Home Propeller Head Plaza

Technical and scientific discussion of amps, cables and other topics.

[ Follow Ups ] Thread: [ Display  All  Email ] [ Propeller Head Plaza ]
[ Alert Moderator ]

I can't see the distortion for all the harmonics .....

mfc,
I probably missed the precursors to this thread but is 'all' we are talking about here follower linearity? The large signal distortion is due to the modulation of the base-emitter/gate-source/grid-cathode voltage by the changing current supplied to the load. For example, the Vbe of a bipolar device increases by 18mV when the emitter current doubles (this is universal to all bipolars, FETs & tubes will vary by device type). The output signal follows the input signal minus a DC offset but that offset is modulated adding distortion to the output. For the bipolar case the output signal will compress by 18mV when the emitter current doubles, the Vbe-Ie characteristic is exponential and odd & even harmonics are generated, a square-law device should generate only 2nd (though I don't think there are any true sqaure law devices - particularly in follower mode if the currents are high as gm will be limited by ohmic resistances. I suspect that transitioning between a square-law gm region and a linear gm region will generate 3rd and higher harmonics).

You're right that the follower circuit performs best with small swings or when lightly loaded because for those conditions the change of current through the load is least - or bias the follower at higher current so the proportionate change due to load current is minimized.

I reccommend that you characterize the circuit thoroughly with single-tone or two-tones before embarking on any more exotic test tones. As I mentioned, I like plotting harmonics and intermods as an intercept graph. Intercept point is an abstract concept but you should expect to see the harmonics rising with the correct proportions as the input signal increases - if you don't then something suspicious is going on. With a circuit as 'simple' as a follower I would not expect it to behave well (ie 2nd harmonic increase by 2dB for every 1dB increase in input, 3rd harmonic increase by 3dB for every 1dB increase in input etc). I think this would make a more useful plot that the spot power level spectra that are published in amplifier reviews - it would take a bit longer to measure but I'm sure it could be automated.
To simulate this make your input signal amplitude vary logarithmically to a variable that will be swept linearly (ie input signal stepped in dB) then run a parametric sweep on that variable over the input amplitude range of interest. Plot the FFT, or DFT of the resulting output transient waveform. Hopefully, you have a post-processing function that returns the magnitude of plot at a specified frequency and use this to plot a line corresponding to the output signal at the fundamental frequency, the 2nd harmonic frequency etc.
Tolerances and accuracy are always important - to be honest I'm not sure if it's most efficient to screw down reltol (etc), decrease time step or average over more periods, though you can see when you have enough dynamic range from the FFT/DFT noise floor.
Simulation is a very powerful tool and I'm sure that anyone who doesn't use simulation does not know their circuits as well as they think they do (for many applications there is on choice but to simulate) - even if model accuracy isn't good enough to accurately predict very low distortion levels it will still give tons of insight into the source of the distortion and the sensitivity to ciruit conditions. Someone already mentioned low frequency settling and that is easy to test by applying an input step or power supply step and probing around the circuit. any high impedance nodes can be tested by applying a stepped current source at that spot and monitor the votage there (check for critically damped settling).

Finally, back to the square wave testing. Again, I think the most important thing is to know what to expect before you run it otherwise the interpretation of the results is open to question. I have another suggestion which should compliment single and two tone sims : set up a two tone test with relatively small fixed tones. Then add a large single-tone at an unrelated frequency - sweep the single-tone amplitude and monitor the intermod from the two tones. You could put the single-tone at a range of different frequencies. It should show how the incremental gain changes as the circuit experiences large signal excursions (I've peaked my own curiosity now - I think I'll run a test case myself).

Regards
13DoW



This post is made possible by the generous support of people like you and our sponsors:
  Atma-Sphere Music Systems, Inc.  

Follow Ups Full Thread
Follow Ups
You can not post to an archived thread.

[ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2025, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]