RCA MK II

On Wed, 18 Oct 2006, Stefan Tiedje wrote:

> I just came across some sites about the first synthesizer, the RCA. I found
> that nice block diagram, and thought why not build a Max patch resembling the
> old RCA MK II. Problem, there is not much more information than that diagram,
> but it seems Brad Garton has the original in his room and could eventually
> pull out some additional infos?

i've got the schematics and the operating manual (ditto-printed in 1957)
in my apartment. it's not the first synthesizer by any means, but it's
probably the earliest example of a programmable sequencer/synth unit.
last time we turned it on it caught on fire, but we've got enough spares
to build two more of them, so we might give it a shot at some point.

eric chasalow made a 'virtual rca' at some point to realize the piece
babbitt was working on when the machine was vandalized. a good chunk of
the machine works fine, though it's not much of a synth to begin with
(4-note polyphony combined + little continuous parameter control + 472
vacuum tubes, half of which are blown = kind of limited). we've got
buchla and serge modular units that are way more powerful in terms of
synthesis... it's the operational logic of the thing that's kind of
amazing.

> After looking on a picture of a punchcard,

the machine uses paper tape, not punchcards (actually two synchronized
tape drives, each controlling two 'voices' of the synth). it synchronizes
to a shellac record lathe for recording of your output.

> it seems all values are coded in 4-bit, which would be a snap in Max,

it's a 4-bit system, though a lot of the parameters only paid attention to
0-11, with the top four values unused. remember that the system was
designed for dodecaphonic music, so most of these guys could only count to
12.

> What I am curious about especially are
> the blocks called "Frequency Source", "Frequency Glider" and "Resonator
> Chains".

the first two are how the RCA codes pitch and portamento time,
respectively. the synthesizer works by heterodyning a bunch of
high-frequency oscillator pairs: a set of fixed tones (still in tune last
i checked) and a voltage-controlled one for each voice... what you hear
from the synth are the difference tones, which you can combine through
this really messy octaver thing to get different timbres for each pitch.
the frequency glider is just a way to do portamento by switching between
different oscillator pairs instead of actually gliding the source
oscillator; i never figured out why the hell you would do this (it's
broken anyway), but my guess is that since the paper tape has no way to
represent a continous function you would use this to 'step' between
different in-between states. the 'resonator chains' are just a bunch of
analog comb filters (short delay banks... very similar to those found on
the buchla 200 series). they also had a big EMT plate reverb wired in
later along with some other gear.

> If there is more info available, it would be a fun project to rebuild it in
> Max/MSP...

go for it.

> here the link to the diagram:
>
> http://uv201.com/Misc_Pages/Misc_Images/rcasynth_block_large.jpg

here's a few other articles:

http://en.wikipedia.org/wiki/RCA_Mark_II_Synthesizer
http://machines.hyperreal.org/manufacturers/RCA/rca_ems.txt
http://www.obsolete.com/120_years/machines/rca/

best.

/luke