nesdev.com

The copy protection only prevents you from writing an entire disk. It does not prevent you from reading self-written disks to play them. If a self-written disk throws DISK ERRORs, then the disk is either truly damaged, or was written with a misaligned drive.

I am trying to add support for all of these models to NintendulatorNRS. What is keeping me back right now is mostly the lack of 1.44 MiB floppy disk images.

Great! The following Nanjing cartridges are of particular interest: Protection by writing PRG A15 to $5100.1: 梁山英雄 (Liang Shan Ying Xiong) (NJ023) 农场小精灵 (Nong Chang Xiao Jing Ling) (NJ025) Protection by reading from $510x: 轩辕剑外传 之 天 之 痕 (Xuan Yuan Jian Wai Zhuan zhi Tian zhi Hen) (NJ045) 暗黑破坏神 (An H...

Warcraft hangs with solid gray screen after pressing joypad button start. This might be the result of uploading a 1 MiB ROM file into a 2 MiB ROM chip. Because of the bit swap with 5300=07, it is not correct to simply double the ROM data. Please try the attached overdumped, then bit-swapped, file a...

Okay, good info. That means that those two games require a different PCB than Harvest Moon.

Interestingly, the 2 MiB Waixing games, such as their Zelda 3 port, will use that same or a compatible PCB, because they *constantly* use $5200.1 to bank-switch.

Yes, just as Yahweasel has found. Good to see that replacing a 2 MiB with a 1 MiB ROM creates no problem with bank re-ordering. The games that use the $5100.1 protection are "Nong Chang Xiao Jing Ling 农场小精灵 NJ025" and "Liang Shan Ying Xiong 梁山英雄 NJ023". Both need the banks re-ordered using my utilit...

Here is my program that can swap the bank order between $5300=$04 and $5300=$07 for any iNES-headered ROM file.

Edit: The A19<->A20 swap is only performed when the ROM has 2 MiB of data.

The game runs with $5300=$07, i.e. with bit swapping enabled. So you cannot use the commonly-available dump from the internet, because that one was dumped with $5300=$07, i.e. with the bits already swapped. Attached file has the banks re-ordered, as if it had been dumped with $5300=$04, and which sh...

Here are pictures of NJ097's PCB (World of Warcraft — 魔兽世界-恶魔猎人 — Móshòu Shìjiè-èmó Lièrén), courtesy of Yahweasel. It should be functionally equivalent to the previous two PCBs.

Yahweasel, who dumped the Warcraft cartridge, has tried it on his NT Mini, and the sprite looks just as garbled there as in emulators except for FCEUX. ANOTHER game with garbled graphics on real hardware. How sloppy can Nanjing become? With Diablo, at least it was on an ending screen, but on a title...

Cool! Then we try can those other games with the $5100.1 protection as well.

You are correct. The $08 write to $5200 is merely the remnant of putting a full bank number plus $80 into A, then doing STA $5000, 4x LSR, then STA $5200. Then something else must prevent the game's sprite (World of Warcraft (NJ097 — 魔兽世界-恶魔猎人 — Móshòu Shìjiè-èmó Lièrén) from displaying correctly on...

Can $5200.3 be seen on any ASIC pin?

Okay, that glitch then occurs on real hardware as well. Sloppy programming by Nanjing. So the mapper must count PPU reads like MMC5? Not necessarily. I get perfect results for all games by adding a simpler condition: If 5200.3 is clear, perform CHR A12 auto-switch normally. If 5200.3 is set, perform...

Ok, so these games must be using a PCB or even a mapper ASIC variation. Another thing that I have noticed is that even with your improved description of the CHR auto-bankswitch, Diablo's ending sequence still flickers a little. Basically, whenever a character is printed, the game clears $5000.7, the...