This module is designed to produce a
variety of voice sounds.
It is based on Ai Cybernetic Systems model
1000 Speech Synthesizer.
It uses an XR2206 IC and a transistor based
noise circuit for the sound sources. These
and a pitch period control are switched on
and off by a DG409 multiplexor.
The circuit requires three gate signals
to operate, perfect for using with various
CGS modules like the Master Divider or
other cmos based sequencing circuits.
The output of this section is mixed and
fed into 7 parallel bandpass filters.
The filters have Q=10 and centre
frequencies of 55, 110, 220, 440, 880, 1760,
3250 Hz.
Each filter has a pot to control the
incoming signal and a switch to select
whether the pots attenuate the signal or
are mixed with a feedback signal from the
output.
There is also an external input if
just the filter section is to be used.
2/7/2010 built two versions of this on veroboard. One used an exponential input cct to control the 2206
, the circuit is a hybrid of one given in Electronotes and the typical
one seen on just about every VCO these days. The other is similar to the 2206 control cct on Thomas Henry's XR2206 VCO, it is the better choice if you want to build this module.
The schematic below is the circuit as it stands, works, doesn't sound anything like someone speaking but sounds pretty good anyway.
The 880 and 3250 Hz filters oscillate at certain settings in resonant mode. TI filter design software was used to design the filters, if
I were to build another I would use higher Q setting and try to get rid of the oscillations.