Musical Soft Clipping Circuits - Tri-to-Sine Waveshapers & the Differential Pair

[Steben]

No mention of CMOS inverter soft clip?
Tri-Sine...

Ok, not perfect, but it's easy (the triangle isn't perfect in the first place). Just a CD4069U inverter amp with very modest gain. x2 in fact.

Wondering the same...an inverter does the same job.
And the video of the differential OD sounded very similar to a single 4049 stage distortion.
For low gains works well, but for high gains you don't want soft clipping..it sounds mushy and congested.

In comes multistage..... IMHO
You have a bit of soft clipping but with a hard clipping element at some point

[teemuk]

Rory and Ritchie are gods but don't tell me they would sound as good as that with a LED clipper through a mixing board.

Of course not, their tone is a sum of many different elements, not in the least their incredible technique and taste. Therefore demonstrating great tones of two great guitarists does very little in demonstrating what audible effects different types of clipping schemes provide, if any, assumed everything else remains equal.

And, by the way, both Blackmore and Gallagher were keen employers of solid state distortion: Blackmore ovedrived the preamp of his (very "hi-fi") Aiwa tape delay unit and Rory frequently used a distorting Rangemaster treble booster effect. And Blackmore used treble booster too, a resonant circuit modified to his guitar. Rory also always had that trademark tone whether he plugged to Fender, AC30 or a transistorized Stramp amp. Same with Blackmore, who happened to prefer modified high power Marshalls instead. They were mostly modified to be louder and brighter. How much you think subleties of clipping distortion weighs in their rig setups and overall tones? We are definitely hearing a lot more in effect than just clipping distortion.

IMO, blindfolded you couldn't tell whether the distortion comes from an overdriven LED clipper, tube preamp, tube power amp, Rangemaster, silicon diode clipper, or an Aiwa tape delay. It's just medium gain clipping with lots of additional processing. But you can probably tell it's Ritchie Blackmore or Rory Gallagher playing the guitar.

[marcelomd]

RE CMOS clipping.

At the other forum there's a guy trying to create a simplified Darkglass distortion. He modelled the FET mu amp and CMOS clipper using opamps and diodes. Very interesting and detailed analysis.

Is it ok to post the link here?

[Vivek]

Talking about Deep Purple,

I was their Official Photographer for their last 3 concerts in Dubai.

[Steben]

Rory and Ritchie are gods but don't tell me they would sound as good as that with a LED clipper through a mixing board.

Of course not, their tone is a sum of many different elements, not in the least their incredible technique and taste. Therefore demonstrating great tones of two great guitarists does very little in demonstrating what audible effects different types of clipping schemes provide, if any, assumed everything else remains equal.

And, by the way, both Blackmore and Gallagher were keen employers of solid state distortion: Blackmore ovedrived the preamp of his (very "hi-fi") Aiwa tape delay unit and Rory frequently used a distorting Rangemaster treble booster effect. And Blackmore used treble booster too, a resonant circuit modified to his guitar. Rory also always had that trademark tone whether he plugged to Fender, AC30 or a transistorized Stramp amp. Same with Blackmore, who happened to prefer modified high power Marshalls instead. They were mostly modified to be louder and brighter. How much you think subleties of clipping distortion weighs in their rig setups and overall tones? We are definitely hearing a lot more in effect than just clipping distortion.

IMO, blindfolded you couldn't tell whether the distortion comes from an overdriven LED clipper, tube preamp, tube power amp, Rangemaster, silicon diode clipper, or an Aiwa tape delay. It's just medium gain clipping with lots of additional processing. But you can probably tell it's Ritchie Blackmore or Rory Gallagher playing the guitar.

Although you are right, we clearly can hear when artists change gear over the years.
And yes, rangemasters were used, but ultimately that half clean tone which gets pushed by hard attack of those days is not the booster giving harmonics but the amp. And both Blackmore and Ghallagher are renouned low gain guitarists.
We don't need their exact setup at all, but again I can't believe you can get their tone with a LED clipper. Unless you combine them, tune them, add resistors and multistage them.

This on the first place a reaction to a comparison of a very high gain chord progression in that video. It is very powerful base material to make anything sound the same.

[Steben]

Talking about Deep Purple,

I was their Official Photographer for their last 3 concerts in Dubai.

owkay. Were you in the position to ask questions and test whether gear memory still stands? 

