A simple circuit with an optocoupler creates a “tube” sound

[merlinb]

I thought this EDN article might interest some folks:
https://www.edn.com/a-simple-circuit-with-an-optocoupler-creates-a-tube-sound/

[BubbaFet]

And from the brief app note...

"As one can see, the second harmonic of a 5 to 15 W tube amplifier dominates the spectrum, and EL34 type lamps are used in the output stage."

...also contributing to that "tube sound".

[Clint Eastwood]

What happens when some punk overdrives this circuit?

[FSFX]

This appeared in one of the FB groups a few weeks ago and was discussed there.
Firstly, what exactly is a 'tube sound'? It certainly is not just defined by the harmonic content shown here.
That is easy to achieve with suitable biasing of any simple single BJT common emitter stage or JFET common source stage.
So the use of an additional opto-coupler and an emitter follower output stage looks a bit of a 'Heath Robinson' type of way over-engineered and over-complicated solution.

Anyway, back to what exactly is a 'tube sound'? Here is some reading material that looks at this question.
https://gitec-forum-eng.de/wp-content/uploads/2019/02/poteg-10-09-1-tube-vs-transistor-sound.pdf

[Steben]

This appeared in one of the FB groups a few weeks ago and was discussed there.
Firstly, what exactly is a 'tube sound'? It certainly is not just defined by the harmonic content shown here.
That is easy to achieve with suitable biasing of any simple single BJT common emitter stage or JFET common source stage.
So the use of an additional opto-coupler and an emitter follower output stage looks a bit of a 'Heath Robinson' type of way over-engineered and over-complicated solution.

Anyway, back to what exactly is a 'tube sound'? Here is some reading material that looks at this question.
https://gitec-forum-eng.de/wp-content/uploads/2019/02/poteg-10-09-1-tube-vs-transistor-sound.pdf

True.
Even a single triode preamp stage has less (even order) harmonics than a jfet stage for example. The Fetzer valve is all about taming the harmonics.
I posted some time ago a paper I found where the harmonic spectrum of a non-overdriven biased triode, BJT and jfet stage were analysed. The conclusion was a classic BJT stage with some (emitter) feedback mimicked the triode stage quite well. Solid state is full of harmonics even the good ones.

[FSFX]

Even with a tube based circuit, the range and comparative level of harmonic distortion can change significantly depending on the value of the cathode resistor and the details of its bypassing as well as the input signal level.
This is pretty obvious as these things affect the operating conditions of the tube in the circuit.
I spent quite some time experimenting with this and analysing the distortion of different arrangements when designing and building a 'guided-missile' pencil tube boost and overdrive pedal a couple of years ago.
When it comes to 'tube sound' and harmonic distortion it really just depends on circuit configuration anyway as a Class A preamp or single ended output stage will create a whole lot different distortion to a push-pull type of output stage.
There are many other factors involved as well as have been extensively documented over many decades.
     

[PRR]

> easy to achieve with suitable biasing of any simple single BJT

I think it IS a frilled-up "simple single BJT". The opto-coupler is highly linear around 4mA. EF, bah, dead-clean.

FWIW: at the extreme a tube will do >5% THD, a JFET a bit more, but a BJT will do 26%. So much that we almost always "need" some degeneration.

[Ben N]

So the use of an additional opto-coupler and an emitter follower output stage looks a bit of a 'Heath Robinson' type of way over-engineered and over-complicated solution.

Check the comments there.

[Steben]

> easy to achieve with suitable biasing of any simple single BJT

I think it IS a frilled-up "simple single BJT". The opto-coupler is highly linear around 4mA. EF, bah, dead-clean.

FWIW: at the extreme a tube will do >5% THD, a JFET a bit more, but a BJT will do 26%. So much that we almost always "need" some degeneration.

The warmth of the low harmonics. 

[Rob Strand]

I think it IS a frilled-up "simple single BJT". The opto-coupler is highly linear around 4mA. EF, bah, dead-clean.

To me the question is where the second order harmonic originates.   In general the IF vs IC relationship of optos isn't very linear so my initial thoughts were the gain characteristic is where the second harmonics originates.   (It's reasonably well known they are non-linear in this way.)

If you look at the CTR from IF = 0.3mA to 3mA (one decade) the CTR varies say from 33 to 130, a factor of about 4.   That's a lot more than even old school transistors, which might only vary 1.3 (30%) over a decade.

The circuit operates around IF = 0.7mA and a swing of about 80uA (assumed p-p), which is in the non-linear region.

As far as the test in the article goes the swing is quite small and the transistors are not saturating so we only need to consider device non-linearities.  If we pushed the circuit the first transistor doesn't really clip until the swing is quite large, due to the diode load, so the opto transistor will be the only part that clips.

So IMHO the CTR non-linearity is what makes this circuit work.

[FSFX]

So the use of an additional opto-coupler and an emitter follower output stage looks a bit of a 'Heath Robinson' type of way over-engineered and over-complicated solution.

Check the comments there.

It is just another non-linear circuit configuration creating some harmonic distortion like the many thousands out there. It is not even analysed thoroughly or scientifically. All it shows is some harmonic distortion using a fixed level and single frequency that is common with many circuit designs. No discusion of IMD or analysis of its behaviour with complex signals.

And, I certainly don't think that the following statement from the article is actually technically correct and certainly does not just apply to tube amplifiers [power amps or preamps, single ended or push-pull, Class A or Class AB, etc.?] in many ways and what exactly is meant by the 'pass-through' characteristic?:

QUOTE FROM ARTICLE
It is known that the main feature of tube amplifiers is the presence of small distortions. In these distortions, the second harmonic prevails. Some optocouplers have a pass-through characteristic close to the pass-through characteristic of the electron tube.
END QUOTE

Anyway, back to what exactly is a 'tube sound'? Here is some reading material that looks at this question.
https://gitec-forum-eng.de/wp-content/uploads/2019/02/poteg-10-09-1-tube-vs-transistor-sound.pdf

As every tube guitar amplifier sounds different then I really can't see that you can define just one 'tube sound'. 

