AW: [sdiy] VC Glide

[Czech Martin]

I did not think this to the end, but it happens that I loked
into the "op-amp cookbook" (Jung) where this topic of offset
compensation was discussed. The example was a high speed, bad
offset circuit which was improved by adding a servo circuit
with low speed , but also very low offset.

If we translate this to the example we need to measure input
and output voltage, derive the difference and use that for
closing the control loop, of course with reasonable filtering,
so that the transients while playing do not count.
Offset will vary with tenperature and aging, e.i. very slow,
at least we can design temerature changes to be slow. This
should all be in the order of minutes. If the error
averaging is somwhere in this time, the usuall glide time
of 1/10s -1s will be far away.

I'm pretty sure that this will work, it may turn out inconveniant
or expensive (low leakage, high capacitance integrator caps).

Also , a chopper design could be taken into account, this
would mean some headache. This boils down to what Juerger 
said about concurently tuning the error away, but not
manually, but automatically.


m.c.


> -----Ursprüngliche Nachricht-----
> Von: jhaible at t-online.de [mailto:jhaible at t-online.de]
> Gesendet: Montag, 1. Oktober 2001 16:30
> An: Jim Patchell
> Cc: synth-diy at dropmix.xs4all.nl
> Betreff: Re: [sdiy] VC Glide
> 
> 
> Jim Patchell schrieb:
> > Take a look at:
> > http://www.oldcrows.net/~patchell/archives/ca3280_7.html
> > 
> >     This circuit uses a CA3280, although, with a few 
> > mods, you can use 
> > either the CA3080 or the LM13600/13700 type OTA's as 
> > well.  
> 
> [...]
> 
> >     Also, If you use a CA3280, you can make the shape 
> > continuously 
> > adjustable by having the Diode Bias Current being 
> > adjustable as well. 
> 
> IMO, CA3080, LM13700 and LM13600 are no good for exponential
> glide. Too much offset for playing VCOs in tune without
> permanent re-adjusting.
> 
> The CA3280 has better offset specs, and I cannot speak for this one.
> But I'd like to ask: Jim, how much offset variation do you get
> with a 3280 in exponential mode (linear mode is never a problem),
> when you vary the time constants across the full range ?
> 
> JH.
>