[dschwartz]

No mention of CMOS inverter soft clip?
Tri-Sine...

Ok, not perfect, but it's easy (the triangle isn't perfect in the first place). Just a CD4069U inverter amp with very modest gain. x2 in fact.

Wondering the same...an inverter does the same job.
And the video of the differential OD sounded very similar to a single 4049 stage distortion.
For low gains works well, but for high gains you don't want soft clipping..it sounds mushy and congested.

In comes multistage..... IMHO
You have a bit of soft clipping but with a hard clipping element at some point

I was just experimenting with 2 cascaded cmos stages preceded by an opamp boosting the input.
I tried:
- simple 10x gain per cmos stage--> compressed and warm at mid gain...mushy on high gain. Sounded like a vintage tube amp not meant for metal.
- flexwave without zeners: sounded great at mid gains, and very compressed on hard attacks, but totally mushy at high gain..too soft clipping. Awesome for bluesy tones.
- honeybee-like clipping (single led shunt clipping within the feedback loop, so gain raises on the clipped lobe, squaring the clipping on that side): a little more suitable for high gain since one side is hard clipped (and tubey because of bias shift) ..sounds very nice, but no dice for metal sounds..my opamps+leds designs sound much tighter and defined..and less noisy.
In conclusion...soft clipping is great for vintage tones...for modern metal, a at least 1 super hard clipping stage is needed.

[Steben]

Absolutely.
I keep referring to the modded BluesBreaker and Mark's WattBreaker.
It is the multistaging that adds a dimension which gives soft clipping at low gain settings and harder clipping at high gain settings because of the treshold you hit in the first stage.
For the moment it is the sum of soft clippers. But a hard clipper in between the stages might give yet another harder clipping very close to "soft knee" power amps still hitting the rail. As I mentioned earlier that "hitting the rail" is crucuial in maintaining the balance. Very high gain combined with quite soft clipping is quite unnatural sounding im(and your)ho.
The multistaging like this might render a very versatile circuit with bluesy low gain and boosting qualities yet better high gain sounds as well.

[teemuk]

...but again I can't believe you can get their tone with a LED clipper.

We would be in the entirely wrong track if trying to explain that clipping creates tone in entirety. No. One sinply can not - and will not - find a The Right Combination of non-linear transfer curves that would suddenly turn everything to sound like Jimi, Rory, Eddie, Ritchie, cranked Marshall, AC30, tweed, and et cetera. No, the recipe to complete tone is so many more ingredients than just clipping.

Unless you combine them, tune them, add resistors and multistage them.

And add filtering here and there, use the right combination of strings, pickups, maybe boost here and there, feed all the stuff to cabinet designed certain way with specific speakers, have the proper picking technique and so forth. I really don't believe whether you have LEDs clipping or some esoteric circuit matters that much.

[Vivek]

Steven

Do softer clipping stages precede the high clipping stages ?
Or other way around ?


How is "hitting rails" different than shunt diodes to ground ?

And how is "hitting rails" different than "soft clipping diodes" fed with veey high gain so the softness is pretty much gone.

[Steben]

Steven

Do softer clipping stages precede the high clipping stages ?
Or other way around ?


How is "hitting rails" different than shunt diodes to ground ?

And how is "hitting rails" different than "soft clipping diodes" fed with veey high gain so the softness is pretty much gone.

In order to see the wood for the trees, it is necessary to see the difference between softness of clipping diodes and softness of the circuit topology.
What I refer to is soft topology with the diodes as hard clipping elements (which they somewhat are in reality yet somewhat not).

Every soft clipping circuit (whether it is with diodes and added resistance as in Bluesbreaker or non inverting feedback loop) is not intended to "hit the rails" or hit ground.
This means there is always a portion of the dry signal given through. Simplified a multistage soft clipping topology "stacks" the dry signal on top of square waves. The more you stack, the less soft the result will be relatively.
Multistage clippers to ground do not stack signals, the treshold is the treshold.

It is however my intuition that even a hard clipping topology with diode to ground for example which are softened with resistors should have the hardest clipping in front.

[teemuk]

Excuse me but this "stacking" of dry signals on top of square waves makes no sense at all. If it was an analogy, I'm not following it.

