[sdiy] microcontroller selection!

[Paul Maddox]

Bret,

> I just completed a prototype of a Doepfer compatible module that uses the
> OOPic microcontroller.  The module is a generic - programmable module for
> logic function.  It can do sequencing, clock division, pattern generation,
> etc.  I'm working on a website that shows it off, so I'll have a demo
soon.
> I'd love to sell it soon and share the joy! ( I may start my own business
> making music gear!!  Woop! )

cool..

> In any case, now that it's finished, I've found that there are some speed
> issues.  Especially when doing a/d conversion.  The module is quite
> functional and fun, but there's a small delay between the time that you
> clock it to when it produces output.  BEST case (very rare), is a nearly
> instantaneous response... but that's the simplest function programmed in
the
> fastest way possible (using virtual circuits in the oopic).

Hmmm, not bad, but I guess it could be noticeable

> Here's my question - - - should I go with this solution, using the
slow-ish
> OOPic, or is it worth researching using a faster microcontroller?  The
basic
> stamps might work better, but wouldn't I still get delays when doing a/d
> conversion?

For my money I would avoid runtime interpreters like the basic stamp, I
would go for a compiler (C would probably be your best move).
this would reduce the overhead.
A lot depends on what speed your PIC is running at and how you want to do
things.
I would suggest finding a C compiler (GCC perhaps) and writing the software
in C and then compiling it and programming it to the PIC, this will be much
quicker than using basic stamp kind of thing.
There are faster micros, AVR, Scenix, but I suspect moving away from an
interpreter and into a compiler will solve the problems.

> I would prefer a much faster family of Amtel microcontroller, but I really
> wanted to make this super easy to program, and super easy to hook up to a
> computer.  Even with an Amtel microcontroller, there's a chance that a/d
> conversion or sloppy programming could cause small timing issues.

16MIPs is plenty fast enough, the delay betweeen the monowave receiving a
complete midi note on and it loading the oscillator/firing the ADSRs is
about 5uS, and thats only running at 8MIPs!

move away from interpreters and into the world of Compilers..

Paul