I'm writing a tutorial episode on PWM, and I've come upon a rather basic question:
Does the term "pulse width modulation" apply only to signals where the pulses are regular - that is, come at the same frequency?
Or, would the term also apply to any irregular ON/OFF system such as the output of a thermostat?
I thought I'd ask here and see what the consensus was.
I would argue that, by definition, it is a signal that contains a variable component encoded (modulated) by changing the width of a pulse.
So for example, a completely sporadic train of pulses with no fixed or even discernible repetition rate or frequency could technically be called PWM so long as the pulse width conformed to a particular function.
Consider a common servo control signal... a 1-2ms pulse width repeated at ~50ms. The 50ms is not all that critical but the 1-2ms sets the output position.
Steve
awneil 
Well, literally, PWM could be applied to anything where "information" is conveyed by varying (ie, modulating) the width of a pulse.
It is not necessarily about fixed-frequency ("regular") pulses - but that is certainly the most common usage.
In the specific case of the thermostat, I'd say it was more accurate to say that the "information" is conveyed in the state of the output (on or off).
If it doesn't have a carrier signal with a fixed frequency, I would call it Pulse Proportional Modulation. Pulse Width Modulation is a special case of Pulse Proportional Modulation.
