Working from memory - U9 is the RAM chip (??). I think the only thing that could cause continual failure is excessive power to it so I'd check the 5v. It's CMOS though so fairly robust - are you sure that simply removing the board and replacing it doesn't also make it start working again, even if temporarily?

If it fails when the game powers up I'd check the capacitors and measure the +5v ripple. Sounds like a power spike is taking out the RAM. On my FT I had +5v ripple at about 250mV until I replaced the capacitors.

Pete - you can measure ripple by switching your meter to AC and measuring the 5v. I replaced C5 on the power-driver board (main 5v smoothing capacitor). The only capacitors on the sound board are in the amplifier section - the digital section only has decoupling capacitors.

It's a strange one.

Sensitive to ESD most certainly, but tolerant of varying supply voltages. It's robust in-circuit is what I meant.

I can vouch for the fact that the boards are almost impossible to find. It's odd because all that feeds the RAM chip is Vcc (power), the address bus, the data bus, Chip select and r/w enable lines. These themselves are all buffered. Unless the 5v is crazy I can't see how anything else could cause the RAM to fail unless it's being incorrectly identified as failed.

Well have you measured the 5v ripple yet on the FT to see if it's excessive? I had very strage sound problems on my FT (playing the wrong sounds, sometimes no sound) which, at great length, I traced to the buffer on the Fliptronics board.

If that's the case simply pressing the reset button behind the coin door with the game powered up would clear it. That causes the game CPU and the sound CPU to restart without any voltage spikes as it's already powered up.