Game: Ms Pac Man
Symptoms: No moving characters, bottom 1/8th of screen "moved up", coin counter clicks on reset.
Got this Ms Pac in a Pac Man cocktail cab. Cab is actually quite nice and as far as I can tell all original. Will clean up nice but first fixing the bad stuff..
The board has been hacked to run Ms Pac using this mod with 1 small twist, the 5E. 5F 2532s were split into 4 2516s..
The mod work was a bit sketchy so I pulled it all, validated it, and put in new sockets and mod wire jumpers. Put it all back together to the same problems. Which I kinda expected so no loss..
Problem 1, no characters.
This has been solved many times over and is typically one of the 2 161s at 1E and 2E. Probing through the pins on the chip I noted that the carry line out of 1E wasn't ever changing. Bad output or input. Output more likely so I opted to change 1E first.
That fixed the no sprites issue.
Problem 2, bottom 1/8th of screen "moved up"
Here's a pic:
This took me a bit to get. I started with thinking that there was some sort of addressing problem in the high order bits and that shifted the bottom of the image to the left (when arranged horizontal). Scanning from left to right the high order memory would be on the right was the thought. Spent a lot of time poking around in the addressing and control signal logic but nothing screamed bad. Time to try a different approach.
I recently bought another Ms Pac board (OEM setup) that was sold as not working. Figured I'd stuff it in and see if I could get it going and then have 1 to compare. The 2nd turned out to just be a bad ribbon cable on the processor board. If I twisted it in a pretzel it would work so I ran with that. This was useful. Something that becomes more important later... I noticed that swapping between the boards required me to adjust the the Hor Hold pot on the monitor. It wasn't always necessary, roughly 50% of the time though.
I swapped the 2 daughter boards, no change. I kept going back and forth with control logic thinking something had an oddball error with a specific combination. I wasn't getting anywhere. Many threads on pac boards, all of them kept pushing back to the VRam addressing boards for stuff like this. However, I knew it was good. So I just started reviewing the schematics moving from the color signals back up the tree. The many Gin-n-Tonics made this quite enjoyable actually. Anyway. I got back to the VRam addresser again. This time it occurred to me that the H and V clocks were feeding this. So I decided to give the clocks another go over this time a bit more closely. So I pulled out the scope and started looking at them. Cascaded counters are relatively easy to validate with a scope as everything just doubles or halves. As I walked up the H signals I noticed that they halved freq but also that all 4 signals on the upper 161 at 3S were not symmetric. Looking at the schematics, the lower 2 lines32/64H should be since the A/B inputs were tied low, but the C input was tied high, so this would change the 128/256H lines. Ok so that would suggest that 3S was bad. Clearly not bad enough to cause the monitor to fail to sync. In fact if the H clocks were off this would explain the need to readjust the Hor Hold pot. So I replaced the 3S 161 and were back in business.
Problem 3, hitting the coin counter on reset.
When I was running the OEM board, the counter worked fine - did not count on reset, but did when coin added. After spending so much time working the board's control signals I figured it was going to be a long haul. Then it occurred to me that maybe it's a glitch in the "Hack". So I transferred all the OEM Ms Pac bits to the board with the mod and everything functioned as normal. So the conclusion I draw from that is the Ms Pack Hack comes with the coin counter activation on reset baked in. Glad I started there as I can't image how long I would have circled the drain before just giving up.
Thanks for listening.