TI 99/4a Cartridge layout

[sjt]

I'm a bit confused about the TI Cartridge pinout.

 

Looking at the docs I've seen showing the cartridge pinout, they don't seem to line up with a cartridge.

 

Cartridge edge connector

The Top of the boards are PINS 36 to 2 (left to right)

The bottom the boards are PINS 1 to 35 (left to right)

 

Looking at a cartridge like PARSEC that has a ROM on it. My assumption is pin 1 of the rom would be A7 followed by A6,A5,A4...A0 then D0,D1,D2, then Gnd.

 

But those pins don't line up the the docs on the Cartridge edge connector.

 

 

As a test I wanted to burn a 2764 with the diagnostic cartridge rom (assuming it was an 8k rom) and put it into one of these boards.

 

But in testing the PARSEC and MUNCHMAN boards... it doesn't make sense.

 

Can someone fill me in on what I;m missing here?

 

Did TI use non standard roms?

 

 

After I posted this , I looked at it again, It occurred to me..what if they flipped it around on both address and data lines Msb & Lsb

 

It appears that is exactly what they did, now it makes sense.

Edited April 1, 2014 by sjt

[Stuart]

Correct. TI reversed MSb<>LSb compared to the rest of the industry.

[+mizapf]

Which was somewhat silly; they had to add address lines AMA, AMB, AMC in the Peripheral Expansion Box instead of naming them A16, A17, A18.

[+Ksarul]

Correct on the buses--but they also used some nonstandard memory chips in a lot of their cartridges (GROMs). The regular ROMs were also in a lot of them, so you had boards with the appropriate set of pads set up to install a ROM. You also have the option of getting one of the after-market cartridge boards of the 64K or 512K bank-switched variety and using that as your test board instead. I should probably build one with a ZIF socket on it for my own cartridge experiments. . .

[matthew180]

Keep in mind that it is only the numbering that is backwards, and not the bit place values. The right-most bit, no matter if you call it bit-15 or bit-0, is still the least-significant bit:

 

32768|16384|8192|4096|2048|1024|512|256|128|64|32|16| 8| 4| 2| 1| bit value
-----+-----+----+----+----+----+---+---+---+--+--+--+--+--+--+--|
  15 |  14 | 13 | 12 | 11 | 10 | 9 | 8 | 7 | 6| 5| 4| 3| 2| 1| 0| industry numbering
-----+-----+----+----+----+----+---+---+---+--+--+--+--+--+--+--|
   0 |   1 |  2 |  3 |  4 |  5 | 6 | 7 | 8 | 9|10|11|12|13|14|15| early TI numbering

[sjt]

Thanks for confirming this.

I put a few sockets on the board to play with it.

Of course, the first thing I found was you can't close the case , well not unless you remove a piece of it to accommodate the eprom socket & chip height.

At the moment I can't seem to find a 2764 around, can I use a larger eprom? will it work if the code is only 8k. Of course I would need to add vcc and address A12 or in TI cartridge A3 pin 24

[matthew180]

I might still have some 2764s laying around. I'll check. You can socket the EPROM, but you have to move it to the part of the cartridge that is under the "bump" in the case. Yes, you can use larger EPROMs, just tie the unused address pins high or low. You can also set up bank-switching to utilize more of a larger EPROM. Maybe Jon, Stuart, or Ksarul will chime in with more details on that.

Edited April 2, 2014 by matthew180

[sjt]

Thanks...I will try one of the larger eproms. I'm looking forward to getting a bank switched board to play with.

I ordered a few 2764's with some other things.

[+acadiel]

You might want to keep up with this topic:

 

http://atariage.com/forums/topic/193163-512k-cartridge-status/

 

Bottom line summary of where we are at with cartridges:

 

2007 or so, Carlos Randolph laid out an 8K ROM Board (v1) from a ROMOX board in ExpressPCB.

2009 or so, I took that layout and made a 16K ROM board (v2) with a 74LS379 (inverted) to do the bank switching. I designed.

2010 or so, I took that same layout and made a 64K ROM Board (v3) (you could also use 8K, 16K, and 32K ROMs with it.) I designed. The Gerber file was released for this.

 

The 64K Boards are the last ones we've done any batches of (we being me, Matthew, and other people like Arcadeshopper). I have no concerns with others duplicating the work as long as my name or website stays on it (i.e. I get credit).

