Do I presume correctly that
HVC-SZROM is HVC-SOROM but with CHR-ROM instead of 8 KiB CHR-RAM?
I started tracing the circuit, but I'm not confident that it's specifically CHR A15 that's selecting between PRG RAM banks on SZROM. Too many traces are hidden under the MMC1 and RAMs.
I am confident that CHR A12 through CHR A15 do go to the CHR socket, however.
I just debugged A Ressha de Ikou. It took me a while to understand how to save the game (START+A+B). It writes $10 to $BFFF to select the battery-backed PRG-RAM bank, and $00 to select the non-battery-backed PRG-RAM bank. This means it differs from SOROM, which uses the $08s bit to switch PRG-RAM banks. Given what you wrote about CHR A12-A15 going to the CHR socket and with four bits below the PRG-RAM selection bit, SZROM must then be able to take 4 KiB*16=64 KiB CHR-ROM.
One thing that I'm really not confident of—say maybe 60% confidence—is that it looks like CHR A16 is probably still connected to the hole on the PRG socket that would be PRG A18. (As well as to PRG RAM ≈A14)
Given that this is a singleton with only 128KiB of PRG, it is irrelevant whether it is thus connected... but on the unlikely chance we ever do find a larger game, perhaps worth remembering.
Or pin 2 is shorted to pin 20 and both are /PRGOE.
I don't agree about the submapper requirement though. Some people may not like ROM/RAM-size based distinctions, but since they are now required for the other fancy SxROM types, detecting SZROM via PRG-RAM/PRG-NVRAM being both 8 KiB each and a non-zero CHR-ROM size (vs. zero CHR-ROM for SOROM) is only consistent. Apart from consistency, if the ROM size is the cause of the board variation, because the standard PCB could not accomodate the particular ROM/RAM configuration, then I find size distinctions actually preferable to submappers.
It's suprising new boards are still found as late as 2019 ! (Well the link is dated 2013 but I wonder whether it's the real date, coming from a broken "forum" with only 2 users 4 threads and 5 posts).
Quote:
Some people may not like ROM/RAM-size based distinctions, but since they are now required for the other fancy SxROM types, detecting SZROM via PRG-RAM/PRG-NVRAM being both 8 KiB each and a non-zero CHR-ROM size (vs. zero CHR-ROM for SOROM) is only consistent.
Agreed, although this new board is now making things more complicated than ever when it comes for oversize MMC1.
There are 51 Licensed Famicom games which are designated to use MMC1 and for which we do not have board shots, and five of them are designated to use SRAM. Until that they are all definitively known, we may still conceivably encounter obscure variants like SZROM. Could there be an SDROM, an SPROM, SQROM, SSROM, SVROM, SWROM or SYROM out there?
Sorry, I miscounted. There are eight with highlights on my list where I could not find the board shot on NES Cart DB or elsewhere and have been indicated to use battery backed SRAM :
Chaos World
Cocoron
Cosmo Police Galivan
Famicom Shougi: Ryuuousen
Jangou
Kabushiki Doujou
Meiji Ishin
Nipponichi no Meikantoku
Great Hierophant wrote:
There are 51 Licensed Famicom games which are designated to use MMC1 and for which we do not have board shots, and five of them are designated to use SRAM. Until that they are all definitively known, we may still conceivably encounter obscure variants like SZROM. Could there be an SDROM, an SPROM, SQROM, SSROM, SVROM, SWROM or SYROM out there?
Sure but if anyone actually played that SZROM game in an emulator, that would have emulated it as regular SJROM, it would be obvious that at some point the game would refuse to save, and/or crash, or something similar. Such behaviour would have made people (especially emulator authors) buy the real cartridge and analyze what's wrong. So it's surprising we only discover it was using more than 8kb of PRG-ROM now.
As for games marked to use PRG-RAM - the iNES header does not specify wether there is non-battery backed PRG-RAM or no PRG-RAM at all. As such, we can't really know whether more games uses non battery backed PRG-RAM (if we didn't try to play through them on emulators as I've previously said).
By Mappers, we are missing board shots from the the licensed Famicom releases :
Mapper 0 - 9
Mapper 1 - 52
Mapper 2 - 18
Mapper 3 - 5
Mapper 4 - 66
Mapper 9 - 1
Mapper 18 - 1
Mapper 23 - 1
Mapper 32 - 1
Mapper 48 - 3
Mapper 66 - 1
Mapper 68 - 2
Mapper 70 - 2
Mapper 80 - 1
Mapper 92 - 1
Mapper 96 - 1
Mapper 118 - 1
Mapper 157 - 5
Mapper 185 - 1
Mapper 186 - 1
Mapper 188 - 3
Total : 176
At some point, someone had to have gone through and assigned mapper information to them and until someone else proves otherwise, the canonical mapper assignments must be relied upon to be correct.
Chaos World, Cosmo Police Galivan, Ryuuousen have 128 KiB PRG of CHR-ROM and thus cannot bankswitch PRG-RAM using any known method.
Cocoron, Jangou, Kabushiki Doujou, Meiji Ishin, Nipponichi no Meikantoku only initialize the CHR registers as one would for CHR-RAM and never when for accessing PRG-RAM, so it cannot bankswitch PRG-RAM to access save data.