Saw Frequency doubler and Phase Shifter

[Magnus Danielson]

From: Harry Bissell <harrybissell at prodigy.net>
Subject: Re: Saw Frequency doubler and Phase Shifter
Date: Mon, 18 Sep 2000 22:36:30 -0400

> I think in real life... you do NOT want the comparator to go rail to rail
> at all... you want it to slew between tightly clamped limits. If you have
> less far to go, you get there much faster. The clamps can establish a reference
> that is more stable than the rails anyway.  Then playing with the  resistor
> values can get you back to the same spot.
> 
> These schemes all rely on PERCISE amplitudes for the sawtooth, if you
> are off the mistakes will show up as fundamental bleedthrough.
> 
> It would also be possible to make a triangle wave first, then make the saw from
> that... there are some very clever glitch reduction schemes (Rene Schmitz comes
> to mind....).   If the amplitude of the triangle is off, it won't affect anything
> as long
> as there is no DC offset  (easier to guarantee I think than the amplitude...)
> 
> Then you can use an invert/non-invert switch to make the sawtooth.  The long way
> around I admit but might be easier ???

Well, my curcuit does in it current embodiment do depend on the rail-to-rail
assumption, but the idea works equally well if it goes 0.3 V below the upper
rail and 0.4 V above the lower rail since you can make rescaling and offset
adjustments fairly easy. I don't think you can attack the core idea that easy,
but there are some practical details to be concerned with, yes.

Also, when I was talking to Jörgen earlier this evening as we where getting
home from a synth-beer event I think we agreed that while using slow op-amps
for comparators maybe was not giving the best waveform on the scope, but may
very well be usefull for music never the less (hmm... maybe I am extrapolating
from our discussion, but we where getting there).

I think we must recall what the goal of the exercise is...

Last thing I saw from Björn he seemed fairly satisfied with the result after
doing practical exercises. I think this is really where a curcuit in the end is
tested, in the process of making a product and how that product is being
apprechiated by some user.

My curcuits are quick and dirties, but may be usefull never the less. They are
simple enougth to be well understood even if some cleverness is behind the
scaling. I think there is a virtue in a minimalistic design too, you need both
that and more elaborate and precise designs.

Cheers,
Magnus