SS analog warmth / tube sound circuit wanted

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I recently had the pleasure of hearing a friend's vintage audio system which I really enjoyed. It's nothing fancy, but I loved the tube sound.

After a lot of reading here and elsewhere, I have learned that the "tubiness" comes from second-order harmonics.

I'm not at the stage where I'm comfortable working with high voltages, plus my budget is limited so good transformers are off the menu. So my question is this - is it feasible to simulate this 2nd-order distortion with solid state technology? I've seen a few mentions that JFETs can be used, but all the designs I've found so far have been for guitars (overdrive, fuzz, specific guitar amp emulation etc).

I also encountered this: https://www.crowdsupply.com/analog-artistry/leo-vintage-analog-sound , but from what I see in the limited technical information available this is effectively just EQ. I may very well look into this, but at a later stage.

Does anyone have a circuit I could try? And is it possible to have an adjustable distortion level :scratch:?

I do understand that this is the opposite of hi-fi, and essentially amounts to an effects box. But I thought it could be fun. Ideally this will be something I could put between source and amplifier with an on/off switch :D
 
Last edited:
Administrator
Joined 2004
Paid Member
You have just opened a can of worms. :Popworm:

Tube sound does not come from 2nd harmonics in the opinion and experience of many people, so you are likely to get a lot of argument about that. If you are a little handy with software, there are wav editors out there that will allow you to add distortion to an audio file. You'll find if that works for you or not, before you build something.
 
Much less in need since my last DAC/amp combo, but I've been tempted a lot to try this: A Harmonic Restorer. OK, it is still tubes, but no Xformers... and you can use for example two small leftover transformers with secondaries interconnected to get about the right supply voltage. Beside two standard tubes, nothing fancy. Keep us informed about your progress!
 
Here is the effect of Harmonic Enhancer in Audacity, it makes obviously big distortions, asymmetrical for even order harmonics, but sounds nice
 

Attachments

  • harmonic enhancer audacity.jpg
    harmonic enhancer audacity.jpg
    169.2 KB · Views: 274
This buffer is what the late Allen Wright called a Super Linear Cathode Follower (SLCF) and I believe he got it from an early very high end Tektronics tube scope.

I built one as one off an output buffer for my Lightspeed Attenuator to drive low input impedance power amps <30kohm.

I must say it's the best tube buffer I've heard, none of that tube tubbiness, clean clear and dynamic, and not much colouration.

I built the original (1st pic) not the improved (2nd pic).

Cheers George
 

Attachments

  • SLCF Super Linear Cathode Follower.jpg
    SLCF Super Linear Cathode Follower.jpg
    31.3 KB · Views: 270
  • Super Linear Cathode Follower Improved .JPG
    Super Linear Cathode Follower Improved .JPG
    31.5 KB · Views: 264
After a lot of reading here and elsewhere, I have learned that the "tubiness" comes from second-order harmonics.

Its a popular myth but I for one don't believe it.

I'm not at the stage where I'm comfortable working with high voltages, plus my budget is limited so good transformers are off the menu.

Transformers can be hand-wound and don't need to be at all expensive (in materials that is). Sure they do take a bit of time, but this is DIY audio.

So my question is this - is it feasible to simulate this 2nd-order distortion with solid state technology?

It might be wise to check your premise. If tube sound is down to 2nd harmonic distortion, this can also be created with solid state. Just build everything single-ended classA and avoid opamps and NFB. But the reality is, with most music intermodulation distortion swamps any harmonic distortion - 2nd harmonic distortion shows up only on test tones.
 
@nironiro - a couple of questions if I may?

What is the purpose of the 10k potentiometer in the schematic (the one drawn, not the modification options)?

The pin numbers don't match the data sheet - do all the op amps use their inverting inputs, or should I go by pin numbers rather?

Thanks a ton!
 
The purpose of the 10k pot is to mix how much clipped signal will be in the output.
Some kind of effect level control.

Didn't check pin numbers on opamps yet, I'm just like you still searching through net for something like harmonic enhancer, in DIY of course.

As I found, there are two "lines":
- first is original signal
- second is high pass filter, usually around 3kHz, than clipping circuit which generate distortions - asymmetrical for even order harmonics

and than those two lines are summed to output.


