[sdiy] ASM1 VCO single sided PCB design tests OK

[Magnus Danielson]

From: Neil Johnson <nej22 at hermes.cam.ac.uk>
Subject: Re: [sdiy] ASM1 VCO single sided PCB design tests OK
Date: Sat, 19 Apr 2003 15:12:35 +0100 (BST)

> 
> Magnus wrote:
> > Charlie wrote:
> > > ...the P supply feeding the op amps has to be "stable DC". If one rail
> > > waivers in voltage relative to ground and the other stays put, the
> > > oscillator seems to go out of scale.
> >
> > Well, the VCO core depends on the +15V to GND potential differance. AC
> > signals there WILL however not only make it to the output but also cause
> > modulation of the waveform.
> 
> Mmmm...that's why, when designing my little VCO core, I tied anything
> related to the core to a stable (and cheap) fixed +5v reference.  The only
> things that connect to the supply rails are op-amp supply pins and the
> inputs to voltage references;  even the negative reference for the tune is
> a cheaper zener (ok, it'll drift a bit with temperature, but way better
> than letting crap in from the -ve rail).
> 
> It did surprise me when I first looked over the ASM-1 that the supply
> rails were directly connected to the signal chain without any form of
> referencing.

Actually, the whole reference chain is a really complex issue. Some stuff is
compensated, but not others. You have to understand these interactions properly
or else the compensation tricks are futile. For instance, the referenec current
voltage and reset-voltage reference should be the same voltage, or compensation
will be flawed. Running only one of the on a stabilized voltage is really
missing the whole point, but then the CV pots need propper voltages or they will
make the whole thing out of tune. You really should have a SEM-like balancing
of voltages, it is wiser than it first looks, since the V+ -> GND and GNDD -> V-
voltages are the same by a control-loop and not my separate trimming (or just
off the shelf components or similar rought estimate).

Cheers,
Magnus - in a hurry so no time to properly handle the subject