This page is a mirror of Tepples' nesdev forum mirror (URL TBD).
Last updated on Oct-18-2019 Download

Analysis of Prototype boards pics.

Analysis of Prototype boards pics.
by on (#70346)
Well for those who don't know, there exists extremely rare carts known as "Prototypes" which leaked from the video game industry when games were in their testing phase.
Not only they are interesting because they contain unpolished versions of games we know with sometimes significant differences.

However, for us they are also interesting because it can give us an idea of how people who were actually working on NES games worked back then. I know some people here don't give a damn, with the argument we are more equipped than they were, but personally I care a lot about how they actually did to make games.

After this amazing interview of Neil Baldwin we were able to learn a few things, in short that they used PCs to compile assembly, that this was horrendously slow, and that they had a card they could link to via some kind of parallel port.

Now let's analyze the pictures of boards found in prototypes cards that I was able to find on the net :

First let's start with the easier, those who are on Bootgod's database, on which we have the most detailed info about them :

http://bootgod.dyndns.org:7777/profile.php?id=2572
This prototype is made of a standard japanese CNROM board with a fami -> NES adapter. Since the game was developed in Japan, it made sense they used a board that they already used for the japanese version and used an adapter to get it on the NES.

Likely whoever developed this game had to compile their source, then remove EPROMs from their sockets, UV-light erase them, program them, and insert them back. Very tedious as opposed to assemble + run emulator on a modern PC.

Because boards with smaller ROMs are compatible with EPROMs, they could use the same boards for large runs of games and for internal testing though.

http://bootgod.dyndns.org:7777/profile.php?id=2694

This prototype is from Sunsoft, which used their own boards instead of Nintendo. Very likely the board is identical to what they used for large runs of games, it doesn't have anything special.

http://bootgod.dyndns.org:7777/profile.php?id=2708
This is a UNEPROM board (prototype specific). This time, they didn't use a fami -> NES converter, but directly a NES board, even though the game was developed by Capcom which is a Japanese company.

Again other than it's compatible with EPROMs, this board is nothing very different from a regular UNROM board. Maybe it also accepts 256 KB EPROMs (UOROM) ? It would make sense as the pinout is the same. Unfortunately the picture isn't detailed engough so we can't tell.

http://bootgod.dyndns.org:7777/profile.php?id=3795
This uses a SKEPROM board (prototype specific). A famicom board for a famicom prototype. Note that since this board doesn't have a circular hole, it wouldn't fit a NES case with a fami -> NES adapter.

Again it's almost like a regular SKROM board, with support of standard pinouts. There is circuitery for battery backup and PRG RAM, but it wasn't used in this game.

http://bootgod.dyndns.org:7777/profile.php?id=2950
This uses a TKEPROM board (prototype specific). It's the most interesting board around. This time it's definitely NOT close to a standard TKROM board.
It has silkscreens, which shows instructions how to use solder pads for different PRG and CHR sizes. Whoever uses this board is supposed to use multiple 128 kB EPROMs to make up the total size, which is why there is 4 slots for PRG ROMs and 2 for CHR ROMs. A 74HC139 adress decoder is used to control the enable lines of those chips.

Again there is circuitery for RAM and battery backup, and this time it's used.

On the back of the board, there is two wires connectings to PRG ROM's A17 and A18 lines. Apprently they modified the board so that they could put a 512 kB EPROM instead of using 4 128 kB EPROMs. Nevertheless the game that is in uses 128 kB, and the EPROM that's on the board is a 27C1000 which is 128kB.

On the back of the board, there is no silkscreen, but there is more instructions about how to use solder pads written in negative on the solder mask. Apparently it says "PRG : ??? Type / JEDEC TYPE". and "CHR : ??? Type / JEDEC Type". So maybe the other type is the same as the regular TKROM boards ?

Because they have all those solder pads, it makes sense they would use this board not only for TKROM games but for all MMC3 games.

Now there is also pics of a Mega Man 3 and Mega Man 4 prototypes form this site.

It shows that they used multiple 128 kB EPROMs to get to the desired PRG/CHR ROM sizes, but also that Mega Man 4, which uses CHR-RAM, can also fit a TKEPROM board. Likely they had a jumper configuration for that or cheated with wires. Mega Man 4 have PRG RAM installed, even if the game didn't used it (at least the final version didn't).

http://kevtris.org/mappers/mmc1/NES_SKEPROM.jpg
This SKROM board, from Kevin Horton, is a MMC1 development board. Surprisingly, it is a quite different from the japanese SKROM board. For some reason it has support for 4 lockup chips (USA, England/Oceania, Contiental Europe, Asia). It has provisions for PRG RAM and battery backup, but this wasn't used here (the final game is SLROM). So chances are that this generic board were used not only for SKROM games, but for many MMC1 games.

The wire connects CPU R/W from the cart edge to the MMC1. No idea why they added it (did they cut the line accidentally ?), but it shows developpers sometimes were dealing directly with wires.

