what clips harder than an LED ?

[Mark Hammer]

If a person wants straight-up-flat-across-straight-down square, you would do well to consider crossover distortion as well.  Diodes to ground simply dictate what the "top" will be like, however there is also what happens on the way to the top (and "it's a long way to the top if you wanna rock and roll"  ).  Diodes in series with the signal will prevent signal passing unless it exceeds forward voltage.  Once it meets that threshold, it goes from nothing to whatever the supra-threshold level is in whatever time it takes for that diode to conduct.  What does it do?  It "flattens the sides" of the waveform.

Now, one normally has to make some judgment calls with respect to how to combine crossover distortion and dynamic clipping of the "classic" type.  Let us say that a back-to-back pair of germanium diodes are used to produce corssover distortion, and a pair of LEDs for dynamic/amplitude clipping.  The series GE diodes will produce a near-flat "side" of the waveform until it surpasses around 200-230mv.  Once the signal passes that point, however, there is no ceiling on it until it reaches about 1.5v, and there is a big distance between 230mv and 1.5v.  A person could use more GE diodes in series, or perhaps SI diodes, to raise the threshold for passing signal and decrease the "distance" between the threshold to pass, and the threshold to be clipped to ground.  The problem that creates, though, is that the higher you make the threshold to pass the series diodes, the narrower the "duty-cycle" of the resulting waveform is, and the thinner and weenier the sound will be.  If you like mosquito buzz, fine, but if you don't....

So, the ideal would likely be to use a combination of low-threshold crossover distortion and  amplitude clipping with a threshold juuust a bit higher, then follow that with some boost and more of the same.  That way you don't lose too much of the ascending and descending parts of the waveform.  In theory, it oughta get you a nice rectangular waveform.  Whether that meets your subjective definition of a "hard" clip is a whole other matter, but if square is the destination, that's one way to get there.

Which now makes me wonder: "What would a BMP sound like if there was a back-to-back GE pair in series with the 8k2 input resistor in each of the clipping stages?"