MMC3 DIP? - NESdev BBS

MMC3 DIP?
by FyberOptic on 2006-08-29 (#16678)

While looking for games on Ebay to cannibalize, one of the carts I ordered was Roger Clemens MVP Baseball. nesmapper.txt said it was 128k PRG/256k CHR MMC3, and while boardtable.txt said it was 53361, boardnames.txt had no information on the board. But I figured as long as it was MMC3, it was probably of use to me.

So I popped it open today, and found that the entire board just seems kind of cheaply made compared to others I've opened. It's very very pale green, and the front doesn't have any sort of material covering the traces (I don't know what you call the stuff). They're just as shiny silver as can be. The back only has just enough of such material as to cover only the actual pathways and not the entire board surface. I'm almost weary to even try desoldering on such a board for fear of tearing the traces off. Fortunately though it uses two 32-pin chips for PRG/CHR, even though the PRG is supposedly just 128kb (and the rom pinouts text only shows a 28-pin version of the 128kb PRG), so that should make it easier to socket.

But here's the kicker. I always thought MMC3 was a small surface-mount chip. But to my surprise, it would seem this has a huge DIP along the top. It's not labeled MMC3, though, it's a 40-pin Acclaim chip, MC-ACC. I thought this was pretty interesting. Is there a pinout available?

In any case, this might be a common thing for Acclaim boards for all I know, and it might be nothing new to a lot of you guys, but if nothing else, I thought it might be of some use for boardnames.txt, since the 53361 board has no info.

http://www.fybertech.com/nes/53361_front_large.jpg

http://www.fybertech.com/nes/53361_back_large.jpg

I tried to cover the flash to keep it from blinding the boards, so one half is a little darker than the other. The silver traces ended up looking gold. But turning the flash off just results in slow shutter mode, which blurs things all to hell without a tripod.

EDIT: Apparently there were two versions of this cartridge according to the NES Cart Database. The first is like I described, and the second is listed as actually using the MMC3. Does this mean the MC-ACC isn't necessarily even an MMC3? Or is it simply a compatible chip from a different manufacturer?

by BMF54123 on 2006-08-29 (#16682)

Huh...apparently, I have the second revision, with the MMC3B:

http://misc.rustedmagick.com/mvpboard.jpg

(the sides are gouged out because it was stuffed into an unmarked Tengen case...yeeeah)

I would assume the Acclaim chip mimics the MMC3, since the ROM revisions are the same (they would've had to rewrite some code if the chip functioned any differently). Acclaim was one of the few companies that manufactured their own cartridges, which is why the board and parts are so much different from "standard" ones. I've seen cheap Taiwanese pirates that are higher quality than that...

by FyberOptic on 2006-08-29 (#16683)

I got to thinkin, since it's just 40 pins, where as the surface-mount version is 44, I'm going to bet that this one doesn't support WRAM. Unless someone has a pinout available, I might have to follow the traces myself and see what I can come up with.

Also, does anyone have the pinout of the 32-pin 128kB PRG roms? I could probably assume what they are based on the 256k ones, but I want to be sure.

EDIT: It would seem that none of the games listed in that NES Cart Database using the MC-ACC have WRAM, so my theory may be the case.

by blargg on 2006-08-29 (#16686)

Quote:
I would assume the Acclaim chip mimics the MMC3, since the ROM revisions are the same (they would've had to rewrite some code if the chip functioned any differently).

Or at least it mimics the functionality and behavior needed to make these particular ROMs run acceptably.

by BMF54123 on 2006-08-29 (#16690)

blargg wrote:
Or at least it mimics the functionality and behavior needed to make these particular ROMs run acceptably.

So Acclaim skimped on more than just their software budget, eh?

So, I wonder if the MMC3B revision is actually the original, with the cheaper (?) MC-ACC coming later? The latter does use a higher board number, after all.

by blargg on 2006-08-30 (#16697)

Quote:
So Acclaim skimped on more than just their software budget, eh?

