SV: [sdiy] Re: (sdiy) English lessons

[Magnus Danielson]

From: harrybissell <harrybissell at prodigy.net>
Subject: Re: SV: [sdiy] Re: (sdiy) English lessons
Date: Wed, 14 Sep 2005 23:33:43 -0400
Message-ID: <4328EB95.C887CD9E at prodigy.net>

> Magnus Danielson wrote:
> 
> > OK, now to get you back on track... remember my sawtooth phase-shifter? What do
> > you guys think needs to be done to make it workable? Should a good op-amp be
> > enought maybe?
> >
> 
> Hi Magnus
> 
> I never got your sawtooth phase shifter to work in simulation. I tried a lot
> of different things (trying to guess where you went wrong).

The problem is that the first op-amp acts as a comparator, so that may confuse
the hell out of you if your model does not saturate. The assumed supply voltage
is +/- 15V, which is "all over the schematic" ;O)

KD tried it ages ago, and then it worked well! ;O)

> I'd say for it to work well, you would need a nice, precise square wave with
> really good speed.  I'd consider triggering a 4000 series CMOS gate with
> a precise power supply voltage.

I have assumed that the comparator needs to be fairly quick, but the second
op-amp does not have to have the same gain-bandwidth product. At least
according my assumptions.

> I have done a couple of sawtooth plase shifters.  You can trim the resistor values
> to get them to work, but change the opamp or the PS voltage and you have to do it
> again.

Indeed. It assumes "perfect" matching. However, a normal oscillator core like
the ASM-1 VCO (which was the intended target) will change it's sawtooth
amplitude linearly with the power-supply, so the phase-shifter will depend on
that too, so then it is really only the precission of the resistors which
becomes an issue. However, it still assumes that the negative rail is a perfect
mirror (or relative perfect mirror) of the positive rail, and this we see in
the Obie SEM. Assuming those conditions, it can behave pretty well! ;O)

Cheers,
Magnus