[Rob Strand]

I did a quick simulation with a transistor model that fairly steep non-linear gain region (hFE vs Ic).   The 3rd harmonic distortion looked worse than a flat hFE curve.  Very puzzling.   It doesn't look like that's it.  Too late here for me to ponder alternatives.

[antonis]

I did a quick simulation with a transistor model that fairly steep non-linear gain region (hFE vs Ic).   The 3rd harmonic distortion looked worse than a flat hFE curve.  Very puzzling.

Interesting enough..

[BubbaFet]

What is the minimum numer of tubes or  transistors required to achieve 'tube sound'?

Is one gain stage enough?

[Steben]

What is the minimum numer of tubes or  transistors required to achieve 'tube sound'?

Is one gain stage enough?

Does a Champ 5W have tube sound? 
it has 2 tubes in the signal chain. Most triodes come in pairs
I guess a champ has "a" tube sound...

It is astonishing how many SS amps are rumoured ot be "ok" in the clean channel. While these channels usually have almost no harmonic content added with their tons of opamp arrays.
Makes me think: they are ok because they lack annoying harmonics.... Combine that with "warmth" of clean triodes (the least harmonic content of discretes)...
Even order.. wasn't that worse for IMD?

Joh mayers sound is usually related to a dead clean tube amp and some drive pedals based around symmetric clipping...
What are even order loaded devices? Many many dirty fuzzes....

On the other hand, "harmonic enhancers" usually are based around mixing some very asymmetric high freq clipped signal to the clean.

[Rob Strand]

I did a quick simulation with a transistor model that fairly steep non-linear gain region (hFE vs Ic).   The 3rd harmonic distortion looked worse than a flat hFE curve.  Very puzzling.

Interesting enough..

I think I screwed up last night. It was getting late and maybe got the traces switched.  No idea.

Here's the results (harmonics levels relative to fundamental)
original article:                       2nd -40dB, 3rd -65dB
transistor with non-linear hFE:  2nd -35dB, 3rd -67dB                ; not far off
common transistor:                 2nd -53dB, 3rd -88dB
second order non-linear:          2nd -38dB, 3rd -77dB

So the non-linear hFE does get close.
The basic transistor isn't close enough

For the second order non-linear, I added second order non-linearity between the LED and base current of a common transistor of the form ILED*(1 - k2  + k2*(ILED/700uA)) = ILED(1-k2) + k2*ILED^2/700uA.   k2 sets the strength of the second order non-linearity and I chose k2 = 0.25 to get close to the article values.

Conclusion: hFE (or CTR) non-linearity of the opto looks like a feasible explanation.

[teemuk]

"Tube sound"... sigh. When will we bury this fallacious concept.

It is always trifling when folks try to explain what it is and can't point a single characteristic that isn't circuit-related or otherwise present in other than tube-based circuits as well.

Even the die hard advocates can't quite agree are tubes great because they are so linear and undistorted or because they are not linear and undistorted. 

Two of my favourite "tube sound" circuits from electronics magazines so far have been:

1. A blatant copy of preamp OD channel from Fender's SS amps. This even had the channel switch FET protection diodes eventhough the magazine circuit had no channel switching and therefore no FETs needing protection either.
2. A circuit that mimicked the weird sounds a tube makes before it has heated to proper operating temperature. Evidently this was the magic of tubes from the point of view of late 1970's Hi-Fi. 

It distorts = tube sound. Problem solved. 

[FSFX]

"Tube sound"... sigh. When will we bury this fallacious concept.

That was my original comment regarding this.

So what exactly is people's definition of 'Tube Sound' and what is so different to 'SS Sound'?

Is it 'Tube Input Stage Sound' [12AX7, 12AU7, etc] or 'Tube Power Amplifier Sound' [EL34, 6L6, etc] or just a combination of these that then involves the colouring of the sound by the Miller capacitance, grid leakage current, output transformer characteristics, etc, etc?

We know that attempts have been made to try to subtly influence the sound with things like the 'Fetzer Valve' and many various tone controls and EQs as well as things like Rupert Neve's class-A transformer coupled designs.

So these days we just use solid state amplifiers with amp simulators and cab simulators and IR editors to allow us tweak the sound to whatever we think sounds 'right'.   

[teemuk]

Is it 'Tube Input Stage Sound' [12AX7, 12AU7, etc] or 'Tube Power Amplifier Sound' [EL34, 6L6, etc] or just a combination of these...

And we are not even touching the circuit architecture aspect. PP is quite different to single-ended for instance. This was acknowledged already back in the early 1970's.

So these days we just use solid state amplifiers with amp simulators and cab simulators and IR editors to allow us tweak the sound to whatever we think sounds 'right'.

Right. I know there have been fantastic simulations of some typical circuit characteristics tube amps - particularly those used as "effect processors" - might feature. But if all the same effects can be achieved with various semiconductor -based solutions (whether analog or DSP) is it even no longer "tube sound"?

Shouldn't "tube sound" be an universal trait? ...Like all tube amps have it? Not like one tube amp has this or that trait and the another doesn't, but then a transistorized amp also does have it. Then it's not "tube sound" IMO.

[Steben]

Sounds tube amps produced all along. But no exclusion of other sources or devices.
The - by now classic - overdrives with clippers in non inverting feedback circuits are so common they are worth mentioning since they are a thing on their own. Very elaborate to mimick with tubes.