SOLVED. Crazy wiring hacks found (of course).
Goal: I’m trying to get a William 1982 system 7 Cosmic Gunfight up and running. The switch matrix has 4 errors remaining in it. I’m starting to troubleshoot switches 9 and 10.
I have two playfields. I’ll call them the “good” and “bad” playfield.
Symptom: When I plug in the good playfield, switches 9 and 10 work fine. When I plug in the bad playfield, switches 9 and 10 report 17 an 18 respectively. Also note that 17 is "not used", and 18 works as expected on its own. Also note that 14 and 16 are reversed, but I'm not troubleshooting that yet.
Conclusion: If a good board works, then the problem is likely in the bad playfield.
Isolation-1: I've buzzed out the switch to the end of the molex, and that seems to work fine for both physical switches 9 and 10 to the corresponding pins on the molex. I haven't checked for a short between various components yet.
Isolation-2: However, when I pull 2J2 off the MPU and check for continuity from the switch to the end of 2J2, I'm NOT good with 9 or 10. This suggests to me that the molex pins for 9 and 10 (red/white and brown/white). When I do this with the good playfield attached, of course I get good continuity. (There's a small chance I'm just not holding the probe in the right place, but I doubt that this would happen to both wires.
Inspection-1: Both 3x5 sets of molex wires look equivalent in both playfields
Inspection-2: Both switch 9's and 10's seem to be wired up the same way
Looking forward to learning how to better track down a short or hack in the wiring.
I'll attach the switch matrix in the next post...
thanks for all help!