VCO pcb worries

Harry Bissell harrybissell at
Sun Mar 12 04:49:39 CET 2000

Good advice Mike....
The resistor value has to be much smaller than the smallest resistor in the
VCO... but
10-100 ohms is usually correct. If you try to reject supply noise you will need
a big cap to get the R/C time constant low enough. This works good for VCF power
supply rejection also... sometimes only the summing amp in a state-variable
needs this treatment...

Another good idea here is to add Zener diodes to make a shunt regulator out of
the decoupling circuit. Zeners are very fast... they are already conducting...
they just conduct a little more...

Diodes in series with the supply can also be good (instead of resistors). You
get a .7 drop but you can't steal current from the "other" vco's supply caps,
the diodes are reverse
biased that way....

Most of these are "fixes" for after the fact... usually to cover up that you
tried to get away with an inferior layout (like you did not "star" the supplies
to the lowest impedance point on the board).

H^) harry

WeAreAs1 at wrote:

> Hello folks,
> I'm wondering if it might help to isolate the supply legs of the various
> circuit sections (each VCO) by connecting their power through a low-value
> resistor (10 to 100 ohms) in this manner:
>                      small res.
> Power Supply leg >-+--/\/\/\/---+----> to VCO #1
>                    |            |
>                    |           ===
>                    |           === decoupling cap
>                    |            |
>                    |          -----
>                    |           ---
>                    |            -
>                    | small res.
>                    +--/\/\/\/---+----> to VCO #2
>                    |            |
>                    |           ===
>                    |           === decoupling cap
>                    |            |
>                    |          -----
>                    |           ---
>                   etc.          -
> I've seen this used to good effect for decoupling noisy or spike-prone
> circuits such as LFO's and digital noise generators.  I honestly don't know
> how well it would work on such a subtle thing as oscillators soft-syncing to
> each other via tiny currents, nor do I know what value the R/C network would
> need to be to effectively decouple - but it might be worth a try.
> Michael Bacich

More information about the Synth-diy mailing list