harmonic content, diode clip vs. jFet triode emulation

matt239 · Post by **matt239** » 04 Dec 2011, 21:00

I've seen analyses of the harmonic content of over-driven tubes vs. transistors.
I've seen similar for clipping op-amps.
The jFet triode emulations are supposed to have a specific harmonic content, similar to triodes; low order, (dominant 3rd?)

What about soft clipping diodes? (w/o op-amp clipping)
Has anyone done/seen a harmonic analysis?

What about germanium diodes with series resistors in op-amp feedback loop?

What's likely to sound most "tube like?"

bajaman · Post by **bajaman** » 05 Dec 2011, 00:33

What's likely to sound most "tube like?"

a tube

bajaman

matt239 · Post by **matt239** » 06 Dec 2011, 00:26

bajaman wrote:
What's likely to sound most "tube like?"
a tube
bajaman

No doubt..

My question was really about the harmonic profile of soft clipping diodes. I know that jFets can be operated in such a way as to produce the desired harmonics, but the headroom seems a bit limited in some of the implementations I've seen. The dynamics aren't quite right, goes suddenly from clean to distorted..
I do find much to like about some of the Runoffgroove jFet/"triode" amp emulations, and several of the cascaded jFet mu-amp designs around, I'm just concentrating now on the "less crunchy" region, from what most folks would call "clean" (but can sound rounded & sweetened through a tube amp..) to just slightly "driven."
I'm building a drive with soft diode clipping. The dynamics are somewhat controllable; series resistance, "diode ladders." but I was wondering about the harmonics.
Not so much even vs. odd, but low order vs. high order.
A device could be built using both, but if the harmonics of the diodes look good, I'd skip the jFet.

Of course there is also to consider non-linearity of the transformer, power supply, and speaker.. but one problem at a time..

RnFR · Post by **RnFR** » 06 Dec 2011, 01:23

you could try FETs at a higher voltage. say 18-24v? I've read that they are closer to triodes that way.

phatt · Post by **phatt** » 06 Dec 2011, 09:41

Hi Matt,,
What your asking is what the best minds have struggled with for years.

IME,, fet circuits even at 37 volts just die when push comes to shove,, most stuff is easier with opamps even though they are not perfect for guitar stuff.

If you really want to get serious stuff happening you need to look at some of *KMG's * work with mosfets.
Very impressive stuff. And don't forget the transformer. that helps!

Even with all that you may still need some form of tone shaping after the output.

That is why I use a ReAmp system (with Valve power stage soaked).
Or I use the tricked up triode in the Tricked up Mesa Vtwin into My DDC,, which is just plain simple Diodes around opamps followed by a cab sim type setup.

IMO,, Everyone focuses on triode curves which is nice BUT a major part of the mojo is the tone shape through each part of the circuit.

I'll even stick my neck out and state that most guitar rigs have more gain/distortion that can ever be used ,,,, What they lack is TONE shaping.

I spend forever trying to get this through to young kids at a local muso jam that tone shaping is supercritical for what they chase. But no they keep spending money on Name Dropped brands that get rave reviews from over rated experts who can play fast apreggios and make crap sound half decent.

Do yourself a great big favor and get some simulations happening.

When you get a handle on how they work simulate some famous *Valve circuits* and watch how the tone shape changes as it passes through each section of the Amplifier.

Now simulate those *hot rod fet circuits*,,, then some *super tricked up opamp based circuits*.

It won't take very long for you to SEE that ,,, they are all VERY different it tone shaping abilities.

One thing you need to get your head around is Gain,,,, At some point you can't add more.

Once the signal gets to square wave,,,That's it there is nothing left.

:(:(

So Fets, opamps or Valves,,,if your sound is not pleasing to you the Answer will likely be found in tone shaping. HINT!!!!!!

My Amp is a crappy old SS 40 Watt Laney linebacker,,, yet the hot rod guitar players at my club Who own far more expensive gear than I will ever own can't believe how my guitar is always the loudest.

