Perhaps I'm coming a bit late with my comment.

Why didn't you exploit the 138 better ? When you only set B (P1.7) to 1, your RAMs are inactive, you address something connected to Y2 and Y3. If you add control of C input, you have access to 6 external units, besides RAMs. If you invert one of them and send to Enable Display, you access LCD just like memory. Use 138's input /E1 or /E2 and attach WR-AND-RD signal (through HC08), and then you write easily to LCD, latches and read and write to RAMs. Also it's faster and simpler to write 8 bits to LCD instead of 2x4