Typical clipping distortion is nothing but instantenous compression that leads to amplitude distortion as signal's peaks are squashed. Amplitude of output signal does not linearily follow amplitude of input signal, instead amplitude of output signal increases in lesser ratio the more amplitude of input signal increases. Whether we have soft or hard clipping is just a matter of compression ratios and knees. This instantaneous  compression does not always lead to brickwall clipping, it will however always lead to distortion.

We can't have dry or clean signal and distorted signal simultaneously. Distortion always means harmonics not present in the original signal and once we have them there the signal is no longer dry / clean.

If we have series of compressing stages, this "multistage" circuit, the following compressor will continue from where the previous one left. It will further compress the compressed signal, and that alters the total compression ratio. The signal is basically squashed more and more. The more we compress the more squashed the signal gets and that's why this distortion turns "harder" in total.

And yes, diode series resistors are exactly about altering compression ratios. In Bluesbreaker the stage gain of 220k / 4k7
turns into 6k8 / 4k7 once those diode forward voltages are exceeded. This means very steep compression ratio if compared to, say, substitung that 6k8 series resistor to another 220k. (Peaks won't brickwall clip in either case but the steeper compression has more resemblence to usual peak clipping). Same in shunt: imagine source resistance of 10k and a shunt diode with a series resistor. If we have a series resistor of 1 ohm we have very steep compression ratio and essentially a peak clipping circuit. If we have a 10k series resistor we have less compression and no evident peak clipping.

[amz-fx]

Excuse me but this "stacking" of dry signals on top of square waves makes no sense at all. If it was an analogy, I'm not following it.

We can't have dry or clean signal and distorted signal simultaneously. Distortion always means harmonics not present in the original signal and once we have them there the signal is no longer dry / clean.

If you look at the typical non-inverting diode clipping stage, like a TS-808, you might expect the output to be the classic square wave. It is not.

No matter how hard the diodes turn on, the gain can only be shunted down to +1. This has the effect of producing a (mostly) square wave with a portion of the input signal riding on top of the boosted/clipped signal. If you use a triangle wave for the test, it is easy to see the pointed peaks of the triangle wave stacked on top of the distorted signal. This has the effect of adding some of the input signal into the output along with the distorted signal, and that is what gives non-inverting clipper stages a softer sound.

Best regards, Jack

[teemuk]

Yes, exactly. The softer sound is again about those compression ratios. In non-inverting format feedback diodes can't reduce gain below unity so once their forward voltage is exceeded the amplitude distortion just shifts from compressing peaks to amplifying the lower amplitudes (below forward voltage) in excessive ratio. (Realistically, a good design should never reach this point). I still fail to see where those square waves are?

[Steben]

Yes, exactly. The softer sound is again about those compression ratios. In non-inverting format feedback diodes can't reduce gain below unity so once their forward voltage is exceeded the amplitude distortion just shifts from compressing peaks to amplifying the lower amplitudes (below forward voltage) in excessive ratio. (Realistically, a good design should never reach this point). I still fail to see where those square waves are?

It is a rough explanation of course.
But when you drive a non inverting feedback style with silicon pair unit real hard with a sien wave the limit is "roughly" 0.7v square + dry signal.
I've seen actual graphs showing such a wave. Vivek?

EDIT:



It's why LED's or higher treshold diode arrays achieve a rougher sound - or "tighter" if you want - sound since the ratio of the "squarish component" is double as high. I do like that much more than a tube screamer. And it's why I added some extra diode selection on my SD-1.
About the good design which would not go there, I might personally agree if it is about a base sound. TS and SD qualities are mostly about driving what comes after. But IF it is about base sound, I suggest the somewhat harder clipping before such a stage which additionaly cuts the signal portion above the treshold. It's what the wattbreaker and my BB mod is all about. Using diode to ground clipping between the two stages as I suggest would render an even tighter sound. The combination would give a more than "soft knee effect" for what it's worth with excellent volume roll of action on the guitar. It all adds some depth which is harder with diode to ground clippers only.
Yet for high gain sounds it is all redundant.

[teemuk]

The curve looks smooth but it's really not that smooth to listen to. The high order harmonics have just shifted from peak region to "crossover" region.

Yes, the compression of non-inverting diode "clipper" is smoother at intermediate gains (near and little past forward voltage) because, once diodes turn to conduct, the gain drops to unity instead of zero like in generic inverting configuration (lrading to that "flat topping" clipping). After that point the higher order harmonics start to appear due to abrupt distortion at middle portions of the waveform. The circuit does not "flat top" but it "flat sides".