Analogue-sequencer

> how are the CV/Gates implemented? i see on the site that 
> they're on an analogue bus, with 16 CV outs + 8 more either 
> gate or CV-in.... this all seems perfect, but what does the 
> bus go to? is there going to be a breakout box as you 
> proposed initially? if so, will it be available as soon as 
> the Cirklon's released? if not, what would you use in the 
> meantime, like a DB24 cable into a patchbay or something?
> 
> another Q, are there any options for things like Hz/Volt + S-Trig?

The factory unit will have two different CV interfaces.
The 'analogue bus' port on the rear is on the CVIO board, which is an
optional extra.
All the electronics to drive the 16 CVs and 8 gates are on the CVIO board,
located inside the sequencer.
A 16bit DAC is used, so the CV outputs can do volts/octave or Hz/volt, with
smooth linear or RC glide.
At the other end of a 25-way parallel cable will be a break-out panel.
It looks like a 1U 19" rack panel is going to be the most popular choice.
PCBs for that are a work in progress, but at the simplest level, are just a
bunch of mono sockets wired to the DB25 socket.
The gate output connections from the CVIO operate from 0 to 5v, with the
same range for analogue input.
Higher gate voltages or S-trig outputs could be achieved with simple
circuitry added to the break-out PCB.
I'm looking at putting the necessary circuitry on there so that it can
jumpered into use.
I'll try to have these available from day one, but it depends on the level
of demand for that option, and how I decide to get them produced.
I've no problem selling a low-cost kit of the PCBs, so that people can stuff
them, and put them behind a Schaeffer/FPE/DIY panel to match their own
system.

The second CV interface, which will be a standard feature, is an alternate
mode for the DIN SYNC output.
To give a bit of background, I have a drum module which I built many years
ago, which has all the original 808 and 909 sounds, copied straight from the
service manuals.
This was pretty much my first synth-DIY project.
I used to drive it via MIDI, but this never sounded quite as good to me as
the internal x0x sequencers.
Now, MIDI gets a lot of stick for latency, but my research led me to
conclude that much of the latency that MIDI gets the blame for is actually
the fault of bloated, non-realtime computer OSes, and over-loading of MIDI
busses.
The cure for that is a hardware sequencer running on a 32-bit CPU with a
hard realtime OS, and multiple on-chip serial ports.

The internal sequencers of both the 808 and 909 run on a fixed timing
resolution of 2ms - and you don't hear many people complain about them being
sloppy.
Timing errors below 2 to 3ms are not audible as timing errors as such. They
are just outside the realm of inter-aural delays - the time difference
between a sound directly to your side arriving at one ear before the other.
Timing differences of that order are perceived as spatial rather than
temporal.
What you most certainly can hear is the difference between two or more
simultaneous drum sounds with their attack phases exactly synchronised, and
the same sounds with their attacks spread over a few milliseconds.
You don't hear this small an error with the sounds triggered in isolation.
IMO it's the spread of attacks, and the changing spectral interaction of
simultaneously triggered drum sounds moving against each other that makes a
MIDI-triggered drum machine sound less tight than the classic Rolands.
This isn't such a problem for synth sounds, since an attack time of even a
couple of milliseconds will mask the spread.

Which brings me back to the Cirklon percussion multiplex.
In this mode, the DIN SYNC output carries 3 digital, and 1 analogue signal.
The digital signals are start-of-frame, clock and trigger. The analogue
signal is accent-cv.
MIDI notes on one channel of the CV port map to up to 32 accent/trigger
pairs, which are clocked out over slightly less than 2ms.
A simple circuit at the other end of the cable does a serial to parallel
conversion of the trigger bits and sample and hold of the accent CVs.
Triggering of all the sounds is then synchronised to the end of the burst.
The result is the ability to trigger many analogue drum sounds with the same
timing precision as the internal x0x sequencers.
You can see this in part 2 of the tutorial video, when I set up the
instrument '909', and assign it to port:cv, channel:10.
There's no MIDI involved - just the multiplex cable, running to this:
http://www.colinfraser.com/tr909/my909.htm

What I haven't shown in the video, is that if I were to trigger all the
sounds in there on the same beat, it really would be at the same time,
within microseconds.
Even with running status, MIDI would spread them over 14 milliseconds.
A mux-receiver PCB, to retrofit an analogue drum machine, will follow when I
get time to do it.
I have future plans to do a new analogue drum module, as a partner
instrument to Cirklon.

Best regards,
Colin Fraser
Sequentix Music Systems Ltd
http://www.sequentix.com