This is less Sega related (still about game consoles) and more of a general electronics question, but we have some very smart people here, so I'll go ahead. 
I have a circuit where the data bus comes from a heavily loaded 74ALS245 which is probably near the limit of it's fanout capacity. It goes to several components on a daughterboard connected through an expansion port (some RAM, various logic) and isn't strong enough to properly drive the inputs of an I/O chip during a write cycle. As a result the value the chip latches is almost always $FF instead of what's really on the bus.
Writes to all other components work fine, as they have more tolerant levels for the required high-level/low-level input voltage.
Anyway, I'm out of spare logic and free space on the PCB so I can't properly add a bus driver to fix the situation. It had been suggested that I try adding pull-up or pull-down resistors to the data bus of the I/O chip, but this didn't change anything.
Any ideas?
What would happen if I wired in another '245 in series with the data bus, where DIR=L and /OE= CPU /WR, so it would only drive the bus (in the direction of the expansion connector -> daughterboard components) during a write? I want to say it would drive the bus strongly, but I'm concerned about having the outputs of the '245 being fed back into the inputs (perhaps the Schmitt-trigger action on the inputs would prevent anything weird from happening?).
I can stand to have the I/O chip be read only, but there has to be a fix of some sort.
I have a circuit where the data bus comes from a heavily loaded 74ALS245 which is probably near the limit of it's fanout capacity. It goes to several components on a daughterboard connected through an expansion port (some RAM, various logic) and isn't strong enough to properly drive the inputs of an I/O chip during a write cycle. As a result the value the chip latches is almost always $FF instead of what's really on the bus.
Writes to all other components work fine, as they have more tolerant levels for the required high-level/low-level input voltage.
Anyway, I'm out of spare logic and free space on the PCB so I can't properly add a bus driver to fix the situation. It had been suggested that I try adding pull-up or pull-down resistors to the data bus of the I/O chip, but this didn't change anything.
Any ideas?
What would happen if I wired in another '245 in series with the data bus, where DIR=L and /OE= CPU /WR, so it would only drive the bus (in the direction of the expansion connector -> daughterboard components) during a write? I want to say it would drive the bus strongly, but I'm concerned about having the outputs of the '245 being fed back into the inputs (perhaps the Schmitt-trigger action on the inputs would prevent anything weird from happening?).
I can stand to have the I/O chip be read only, but there has to be a fix of some sort.
