So, a gentleman reached out to me right before my fusion surgery asking me to have a look at his Ms. Pac-Man game. It was not showing sprites but was otherwise behaving correctly – it passed self test, and you could still see dots being eaten during attract mode, etc.

Initial instinct was something wonky with the VRAM Addresser Board, so I gave him a known good one and sacrificial socket to test with, but this produced no change. He dropped off the board and I gave him one of my spares so he would be able to at least play his game while I was in recovery.

When I was able, I took a closer look at the board and did not see anything obvious as far as lifted traces, bent pins, etc. I looked at all of the troubleshooting information and repair logs that I could find, including the Mowerman docs, but none of the solutions for the same identical problem seem to apply here.

Using the troubleshooting guide for some theory of operation info, I started working my way from the Attack RAMs all the way to the video output. I eventually stumbled upon a 74LS161 @ 1E whose outputs were not following its inputs: they were essentially floating.

I piggybacked a replacement ‘161 on top of it:

– and this made the game draw the sprites correctly, so it was a simple matter of clipping the chip, desoldering its pins, putting in a socket, and sticking in the replacement ‘161.  After that, I confirmed the sprites returned and were behaving correctly.

Board repaired, and a new(?) solution for the “missing sprites” problem.

As it turned out, the board that I let him borrow happen to have the Doc Cutlip cheat on it where you hold the one player start to become invincible and the two player start to go fast (or vice-versa?) and he liked it, so I planned to put that on his board before I gave it back to him.

Originally I was going to erase and burn his existing 2532 @ 6F that was already on his board.  But even though it erased correctly, it would not program consistently and would fail verification at different (progressive) locations. I tried another erase/burn cycle and got the same behavior so I binned it and burned one from my stock and used it instead.  (Working versions of these EPROMs are getting harder and harder to find without breaking the bank.)