 

Others that have been done but not mass released:

• V4B - 128K x 4 board - a prototype 128K ROM board that used all the inputs/outputs on the 74LS379 (27C010 EPROM). Ksarul helped design. Had space for two switches (i.e. four ROM banks of 128K).
• V5 never left the ground. It was going to be a GAL bank switched 512K cart with a 49F040 512K EEPROM. One rough draft exists in ExpressPCB.
• V6H - Uses a 49F040 512K ROM, a 74LS378 to bank switch, and an Atmega 1284P for GROM simulation. I believe this is close to the current iteration of the one that we are going to produce.
• V7D - Did get designed and printed; this used a GAL to replace the 74LS378/9, the 49F040 512K EEPROM, and the Atmega 1284 GROM simulator. We had issues with the GAL and decided to go with a 74LS378. Ksarul can elaborate why we moved from the V7 to the V8; I believe it was due to a different pinout on the ATMega 1284P.
• V8J - A 512K ROM board (49F040 - PLCC EEPROM) with an optional Atmega 1284 controller that can simulate the GROM in conjunction with the ROM with the GAL 74LS378/8 type bank switching. Shelved due to issues with the GAL.
• V9a - A 512K ROM board that uses a 74LS378 to bank the whole 512K (2764/128/256/512, 27C010/020/040) in 8K segments. An evolution of the V4B board with the 74LS378. Ksarul helped design.

Ksarul can help correct anything that I missed.

[+Ksarul]

On the V7 and V8 boards, both of them were GAL variants, with the V7 using Tursi's original spec for the Atmel pin-out, and the V8 using an updated pin-out. A small number of an earlier iteration of the V6 boards using the original Atmel pin out were also produced (20, all used as test articles). When the GAL on the V8 boards (no V7's were ever produced, but the layout is complete) exhibited issues, I went back to the V6 board and updated it to the new Atmel pinout and made a few other board optimizations we'd incorporated into the V8 boards. This generated the V6H boards (I made 50 of those, and I've used about half of them as test articles). Of the V4 boards, Acadiel made 4 of them and I made another 23, of the V9 boards, I made 52. There are several hundred of the V3 boards in circulation.

 

I also incorporated a few lessons learned into the V9 boards (and in my current iteration of the V6 board), as I've shaved a very small amount off of the cartridge connector to make it a little easier to insert and added a bit of an angle to the connector corners to make it seat more easily.

[Asmusr]

I have no concerns with others duplicating the work as long as my name or website stays on it (i.e. I get credit).

 

What if someone wanted to produce and sell cartridges based on your board layout, would that be legal and free of charge?

[+acadiel]

 

What if someone wanted to produce and sell cartridges based on your board layout, would that be legal and free of charge?

Sure, Matthew H and Arcadeshopper have already done so. :-)

[+Ksarul]

Acadiel's already answered for the V3 boards, and I put the V4 Gerbers up here on AtariAge, so those are available for folks to make as well.

[+acadiel]

Acadiel's already answered for the V3 boards, and I put the V4 Gerbers up here on AtariAge, so those are available for folks to make as well.

Can we put up the V9 Gerbers since they handle the whole 512K without the two switches? Folks will probably want to make those over the V4's.... Thoughts?

[+Ksarul]

I can do that once some of them have been thoroughly tested, Acadiel. I haven't got anything larger than an 8K ROM image to work with right now. . .I'm good at designing things, programming, not so much. . .

[Asmusr]

I can do that once some of them have been thoroughly tested, Acadiel. I haven't got anything larger than an 8K ROM image to work with right now. . .I'm good at designing things, programming, not so much. . .

 

I can make a small Java program that will invert any ROM image. Applied to the 128K image with my three games it should make a good test case.

[+Ksarul]

That would come in useful, Rasmus--and for much more than testing. It would make builds for both Inverted and Non-inverted cartridges a snap, as you'd just have to run something built one way through the Java app to turn it into the other. . .

[Asmusr]

That would come in useful, Rasmus--and for much more than testing. It would make builds for both Inverted and Non-inverted cartridges a snap, as you'd just have to run something built one way through the Java app to turn it into the other. . .

 

This should do it. Just unzip the jar file and double click to run it. I haven't been able to test it, of course...

bankinvert.zip

[+Ksarul]

Many thanks! I'll have to set up my burner and do some tests later this weekend. . .or next, if the family doesn't leave me time this weekend.

 

I just got a Powertran Cortex and a Marinchip S9900 in the mail. This is one really happy retrocomputing weekend for me, as I've been trying to find a Marinchip since 1987 or so, when I first found out there was actually an S-100 computer using the TMS9900. . .