[sdiy] CMOS Synth

[jhaible]

Very nice stuff, Sam !

JH.


-----Ursprüngliche Nachricht-----
Von: Osamu (Sam) HOSHUYAMA <houshu at rogers.com>
An: synth diy <synth-diy at dropmix.xs4all.nl>
Gesendet: Sonntag, 2. Februar 2003 04:40
Betreff: [sdiy] CMOS Synth


> Hello List,
>
> Finally, my CMOS modular synth (MOSular ;-) began to work as a system!
>
>     http://www5b.biglobe.ne.jp/~houshu/synth/MOSular01.JPG
>
>
> Frankly, it's a toy project, but not so bad as a toy.  I know that
> plastic lunch boxes can't last so long.  However, they are extremely
> cheap (1 CDN each), and drilling and lettering panels are very easy.
> (Drilling a box took less than 1 hr.  I used 4 sheets of inkjet-
> printed paper for the panels.)  The boxes are not robust, but much
> better than bare boards.  I think this is the best choice for my
> limited tools and time in Canada.  I first planned wall-mount modular,
> however, the panel is too weak to support so many plugs unfortunately.
>
>
> Most of the schematics including block diagram are available here:
>
>    http://www5b.biglobe.ne.jp/~houshu/synth/index.html
>
>
> Features (that I'm proud of):
>   * All the audio signal paths including VCF consist of CMOS 4000
>     series.  No OTA is used.
>   * Very compact including Midi to CV converter (Synhouse MidiJack:
>     http://www5b.biglobe.ne.jp/~houshu/synth/MOSular02.JPG)
>   * VCOs are temperature compensated using NTC resistors.  Their
>     stability is comparable or better than my PAIA 9700 synth.
>   * Almost every module uses just one chip except for VCOs and VCF.
>     Very simple in the schematics.
>
>
> Problems (a little shame):
>   * Too many modules are packed in small boxes.  Complex patching is
>     not easy.
>   * Power consumption is large because all the amplifiers (inverters)
>     are operated in class A.  (Heatsinks are needed for regulators.
>     I used 2 sets of 3 terminal regulators to distribute the heat, and
>     to reduce noise with VCF and soft sync with VCOs.)
>   * VCF and VCA are a little noisy and have some thump with fast attack.
>   * S&H's performance is very poor.
>   * LFO can't generate real sawteeth waveform (0:100).
>   * A lot of resistors need trimming, which depends on the chips.
>     I don't want to change the chips. ;-)
>
>
> Anyway, my absurd but challenging project has been almost completed.
> Thank you very much.
>
> Special thanks to Rene Schmitz and Scott Gravenhorst for their
> inspiring schematics, and to Harry Bissell, Mike Irwin, and Peter Snow
> (and Mark Hammer and Gilles Caron) for their encouragement on the
> packaging.
>
>
> My deepest sorrow to Columbia's accident.
>
>
> Sam Hoshuyama
> -- - ---  - -- ------ - -------- -- - ---- -
> Osamu (Sam) HOSHUYAMA      houshu at rogers.com
>                               Ottawa, CANADA
> --------------------------------------------
>