new information about the breakdown

Hello All! Maybe this information will tell someone something and someone will help figure it out with this tricky breakdown? I installed something else today: the glitch is not affected by the state of the latch memory port decoders. Glitches occur in tests during verification data bus stability. The firmware is turbo-90, so when resetting the testers memory is not cleared; print usr 0 also does not affect the interference, i.e. they clean up either reset or new. But demos with Mouzon play fine... I don’t even I’ll understand which way to dig... Overloaded or not taking enough time some kind of bus that causes such a glitch? I disabled the shadow RAM circuit. However, it also works strangely: the area corresponding to fonts a-q (from 0,2,5) turns out to be damaged pages do not intersect, tested in 48 mode). Sincerely, Yura AKA d00RsFan [ZX] [D00RS] [IRCNet#Z80]