To clip or not to clip?..that is the question.....

[dschwartz]

Hello everyone here...
I´m having some fun designing a new distortion. My intention is to make it really 80´s hair metal/hot rodded marshall style

I have designed many distortions before, but always following the "what it is supposed to be" approach. By this i mean trying to have a "correct" distortion/OD, plenty of soft clipping, controlled harmonics, transparency, warmth, pleasant decay and everything that everyone talks about when describing good distortion..

But lately i have found that the tones i most like are hard clipping fizzy crunchy raunchy distortion. But with nice decay..I found these kind of tones more dynamic and present in live situations and responds better to pinch harmonics and aggressive fifths, instead of the typical compressed metal tone.

I have experimented with slightly cold biased NPN´s in cascade (2 stages) slamming an lm386mx1. It could be such a perfect tone if it wasn´t for a nasty buzz when palm muting, and matching impedances between npn´s is a nightmare..

logic tells me that i could have a series of soft clipping stages , followed by an asymetrical hard clipping transistor stage.. explosive harmonics, but the decay could be too harsh and abrupt..

my intuition tells me that i should use the hard clipping asymmetrical stage before a couple of typical soft clipping stages, but that could yield an harmonically rich, but too compressed output...

i love distortion design, i think it is a fine art, and i would love to start a conversation about it here..
a penny for your thoughs

cheers!!!

[Derringer]

how about a big clean, adjustable boost stage, whatever tone-shaping you like in that boost (accentuate the highs, etc) into a stage designed to clip really hard at the max boost setting?

Or have what you spoke of, two or more softer clipping stages that can be overdriven by the clean stage to produce hard clipping.

[dschwartz]

Maybe a design that clips one side softly with a lower threshold, and the other side harder with a higher threshold, then, into an inverting stage that does the same..that way each side of the wave gets hard clipped when pushed, and then morphs into soft clipping..much like Triodes do..

I am now playing with joe davisson black fire idea ( buffered npn boost stages) ..it seems that they clip softer on saturation and harder on cutoff, also you can play with the bias just like tube preamps. Interesting! And no jfet matching or trimpots...

[antonis]

Haven't you try clipping with low voltage Zener..??
("noisy" items result sometimes in interesting sound..)

[dschwartz]

Haven't you try clipping with low voltage Zener..??
("noisy" items result sometimes in interesting sound..)

I haven't tried them, but i have studied that option. Using one at the feedback of an opamp (like amptweaker designs) is very asymmetrical, maybe too much. Playing with the opamp Vref maybe can get closer to what i want..

[deafbutpicky]

Maybe a design that clips one side softly with a lower threshold, and the other side harder with a higher threshold, then, into an inverting stage that does the same..that way each side of the wave gets hard clipped when pushed, and then morphs into soft clipping..much like Triodes do..

I am now playing with joe davisson black fire idea ( buffered npn boost stages) ..it seems that they clip softer on saturation and harder on cutoff, also you can play with the bias just like tube preamps. Interesting! And no jfet matching or trimpots...

The gain stage I used for the "fat gary" would be worth further experiments. The gain and decay is certainly improvable but the basics you describe are there:

http://www.diystompboxes.com/smfforum/index.php?topic=109002.msg995283#msg995283

As the links are gone here's the latest schem of it:

[dschwartz]

Looks interesting..it is like 2 fet/pnp fuzz face stages..
But it uses jfets and germ pnps ...
Do the jfets need trimpots?

[deafbutpicky]

take the fet and pnp as one clipping stage. Bias is what shapes the symmetry of the clipping and decay(crackle), feedback (R11/R17) sets the gain. The interesting part is in setting the Q-point for the fet at the gate with Rs and Vc for the pnp with Rc in interaction with the feedback R. Pros are high input Z and germanium goodness, the rest is tinkering.
Comparing it to a FF would be a long shot though there's some similarity when it comes to clipping. It's not sounding as much compressed and wholly.
The idea of the original circuit was to get maximum linearity over the max frequency spectrum in a VCR gain stage and I just adapted it to it push it to the limit to see what interaction of a fet with germanium would be.
Properly pushed, biased and shaped it would be an interesting alternative to usual clipping stages and it retains the somewhat stiff behaviour of hard clipping.

[dschwartz]

i came up with this one:




First stage clips very hard o the rails, second stage uses leds and a voltage divider to get soft clipping and higher threshold, so it will also clip hard if pushed hard enough..
note the vref, is about 5.5v to change the symmetry of both stages. usual wstc tone and volume at the end..
note how tube-like is the clipped wave and the transition is very smooth (in theory at least)

[teemuk]

I have designed many distortions before, but always following the "what it is supposed to be" approach. By this i mean trying to have a "correct" distortion/OD, plenty of soft clipping, controlled harmonics, transparency, warmth, pleasant decay and everything that everyone talks about when describing good distortion..

