Discussion of the musicians preference of the older type instruments "sound quality"

Audioguru again

Joined Oct 21, 2019
6,674
The graph in post #37 shows the 1kHz fundamental level and the levels of the harmonics. The levels of the 2nd and 3rd harmonics are each at about -90dB (0.015%) which is good but the input level is very low at only 10mV.

A linear amplifier is linear if the load impedance does not overload it. A non-linear load has distortion built-into it.

Your two-way woofer-tweeter speaker should have a crossover at the recommended high-cut for the woofer and at the recommended low-cut for the tweeter. Adding a midrange speaker to the crossover messes-up levels and frequencies of all the speakers.
 

Ian0

Joined Aug 7, 2020
9,677
Now that you've asked the question, I'll do my best to make sure you don't leave without understanding the answer.
When we test an audio circuit, the first thing we generally do is test it with a 1kHz sinewave. 1kHz is nicely in the middle of the audio band, where it won't be affected by low- and high-frequency roll-offs, and we use a sinewave, because it doesn't contain any harmonics.
The graph is a spectrum of the signal - x axis is frequency and y axis is amplitude.
If I were to plot the spectrum of the input signal, there would be a peak at 1kHz that reached -40dBV (which is 10mV)
In theory there would be a vertical line at 1kHz reaching -40dbV, and it would be -∞ everywhere else, but that isn't the case. Thermal noise puts the base line at -180dBV rather than -∞, and the peak has a finite width instead of being infinitesimally narrow, but anyway, the input signal would be a single peak at 1kHz.

The plots shown are the output signal. If there amplifier were perfect, then there would be a single peak at 1kHz, and its height would be the amplitude of the output signal. The circuit had a gain of 28, so the signal would be 280mV (-11dBV). However, be the amplifier introduces harmonic distortion, and the second and third harmonics are at 2kHz and 3kHz respectively.
So the distortion shows up as additional peaks at 2kHz and 3kHz etc.

To see how much distortion measure the difference in heights of the peaks. If the 3kHz peak is 40dB lower than the fundamental then the 3rd harmonic distortion is -40dB, which is 1%.

On the first plot, where I used two cascaded (not cascoded) FET amplifiers, you can see that the 2nd harmonic distortion (at 2kHz) from the first stage is higher than the 2nd harmonic distortion after two stages, so adding the second stage reduced the distortion by straightening out the curve, just as you suggested at the very beginning. But it didn't do any good for the 3rd harmonic.
 

Audioguru again

Joined Oct 21, 2019
6,674
I see two different circuits
IanO is talking about the fairly high distortion high input level circuit in post #36 and I was talking about the low distortion low input level circuit in post #37.
 
Top