One more schematics from net, this looks more interesting, because it makes asymmetrical clipping, (one diode), but, as the first one, I did not try it yet.

http://experimentalistsanonymous.co...apanese Electronics Book/harmonic-exciter.gif

And yes, abraxalito is right, "tubiness" is very popular myth on the net last years, and if you make normal, properly engineered tube buffer or preamp, you'll don't hear any distortions.
 
Last edited:
Just for the record: it is said that the “tube sound” comes from enhanced 2ND order harmonic distortion, but this is a convenient ruse, which obfuscates what is really going on.

In a nutshell, the ideal (not for musicality, but theoretically) amplifier would offer this:

Voutput = Vinput • gain

And that's that. A straight, linear, ideal 'curve'. Because semiconductor amplifiers have long been designed to adhere to this relationship, their “distortion” figures are oft bandied about at arbitrarily low values: 0.01%. 0.003%. 0.00027% and so on. They are - if you're looking for straight, crisp idealism - as close to ideal as one might reasonably imagine. But their doppelgänger is the vacuum valve amplifier.

Vacuum tubes are weird devices (and I mean that, lovingly) in that they “kind of, sort of” are linear when used in some circuits in special ways. In truth, they're pretty nonlinear, and are actually rather “polynomial” in transfer function. A triode, for instance, might be grossly modeled as:

I = gain • V²

on the grid. (this is a really large approximation, and the power isn't really a square, but it is definitely a power, and not a linear straight line!). But the gain part of that equation is also dependent on the VA-K or voltage difference between cathode (emitter side) and anode (collector side) of the device. The larger the VA-K, the larger the nominal gain.

Well, in real world circuits using inexpensive and pretty darn linear devices like resistors, to load the anode (to convert amperes to volts, thus “amplification”), as the tube's transconductance (multiplying grid voltage into anode current) amplifies the input signal, there's a phase reversal as far as the plate is concerned (a drop in voltage at the plate with increased grid voltage).

This antiphase plate signal is a kind of negative feedback which serves to linearize the tube's power-law conduction. Got that? Its true. So, our little triode that kind-of-sort-of followed

Current = transconductance gain • Vgrid²

Becomes

Voutput = gain • Vgrid≈1.0-1.3

And the world is a happy place. Thing is … the power isn't exactly 1.0. The absolutely linear and ideal straight gain line … is still curved.

It is the curvature of that gain line which distorts the output signal to not be a simple scaled version of the input. This is important for assessing the “tube sound”.

MOREOVER, tube circuits, like all amplifier circuits, have an output range beyond which they are physically incapable of achieving. This is called “clipping”. Transistor-type amplifiers tend to be school-straight ruler amplifiers until they reach clipping, then just STOP. Sounds terrible. Like someone popping little firecrackers (its the Year of the Monkey!) during a concert.

TUBE amplifiers however, either though intrinsic nature or intentional engineering, almost always have a much broader shelf of “clipping” and nonlinearity leading to their saturation levels. Moreover, the nonlinearity is often asymmetric about the polarity axis, meaning that positive excursions and negative excursions will have quite different clipping levels, waveforms and recovery modes. This - when deliberately engineered in - is what contributes to much of the wicked “scream” of legendary solo-guitar amplifiers when mercilessly overdriven. The power supplies are intentionally underrated, so as to enhance time-varying distortion and odd recovery. And yes… there is a certain unmistakable fingerprint of musicality to it too.

SO… getting back to the leading paragraphs … the only “harmonic” component to the distortion of tube amplifiers is related to the distortions to waveforms that happen when only considering sinusoidal input signals. The “harmonicity” is something which is a mathematical consequence of distorting sinusoidal waveforms. Can be analyzed with Fourier transforms ("spectrum analyzers") and so on. Engineers love 'em. But oddly enough, in the real world, the input signals are far, far, far from being idealized sinusoids.

So what then is the “warm tube sound”? It is the hard-to-define shape of the nonlinear amplification curve of an amplifier. And yes … if one is intending to “simulate” the warm tube sound, it is really as simple as having a voltage-and-time varying nonlinearity curve applied to the signal.

But that takes more than an undergraduate degree to come to grips with.

GoatGuy
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.