hi guys,
one of the first decisions would be to define the border between existing SSL and new electronics
i could see the use of existing SSL interface cards rather than designing an entirely new system - here's why
the existing computer is actually a hard disk recording system with 12bit resolution and up to 128(?) i/o lines
the voltages on the in and outputs are analogue and range from 0V to +10Vdc
0V (fader down)
5V (fader up)
10V (channel cut)
so you could probably use any existing daw (with slight modifications) to substitute the computer!
some years back i thought of using some of the SSL computer cards and interfacing them to a new box, rather than starting from scratch at the ribbon connectors
hardware in small quantities (even a couple of 100 units) is costly
to incorporate the SSL ribbon connectors onto a pcb takes up lots of space (which is the factor that defines it's price)
and everything is there already
the ribbons carry analogue voltages from/to the console, so for a fully digital (cheap and easy) solution you'd have to replace the vcas in every channel for digitally controlled gain elements
this would be expensive, time consuming and change the sound of the desk dramatically - but not for the better
you'd need to move digital serial information through the fader lines which would be audible - you can probably hear what i mean when using total recall to scan the desk
i personally would not consider any solution down that line
the actual interface cards for analogue i/o and leds+switches (trim/abs leds + their adjacent status switch switches) are widely available for reasonable prices and can be serviced easily
most folks already got them anyway
the analogue output cards contain sample and hold circuits for every channel - to come up with a substitute could be costly
they were designed to respond quickly enough (= they support fast fader moves) + also cater for the largest number of channels
the 'sample and hold' circuit buffers information (in our case the level of a channel)
the automation computer outputs updated information periodically (i.e. it updates every channel within specific intervals)
i guess the value of that frequency relates to the channel count which has to be set in the setup pages of the e/g series software
i actually thought that very large consoles respond more sluggish
maybe a modern processor could get away without s&h..?
for me the 12bit (1024step) resolution is no bottleneck either - so many of the worlds best sounding tracks got away with it
like i said, these i/o cards can easily be maintained
it's the processor and floppy drives (+ controllers) which eventually could create problems
i'm quite confident we could come up with an interface circuit which translates the information to a commonly used protocol - based on ethernet or something?
but then there's the software bit..
to make use of this information, sequentially record it with sufficient resolution, rock solidely linked to smpte t/c, represent it graphically + provide a gui for editing and control the transport of the daw
needless to say, another requirement is to provide up to date insults on user errors
just my thoughts..cheers, werner