The Bass player has a 400 Watt rig,, the other guitar player has a Carvin Legacy,,, another has a Cyber Twin,,,,and when push comes to shove I'm the loudest player...No Contest

LOL.

The answer to that is TONE shaping,,,,, It is by Far the most critical aspect.

As I have taken 30 odd years to learn,,,I'll take 30 watts of great tone over 100 watts of MUD BS Brand name crap anyday .LOL

Oh As to Harmonic content,,,, just get the tone shaping right and the harmonic stuff will take care of itself,,,,winky

End of rant.

Phil.

FiveseveN · Post by **FiveseveN** » 06 Dec 2011, 10:00

What exactly do you mean by "tone shaping"? Would that be pre- post- and inter-stage filtering? And if so, what makes topologies using various active devices "VERY different it tone shaping abilities"? A filter is a filter, it's a linear function.

phatt · Post by **phatt** » 06 Dec 2011, 11:19

FiveseveN wrote:What exactly do you mean by "tone shaping"? Would that be pre- post- and inter-stage filtering? And if so, what makes topologies using various active devices "VERY different it tone shaping abilities"? A filter is a filter, it's a linear function.

"A filter is a filter, it's a linear function."

Ok go test it with real working circuits.
Use a Valve, a fet , then an opamp.

Don't be surprised if they ALL sound different. Winky.
Phil.

FiveseveN · Post by **FiveseveN** » 06 Dec 2011, 12:12

1. The fact that they "sound" different might have more to do with non-linear aspects (the active devices' distortion) than what I'm talking about.
2. In real life no component is pure gain, pure resistance, capacitance etc. Each active device has non-infinite input impedance, non-zero output impedance, parasitic capacitance etc. And these properties may vary with operating point. But we know that. I was assuming the designer is aware of these things and will compensate for them when it matters.
And in the narrow-bandwidth distortion-full land of guitar amplifiers and effects, does it really matter unless you really screw things up?
That being said, emulating a device's static characteristics is a piece of cake. It's the dynamics we have trouble with, and shrouding some basic filters in the Holy Veil of Mojo isn't helping anyone.

phatt · Post by **phatt** » 08 Dec 2011, 08:53

FiveseveN wrote:1. The fact that they "sound" different might have more to do with non-linear aspects (the active devices' distortion) than what I'm talking about.
2. In real life no component is pure gain, pure resistance, capacitance etc. Each active device has non-infinite input impedance, non-zero output impedance, parasitic capacitance etc. And these properties may vary with operating point. But we know that. I was assuming the designer is aware of these things and will compensate for them when it matters.
And in the narrow-bandwidth distortion-full land of guitar amplifiers and effects, does it really matter unless you really screw things up?
That being said, emulating a device's static characteristics is a piece of cake. It's the dynamics we have trouble with, and shrouding some basic filters in the Holy Veil of Mojo isn't helping anyone.

Sounds good to me mate,

May I say, a very good simple explanation of whatz happnin inside these things.
Maybe you can explain it all to our friend Matt who's is havin a hard time trying to get his head around it all.

I just think folks are better off if they proto a few test circuits out for themselves and over time (with some good reading) a few things will fall into place.
Phil.

matt239 · Post by **matt239** » 09 Dec 2011, 04:10

The more I think about it, the less sense my original question makes..

The diode clipping will have different harmonics with different wave-shapes, depending on the circuit, & levels.

Still would be nice to see some graphs under different conditions.

kleuck · Post by **kleuck** » 09 Dec 2011, 08:19

A friend did some tests with some of my circuits. (in French)

kleuck · Post by **kleuck** » 09 Dec 2011, 11:02

Will not give the exacts schematics, but the BTB is fet/mosfet based, the SRT derives from the craig Anderton's tube sound fuzz, so CD4049, and the differences are mainly the voicing : aka the harmonic content.

matt239 · Post by **matt239** » 10 Dec 2011, 18:12

Tres bien. Merci!
- Those are some good looking curves!

Checked out some Youtube vids of those pedals; Nice! especially the BTB.

