[sdiy] Control volt into computer over soundcard?

[Iain Duncan]

> I would think that all PC soundcards are AC-coupled, so no chance to get
> a DC signal in.  However, I have seen modifications on the net where
> people have made them DC-coupled in order to use as a scope, voltmeter,
> etc.

Thanks, I'll look into that.

> If you want to modulate the amplitude of an AC signal with a DC control
> voltage .. that's done with a VCA.  If you want to modulate the pulse
> width of a pulse wave with a DC control voltage .. that's done with a
> VCO with PWM input.  But of course you knew these things already ;)

Yeah, but I'm wondering if there is perhaps another way that I'm
missing. I know a guy who made a radio drum input into a computer using
sine waves that encoded the four control signals into one composite fm
signal and sent it into the computer that way, he said it worked quite
well. But what would be cheapest/easiest way of making a reliable
oscillator that I could use to encode the CV signal over audio? The
thing is, I don't want to *listen* to this wave, just take it apart. So
it doesn't need to sound nice, just be reliable in its specs, and
probably at a very stable frequency so that taking it apart at the
computer end would be easy.

I was thinking, perhaps an amplitude modulate pulse wave at the nyquist
would be good because on the computer end, then every second sample
would be the correct data ( either 0 or peak amp ) making riding the
signal simple ( and need a very small window ). I have no idea how
practical this would be to generate in electronic terms though, any
suggestions?

> Also, perhaps consider using MIDI CC pairs to get the data in - this
> will give you 14bits of resolution.  The data rate will still be
> relatively slow though: 14bits data per ~1.5ms (running status mode).

Yeah, I could do that. I know how to do midi, but would rather avoid it.
Slower, and I still have to make a CV-to midi so I don't see that it
solves any problems really. I will probably do CV to parallel port with
my own drivers on both ends, but I think that will be more complicated
that the sound card solution, and getting budget sound cards for peanuts
is easy, and on linux I can service and analyse the audio input with
under 1ms latency.

Thanks
Iain