Through-zero FM of audio signals (was: Module mania!!!)

[Sean Costello]

david lacey wrote:

> I've sort of hacked a module I'd like to see done better:
>
> Audio FM Module, for Modulating the frequency of any audio.
>
> You can do this with an Oscillator patched into the 'control' input of
> a
> Delta Labs Effectron II (and this is about the only thing I ever use
> mine
> for.)  After I stumbled onto that, I hacked a 1/2 rack Boss delay for
> FMing4
> as well.  They both give some really over the top results, but I think
> a
> module designed with high a quality ADC & a DAC (no, not the DAC who
> so
> wisely started this thread in the face of some horribly off topic
> ones)
> driven by a clock whose frequency is linearly controlled by input
> (audio
> rate, varying around zero volts as LFOs would keep the playback rate
> high
> too long and the playback would get ahead of the recording) voltage.

Great idea.  Could you do this with a phase shifter?  I know that my
Voodoo Vibe (a real primative phase shifter) does a pretty nice vibrato;
maybe a voltage-controlled phase shifter could produce a type of FM if
only the phase-shifted signal is heard.  A simple implementation of this
could be made using a single LM3900 (check out US Patent #  4,306,480
for an implementation of this by Frank Eventoff and Serge Tcherepnin -
the full-wave rectifier used in the Serge Wave Multiplier module is also
explained in this patent); a more complex and useful version could use 4
(or more) OTA/buffer stages, and be switchable between a phase shifter
and a lowpass filter.  The capacity for linear FM would be nice, so that
the frequency of the note doesn't shift with modulation index (the
standard Electronotes exponential converter would be useful for this).
Through-zero FM would be VERY cool, but I would have no idea how to
implement this in a phase shifter.

One of my dream modules would be a combination filter/phase shifter,
with capacity for self-modulation (controlled by a VCA), through-zero
linear FM, 1V/oct control, VC resonance, oscillator sync, etc.
Theoretically, this could be useable as a filter, timbre modifier,
oscillator, etc.  Of course, it may be physically impossible to do
through-zero FM and oscillator sync with a filter.  Anyone have any
ideas?  Has anyone done this?  Is it impossible?

Are there any other ways to implement through-zero FM of audio signals?
How about a frequency shifter (i.e. the Electronotes design)?

Sean Costello