Well done.

kleuck · Post by **kleuck** » 10 Dec 2011, 22:38

Thanks (Docloco is THE guy i think to demonstrate a booster, helps a lot), i posted the link to my friend's blog, as he thinks that the "tubey sound" of these two has a lot to do with the limited bandwith, and indeed when working on the SRT (derived from the tube sound Fuzz) each new prototype had a lower response in the highs than the previous, the whole trick is to find the sweet spot, enough, but not hearable as a loss of brightness.

artifus · Post by **artifus** » 11 Dec 2011, 01:21

freestompboxes.org dedicated stompbox dissection since 2007

Les circuits restent secrets car, à force de réflexion et d'essais, Kleuck a trouvé quelques astuces inédites en matières de composants, de polarisation et d'architecture.

Google Translate:

The circuits remain secret because, by dint of thought and testing has found some tips Kleuck unpublished material components, polarization and architecture

kleuck · Post by **kleuck** » 11 Dec 2011, 01:41

A better translation would be "Schematics remains secret, as Kleuck found some new tips and tricks regarding bias and architecture" (no special components, that's not true)

The BTB is a "brand new" discrete design, so i do not want to reveal it (yet), the SRT is not, deriving from the Anderton's Fuzz (or Red LLama if you prefer) so i won't hide that, even if it's far from the original in sound.
My "Ge" BTB, more classical is not really secret neither, as i discussed and published my prototype's schematic on a French forum (not the actual i'm selling right now, but there's not a lot differences, only optimisations, but "the devil is in the details" sometimes

I'm not here to publish all my schematics if it is your idea, though i will for some of them.

phatt · Post by **phatt** » 11 Dec 2011, 12:55

kleuck wrote:Thanks (Docloco is THE guy i think to demonstrate a booster, helps a lot), i posted the link to my friend's blog, as he thinks that the "tubey sound" of these two has a lot to do with the limited bandwith, and indeed when working on the SRT (derived from the tube sound Fuzz) each new prototype had a lower response in the highs than the previous, the whole trick is to find the sweet spot, enough, but not hearable as a loss of brightness.

BINGO!!!!!
Nice one Kleuck,,

That one line covers the basic concept of what folks need to get inside their heads before they build stuff.
Whether you use glass or silicon,,This Whole game is won or lost with TONE SHAPING.

Phil.

kleuck · Post by **kleuck** » 11 Dec 2011, 20:41

I have another example
I made a clone of an old DO -i usually do not make clone, but a nice guy asked me to, as he's a Malmsteen fan, and there's not a lot of clone, most expensive, i accepted indeed because my first pedal ever, and the beginning of my interest in electronics was a Disto+, almost the same circuit-
He wanted the most realistic in sound possible, old ones used Panasonic or Ceramic caps, sometimes the two, which i choose to do too.
One of the interests of this circuit, is that the use of a very bandwith-limited 741, along with the topology of the nfb, reduces the highs and the low as you are turning the gain pot clockwise, making it really as advertised on the box a "Booster-overdrive" more than an overdrive.
That's already a good thing, but i did a mistake, and in the clamp section used a 10nf cap as a filter, where there should be a 1nf !
You should think that the sound would become dull, but not at all, it turns the DOD in a very very good OD, perfect for blues and classic rock, touch sensitive and natural-sounding (yes, with the original SI clamping diodes) !
I would not play an original, but this simple mistake radically changed the sound of the pedal, for the best i think.
Replacing the clamping resistor too for a 8K2 instead of 10K made perfection (to me)

matt239 · Post by **matt239** » 11 Dec 2011, 20:51

RnFR wrote:you could try FETs at a higher voltage. say 18-24v? I've read that they are closer to triodes that way.

I am starting to consider @ least +/- 9V supply for whatever device. Single 9V supply just doesn't allow much wiggle room..

kleuck · Post by **kleuck** » 11 Dec 2011, 23:01

For me, with fets (or Mos), the main issue with a low voltage supply is the low voltage gain you can achieve, BUT, using them in a SRPP or µ-follower topology can give real high gains.