Finally got the scope rigged up and looked at some waveforms. Here is what I typically saw coming out of the buffer boards:

http://s282.photobucket.com/user/pac...7354c.jpg.html

This was measured both with and without the Z board outputs connected, and at the good and bad (dead) half of the screen. It was the same under all conditions, although it changed slightly depending on the scene (menu brought up, etc.)
BTW here is what the TV picture itself looks like:

http://s282.photobucket.com/user/pac...db3a1.jpg.html

A faint, red image can still be seen with the Z board disconnected, and it too cuts off just above halfway up the screen.
I can't think of anything other than a bad PDP at this point, but I welcome any advice...