[sdiy] DCO theory

phillip m gallo philgallo at attglobal.net
Tue Jun 18 01:02:11 CEST 2002

Exactly how do the Roland style DCO's work?

I assumed that they just converted a pulse train into a square wave
and then mutilated in analog from there to the few waveforms they
offer ( speaking from the viewpoint of a Juno6/60/206 owner ).

-->Using the JX3P for example:  A fixed Hi-Freq LC oscillator is input to a
Timer/Counter Chip (i8253) a count is programmed into the timer (correlating
to the "Note" frequency") that will cause the chip to output a short pulse
when the count is exhausted and the Timer Reloaded.  This pulse is used to
"reset" the standard sawtooth integrator circuit. Since 16 bit's won't span
the entire freq. range of interest, the ref osc is divided a few times to
allow some "agility" across octaves.  Two oscillators implement per "Voice"
and they each have their own Timer but are fed the same ref osc/fz division
so they can only be "detuned" from each other but do not differ beyond the
value that can be input into the timer (16 bits).

-->The integrator, waveshaper, filtering and amplitude scaling are all in
the analog domain. Of note, Roland "built out it's 12 bit
convertor/multiplexer which drives the VCF/VCA's.  The non-continuous
frequency nature of this form of DCO makes implementing "smooth" portamento
(which Roland didn't try to implement) and certain other kinds of modulation


Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.371 / Virus Database: 206 - Release Date: 6/13/02

More information about the Synth-diy mailing list