[sdiy] Analysis of frequency variation in analogue synths

Magnus Danielson cfmd at bredband.net
Thu May 3 09:49:18 CEST 2007

From: "Paul Perry" <pfperry at melbpc.org.au>
Subject: RE: [sdiy] Analysis of frequency variation in analogue synths
Date: Thu, 3 May 2007 12:53:18 +1000
Message-ID: <003c01c78d2e$361a6520$8501a8c0 at ice9>

> I suspect the effect might be too small to detect easily with a FM
> discriminator.. at the very least, the FM unit would need to have
> electronics more stable than whatever was producing the analog signal!

A normal of the shelf crystal oscillator should be stable enought for these
exercises. You can do better ofcourse and I can overkill in that area without
much of an effort.

A phase detector, such as that you find in 4046 may be handy, but a propper
counter is much better. The hard thing with both a counter and a digital
phase-detector such as 4046 is to be able to have enought gain as they belong
to the through-zero types of analyzers. Another method is to use multipliers
and multiply the input signal with the sine and cosine from a stable source.
You can do with only one multiplier, but it is usually wise to use two, since
then you can do propper phase-analysis and we are discussing phase deviations
here. If you don't have a sine/cosine source, but only a sine source, then
don't dispare, the polyphase filter setup is easy enought to build and that
will do the phase-separation for you. What you now got is the basis for a
spectrum/vector/network analyzer. Naturally you need to lowpass filter the
output of the multipliers to remove the sum frequencies.


More information about the Synth-diy mailing list