Eventually, this thread is a goldmine for prototype pics.
In addition to some non-NES stuff, and to what was already shown above, we can see :

http://i113.photobucket.com/albums/n223/mrmark0673/nintendo/100_0270.jpg
A Mega Man 2 proto which uses a fami -> NES adapter. The board is HVC-SNROM-TEST-01 ! It uses a 3-mil wide CHR-RAM chip instead of most commonly seen 6-mil wide ones, a surface mounted (!) PRG-RAM chip (Mega Man 2 doesn't use it), and two 128 kB EPROMs, with a surface mounder adress decoder (likely 74HC139). Has circuitery for battery backup, unused here.

http://i113.photobucket.com/albums/n223/mrmark0673/nintendo/100_0990.jpg
This ones uses a normal NES-TSROM board, modified with wires to accept EPROMs. Which leads me to suspect it's fake, but since the guy has real prototype why fake one ? Or maybe someone else made a fake and sold it as a prototype for an amazing price ? We'll never know...
A few others prototypes on the theads uses apparently modified standard boards as well.

http://i113.photobucket.com/albums/n223/mrmark0673/nintendo/100_1240.jpg
The right part of the image is (if the label on the cart is true...) a SKWEPROM board, which apparently allows for 2x128 KB PRG ROM (Bases Loaded II effectively uses 256 kB PRG ROM).

http://i113.photobucket.com/albums/n223/mrmark0673/Protos/100_1617.jpg
This appears to be an AOEPROM boards. With 2x128 KB PRG ROMs again, the two chips on the left are likely a 74HC161 and a 74HC?? which is used both as an adress decoder and a bus-conflicts preventer.
You can bet Rare used almost exclusively this board when developing games, with their insane fidelity to mapper 7.

http://i113.photobucket.com/albums/n223/mrmark0673/Protos/100_1625-1.jpg
Another one with a fami -> NES converter. Apparenty it's a SMEPROM board, it's functionally like SGROM but with 2x128 KB of PRG ROM again. The address decoder is made with resistor/diode/transistor logic instead of using a 74xx chip !

Wow this was long, and hopefully interesting.

by on (#70363)
I would love to see some of those test boards reproduced so that we don't have to desolder several different boards of the same mapper to test our eproms. Especially something as common as MMC3.
Re: Analysis of Prototype boards pics.
by on (#70367)
Bregalad wrote:
http://bootgod.dyndns.org:7777/profile.php?id=2950
This uses a TKEPROM board (prototype specific). It's the most interesting board around. This time it's definitely NOT close to a standard TKROM board.


I actually saw these on ebay once, it must have been 12 years ago I think, I didn't know what it was, and didn't have much money anyways. But it was a decent-sized lot of maybe 8 or 16 unused TKEPROM boards.

marvelus10 wrote:
I would love to see some of those test boards reproduced so that we don't have to desolder several different boards of the same mapper to test our eproms. Especially something as common as MMC3.


How long you think it would take to use 100 of them? I wouldn't know, but it would be interesting to have a board like that.

by on (#70550)
Quote:
I actually saw these on ebay once, it must have been 12 years ago I think, I didn't know what it was, and didn't have much money anyways. But it was a decent-sized lot of maybe 8 or 16 unused TKEPROM boards.

If they came with MMC3's soldered on then wow ! Otherwise, they're totally useless.


Quote:
I would love to see some of those test boards reproduced so that we don't have to desolder several different boards of the same mapper to test our eproms. Especially something as common as MMC3.

Well once you did the rewiring and added socket it basically becomes the same, except that you can't get multiple slots for 128 KB EPROMs.

I'd LOVE to see a MMC5 development board. If they used all slots of 128 KB EPROMs, and supports up to 1 MB for both PRG and CHR, that would make 16 EPROM slots on the same board !! Definitely would never fit a NES case.

Some thing weird is that apparently they used surface mounted componants for their SNROM-TEST board, but they used all DIP componants for their large production boards.
This is the total opposite of what you would do by common sense. Since surface mounted takes less pace and is so easy to deal with robots who does the assmembly, you'd want to use that for large productions. When you have to deal with manual soldering etc... it's good to avoid surface mounted stuff.

Also the fact developers didn't hesitate to deal with wires directly explain how some mapper's variants were created. Typically, you could imiagine that they had the idea to use one CHR-ROM and one CHR-RAM chip on a TKEPROM board, and that this became basically TQROM.

Also I wonder how did developers to debug their code. None of those boards appears to have an extra connector or anything that would connect to a debugging device that could trace the adresses of code executed, etc...

by on (#70551)
Bregalad wrote:
Also I wonder how did developers to debug their code. None of those boards appears to have an extra connector or anything that would connect to a debugging device that could trace the adresses of code executed, etc...

Probably a 6502 in-circuit emulator.

by on (#70608)
Bregalad wrote:
Quote:
I actually saw these on ebay once, it must have been 12 years ago I think, I didn't know what it was, and didn't have much money anyways. But it was a decent-sized lot of maybe 8 or 16 unused TKEPROM boards.

If they came with MMC3's soldered on then wow ! Otherwise, they're totally useless.


They were totally empty, so I probably also thought they were useless. But now I know that removing a donor MMC3 would be easy enough with a heat gun, you're basically scrapping a board by removing an MMC3 anyways, so you could really toast it. I told some NES collectors at the time too, that I thought that could lead to fake prototypes if there were a lot more of those boards around. But those could have been the only ones, for all I know.

Quote:
Some thing weird is that apparently they used surface mounted componants for their SNROM-TEST board, but they used all DIP componants for their large production boards.


Yeah, that is weird. I'm guessing that DIP package was just more common or cheaper for SRAMs. It also seems that they stuck with DIP packaging with their SNES cart SRAMs as well.

Quote:
Also I wonder how did developers to debug their code. None of those boards appears to have an extra connector or anything that would connect to a debugging device that could trace the adresses of code executed, etc...


I'm sure they would have had something like an EPROM emulator, if not a rewired 6502 emulator. A ROM emulator would be relatively easy to build, and a natural extension to using EPROMs. Often you see those designed to be chained together, so with 2 of them one could emulate PRG and CHR ROM (still using 1 PC port). And lots of arcade boards of course had rows of EPROM chips.

by on (#70638)
Since I'm back, I just want to say this is awesome! I love seeing prototypes, prototype hardware, and everything of this nature. More boards and hardware would be awesome, especially to see how they worked and were sorted out and such. Keep 'em coming? :)