AW: LM 13600 based filters
HJ2743 at denbgm3xm.scnn1.msmgate.m30x.nbg.scn.de
Thu Dec 7 21:40:00 CET 1995
>Without the diodes signal gets distorted earlier but still the distortion
>does not increase very rapidly with the signal (soft clipping).
I happened to build a SEM-like filter for my Modular some months ago,
and I used a 13600 instead of two 3080s. I also made use of the diodes
to reduce noise (with re-calculated input divider, of course), but later
I removed the diodes and I think I like the filter better now.
You don't get any improovement in SNR at a point of 10% distortion
with the diodes; its only that the 10% come suddenly with the diodes, and
smooth without them. For Hifi and Mixing stuff they are ok (but *there*
I prefer other VCA's, anyway!), but for high level signals a smooth
overdrive is always better.
There is another point *for* the diodes, however: If You have to match
the gain of more than two OTAs, You don't have to use trimmers or select
chips when You use them: The diodes not only compensate for nonlinearity
of the gm cell, but also for gm tolerances!
>bad VCA desing may destroy the
>sound of an otherwise good filter (at least I fear that with the MS-20).
Amen to that! And not only in the MS-20. Almost *every* Korg synth
used these ultra-low-cost VCA's. You should have heard my (former)
Mono/Poly, when I had replaced the crap VCA with an SSM2020 !
More information about the Synth-diy