MB cartridge with 1K of RAM?

[+retroclouds]

Just for fun I cracked open my Milton Bradley "Meteor Belt" cartridge.

 

To my surprise this cartridge has 2 MCM2114P20 chips on it.

I looked up the specs and found out each chip is 4096 bits (512 bytes) of static ram.

it means this cartridge has a total of 1K of RAM onboard

 

I wonder how this memory is paged in.

 

Guess that the big black chip is the 8K (G)ROM ?

Note that this game skips the TI-selection screen (1 FOR BASIC, 2 FOR ....) and starts the game right away.

 

 

EDIT: Here's the datasheet datasheet.pdf

[Willsy]

If you could post some photo's or scans that would be nice. Can you see a 74xx series chip there, too?

[+retroclouds]

If you could post some photo's or scans that would be nice. Can you see a 74xx series chip there, too?

 

Some photos added. Note the 6 SN74..... chips.

[ti99iuc]

very, very interesting....

[+acadiel]

Apparently, it looks like there are several 74LS logic chips, a GROM, two 2114 (512 byte) ROMs, and the ROM (can't find any specs on it.)

 

Look at page 8, here. It states how MBX carts work: http://pdf.textfiles.com/zines/MICROPEDIUM/mp970304.max.pdf

 

"The MBX method is a semaphore method to switch the ROM bank. You write the bank number directly to location >6FFE. In any MBX module, you will find a RAM of 256 bytes. (Why is there 1K in this one?) This is decoded to ROM location >6C00 through >6FFF. So you can read back the bank number from this RAM, because it will be written in parallel to the switching register. Nice feature, but incompatible with all other banking methods. This is the reason, other programs will not correctly read out the ROM banks of MBX modules. You have to write 00, 01, 02, or 03 to the banking register."

 

Also, here: http://tech.groups.yahoo.com/group/ti99-4a/message/3078?threaded=1&var=1&p=2

 

MBX banking seems to keep >6000-6FFF static, while banking >7000-7FFF. The Gramulator accomplishes this with 20K of RAM, which I take to mean byte 1 values of >0B-0E make bytes 4,5 the value >7000. I am not sure how MBX

banking works.

 

BTW, here are all the MBX carts dumped (HSGPL format) as well as some example code used to generate the headers for them. There aren't any other places on the 'net that have these, so I'm sharing the wealth...

 

All of the carts are here except 3154 (Terry Turtle's Adventure). Also, PHM 3149 (Space Bandits) doesn't work. I don't think it has the proper header, because it loads as an autostart cartridge.

 

I haven't tried these in MESS with the HSGPL yet, but on a real HSGPL, to get these to work, all I did was this:

 

Load the files into bank >10

reset

Go into TI BASIC

CALL GRAM

CALL MBXON

bye

mbx.zip

Edited July 10, 2011 by acadiel

[+retroclouds]

From the article, page 8:

 

The MBX method is a semaphore method to switch the ROM bank.

You write the bank-number directly to location >6FFE.

In any MBX module you will find a RAM of 256 bytes.

This is decoded ROM location >6C00 through >6FFF.

So you can read back the bank number from this RAM, because it will be written

in parallel to the switching register.

....

 

 

Don't get it. Is my math wrong here?

 

The specification says 4096-BIT STATIC RAM.

So 4096 / 8 = 512 bytes. We have 2 such chips, so that's 1K of RAM.

 

But according to the article there are 256 bytes of RAM available ?

Edited July 10, 2011 by retroclouds

[+retroclouds]

Apparently, it looks like there are several 74LS logic chips, a GROM, two 2114 (512 byte) ROMs, and the ROM (can't find any specs on it.)

 

That would read: it looks like there are several 74LS logic chips, a GROM, two 2114 (512 byte) RAMs, and the ROM (can't find any specs on it.)

 

Where is the GROM located ?

[+acadiel]

Apparently, it looks like there are several 74LS logic chips, a GROM, two 2114 (512 byte) ROMs, and the ROM (can't find any specs on it.)

 

That would read: it looks like there are several 74LS logic chips, a GROM, two 2114 (512 byte) RAMs, and the ROM (can't find any specs on it.)

 

Where is the GROM located ?

 

I'm guessing the CD5256 chip is GROM. The other five on the right are 74LS logic chips, and on the left is the ROM and the 512byte x 2 RAM.

[John Doe]

Edited February 22, 2014 by Tony Knerr

[marc.hull]

From the article, page 8:

 

The MBX method is a semaphore method to switch the ROM bank.

You write the bank-number directly to location >6FFE.

In any MBX module you will find a RAM of 256 bytes.

This is decoded ROM location >6C00 through >6FFF.

So you can read back the bank number from this RAM, because it will be written

in parallel to the switching register.

....

 

 

Don't get it. Is my math wrong here?

 

The specification says 4096-BIT STATIC RAM.

So 4096 / 8 = 512 bytes. We have 2 such chips, so that's 1K of RAM.

 

But according to the article there are 256 bytes of RAM available ?

 

No your math is good. They must have gotten a deal on the 1024 nybble SRAMS. Doesn't make sense to use them otherwise. Of course this is pure conjecture. .

 

Although it would seem a shame to waste a good 750+ bytes huh...

Edited July 11, 2011 by marc.hull

[matthew180]

74LS175 - Quad D flip-flop

74LS133 - 13-input NAND

74LS10 - Triple 3-input NAND

74LS02 - Quad 2-input NOR

MCM2114 - 4096-bit SRAM - 1024 words by 4-bit organization

CD5226 - Custom TI GROM

8341A - Custom ROM

 

The 74LS logic chips are probably doing decoding for the RAM vs ROM, and possibly a bank switch if the ROM is big enough. You would have to trace out a schematic to be sure.

 

The CD5226 is almost certainly a TI GROM, it is wire like a GROM too, with two empty locations for 2 more GROM chips. Again, a quick trace of the board to see if the GROM lines from the cartridge port go to that chip would confirm what it is.

 

The *pair* of 2114 SRAMs provide 1K *byte* of RAM.

[+RXB]

If I remember correctly all the Bank switching in the MBX carts does so in the GROM not the ROMs.

 

So if the GROMs were disassembled and we modified the GROM bank switching method there(in GPL), we could make them more like normal cartridges.

 

Just a thought to make a better attack on the problem of compatibility for these carts.