JFET overdrive, how to preserve/enhance dynamics?

[samhay]

^I think that what is lacking in all those JFET distortion makers is IM distortion

I don't follow. As soon as you have a box that makes square-ish waves out of your signal (and that's what most of the FET OD's seem to do) then you will get plenty of IM distortion if you play more than 1 note at a time through it.

[tca]

Before the square-ish it and how it is square-ish-ed.

You are right but... not all frequencies are squared the same way and at the same time (same phase) and with same duration (knee).

You will have to simultaneously balance clipping (which is frequency dependent), amplification (again freq. dependent, even when non clipping, think about distortion when the quiescent point is moving due to a large input signal) and mixing (IM) (and again freq. dependent).

It is truly a gourmet secret formula to balance all the ingredients!

Cheers.

[Steben]

It would have been done over and over already by commercial builders already.

In a strict logical sense that sentence is false, i.e., that's not the all truth.

I think that what is lacking in all those JFET distortion makers is IM distortion:  The elephant in the dance floor that nobody sees!

Cheers.

What do you mean, intermodulation?
Even a crappy opamp has that....

I think the major problem is the addition of power sag and output transformer saturation.
That is what all the mid drive tube output distortion is about.
very low drive is easy, very high gain is easy. But designing solid state that does all the range, inlcuding dynamical feeling by the player is very very very difficult.
Power sag is not just compression, it is about lowering the headroom as well, not just the gain (otherwise again would be easy).
OT saturation is not just clipping, it is frequency dependent as you said.
Ideal emulation would include when raising signal a very slight loss of gain, a certain rate of headroom loss and loss of bandwith (especially at the higher range).
On top of that, there is the "hysteresis" of the transformer, which brings us back a bit to your IM suggestion. the higher you drive a transformer, the more it will "lag behind".
This means the crossover distortion share in the total rises when going to higher saturation, instead of solid state, were the crossover distortion is the only distortion at lower levels.

That's why every simple SS design only features a part of the total pallete.
- TS/OD ciruits feature nice soft distortion, but they lack both the brute raw full gain character as the dynamic frequency response, which brings us to describe TS as a "midrange horn" at low to clean settings and tooo smooth at high gain.
- opamp clippers tend to have a decent low budget high gain emulation, given their slew rate (loss of highs at high gain) and crossover distress. At low gain however, they are the worst possible circuit, because all the high gain distress factors don't go away. We ten to call this "Flabby, farty, undynamic"

The best emulation therefor "combines" many circuits, where brute opamp clipping usually gets pushed in at the highest gains.

You are right about the "gourmet", but i don't think it is IM.

[tca]

... but i don't think it is IM.

I agree with all that you say. What I mean about IM is something like this:

Most sims include a wave form graph, say, at 1Khz, and say what happens to the output signal, it is clipped so and so...

Take for instance, not a FET but anyway, the electra distortion (same thing for any 3 legged device, FET, MOSFET or tube):



One note at 1Khz and two notes at 1Khz+1.2Khz. The first is non-clipping the second is clipping, there is more on the 2nd plot than the two freqs. taken separately (even if clipped). There's much more than clipping... but clipping and IM go hand-in-hand.

P.S (edit)
Added a non-clipping sim for the two notes case.

P.P.S.
Clipping => IM (the reverse is not true)

[Steben]

I think I'll make it a bit clear as well: IM happens, yes. But it happens in all circuits. It even happens in clean signals. Think of tuning your guitar. Even when two strings are not tuned well, you get IM notes popping up.
The way a certain circuits responds concerning IM is largely connected to the way it clips, however. The softer the clipping, the smoother your spectrum analysis goes. A TS-style circuit has less harsh IM notes.

[DDD]

To be correct, IM distortion takes place only in non-linear systems.
Guitar itself is a mechanical non-linear system, so it produces IM tones... as well as human ear, by the way.