This is the fft (fast fourier transform) of the diode tube clipper I am currently building, when fed with an 1V 440Hz sine wave. On the graph the fundamental frequency can be seen on the far left, it is 440Hz (just an A) and the peaks are the harmonics produced by clipping the waveform. I have marked them with datapoints and they are off by a few Hz (due to measurement errors and such) but if one compares them with the actual notes frequencies, he'll find that they are A4, A5, E6, A6, C#7, E7 and so on.
What I found quite impressive was the fact that the harmonics were singable and clear.
I do not show it here, but due to the voltage transfer characteristic of a diode tube clipper, the outcome is very dynamic, offering a more natural sound, giving the guitarist the ability to transfer his expression from the guitar, to the speakers. Those that have been lucky enough to use (or even own one!) a tube guitar amp will already know that. I of course have no means to actually test that (since I don't have a guitar), but this is what my measurements say will happen when we plug in a guitar. We will do so eventually, and I'll try to provide a sound file.
If I were to do this with a silicon diode (and I have tried it), the result is much less musical, and my problem with that is that almost no dynamics and expression can be passed to the output, due to the nature of the silicon diodes (producing hard clipping, as opposed to the soft clipping we have here).
You can propably understand that I am very excited about this little project, and it was a perfect opportunity to test my new oscilloscope and function generator (the fft was produced by the oscilloscope and imported to matlab to be examined more carefully).
And last but not least folks.
TUBES FUCKING GLOW.
too bad I can't take a picture.