Some quick tests, and I've managed to find the cause of the issues I was facing.

See, this is a pretty much normal PHI2

And this is PHI2, when I added the three boards.

So awful!

Of course, nothing running on PHI2 worked. Not CIA#1, not CIA#2,  probably not the SID, not even my 'thing'.

This is caused by PHI2 being fed into too many inputs, up to 18. Old NMOS technology can't pull up fast and hard enough, so the rising slope of PHI2 was slowed too much.

Adding a single non-inverting buffer to pass PHI2 through has been enough to fix the problem, and now I have all 3 boards working.

Next task... prepare an intensive test suit for the Timers. There are just toooo many quircks about them!

Cheers!