[sdiy] Digital pots for polyphonic portamento..

[Colin f]

> Software method.. Good idea! But I don't want to load
> the 'analysis' time for each received MIDI bytes to CV outs.. 
> wondering..
> I'm using a PIC16F877 at 20mHz.

Analysis time isn't an issue if you're receiving MIDI data in an interrupt
routine which then updates a state table for the code generating the glide.
The glide code must run at a fixed sample rate - easily done in a timer
interrupt, with a lower priority than the serial interrupts.

On a current project, I'm using a PIC 18F4520 with a 16-bit DAC.
The DAC is demultiplexed between 8 outputs.
The PIC is calculating an RC curve at 16-bit resolution for each output,
with a total refresh time for all channels of 1ms.
There's plenty CPU time left over for MIDI handling, voice allocation and so
on.
A PIC 16F877 only manages half the speed of the 18F part, but for 12-bit
values, and fewer outputs, it should manage the same sort of refresh speed.

Not only could you save on the hardware glide circuits, you could save on
the DACs by swapping all but one of them for a 4051 and some op-amp buffers.

Cheers,
Colin f