You're right, the first two bits are used as the bank select for patch memory. The last one was used in previous versions of hardware to provide a RAM bank select. But, it cannot be used for anything due to incompatibilities with the older version boards.
The three new latches are available for anything right now. I haven't written any code around them. But, I would imagine two could be used for LFO flashing LED's. And one was intended for note hit trigger.
But, there are lots of spare chip select lines available for your multiplexer needs. If I recall correctly there are six spare chip select lines. The latest version boards also use a buffered write signal. So, it is very likely that we can add up to 6 additional input buffers or output latches to accommodate additional hardware mod's.
Mike
From: bimmerfan222 <bperkins211@...>
To: korgpolyex@yahoogroups.com
Sent: Tuesday, July 3, 2012 1:04 PM
Subject: [korgpolyex] Re: HAWK-800 - 100+ units shipped...
--- In korgpolyex@yahoogroups.com, Michael Hawkins <korgpolyex800@...> wrote:
>Basically, the code can write values to the six bits and the latches keep them in that state after the CPU moves on to other tasks. I know someone wanted a signal that indicated a note hit. So that's one bit latch taken for that already. Then, your encoder project was probably going to use one or two latches. Adding a scanning latch to HAWK would not be too difficult so it would seem unnecessary to steal keyboard inputs.<
So can the code be altered to release that note hit latch?
I'm guessing the first three used up are required to pick between memories or something...
Am I following that it's possible to add in another 3 to 8 demux to create another scanning routine to a new set of matrix's?