If this was the cause of the previous problem, there may still be some remaining flux under the SK1 connector causing some conductivity and pulling the voltage of the pin 15 too low and cause also voltage rise on pin 17, due to possible leakage current.
I've desoldered SK1 (which was super difficult to do, but I finally managed), cleaned everything and resoldered it.
Unfortunately it makes no difference.
You may "gently" try to force the pin 15 to higher voltage by following way:
- Solder (about) 10k resistor in parallel with diode D8 on the bottom board.
- Attach the top board and check the voltage on pin 15 of SK1.
If the voltage increases, this may indicate there is still some excess flux on the board.
I started with a 10k resistor, which slightly increased the voltage, but I had to go all the way down to a 180 Ohm resistor to get 3.2V on pin 17.
Unfortunately there's still nothing on the LCD, although I've noticed one or two horizontal lines with some black pixels in one occasion.
They have disappeared though, and the LCD shows nothing again.
Maybe also on this board there is still some flux, but the main cause of the problem may be on the top board.
I've cleaned everything very carefully, except underneath the IC sockets, the relay and the SK7 connector.
I'm afraid to damage these parts trying to desolder them for cleaning, so I've measured the resistance between adjanced pins with a multimeter instead to find short-cutting pins, without finding anything.