Very exciting!

There is some progress.
All power sources work properly (+ 5V, + 3.3V, -5V).
But I found a few minor mistakes.
First of all L-connector has a shifted connector EX2 by one pin to the right. I will have to correct and order again
Pads for SMD quartz generators are for version 3.3V instead of 5V, so they can not be used. On this PCB may be used only generators in the DIP housing.
And at last, was supposed to be a socket for the CR2032 battery, and by mistake it is for CR1632. This is the smallest problem.

My garage is so cold that even the camera on the smartphone refused to take pictures

I'll manage somehow. I have to.
I have yet to mount the CPLD and move to my home.
I'll start to write tomorrow a firmware.

I have not seen this before. It's a pity, because it's a bit too late for such modifications, but ...
Has anyone tried to insert capacitor of value 47 or 33pF as C7 (the R12 resistor should remain in place) ?
How then will go at 10 MHz ?

short circuit R12... removing C7.

I do it always.

Not a big problem, ghost errors only happen on not installed memory.

This is because no pull ups on the data bus. Then totally undefined which data can be read from unused memory (and I/O) space. Usually FFh, but not guaranted. It mainly depened about the CPU. With CMOS Z80 are usually come the floating bus problem (ghost bad segments). But in some machines still exist the problem with the original NMOS CPU.
If the floating bus problem present then some games can hang up. Because using IM2 mode with non complete jump table (assume FFh from the non definied data bus).

The right solution add pull-up resistors to the data bus. As I wrote it here: https://enterpriseforever.com/hardware/6mhz-and-more/msg37473/#msg37473

There are signs of life
Part of the circuits are already working.
There are correct clock signals for the microcontroller, EXDOS (after reset 8 MHz) and a control signal of 1 MHz.
Somewhere is a short circuit between the LED control pins and the clock output for the system.
The CPLD assembly with such a density of legs is a real nightmare (especially stiff with cold fingers).
Each one pin is a new gray hair
Unfortunately, I have to wait with the amendments until the frost will let go.

Taking advantage of the opportunity that the frost for a moment has eased (in the garage was the temperature close to the top shelf in the refrigerator ) I removed weak connections and short circuits.
All signals from the CPLD are generated correctly
Unfortunately at the first connection, I mistakenly connected the power ground and damaged the buffer (this is here to protect the rest of the components and as you can see works). I need to replace it
Finally, I can start to fun with the firmware.