But lately i have found that the tones i most like are hard clipping fizzy crunchy raunchy distortion. But with nice decay..I found these kind of tones more dynamic and present in live situations and responds better to pinch harmonics and aggressive fifths, instead of the typical compressed metal tone.

Well, the problem is that "what is supposed to be" approach is usually erroneous. Subtle distortions, such as soft clipping, may not generate enough of those high order harmonics that are required to generate certain timbre for the distortion. If you only soft clip harmonics inherent to harder clipping will remain low in magnitude. In turn, distortion does not generate harmonics required to generate those traditional "pinching" squels or power chord tones.

You just discovered that.


What those subtle distortions introduce, however, is greater magnitudes of overall harmonic distortion, though somewhat low in order. When you amplify complex signals that harmonic distortion creates intermodulation distortion (at frequencies un-related to fundamental) throughout the effective bandwidth and if IM increases too much it severily degrades "note separation" turning everything into fuzzy mess. With soft clipping IM may not create high order harmonics but that just results to "darker" timbre of distortion, which has "fuzzed" detail and sounds discordant while chording (complex signal input). So, if you're going to ovedrive a lot then you need to reduce intermodulation in order to retain signal intelligibility.

Best way to reduce intermodulation distortion is to reduce harmonic distortion, which means clipping schemes that introduce least amount of THD may be preferred. One clipping scheme that introduces almost no harmonic distortion at all - below its threshold - is, surprise, surprise, hard clipping. Hard clipping creates very little IM below clipping threshold, and above clipping threshold a plenty. Yes, the distortion turns more abrupt and higher magnitude of higher order harmonics make it sound brighter. Then again, the bright tone cuts through the mix much better, and below clipping threshold there is less IM to turn the signal into fuzzy mess.

This is also where we get to symmetric versus asymmetric distortions. Symmetric distortion creates largely only odd order harmonic content, so effectively intermodulation is restricted to modulation of odd harmonics only. Asymmetric distortion introduces BOTH even and odd order harmonics so it introduces more overall IM because modulation covers now combinations of even and odd harmonics. So once again, if you overdrive too much then asymmetric distortion with higher harmonic content will cause intermodulation (and its effects) to have a greater magnitude.

So schemes that introduce subtle distortions, and distortions with more harmonic content overall, may not work too well if you aim for "high gain" tones. But they can nicely warm up "cleaner" tones without IMD freaking to much havoc. If higher gain, and associated timbre is required, the distortion must be able to change its characteristics. The harmonic design of distortion circuits typically aims towards exploiting such operation: You may find that the clipping is "soft" and "asymmetric" at low gains, while higher gain push the clipping to be harder and more symmetric. Since the shifting is usually cauased by DC bias shifting it will have discernible time-variant response which one senses as "feel" or "touch sensitivity".

[TejfolvonDanone]

Haven't you try clipping with low voltage Zener..??
("noisy" items result sometimes in interesting sound..)

I haven't tried them, but i have studied that option. Using one at the feedback of an opamp (like amptweaker designs) is very asymmetrical, maybe too much. Playing with the opamp Vref maybe can get closer to what i want..

I have a design with a 2V Zener in the feedback loop of an op-amp.
http://www.aronnelson.com/DIYFiles/up/Z_DIST.pdf
It works really well but needs some tweeking:
1) Add a 100 Ohms resistor in series with the 9V supply.
2) Omit the C4 and the tone switch (it really decreases the gain and interacts with the gain control in a not so good way)
3) Add a 100 Ohms resistor in series with the Zener.

This goes from goes from a clean booster to an aggressive and all the way to a nasty fuzz.

[anotherjim]

Just wondering Daniel, if you know of a pedal design that comes close to what you want?

And to be careful, I reckon the amplifier & cab used might have significant impact. Even a clean channel isn't necessarily clean in all respects. A tone stack with a strong and permanent mid-scoop can have a severe impact on things. If you want the pedal to do most of the harmonic balance, an amp channel that can be set flat (at least 100Hz-5kHz) is maybe vital?

Maybe also, a sealed cab type is to be preferred. It seems to me that a heavily distorted tone makes more sense tied to an identifiable point-source in the room. An open back lets it wash all over the place and the brain identifies certain aspects as background noise with no obvious origin.

[dschwartz]

@Teemuk:
I've read your book so many times, most of my new designs always have something I've learned from you

I don't know if you know I'm a full-time pedal designer/builder now. I'm​ quite experienced on designing distortions, but there's always something one can miss..like IM distortion, of course!..

