Problem

Unstable sprites.

Diagnosis

Examining the PCB, it appeared a component had been replaced and socketed:

The PC0900J is a sprite generator used by various Taito games of the time. IC6 is used to derive a /CE signal from a 13MHz clock for the 4 SRAM ICs connected to the PC0900J:

Notice how IC6 is specified as a 74F00 on both the PCB and schematic. The 74F series is a Faster variant of the 74LS series, with shorter propagation delays. Can we get away with using a 74LS00 here? The answer is no. Why not?

Using a logic analyzer (or an oscilloscope), we can view the input and output signals of the timing circuit formed from IC6. This is what we should see:

Nothing terribly exciting; just a duty cycle reduction of the 13MHz clock input from 50% to 30%.

What if we use a 74LS00 instead?

Suffice to say, the sprite hardware does not function correctly with this output signal.

Fix

Replace 74LS00 at location IC6 with 74F00.

(Insert shaky iPhone video of Rastan working here.)

Philip Bennett