What is "amp-like distortion" anyway?

[Ben N]

hMMM... Jack seems to have used non-inverting stages--which would seem to kind of defeat the point, no?

[merlinb]

hMMM... Jack seems to have used non-inverting stages--which would seem to kind of defeat the point, no?

I doubt it makes much difference tbh. Shame he didn't bother to make it work on 9V though.

What bothers me about that circuit is that it starts bias shifting long before it gets close to clipping, which defeats some of the effect and is not how a tube amp behaves. The patent does mention this and show a second variant with an extra zener, but it's still a bit of a kludge.

[teemuk]

That St. Louis Music patent circuit looks like it might be fun to breadboard.

Note that Crate and Ampeg amps typically will not employ just one of such clipping stages. They typically cascade at least two.

The patent also doesn't factor in how those amps are "voiced" overall. The circuit alone will not sound like VH-140, for instance. Not even remotedly like one. 

[Ben N]

I doubt it makes much difference tbh.

Just going by the patent claims, which seem to argue that the inversion (at least of the second stage) matters to the overall effect.

Note that Crate and Ampeg amps typically will not employ just one of such clipping stages. They typically cascade at least two.

Do you mean two stages the way they appear in the patent application, or two pairs of stages?

The patent also doesn't factor in how those amps are "voiced" overall. The circuit alone will not sound like VH-140, for instance. Not even remotedly like one. 

Yeah, for pedal purposes I would assume that at a minimum a buffer, some HP filtering and maybe some voltage gain would be needed before this circuit, as well as some kind of eq after. The patent application says as much. But I also note that Jack's version doesn't seem to have bowled anyone over, and that he presented it more as an interesting idea for an alternative to yet another iteration of two diodes and an opamp.

[teemuk]

Do you mean two stages the way they appear in the patent application, or two pairs of stages?

The patent image is largely a simplification of overall idea. There are many different versions of the circuit; Zener voltages may differ (to achieve different kind of asymmetry of clipping) and some embodiments employ "discrete" Zener circuit, for example. The schematic shows the overall idea how the circuit works and I guess the patent probably lists a number of alternative embodiments how the circuit could be built as well.

Yes, I mean two pairs of clipping stages, each employing the patented DC shifting circuit. There are also versions with just a single shifting stage cascaded to one "conventional" clipping stage. IIRC, "Powerblock" featured just a single stage for distortion (but it doesn't have "OD" channel either and the circuits is employed for "clean" signals as well). Perhaps try downloading some Crate schematics. This circuit has been exploited very often ever since it was conceived. There's quite a lot of variance how they are actually put together (gains, clipping thresholds, etc.) but none of that changes the overall idea.

Here's (partial) circuit diagram of Ampeg VH-140C preamp for reference:

Note the great deal of pre and post distortion tone enhancenment employed. IME, that contributes a lot more to overall tone than these distortion stages. Asymmetry and dynamic offset shifts of distortion are just some extra spicing to make the effect less "static".
Here's (partial) circuit diagram of Crate GX-120:

Notice a conventional clipping stage cascaded to patented circuit but with "discrete zeners". More importantly, notice the use of similar enhancing circuitry architecture to shape tone pre and post distortion.

[Ben N]

Asymmetry and dynamic offset shifts of distortion are just some extra spicing to make the effect less "static".

Yup, that's what I took from it.

[teemuk]

It's been a long time since I read that patent the last time. I can't remember all the details about it. Luckily, anyone interested can access it by Google Patent Search in just few seconds.

Anyway – and this is stuff you more prominently learn from circuit diagrams, not from patents - SLM approach to "voicing" of the high-gain and clean tones isn't all that different from what other amp makers employ as well. (After all, they are practices that overall seem to "work" and are also provide results that are quite popularly favoured by guitarists). The high-gain channel introduces pre-distortion enhancement that is basically a rather distinct band-pass filtering at mid-range frequencies, centered around approximately 1 kHz. Lower frequencies are attenuated radically because IMD in great magnitudes (i.e. "high gain") would produce a very unintelligible "farty" or "muddy" tone, instead of that "tight" response we've become accustomed since, well, mid 1970's Marshall amps.
The VH-140C cuts higher frequencies (pre-distortion) perhaps a bit more than "usual" high-gain voicings but the audible effect is not all that radical because higher order harmonics of distortion subsequently re-enforce those attenuated higher frequencies (and add a fair bit of new ones as well). Rolling down high-end response also helps to reduce resulting (high frequency) intermodulation and related "fizzyness".
Overall, IMD is dominated by the mid-range frequencies, which results to that "gnarly" Marshall-y -type of tone instead of that "fat" but somewhat less "focused" and less "tight" tone of, say, overdriven Orange amps. So, this an approach to produce that "modern metal" tone, not that "fuzzier" high-gain tone dominant in genres like pre-metal (e.g. Black Sabbath) or stoner rock/metal.
---
In post distortion enhancement the low frequencies are somewhat boosted back again. This brings some nice authority to that "chugging" palm mute –style of playing and overall helps to balance the sound making it less "thin". Higher frequencies are somewhat accentuated, or I guess one could alternatively say that mid-range frequencies are "notched" a little. If you think about it, this is also what a power amp with poor damping factor does when it drives a (reactive) speaker load. Upper high frequencies may be attenuated (to reduce fizziness of distortion) but sometimes this is left as a task for the low-pass filter introduced by the speaker system. It is courteous to provide some sort of controls for fine-tuning these three bands (bass, mid-range and treble) because it greatly affects timbre of the distorted tone.
---
If you research this, you will find that in general such voicing practice applies to several "high-gain" amp designs. In what magnitude they do this, and what cutoff frequencies they employ, characterises their tone and provides each design a somewhat unique timbre. But overall the voicing practice follows that generic approach: Cut lows (and possibly highs in preferred magnitude), distort, boost back lows (and possibly highs in preferred magnitude).
---
"Clean" channel, at least in VH-140C, has enough gain to "Crunch" as well. (Basically with high gain settings it provides more gain that "OD" channel at lower gain settings). However, distorting "clean" channels are typically subject to less drastic voicing practices. The bandwidth of the signal is, for example, less limited before signal gets distorted. Because distortion (and more specifically intermodulation distortion) is generally produced in lesser magnitudes for "clean" tones this is a tolerable approach. With too much bass reduction in effect the resulting tone would be way too "thin". Likewise, there is less need for post-distortion enhancement to "balance" the tone (e.g. no need for bass boosting). Mid-range notching may be featured (and is in Ampeg's case and many other amps) for overall purpose of producing tones people have been familiar with since about silverface Fender amps.