I have spent like 2 years trying to design the perfect distortion and came up with this:

I absolutely think it is amazing.. but last week i was playing with my burning crunch and i was delighted with that extra crunch..square wave goodness. The drive maker is very amp like and can get those jcm tones..but in a more "educated" manner..

The final experience was last rehearsal with my band of friends. The room had this 40w tube amp with busted power tubes..i couldn't get a single clean note at drum volume..so i split the signal and went into the PA with the omnicabsim..
But the PA preamp clipped very hard when i used the drive maker booster and it sounded AWESOME on the mix, and on the fingers!! I felt like playing with the best marshall amp...and could control everything with the volume on the guitar..

That's why I'm​ beating this horse again..maybe it is not completely dead...

[deafbutpicky]

@Teemuk
now that's how to sum up half a book. Excellent!

@Daniel
Just an idea about dynamics:
Did you try to change the order to soft clip first and then driving a hard clipping stage?
Clipping it hard first will limit any headroom after the signal hits the rails and the transition to soft clipping is mostly controllable with the gain knob. I think the other way round one could push the amount of soft to hard clip better with playing dynamics as
the signal amplitude is still reactive to playing when the first crunch sets in. Somewhat like a tube amp entering saturation.

[dschwartz]

@Teemuk
now that's how to sum up half a book. Excellent!

@Daniel
Just an idea about dynamics:
Did you try to change the order to soft clip first and then driving a hard clipping stage?
Clipping it hard first will limit any headroom after the signal hits the rails and the transition to soft clipping is mostly controllable with the gain knob. I think the other way round one could push the amount of soft to hard clip better with playing dynamics as
the signal amplitude is still reactive to playing when the first crunch sets in. Somewhat like a tube amp entering saturation.

Thats exactly where my doubts are..usually i use soft into hard clipping..it sounds really good, but since the hard clipping is after the soft one, what you get is mostly hard clipping.. so, when you play softly, still you are getting hard clip mixed with soft clip.
The circuit i proposed changed the order..if you play softly, you get mostly soft clipping, and hard clipping occurs during loud notes. Also, letting the first stage clip to the rails should be beneficial noise wise, and the soft clipping stage has a higher than usual threshold.

Good thing is i can change thresholds and hardness per stage unitil i find a nice combination.
I remember using a diode in parallel to the second stage inpur resistor on previous experiments and i got pretty wild pinch harmonics friendly tones, but the decay was awful..
I will try that also...

[bool]

R-R clipping is a B.S. imho.

It's (at least) extremely dependant on supply rail voltage. If you are concerned with repeatability (and not cost-effectiveness), this is imho a bad approach. And not all batches of opamps are *exactly* the same with regard how close to rails they go ad how do they stick to the rails when they get there. The bonus point is however that you can "mod" the tone with simple opamp swapping.

If you want a semi-foolproof approach that's a little "harder", try LEDs in an inverting stage. Red + green LED for a bit of assymetry. The trick is to drive them "harder"; from a relatively low-impedance. You'll have to find the sweet spot, but it exists.

So an input buffer/pre-gain stage, maybe with some EQ curve; into an inverting LED clipper; into an inverting "softened" standard diode clipper.

Filtering around clipping stages always brings the noise down a bit. So you can incorporate some inter-stage filtering to combine the "tonal" frequency shaping as well as improving the noise performance.

[BubbaFet]

My goodness... THAT does sound like the perfect distortion! Nice!

[robthequiet]

The Rockman Sustainor has an attenuation setting ["Rhythm Vol'] so that you can switch the output volume up and down with a footswitch. In theory, it lets you keep your loud crunchy tone but reduce the volume, hence no change in timbre or dynamics, so as to duck under the vocals in a verse and switch back up for the bridge. I found that you get a more boring version of the crunch tone when you do it, because the ear expects the timbre to change when the volume goes down, at least imo. There is some psychoacoustic property to the sound that comes into play.

I think that a subtle compression +boost might help with the sparkle at low distortion levels, but what I would like to test is putting the higher threshold LED-type clipping right at the front, then have a 4148 stage then a Ge diode stage successively, because I think that once you drive the LEDs into clipping, you still get some rounding of the tone from the lower threshold stages, depending of course on how much gain each stage has. Does this make sense?

[bool]

More than 2 stages is also too much (imho).

I think that the recipe would be to buffer (or with added "800Hz" eq; since we're already heading into Sholtz tone territory); then led hard-clip (with band-pass to kill the content upwards of something like 1.8kHz); then into a soft-clip standard silicon diode stage.

The led hard clipping stage would act as a sort of "limiter" here; in contrast to pre-compression as in Scholz gadgets. But so should input and the following stages contain band-passing filtering to get to the 80s tone.

[dschwartz]

My goodness... THAT does sound like the perfect distortion! Nice!

Thanks bubba, it was such an obsessive task!!