PolyModular

[Edward Leckie]

     OK, I have another idea. How about using cross-point switches and OTAs 
     to route signals? But that loses the beauty of a modular, the manual 
     routing of patch cables. So how about using some sort of simple 
     digital I/O to communicate connection between modules. Each output 
     could transmit a let's say a 5V pulse sequentially, and each input 
     could sense the pulse and configure the n (= number of voices) 
     cross-point switches accordingly. That way, you can use single 
     conductor patch cables and VC modules for a poly synth. Pots could be 
     used to control VCAs of signals entering or leaving the cross-point 
     switch matrix. To generate the serial data one simple uC and some 
     decoder ICs. The beauty of this as well, is that a patch setup can be 
     stored!! To interface to the outside world, an external CV inputs to 
     crosspoint switches should be included.
     
     Ed.
     


______________________________ Reply Separator _________________________________
Subject: RE: RE: PolyModular 
Author:  Barry L Klein <Barry.L.Klein at wdc.com> at Internet
Date:    04/5/99 9:51


     
     
Hi Barry
     
thanks for your critical points - you're certainly right in some way. 
Only that *I* cannot do microcontroller stuff. (And aren't uC's 
somewhat outdated as well, in the age of DSPs ?)
BTW, I hope I wasn't too abrupt with my reply and suggestions - I was limited in
ime to
get the reply out before I left to pick up my kids at school.  I have familiarit
with the
hardware aspects of microcontrollers, but not with the programming aspects.  But
his is
what I want to learn about and such a project would be an opportunity to learn 
it.After
all, I learned the electronics I  DO know from building analog modulars.  I susp
t this
is true of many of us.  The costs to include uC's in designs is pretty low.  Mic
chip,
Zilog, and others have chips that are under a dollar in quantity that could impl
ent
autotune, one or multiple sequencers, multiple envelope generators, etc. etc.  T
se chips
will have much longer longevity than discrete counterparts.  We already see diff
ulty
getting CMOS, LS, and analog components.  But uC's are available all over the pl
e.
     
I don't see much reason for a uC within the modules. But if someone 
can make a keyboard interface that has *all* the old OB-FVS 
features (plus velocity added ?), or a Midi interface like this, 
using a uC, I would be very pleased. 
Can be done.  Plus, using a uC should allow a much more complex and capable modu
.  The
emphasis as schools/colleges is in training on how to use them and other FLPGA's
tc.
rather than basic electronics.  So lets put those "kids" to work!  What these 
kidsdon't
understand is what we want them to design.  We can still contribute towards 
that.
     
As for tuning, I think with linear VCOs I can do well without 
autotune. My CS-50 is rock stable for example. I have not 
tuned it for years, and the tuning is still perfect.
For the Modular, 5 tiny tune pots will do for a VCO module.
Myself, I don't care if it uses one or not if it works decent.  But I hate to us
so much
panel space for one oscillator.  And then only have the standard waveforms.  I'm
ored
with them.  I want a microwave in a module that I can modulate/control with the 
her
modules in my system, as well as download sounds from my pc into each one.  I 
wantto do
all this trippy stuff like granular and physical modeling, more exotic FM/AM, 
etc.
     
What I'm getting at is that if we continue to think the same way about the archi
cture we
will get the same basic results/sounds.  Lets try and use new ideas and concepts
o go
further (before the software synths clean house).  
     
I do see that there is an immediate interest in an inexpensive analog modular pr
ect, and
much have what's gone on in this list pretty much supports that interest.  But w
really
haven't emphasized utilization of 90's technology into these systems.  Many of u
do this
in our jobs every day but then come home and discuss sequencers using 4017's or 
atever. 
I don't get it.  I have access to a Data I/O at work and would have no problem b
ning
micros at my lunch hour to distribute to list members.  Others have access and f
iliarity
with other programmable chips.  Others think Visual C++ or Basic all day and 
couldwrite
really cool pc interfaces to all our stuff.
     
And no, I don't want to start a "digital_synth-DIY" list.....  :-)
     
Is this doable - could we come up with some sort of specs to make it all work?  
ts hear
YOUR off the wall desires you guys!
Don't leave me out here all alone.....
     
Barry