If I remember correctly, the patch corruption is caused by a stack overflow that pushes the stack up into the patch memory when MIDI comes into fast for Poly to process.
And sysex was left out of the Poly code because it can't fit with the keyboard and joystick code. So the EX-800 had sysex because it had no keyboard or joystick. I calculated fitting some sysex functions in to where the tape code was. But the tape code is actually quite small. So no luck there either. Bottom line is, there is no room.
And as you all know, I myself poured over the assembler years ago contemplating these problems. There is literally only 6 bytes unused out 8,192. It's a real shame that Korg chose to design the Poly with only an 8K ROM. Had they added just one socket for another 8K ROM, they could have dramatically improved the Poly and charged say $100 bucks for the upgrade.
So due to the space limitation, neither of your wishes can be granted.
/Mike
From: "Gordonjcp gordon@... [korgpolyex]" <korgpolyex@yahoogroups.com>
To: "Michael Hawkins korgpolyex800@... [korgpolyex]" <korgpolyex@yahoogroups.com>
Sent: Friday, May 29, 2015 8:58 AM
Subject: Re: [korgpolyex] Demodulating and Encoding the Audio/Tape Data
On Fri, May 29, 2015 at 12:56:04PM +0000, Michael Hawkins korgpolyex800@... [korgpolyex] wrote:
> Here is the tape disassembly as far as I took it. The main routines are labelled but the guts of each function is a mystery to me. Personally, I didn't see the point in trying to continue to use the tape function at all. So I didn't pursue disassembly any further than the basic entry point, functions and exit.
>
> If anyone wants to pursue this I can provide a memory map of HW IO and memory.
> But I would seriously suggest that any time spent on Poly would be better spent elsewhere. Just my two cents.
>
Tell you what I'd like to do is fix the patch corruption bug in the Poly 800 original firmware, and maybe graft in sysex patch dumping. I'd like to keep the rest of it the same though.
--
Gordonjcp MM0YEQ