I think you mixed up U14 and U13. try this one:

(1) connect a wire from one column drive to all 8 inputs of J208/209
(2) measure the 8 outputs of the LM339, they should be pulsing
(3) same signal should enter the inputs at U13, measure them too.

CPU_zaz.png

Remove all connectors from J205-J209 and J212, power up the CPU-board.
Do you measure 5 Volt at all red marked pins of U13 ? (pin 2,4,6,8,11,13,15,17)

do I understand correct:
-With CPU in the machine and no switch inputs connected (on J205-J212) still give row short 1+2+4 , also when U13 is out of socket ?
-There is continuity from D0-D7 U13 to D0-D7 U12, measured on IC pins
-The pull-up resistors are connected and working because you measured 5Volt at U13 inputs 1-8.

If machine detects a 'ground row short' at row 1,2,4 and the inputs at U13 are Hi-level, all data lines are ok, all (8) direct switch are ok, then my last idea would be to try a new 74LS240 for U13.

Is there continuity between pin 1 and pin 19 of U13 ? this IC has 2 separate enables (but that wouldn't explain why row 3 has no problem).

applause for all !
Glad you found it because this was the last thing I could come up with.

Sascha

edit:

Yes, I see what you mean:
U13_1-19.gif

I don't know, row 2 and 3 look identical to me in the tech-chart with some mechanical and optical switches combined.

No, not really. I've no schematics / boards / machines around to test or check. Maybe after I fully understand WPC I'll dive into other systems but that is still far away.