the most strange MMC3 i ever seen

This is an archive of a topic from NESdev BBS, taken in mid-October 2019 before a server upgrade.
View original topic
the most strange MMC3 i ever seen
by on (#160533)
it was worth for chinese bootlegers to make profit with their pirates?

big image:
http://i.imgur.com/pPZK61o.jpg
Re: the most strange MMC3 i ever seen
by on (#160534)
Maybe the cost of a 74ls670 was comparatively cheap enough at some point in the past to justify using five of them?

Would you be willing to share a picture of the back? I'd like to trace it, to see just what it's implementing. (Whether it's a full MMC3, or some subset between that and an N108, or something else altogether)
Re: the most strange MMC3 i ever seen
by on (#160535)
Wow, Very interesting cartridge.
I like to see the back as well.
What is the exact name of the game?

I found this a while ago : Downtown - Nekketsu Monogatari with TTLs!
Re: the most strange MMC3 i ever seen
by on (#160542)
Wow, that's interesting ! I however doubt that a whole MMC3 is implemented with so few chips, probably just a subset of it. Considering most MMC3 registers are 6-bit and the 74xx161 is 4-bit, it means two of those would be required for each MMC3 reigster if you wanted to simulate the full chip, something unpractical.
Re: the most strange MMC3 i ever seen
by on (#160546)
served!
http://i.imgur.com/J7GnZP1.jpg

i never test the game before, i am trying to collect all wirlwind manu carts (LXX). everytime i bought lots of, i do test the cart into my familygame to see that game works and finally the cart gets archived in my W-M box (i have alternate carts to play, such as devcarts for each mapper and some coolboy/supergame/pocketgames series).

the surprise is:

i just saw that the game has "mirroring issues"
http://i.imgur.com/Yy1qPCb.jpg
the garbage at the left place is changing according your advance in the game. (remembah fancy mario?)
also take note that the game just has sound for intro, the whole gameplay is totally mute (no bgm, no sfx).
EDIT: it occurs while i am writing here: unexplainably, the game returns to play music as well lol.

so, i have another curious boards to see:

top gun (a simple UNROM, but WTF!! 4 x 27256!!)
http://i.imgur.com/uh18Lrq.jpg
i have the same board and the same qty of ICs, but the game is a hokuto no ken II.

TKO Boxing (please note the impudence of chinese bootleggers in the shell. taiwan is beautiful)
http://i.imgur.com/N6iMw3H.jpg
the small IC is 7453

it comes to me without shell, sorry. but is a dragon ball 3 from bandai (pirate,obviously), with full working save support.
WTF!!?!? is this the custom chip used as mapper for bros. 8? or i am wrong?
http://i.imgur.com/vnPG9qW.jpg
the game works ok. saves are OK but i am trying to convert into english. please help me as guide:

i am doing:
a patch between the prg from the original dump vs. the prg dump from this cart (in order to extract the mapper hack only)
applying the patch to my favourite translated rom (just prg)
i might this is right. i do with my dodge dampei AX5208C Hack (pictures soon) and all is OK, but i dont have the same luck with this board. the game runs without problems, and text seems to be the same ( i know that i need to replace chr, but the lines, spaces and all this sh*t are identical than original) what to do?.

fire emblem gaiden (AX5208C)
http://i.imgur.com/81D40Jq.jpg
wow.


EDIT:

Japanese Contra:
http://i.imgur.com/BhxTw4d.jpg

HAHHHHAH
http://i.imgur.com/fPDubrj.jpg

wait a moment..... kaiser shell... kaiser board.... but whirlwind manu label? i thought that whirlwind manu and kaiser were independient bootlegers!! any suggestion abount this?

i have a lot of boards that escapes from the standard. i will send tons of pictures from each board (with shell), the problem is that i am full of lazyness haha ok no. all my whirlwind manu games are in a box that is unaccesable atm.

sorry for my basic (and probably wrong) english lol.
Re: the most strange MMC3 i ever seen
by on (#160553)
Snacks wrote:
it comes to me without shell, sorry. but is a dragon ball 3 from bandai (pirate,obviously), with full working save support.
WTF!!?!? is this the custom chip used as mapper for bros. 8? or i am wrong?
http://i.imgur.com/vnPG9qW.jpg
Sure looks like it. The 23C3662 is a VRC2 clone.