I don't know. My comment was simply that it would be unwise to assume that Acclaim's MMC3 substitute is a full MMC3, since it only had to work on for the particular games they used it with. You could spend the time removing it and determining the pinout, only later to find that it's only MMC3-like.

by Bregalad on 2006-08-30 (#16698)

Industrially talking, I don't say the interest Akklaim had in devlopping their own boards and ICs when they aren't basically a corporation specialised in electronics but only in video games publishing, while they could just let Nintendo do all this work and let them gain a part of the buisness (wich they were most probably forced to do anyway).
The same goes for Konami, but Konami is definitely a bigger company than Akklaim.

by tepples on 2006-08-31 (#16732)

special-ised becomes "my mom" because it contains the name of an ED drug (Cial-is). Please use -ized on this board; it's closer to the original Greek form anyway.

But then Konami also made arcade games, which at the time used heavily customized hardware.

by Memblers on 2006-08-31 (#16753)

Well they definitely went all-out with that game, I know it was released on SNES too. Actually I thought it was pretty good on NES, especially since they usually released their worst games under the LJN label. I don't understand why they had their own mapper, but it must've been cheaper in the quantities they were making. Could also be that it was cheaper for them to do only through-hole assembly with what they have (wave-soldering I'd guess, since the vias are filled).

Pretty amusing how there's only solder-mask on the copper, rather than the whole bottom of the board. I guess that'd save a bit on materials somehow (but wouldn't it just go on the stencil or something?). I've never seen that done before.

by rbudrick on 2006-09-01 (#16786)

That board is super ghetto. Fucking Akklaim, man. I'd almost say that board was actually made in Afghanistan, but that would be insulting to the Taliban.

-Rob

by sevast on 2006-09-01 (#16792)

Too bad it's a standard CIC. A large die CIC might have been very helpful..

by FyberOptic on 2006-09-01 (#16795)

The horribly tedious work of continuity testing has given me a preliminary pinout:

01 - GND
02 - PRG_ROM #2 (A16?)
03 - PRG_ROM #3 (A17?)
04 - PRG A14 (n)
05 - PRG A13 (n)
06 -
07 -
08 -
09 - PRG A0 (b)
10 - PRG D0 (b)
11 - PRG D1 (b)
12 - PRG D2 (b)
13 - CHR_ROM #2 (A16?)
14 - CHR_ROM #3 (A15?)
15 - CHR_ROM #29 (A14?)
16 - CHR_ROM #28 (A13?)
17 - CHR_ROM #4 (A12?)
18 - CHR_ROM #25 (A11?)
19 - CHR_ROM #23 (A10?)
20 - GND

40 - 5v
39 -
38 - PRG_ROM #30 (A17?)
37 - PRG_ROM #29 (A14?)
36 - PRG_ROM #28 (A13?)
35 - PRG R/W (n)
34 - PRG /CE (n)
33 - PRG_ROM #22 (/CE?)
32 - PRG D7 (b)
31 - PRG D6 (b)
30 - PRG D5 (b)
29 - PRG D4 (b)
28 - PRG D3 (b)
27 - VRAM A10 (n)
26 - CHR_ROM #30 (A17?)
25 - /IRQ (n)
24 - CHR A12 (n)
23 - CHR A11 (n)
22 - CHR A10 (n)
21 - O2 (n)

(n) = to NES only
(r) = to ROM only
(b) = to both

The IRQ line is in fact connected, which suggests it supports that functionality. Four pins appear to be completely unconnected, however. I might go as far to say that this chip does actually support WRAM from those, but since it's not connected to any, I'd be hard pressed to figure out what does what at this point.

I emailed that Kevin Horton fellow and asked him about it, since he has such a collection of mapper information on his site. While he doesn't seem to have any more info on the chip than I do, an interesting tidbit that he passed along was that the roms on this board use standard eprom pinouts, meaning basically I should be able to take off the old ones and just put sockets in their place, which is pretty nice. From a brief check of some of the pins, that seems like it might be the case, but in the off chance the pins aren't an exact match, I still listed the pin # in the pinout above, with the possible signal in parenthesis.

Upon socketing it, he also suggested running blargg's MMC3 test roms, which sounds like a good idear, though I'll have to get some more eeproms. I still only have one.

I'll probably go back and try to double-check the pinout, since it's easy to make a mistake when you can't physically see what's connected where, with the chips in the way and all. But I'd say that's pretty accurate at this point.