[sdiy] XOR as 'digital' ring modulator

[Tom Wiltshire]

Wow, loads of responses.

Let me answer a few things in the order they came up;

> You're missing non-linearities in the analog multiplier.


Cheater; No, not missing - ignoring. Fair enough an analog multiplier will be significantly imperfect. But I'm interested in the theoretical question about whether an XOR is the same as a Ring Mod for a limited set of signals.

> Now that "pwm" issue has me a bit perplexed.
> Are you thinking you might input analog waveforms (encoded as PWM)
> and then use the X-or, then decode (filter) the result back into
> the analog domain ???

Harry; No, I wasn't thinking that. I was just thinking that a couple of pulse waves with PWM are fairly 'unsquare', even if you're not playing Dave Brubeck.

> Maybe that the term "square wave" doesn't necessarily imply 50% duty cycle?

> my reading of Chamberlin's statement is that
> he means "square" in the "as opposed to sine, saw, triangle, or other
> analogue waveforms" sense


Assi, Ben; Yeah, you might be right. I think maybe Hal Chamberlin *was* meaning it that way - That 'square' includes all the 'pulse' and 'digital' signals we're talking about and 'non-square' is things like ramps and triangles. But that wasn't how I read it at first. My bad, I guess.

> For very little extra parts count, you can build a ring-mod with one input
> that is square wave-only, but another than can take any waveform.

Colin; Good point. Depends on the budget and available space I guess. The XOR route gets you 4 ring mods in one chip!

In summary, if you consider a digital signal as one that has two levels (call them -1 and +1) then an XOR gate is identical to a ring modulator.

Thanks all,
Tom