[sdiy] LFO VC Skew?

Oakley Sound oakleysound at btinternet.com
Tue Feb 18 10:31:19 CET 2020

I built Don's many years ago. I'll have to dig up the details but I 
found that the steepest slope of the waveform could not be made fast 
enough for my purposes. That is, the flyback part of the sawtooth/ramp 
was not steep enough to sound like a sawtooth/ramp proper.

In the end I used the same two VCA sections as Don but used a different 
minimum extractor circuit with a much larger inherent gain to allow for 
a wider 'mark space' ratio. The problem I then ran into, not 
surprisingly with such a huge gain, were the unmatched offsets from the 
OTAs (I used a LM13700) which led to obvious glitches in the output 
waveform. Using a 2164 probably would have been the better approach, 
although Don used the 3280 which would have been better too.

The benefit of Don's waveshaper idea, and any that uses the core of the 
VCO, eg. DUSG and others, is that you have a system that morphs smoothly 
from sawtooth through triangle to ramp. The downside is that you are 
unlikely to achieve fast flybacks as you can't get that close to the 
100% or 0% mark space ratios without a big and unwieldy spread of VCA 
gains. It's often good enough for an LFO though.

Those systems with the gain stages in the VCO core are also likely to 
suffer from pitch deviations as the shape is changed since we may not be 
able to produce perfect complementary gain control to the pair of VCAs. 
David's idea with the 2164 may prove that to be less of a problem.

The VCS3, and others based around the same offset and rectification 
idea, use a different morphing arrangement whereby you get an obvious 
discontinuity in the waveform as the shape is changed. You get a perfect 
sawtooth and ramp, with ultra fast flybacks at the ends, but the 
triangle is only half the amplitude and it's not really varying mark 
space ratio at all. It is, however, a very interesting waveform I think 
is probably more useful at audio frequencies than the true morphing 
system. I wish when I implemented it in the Oakley SVCO I had added a 
VCA in there to keep the overall amplitude of the waveform the same for 
all shapes. I have done so for the Oakley Mysterious Audio Plaything 
(MAP) project which has yet to be properly finished - but I am impressed 
with its audio qualities.

Digital solutions may be the best though. The saw-tri-ramp waveshape 
control on the audio oscillators of the venerable Alesis Ion does the 
job well.



More information about the Synth-diy mailing list