Lostgallifreyan wrote in
:

Only when the pulse width ratio becomes significantly
asymmentrical do you get significant occurence of even harmonics in a ramp.


WHich also applies to a 'flat' wave, rectangular, pulse, whatever, of which
square is the uniquely symmetrical case as the triangle is to the ramp.

Incidentally, when you skew the symmetry fully to one or other extreme, the
resulting sharp transitions have a full 'natural harmonic' series, but the
closeness and timing of the transitions in the flat wave complicate the
result and eventually all but cancel as the pulses get extremely short, while
a ramp gets a full energy, full natural spectrum set of harmonics. Even then
it's just a special case of sawtooth, natural sawtooth waves have lots of
nonlinearity.