Oooh, and it has a working Microwire save -- it's got a 93C56 on it.
Re: the most strange MMC3 i ever seen
by on (#160565)
FARID wrote:
What is the exact name of the game?

the same game you mentioned
Re: the most strange MMC3 i ever seen
by on (#160567)
Snacks wrote:
it was worth for chinese bootlegers to make profit with their pirates?

big image:
http://i.imgur.com/pPZK61o.jpg


Looks awesome. I love these discrete logic versions of mappers.
My only guess is that designing a board without a CPLD is less labor intensive. you don't have to program it so you dont need the extra hardware or time to do so. And the other option of having a new IC produced would also not be worth the time effort and money since they probably planned on making a handful of these as opposed to thousand like Nintendo would have.
Re: the most strange MMC3 i ever seen
by on (#187928)
Snacks - could you please check with continuity tester, where traces from top layer (under chips) go on the cart from your first post (MMC3) ? I am really willing to rev-en this PCB, alternativelly maybe you're willing to sell this cart?
Re: the most strange MMC3 i ever seen
by on (#205980)
Ok, I tried (with success) to restore all tracks and do schematics. It is way more fun that when solving Sudoku, especially if all you have is only two photos =)

Image Image Image
Code:
------------------------------------------------------------------------------------------------
CPU:
$8000-$9fff | $a000-$bfff | $c000-$dfff | $e000-$ffff
   $8000         $8002         $8001          -1
------------------------------------------------------------------------------------------------
PPU:
$0000-$03ff|$0400-$07ff|$0800-$0bff|$0c00-$0fff|$1000-$13ff|$1400-$17ff|$1800-$1bff|$1c00-$1fff|
   $8010       $8011       $8012       $8013       $8020       $8021       $8022       $8023
------------------------------------------------------------------------------------------------
REGISTERS:

[.... PPPP] $8000/$8001/$8002 (mask $8073): Set PRG bank
      ||||
      ++++- PRG bank to set

[.CCC CCCC] $8010/$8011/$8012/$8013/$8020/$8021/$8022/$8023 (mask $8073): Set CHR bank
  ||| ||||
  +++-++++- CHR bank to set
 
[.... ...E] $8020 (mask: $8070): Counter enable/disable
         |
         +- 0: disable, 1:enable

[VVVV VVVV] $8030 (mask: $8070): Set counter value
 |||| ||||
 ++++-++++- set low byte of counter to the following value
 
[VVVV VVVV] $8040 (mask: $8070): Set counter value
 |||| ||||
 ++++-++++- set high byte of counter to the following value
------------------------------------------------------------------------------------------------
Counter operation:
There is 16 bit counter, which counts up and is clocked on every falling edge of M2 (it can be enabled/disabled via $8020)
If it rolls from $ffff to $0000, IRQ is asserted (counter counts from $0000 to $ffff again)
IRQ will be acked automatically after 16 CPU cycles (if counter is still running) or when writing to $8030.
Re: the most strange MMC3 i ever seen
by on (#206022)
That's great!!
Would be too complicated to implement VRC3 in TTL?
I really would like to build a Salamander repro, since the original is very expensive!
Re: the most strange MMC3 i ever seen
by on (#206027)
8x74161 for counters (4 for storing reload value and 4 for storing counter).

Because reloading is one clock operation so you need to do it in one cpu cycle and thus having access to all bits at once
and storing reload value is by 4 bit at once, optimizations like 1x74670 for reload value and 2xsome 8 bit counter won't work.
74138 for decoding, any 3 bit register with 3 state outputs + resistors (74374/74574) or 2 state outputs and mux (74157)
Something for switching counter mode from 8 to 16 bit, probably 2-3 x 7474 for irq signal/ack.

Around 15 chips, thinking how to connec them will take probably a lot of time, routing PCB, etching, drilling, soldering also a lot of time, debugging might be pain in ass (because nothing works since 1 run).
I suggest doing it in FPGA.