paprium is a home-brew game on sega genesis platform, but never released till now.
i believe that its prototype has been already finished and can run on some emulator pretty well.
but there is something wrong while mass producing, most of cart providers can not design a set of hardware that supporting 80M address accessing.
i designed a full range address accessing expansion chip, supporting another 80M homebrew game of UMKT(ultimate mortal kombat trilogy). maybe can help original designers and let dream come true
The SMD 10MB map requires that the cartridge drive /DtAck=B20 when the game reads from memory in the $800000-$9FFFFF region.
However, I tried to build a cartridge that supported this (for UMKT) but it just crashed, so I clearly missed something else.
The 10 MiB map also messes up if a 32X or Sega CD is present, so I would recommend against using it in new productions.
tepples wrote:
The 10 MiB map also messes up if a 32X or Sega CD is present, so I would recommend against using it in new productions.
There's no way to bank the 10MiB to avoid conflicts with these peripheral addresses?
Bankswitching is fine, yes. But the 10MB map is a flat map, not bankswitched, which sets it up to collide with the CD (in the $400000-$7FFFFF region) and the 32x (in the $800000-$9FFFFF region).
(edit #1: For example, the 8MB homebrew RPG Pier Solar uses unreasonably complicated bankswitching, dividing the 4MB dedicated for cart memory into eight 512KB individual banking regions)
(edit #2: I just want someone to explain how the UMKT 13MB map is supposed to work)
Can Megadrive use bankswitching techniques as simple as a latch or a flip flop?
How usefull would that be?
Yes, some pirate multicarts did.
lidnariq wrote:
the 8MB homebrew RPG Pier Solar uses unreasonably complicated bankswitching, dividing the 4MB dedicated for cart memory into eight 512KB individual banking regions)
I was under the impression that Sega's official mapper, used by Capcom's
Super Street Fighter II, supported the same "unreasonably complicated" window set despite
SSF2 using only a portion of its functionality. According to
this description of Sega's mapper, the first window (containing reset vectors among other things) is fixed and the other 7 switchable.
PSolar devide 32Mb into 8 banks. init them 0 to 7 after reset
bank0-4 are fixed.
bank5-7 are swapable.
lidnariq wrote:
Bankswitching is fine, yes. But the 10MB map is a flat map, not bankswitched, which sets it up to collide with the CD (in the $400000-$7FFFFF region) and the 32x (in the $800000-$9FFFFF region).
And apparently the reason GEN Virtua Racing doesn't work with a 32-X attached, though looking at these docs the clash doesn't jump out at me.
https://notaz.gp2x.de/docs/svpdoc.txthttp://devster.monkeeh.com/sega/32xguide1.txt
Maaaaaybe either the 32X or the SVP don't treat /DtAck as the open-collector input to the 68k that it should be? (Needed for the 32X in $800000-$9FFFFF range; needed for the SVP in $A15000-$A15FFF range?) So they get in a fight, one of them wins by pulling it up when it shouldn't, and the 68k thus hangs waiting forever for the data bus to become ready?
If this is what's wrong, you should be able to fix it by putting a diode in between the Genesis and the relevant poorly-behaved coprocessor.
32X doesn't pass top slot !DTACK and that's why SVP/Virtua Racing doesn't work, a simple mod will restore compatibility.