[sdiy] New Thomas Henry VCO-1 Page Up

[Magnus Danielson]

From: Neil Johnson <neil.johnson97 at ntlworld.com>
Subject: Re: [sdiy] New Thomas Henry VCO-1 Page Up
Date: Sun, 27 May 2007 20:18:06 +0100
Message-ID: <13011C9D-9B34-4D38-A6C7-465CB574DB22 at ntlworld.com>

Neil,

> Ian Fritz wrote:
> > In 2000 I made additional improvements, including using good modern  
> > opamps (OPA27 for the CV summer, OPA134 for the CV servo and the  
> > Schmitt trigger).  At the same time I added added temperature  
> > compensation for linear temperature drift and tweaked up the scale  
> > factor drift.
> >
> > My version of this VCO is now reasonably stable, although the use  
> > of the supply rails for critical voltages is something I would  
> > avoid in a new design.
> 
> It is a continual source of amazement to me that so much effort and  
> expense is put into the choice of fast comparators, low-offset  
> opamps, esoteric temperature compensation schemes, etc, only for the  
> design to then rely on the supply rails for reference or other  
> critical voltages.  Especially given the low cost and ease of use of  
> even half-decent bandgap reference diodes available these days.

Well, the frequency tracking does actually not depend on it, since the expo-
pair reference current dependence balance out the reference voltage dependence.
I was also thinking like you until I derived the whole thing myself.
Modulations on the powersupply will however leak-through, but this is a high-
frequency phenomenon due to the integrating part and that DC wise it cancels.

The feature that does change with supply voltage is the amplitude.

If you want to use a bandgap reference or similar, you need to treat both the
current reference and the reset-reference voltage. If you do not treat both but
only one, then you have actually made things worse, since then you actually
brought the balance out of operation and you *will* depend on the supply
voltage.

This explains how we have been able to handle it fairly well as it is.

There is however a voltage dependence you can run into, that you acheive by
having the tune-knob work between the positive and negative supply if these
does not track each other. This is actually one of the hidden smart things in
the Oberheim SEM, where the negative supply is made to mirror the positive
supply. Even if the resistors isn't perfect match, that is not a problem for
the tracking.

So, while it is not ideal, it is actually better than one can first assume.

Cheers,
Magnus