide range of reactions you can get like different levels of compression, how the compression releases over time and signal input, and how you can change the symmetry.

If anyone else has creative techniques to saturate their signal, feel free to share.

[vigilante397]

This is hands down my favorite clipping option. Creative? Maybe not. Awesome? Always.



I'm a simple man with simple tastes

[Reinerterig]

This is hands down my favorite clipping option. Creative? Maybe not. Awesome? Always.



I'm a simple man with simple tastes

just full-on hard clipping man!

I love the Fulltone hard clipping circuit.

[GGBB]

[PRR]

> how the compression releases over time and signal input

As said in the other forum: there are no time-constants in the clippers you present, so it is hard to understand your "over time" observation.

[Reinerterig]

> how the compression releases over time and signal input

As said in the other forum: there are no time-constants in the clippers you present, so it is hard to understand your "over time" observation.

I can see how that can be confusing, and I agree that "release over time" might not have been the most accurate way to describe it. A come accurate way to explain this is to talk about the clipping threshold.

In some orientations, the diodes ( I assume ) are causing the opamp to considerably increase its gain, but the diodes are clamping that signal down ( at a voltage set by the value of the diodes ) without hard clipping (compressing the signal to some extent). In some cases where the gain of the opamp is extremely high most of the signal is amplified above the threshold resulting in a great deal of sustain.




Here is another example you can see V(h3) is a lot more compressed than V(h2) especially when it comes to the "attack" or higher voltages of the signal but there is a point where they "blend" together and react the same.



Keeping V(h3) as a reference here we can see that they have the same clipping threshold and are almost identical when it comes to larger signals, but V(series) drops below the threshold much sooner ( hence me referring to it as a release). this is probably because the gain of V(series) is a bit lower.



It is just interesting to me how you can make a circuit react in different ways using the same parts and values. I think understanding these small changes can really help to zero in on a specific sound or feel you want to achieve, So I am grateful for everyone clearing up any misconceptions and anyone willing to share some of their knowledge and experience.

I hope this clears it up a little bit,
Thanks.

[teemuk]

Yes, dynamic gain ratio characteristics explain differences in clipping behaviour.

This has nothing to do with "release time" of compression, though. Gain compression is instantenous in each and every circuit but dynamic gain ratios of those different circuits are different.

As expected and as is the purpose.

Yes, still a very enlightening experiment in how clipping characteristics can be shaped drastically with generic components and different methods to connect them together.

[Gus]

https://www.diystompboxes.com/smfforum/index.php?topic=100633.msg886103#msg886103

[Kipper4]

http://www.muzique.com/lab/sat2.htm

http://www.muzique.com/lab/warp.htm

http://www.muzique.com/lab/zenmos.htm

[thermionix]

[EBK]

I'm currently working on a circuit that does have a time constant involved in its clipping behavior.

Here are some preliminary graphs, with the different green lines representing different gain pot settings.  Note the DC shift and the asymmetry, particularly at lower gain settings.

Initial behavior:


After settling:


The somewhat-lesser-known Fetzer 2 paper describes how this can be done with a diode network at the input (gate) of a JFET circuit (I'd have to refind the link [Edit: here (scroll down on this page): https://www.researchgate.net/publication/290534843_Triode_emulator_-_Part_2 ]), but I'm doing it in the feedback networks of two op amps. Jack Orman's site has a generalized version of this idea:
http://www.muzique.com/schem/louis.gif (the gain pot in the circuit I'm working on changes the gain of both stages).  I'm borrowing very heavily from Fig. 4 of the patent Jack cited:
https://patentimages.storage.googleapis.com/pages/US5032796-4.png

Of course, a tube amp will do this too....

[Reinerterig]

http://www.muzique.com/lab/sat2.htm

http://www.muzique.com/lab/warp.htm

http://www.muzique.com/lab/zenmos.htm

muzique and RunOffGroove are just goldmines!!

I'm currently working on a circuit that does have a time constant involved in its clipping behavior.

Here are some preliminary graphs, with the different green lines representing different gain pot settings.  Note the DC shift and the asymmetry, particularly at lower gain settings.

Initial behavior:


After settling:


The somewhat-lesser-known Fetzer 2 paper describes how this can be done with a diode network at the input (gate) of a JFET circuit (I'd have to refind the link [Edit: here (scroll down on this page): https://www.researchgate.net/publication/290534843_Triode_emulator_-_Part_2 ]), but I'm doing it in the feedback networks of two op amps. Jack Orman's site has a generalized version of this idea:
http://www.muzique.com/schem/louis.gif (the gain pot in the circuit I'm working on changes the gain of both stages).  I'm borrowing very heavily from Fig. 4 of the patent Jack cited:
https://patentimages.storage.googleapis.com/pages/US5032796-4.png

Of course, a tube amp will do this too....

