The more I analyze the discussion on this thread, the more I realize what I am doing to contribute to the problem.
First, I am using layers with at least 50% of my leads.
Second, I am playing 4 note chords at least 75% of the time.
Thirdly, I am using Chordal Harmony at least 40% of the time.
I think, and this is just speculation on my part, the reason this problem exists in the PSR9K and not on the PSR2k is because the chip on the 2k may be faster. Maybe the PSR9000 is too slow to handle this many demands at once.
By the way the X1 occasionally "dropped out" under my style of playing, perhaps this contributes to the chordal harmony "bug" that only me seem to have experienced on this machine. Who knows?
Thanks guys for your contributions to this problem, which does not seem to have a solution at this time.
Mario

ps. Can't wait to try the Tyros under similar conditions.