Alex,

I know that with an electric guitar amp it does look like the signal is clipped off with scissors, but with slightly rounded edges. I would guess that a power amp into a speaker has more going on.

But a square wave, though flat at the top/bottom is not the same as DC. As I recall it's a series of frequency spikes. Let's say the signal is a 1kHz square wave. The lowest frequency component is still 1 kHz. After that you get the odd harmonics - 3 kHz, 5 kHz, 7 kHz etc... Each harmonic is weaker than the last, but for a perfect square wave the harmonics go on to infinity. (Which means that there are no perfect square waves in real life - that would account for the high frequency oscillations.)

BTW, tubes tend to distort in a way that produces even harmonics. That's why tube sound different (and better) than transistors when pushed hard.

But that's the frequency domain view (and that's how we know square waves will pass through the x-over). From the amplitude view the speaker sits there in a DC state between the transitions. It burns power and gets no cooling effect since it's not moving air. Not good.

But I think the shrink idea is excellent. I, like lots of people here, am obsessed with deep, clean bass. The only cure I've found so far is a pair of 2235Hs.