I am trying to fix a Bally -17 MPU board, and it's becoming a real pain in my backside. It had previously had work done to clean up battery corrosion, but according to the owner, it had been working fine and just quit. (Though I'm told it quit around the time of a lightning storm that damaged some other electronics.)
When I got it, symptoms were that the on-board LED was locked on. Following standard diagnostic procedures, I determined that the original U6 was bad. Had an internal short to ground on the data line. I made the necessary changes to set the board up for 2732 EPROMS, and popped Leon's Test Rom (V4) in U6 and pulled all other socketed chips. Traced down bad U15 & U16 and replaced them.
CPU now boots, and I get the expected blinking from LED on U9 Pin 15.
Added in U10 and U11. On-board LED now blinks as well. Checking outputs of U11, I find several that are stuck high. Checking outputs of U10, none of them are changing. Swapped U10 and U11. The chip that gave no outputs in U10 socket now gives results similar to the original U11 chip in socket U11. New U10 is behaving the same as previous U10. (Should mention that CPU and these 6820s are from a known working board.)
So it's like something isn't allowing U10 to be selected. I've checked all the data lines D0-D8 with a logic analyzer at U9, U10, and U11, and get the same pattern at all locations.
And then there's the other weird symptom... If the board flexes at all, the LED will either go back to locked on, or to quick pulses at about 1 second intervals. Whereas the normal Leon test pattern is a steady on-off-on-off at roughly one second intervals, this pulsing is a quick dim flash followed by a about a half second off period, then another quick flash, etc.
My gut says there's either a bad solder joint or a broken trace somewhere, but I'm not seeing anything jumping out at me. Any ideas where to look?