Anyway, these sorts of enhancements are what amps generally do to mangle the signal. Today it is also getting somewhat vague concept to distinct amps from several distortion effects because those effects are more and more starting to resemble generic preamps and may also introduce elements of characteristics typical to power amps and even speaker systems. As I mentioned earlier in this thread, that is quite different approach in comparison to pre mid 1970's effect designs. A vintage distortion effect necessarily did not introduce much enhancement or voicing. In simplest it was just an overdriven gain stage or perhaps a gain stage driving a pair of diodes into clipping. You wanted that "fuzzing" effect (or even amp) to sound brighter and tighter...? You drove the thing with a treble booster pedal. Soon designers implemented that treble boosting feature permanently to amps. Then to effects. And so on. When effects in the late 1970's and onward started to mimick amp sound´s more closely they started to integrate amp-like features as well. And vice versa, amp circuits copied effects. And now the difference between the two (amp vs. effect) isn't so obvious any more: Rockman amps, early SansAmp effects, and all alike, heck even the Boss Metal Zone are basically full-blown preamps with very "amp like" characteristics. Most amps are basically effect processors with a power amplifier, which in many cases also works like an effect processor. Hardly anyone anymore uses a conventional "amplifier", which those things generally still were up to early 1960's.

[bool]

Ihmo you could leave the rockman "processors" out of this grouping. Afaik the whole rockman concept was to deliver a tone that already "sounds (almost) like a record" instead of delivering a tone that sounds like a spot-on amp emulation. Back in days there were rumours of people requesting that srd mods their rockmans to switch off some of the processing to make it easier to work with it in studios etc etc...

[garcho]

also provide results that are quite popularly favoured by guitarists

That is a huge consideration for businesses that have a loyal customer base. I'm sure a million young employees have presented "fixes" or "something new" only to be told by brass that customers want the unfixed version, so piss off.
And brass is probably right. I've witnessed so many guitarists who normally play say, a Fender amp, for so many years that playing through say, a Marshall, sounds "wrong" to them. They fiddle a moment, frown, and then stare at the knobs like, "what do these do again?". Is the Marshall amp actually wrong for "their tone"? Most likely not, they just don't know how to instantly "dial it in". I see that happen with outboard gear, pedals, synths, mixing boards, consoles, etc. How's the cliche? Some people know what they like, some just like what they know.

deliver a tone that already "sounds (almost) like a record" instead of delivering a tone that sounds like a spot-on amp emulation

This is such a huge thing, I can't believe how little people around here consider that difference as they argue about slew rates and soft knees. When people start rattling off big name artists and their gear and tone, I always think "their tone from the studio on a Neve with a Fairchild, a 414 et al? Their tone coming out of your stereo speakers? Their tone from the 5,000 seat auditorium with ultra complex acoustics/bodies, a PA, stage sound bleeding in the mix, etc.?" If your sound sounds good to you, it's good. Chasing someone else's sound is impossible, unless you've spent time in the rehearsal space with them, listening to nothing other than their guitar, chain, and amp. Even then the room can really change things. If you're chasing tone from recordings or big venue shows, you'll just end up chasing your own tail.

[Fancy Lime]

Hi there,

sorry for my lack of participation. Super busy at the moment. This thread has turned out much more educational than I could have hoped for. Noice!

Anyway, I want to share a schem for what I breadboard to start experimenting:



The R and C values are mostly just pulled from where the sun don't shine. Have to tinker with those. I went with a dual CMOS inverter drive stage. The first inverter can be biased symmetrical or asymmetrical, the second one runs symmetrical and without additional gain. This allows to dial in the envelope-dependent symmetry if the user so wishes. A passive midcut before and baxandall/james tonestack after the drive stages are something that has worked well for me in the past. Simple yet flexible. I left out a "proper" cabsim and just included a 2nd order highcut at 5.8kHz. A proper cabsim may be added along the way. As may one or more effects loops, headphone out, symmetrical line driver and possibly a semiparametric mid badn eq.

Cheers,
Andy

[Ben N]

Intuitively, Cmos seems like a great way to approach this. Cool, Andy.

[Eb7+9]

the authors of this paper provide illustration of a large-signal equivalent
transfer curve for a push-pull power tube stage

https://ant-s4.unibw-hamburg.de/dafx/papers/DAFX02_Moeller_Gromowski_Zoelzer_measurement_nonlinear.pdf

a different style curve would be extracted for a single ended amplifier
ie., a Champ

[Steben]

the authors of this paper provide illustration of a large-signal equivalent
transfer curve for a push-pull power tube stage

https://ant-s4.unibw-hamburg.de/dafx/papers/DAFX02_Moeller_Gromowski_Zoelzer_measurement_nonlinear.pdf

a different style curve would be extracted for a single ended amplifier
ie., a Champ

... and a different curve from a feedback style AB amp....

[Eb7+9]

yes, good point ...
and curious to see how it would differ then