midi to Atari2600 cart interface idea

[yogi]

Hi,

Been thinking about Jarek Lupinski's kickstarter project, Chip Maestro http://www.kickstarter.com/projects/jarek/chip-maestro-an-nes-midi-synthesizer-cartridge

After going through his proof of concept design, it occurred to me a similar approach could be used on the A2600, as well as other rom cart based platforms (thinking of my A800Xl). Little Scale has demonstrated a similar design for the A2600 quite a while ago on his blog; I haven't talked to him direct, but he too, seems to be using an Audrino.

I have Paul Slocum's SynthCart and a HiLiq Midi2600 interface, but the player port interface approach has limits, 12 keys/notes per port/channel at a time. So rapidly switching AUDF, AUDV or AUDC values in real time is not possible. In addition, the SynthCart relies on predefined note tables; to achieve scales within the limits of the TIA chip. Don't get me wrong, this setup is very fun and playable considering the platform.

To be able to use the full range of valid combinations of TIA settings, you really need to sequence the 'song' data for a kernel based sound engine. SO not to good for live play.

So a true midi port, mapped into the 6507's 4K of memory; seems like the best option. My first though was a PIC based design, not that I don't like Arduinos, just have a lot of F452s on hand and PICs fit me like an ol' glove.

First design idea- Mapping the PIC's PSP into a small block of memory adrs and running a simple Atari kernel (batariBasic quick an dirty), on a EEPROM in the other part of the 4K; reading the PSPort as if it were 8 address', a basic state machine. Would need a small amount of glue logic for Adr decoding, but it would be very similar to a A2600 bank switching carts' ('magic read' adr switching. RD control only at the cart port). PIC could apply ASDR and LFO type effects to the Atari notes on the fly! Snd FX patches, tempo sync!

Then I thought MIOS! As I research, I think I can:

Use MB SRIO h/w model to interface to the 6507's bus and not use the PSP h/w. Pic controlling the HC595's SC and the 6507 controlling the RC & /OE; or let the PIC control both RC and SC as normal, and tie /OE to the 6507's bus. Glue logic needs to deselect EEPROM and gate RD signal to /OE of selected HC595 (HC138 controlling 8 HC595s). Still thinking about the timing, but I'm confident that the PIC @ 10MIPS will be more then enough (a small CPLD is starting to look good).

So does this seem like it will work? Any thing I've missed?

Yogi

Edited December 7, 2012 by yogi