[sdiy] VCO Linear FM input - DC or AC coupled?

[Neil Johnson]

Hi,

On Mon, 26 Jun 2023 at 19:51, Mattias Rickardsson <mr at analogue.org> wrote:

> On Mon, 26 Jun 2023 at 20:39, ben gebhardt via Synth-diy <
> synth-diy at synth-diy.org> wrote:
>
>> Isn't this just the difference between phase modulation and frequency
>> modulation?
>>
>
> No, there's a derivative involved.
> Frequency is the derivative of phase. Hence,
> PM with a signal is the same thing as FM with the derivative of the signal.
>
> Having AC coupling on linear FM inputs is because any DC component will
> bend the pitch away. Ideally I guess you want low enough DC-cutoff to leave
> your modulating signal waveforms untouched, but fast enough to get rid of
> DC components introduced while patching around... or something like that.
> :-)
>
>
A DC-coupled FM input is a true FM input: you can see this if you modulate
it with a square wave - the output frequency instantaneously flips between
two frequencies corresponding to the modulating voltage.  The voltage is
directly modulating the frequency, hence FM.  The downside of FM is that
_any_ DC offset is going to detune your carrier oscillator.  Where do these
DC offsets come from?  Two places: the real components (e.g., opamp
offsets, which also drift with temperature) and from the modulating
waveforms themselves.  For example, a PWM signal has a shifting DC
component that corresponds to the duty cycle.

When you add a DC blocking capacitor to the summing opamp input (assuming
you're using an inverting opamp circuit) what you have done is create a
crude differentiator.  It's crude because, unlike an ideal differentiator,
the gain at daylight is not infinite.   Certainly the gain at DC is 0.  How
good a differentiator it is depends on the resistors and the opamp (and for
lols, opamps are crude integrators).

You can really hear the difference with a slow square wave.  FM sounds like
a police siren (neee-naaahhhh-neee-naahhh).  Whereas PM is more interesting
(more like "bing...boing...bing...boing").

Purists will pop up and say that when the modulating signal is a sine wave
there is no difference between the two, and that's fine (lets ignore the
pi/2 difference between sin and cos, eh?).

Neil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20230626/aa43a6a1/attachment.htm>