(recycled thread): a home for A20 on a 2 chip MAD-1 pcb?

[nintendo2600]

I'm recycling this thread as it' a similar topic to my previous use
of it. Details a few posts down.

Hello All,

I have a FF4J that I bought with the intent of using it to translate
it to english so I had a real FF4 cart with the english translation
on it. Turns out this pcb only has a dip 32 chip socket and I need
to find a 36 pin compatible alternative as the chips I'm using need
the A20 address line and there just isn't on on this board. FWIW I'm
using a 27c160 in 8bit mode with the address lines shifted. I've used
this method with other games and it worked perfect so I'm sticking
with it.

I already tried tying A20 to edge connector pin 46 (CPU A21) but
no dice. If anyone knows the name of the PCB I'm after and or the
titles of the games I could use I'd be very greatful.

Thanks In Adance!

[getafixx]

Well the cheapest you can use are any NBA live 96-98 game, or FIFA Soccer 96-98. With the NBA Live 96 games, some of them have 2 32 pin maskroms instead of 1 36 pin one, so make sure you can check before you buy it.

[nintendo2600]

Well the cheapest you can use are any NBA live 96-98 game, or FIFA Soccer 96-98. With the NBA Live 96 games, some of them have 2 32 pin maskroms instead of 1 36 pin one, so make sure you can check before you buy it.

Thank you very much!

[Hojo_Norem]

Funnily enough I have just done the same a few days ago, the only difference being that I used a 27c801 and a 27c4001. The PCB that FFIV lives on uses a 74LS139 to do it's address decoding. The 139 is wired in such a way that it uses A20 and A21 to map up to 3 8Mb ROMs and a SRAM. When A20 and A21 are low the first rom is selected but when either one is high (not both) one of the other possible roms are selected. On the 1A3B those two additional ROM selects are not connected. In my repro all I had to to was connect the 27c4001 to pin 5 of the 139 to get it to work.

In your case if you want to use the original pcb you could try a simple diode OR gate to tie pin 4 & 5 of the 139 together to make the single select signal you need for your 27c160. I don't know if it will work I can't see it hurting anything by trying if you have a couple of diodes and a resistor spare.

[MottZilla]

So he could just use a 74'08 to AND the first two outputs of the 74'139 so that whenever the output is not 11 (meaning sram enable?) that it will enable his EPROM instead?

[Hojo_Norem]

Not quite. The 139 is a dual 1-of-4 decoder/demultiplexer. A OR gate is what we need, either a diode based one or you could use a 74'32. (A bit of a waste for just one gate).

[MottZilla]

Why an OR gate? Maybe I'm just too tired but let me explain what I think I saw earlier.

Inputs A and B are A19 and A20. So When A and B are both 0, ROM Socket gets /OE.

When A=1 and B=1 then RAM Socket gets /OE.

So the important thing is when A and B aren't both 1 (high) then we want to enable our one and only ROM. So with the AND truth table:
http://en.wikipedia.org/wiki/AND_gate

When either A or B is 0 then /OE to the ROM is Low. In other words If ROM 1 or ROM 2 is getting a low /OE signal to enable, your ROM enables.

If you use an OR gate then when either A or B is high the result is high meaning that your ROM would never get a low /OE signal.

[Hojo_Norem]

The 139 in FFIV is wired to proved 3 8Mb ROM /CE lines and a /CE for the SRAM, however only the first ROM /CE (call this /CE0) is connected so when the code in the translation tried to access the upper 8Mb of it's ROM it gets a open bus instead. /CE1 and /CE2 on the 139 are not connected but still function in relation to the address decoding so what we need to do is OR /CE0 and /CE1, connecting the result to the 27c160's /E pin.

[MottZilla]

I understand that, but I think you have your operation mixed up. 0 OR 0 = 0. 0 OR 1 = 1; 1 OR 0 = 1. 1 OR 1 = 1. So only when both /CE signals are low which will NEVER happen from the 139, will you get a low /CE signal to the ROM chip. So if you used an OR gate you'd actually never access rom.

You want an AND gate. Because if CE0 = 0 or CE1 = 0, regardless of the other the AND operation will make the product of those 0.

0 AND 0 = 0. 0 AND 1 = 0. 1 AND 0 = 0. 1 AND 1 = 1. That means that ROM will be enabled unless BOTH CE0 and CE1 are HIGH which means that neither CE0 or CE1 is selected because they are Active Low. I don't believe I've gotten that backwards, I think you did.

[Hojo_Norem]

Dang it! If I had a had I'd eat it...

When you pointed that out I only just remembered that the /OE signals are active low. Thats what the '/' is all about but at the time I just didn't realise it and I feel silly trying to argue about it.

[MottZilla]

Well I'm just glad we have an agreement on what he could do to fix the issue with that particular board, a single AND gate for the two enable signals. I suppose if you needed 24mbit you could use two AND gates.

[nintendo2600]

I ended up getting an NHL 97 and using that. If I get bored I'll try
using the original FF4 board but for now I'm happy with transplanting
the NHL board into my FF4 shell

[nintendo2600]

On a similar issue but this time with a 2 chip version of the same board but controlled
by a MAD-1,

How would I go about placing A20 (shifted from A19 on a 27c160 in 8bit mode) if I want
to use a single chip on this 2a3M (2 chips, 8meg on the first, 4meg on the second pic below)
if it's controlled by a MAD1? I have not a clue about how the MAD-1 works or how it decodes
signals to enable chips.

The U1 and U2 of this board are both 32pin slots. I only want to use the first one but I have
to find a home for A20 on my 27c160 (the highest address line) and also make it so the
MAD-1 doesnt go trying to enable the other chip that won't be on there.



Thoughts from those who know more than me please?

[MottZilla]

Did you not get my message from earlier?

[nintendo2600]

Did you not get my message from earlier?

Yes I got your message, looking in some other threads right now too as
I think I recall a similar topic from another user being addressed before.