The GEM amplifier

[petr_2009]

If you don't mind me asking, what does the distortion spectrum look like at 100Hz?

Note the loop gain - it is constant throughout the audio range. This guarantees a constant output impedance over the entire audio range and a constant distortion spectrum.

 

[Hierfi]

That was always clear TNT.

A hint to answer the questions posed in #19 is that:

For any symmetric input stimulus that traverses up and down the same upper or lower half of the transfer curve requires that the shape of the non-linearity be symmetric around the peaks of the input. In the case of a sinusoidal input this places the peaks of the second and third harmonic in line with the peak of the input sinusoid.

 

[Hierfi]

Just so you know, that was aimed at petr_ hence the suggestion for the name...

Syn08 could actually be a contender for a guru for me 🙂

//

Sure TNT. To be clear I also have great respect for petr_

 

[traderbam]

Note the loop gain - it is constant throughout the audio range. This guarantees a constant output impedance over the entire audio range and a constant distortion spectrum.

You state this like it's a good thing. Why?
It is known that there is a poor correlation between THD and sound quality.
A rising output impedance is like having an inductor in series with the output. The speaker cable has inductance and the speaker is highly inductive too.
What is your reasoning?

 

[petr_2009]

In the quest for a guru, perhaps as yourself:

There are four basic types of curved transfer function shapes of an amplifier being in relation to the midpoint of an input stimulus.

1. Upward expansive/Downward expansive
2. Upward compressive/Downward compressive
3. Upward expansive/Downward compressive
4. Upward compressive/Downward expansive

One and two generate third harmonics, three and four generate second harmonics. So the question is:

1. What is the wave shape and position of the non-linearity alone (in the time domain) resulting from a sinusoidal input stimulus that only generates a second harmonic in the frequency domain?

2. Under the same circumstances, that only generates a third harmonic?

for clarity, your questions can be answered graphically
asymmetric compression and expansion will lead to a wide range of harmonics