SEM VCO

[gstopp at fibermux.com]

     Nice report Juergen. So I take it from your description (transistor 
     array expo transistors, FETs) that you are talking about an exact 
     duplication of the Oberheim SEM VCO, with a discrete transistor 
     schmidt trigger and dual FET zero-offset buffer. I'm quite curious 
     about a few points:
     
     Did you implement the high-frequency track trim?
     
     Did you use a 3046 or 3086? Will the 3046 work?
     
     What is the reset time of the sawtooth? And a closely related question 
     - what is the maximum frequency of the VCO?
     
     Were the problems you encountered due to high frequency oscillations 
     around the reference op-amp in the exponential converter, because you 
     used FET-input op-amps? How did you fix the problems, extra caps?
     
     I'd like to compare the performance of the original SEM VCO to the one 
     I implemented in the ASM-1. They are quite similar in many respects, 
     but I never had tracking or converter problems.
     
     - Gene
     gstopp at fibermux.com


______________________________ Reply Separator _________________________________
Subject: SEM VCO
Author:  Haible_Juergen#Tel2743 <HJ2743 at denbgm3xm.scnn1.msmgate.m30x.nbg.scn.de>
at ccrelayout
Date:    5/15/96 2:24 PM
     
Recently I built a SEM-like dual VCO for my Modular. 
Made minor changes (FET input OPamps instead of 
discrete FETs etc.) and ran into problems. I have found 
some workaround, but nevertheless I decided to
rebuild the original - so this is another pcb that was 
carefully designed, but is now scattered with additional 
components on the copper side - oh well.
But now it works as intended, and it sounds *great*.
I have still some minor changes compared to the original 
SEM VCO: I used LF353 for all opamps, and I used
BF245C FETs instead of the original ones. I used
a 560R tempco resistor instead of the original 1k, 
and altered the other resistor of the divider, too.
I bent the tempcos over the transistor array, and filled 
the space in between with much of this heat sink paste. 
As the wheather is quite moody at the moment in germany,
I had the chance to test it with different room temperature. 
I found that under these conditions, temperature stability 
is excellent, nothing short of a 3340.
It sounds very, very nice. And it sounds quite different from 
a 3340. A lot more high end in the discrete design!
(I did a 1:1 cross check with my 3340 based dual VCA, adjusting 
the amplitude of the SAW wave to the same level, and
switched between the two VCOs - so the remaining signal chain
 - VCF, VCA - was identical for botth cases.)