EBK, this just looks beautiful, the reason I started playing around with this in the first place was to see if I can get a deeper insight into diode clipping and how they can help emulate triode like behaviors. I haven't even gotten to LCRs yet to experiment with time constants.

[Mark Hammer]

I always thought that what the late Fred Nachbauer implemented as a limiter in his Dogzilla amp design might be put to creative use as a clipper, perhaps using Schottkys instead of higher Vf silicons.

Imagine we have the upper 3 sets of diode pairs in a feedback loop, like a Tube Screamer, or perhaps usedfor hard clipping.  R67 is a suitably-low value to permit audible clipping, or at least coloration.  R68 and R69 are replaced with a dual-ganged pot, with some selected value, and a selected parallel fixed resistance for each pot section, to provide the same sort of proportional change.  In theory, a person ought to be able to go from simple coloration to more intense clipping.  As the equivalent or R69 approaces zero ohms.

[pinkjimiphoton]

here's the weirdest clipper i came up with, using a ge q in the feedback loop of an opamp. the weird thing is the clipper goes from passive to "active" clipping when juiced.

different from any other application i've seen AFAIK

[ashcat_lt]

I've occasionally wondered what might happen if we were to add gain in the negative feedback loop of an opamp (or I guess about any amplifier circuit) such that the opamp itself applies less and less gain as the input increases.  This should be able to create something like hard clipping, but also I think could get us into "foldover" territory where the output for higher inputs is less than that for lower inputs and the peak of the wave dips back down rather than just staying flat.  This can look like an octave-up on a pure input, but ends up sounding just horrible in most cases.  Might be fun to mess with, though?

[EBK]

You almost described a compressor.
Definitely worth experimenting with though.

[ashcat_lt]

You almost described a compressor.

A compressor shouldn't really fold over, though.  At least not IMO.  Like, if the input is above the threshold, the output really shouldn't ever be below it, right?  It doesn't go as far over as it otherwise might, but it will never actually end up under.

[pinkjimiphoton]

dude, serious, try that flying spaghetti monster on a breadboard. i think you'll be both surprised AND amused.
its a very different sounding way to make it clip,  and you can get some pretty extreme distortion out of it, that kinda blooms into multi octave madness at points. sounds likke yer playing lightning bolts when its juiced with the bloom n boost switch on.

https://www.youtube.com/watch?v=G2IXNrDZXZ4

[Reinerterig]

I've occasionally wondered what might happen if we were to add gain in the negative feedback loop of an opamp (or I guess about any amplifier circuit) such that the opamp itself applies less and less gain as the input increases.  This should be able to create something like hard clipping, but also I think could get us into "foldover" territory where the output for higher inputs is less than that for lower inputs and the peak of the wave dips back down rather than just staying flat.  This can look like an octave-up on a pure input, but ends up sounding just horrible in most cases.  Might be fun to mess with, though?

This sounds like an interesting idea, I'll be sure to give it a try. Maybe it is just a compressor or just reduces the gain of the first opamp as the signal increases (to it starts to clean up).

[Transmogrifox]

Here's one thing I was playing with for a while. 
The obvious part of the clipping implementation is J5, which clips the negative-going part of the signal at which point the JFET is fully turned on in the capacity of a gate-follower.

The less obvious part of this circuit is designed into the current-limiting threshold which is coordinated by R10, Q2, J2, R11, R10, R14, Q1.  When J2 is driven positive, this causes an increase in current steering between J2 and J3.  You get the soft pinch-off as J2 drives J3 off, but it reaches a hard limit which is set by Q2.  With the right resistor combinations you can adjust the symmetry of this circuit.

Finally there is a dynamic element added to this by C8, since this cap is charged/discharged to an offset determined by the amount of asymmetry programmed into the whole network.

J1 clips if you drive it hard enough at the input.  This does something...but it's subtle.  It's mostly there for gain.

Needless to say, it is very interactive and interesting from a technical point of view.

How does it sound?  Like an overdrive pedal :/ .  The overtones added by the interesting circuitry are only subtle -- the sound that reaches out and grabs you is dominated by the pre-emphasis EQ and tone shaping at the output...just like any other clipper circuit.  When I was tweaking on the breadboard, I spent more time on the EQ and tone stuff than I did playing with the clipper.

I will eventually box one up because I need a good crunch overdrive pedal, but not because the method of clipping does anything amazing.

Here are some audio demos, anyway:
http://www.cackleberrypines.net/transmogrifox/Misc/Distortion_Experiment/demo/