[sdiy] Sawtooth vs. Triangle core VCOs

Neil Johnson neil.johnson97 at ntlworld.com
Fri Jan 25 08:59:50 CET 2013


Paul McLean wrote:
> I've been wondering about the pros and cons of these two approaches.
>
> Would folks be willing to comment and the advantages and disadvantages of both?

As I wrote in a thread here:

http://www.muffwiggler.com/forum/viewtopic.php?t=47713

"It really depends on what the goal is. A saw core is very simple and 
very cheap to implement (one op-amp, one comparator, a transistor and a 
capacitor). As a basic tone generator it is hard to beat on price. But 
it has problems tracking at high frequencies due to the finite reset 
time. Franco compensation helps, as do other compensation tricks. And a 
sawtooth is not the best starting point if you want triangles, for example.

Triangle core is a bit more expensive, but you do get lovely triangles 
from it, and from that its easy to bend those into sines. Because there 
is no reset time a suitably-designed triangle core can run up to very 
high frequencies (MHz).

Through-zero just adds a few tweaks to the core to handle negative 
control currents, and can be applied to both triangle and sawtooth cores.

Then you get into more esoteric designs. As mentioned you can make a 
filter oscillate which will produce decent sine waves with varying 
amounts of phase shift between them. The typical 2-pole 12dB state 
variable filter can produce two quadrature sine wave outputs 90 degrees 
apart, while the main LFO in the CESYG DuaLFO is a 3-phase sine wave 
integrator-ring oscillator, giving three sine waves 120 degrees apart 
(the auxiliary LFO is a triangle core for comparison).

Other designs include astable multivibrators (square waves), inverter 
ring (mostly high frequency), gyrator tanks (high-purity sine VCO), UJT 
relaxation (exceptionally simple), and so on. It really is a case of 
choosing the method that best solves the particular problem. "


While I agree with Ian that with proper design you can minimise the 
reset time to sub-microsecond, it does put a lower limit on frequency 
range than you have with a triangle core.  Example: HP 8165A covers 1Hz 
to 50MHz using a triangle core and only changing integration capacitor 
for range switching.

Neil
-- 
Modules and more :: www.cesyg.com
Homepage :: www.njohnson.co.uk ::



More information about the Synth-diy mailing list