But the problem occurs when CH_PD is actively pulled down by the push-pull output of the 7555.
As I will hopefully succeed in resolving this issue, I will make a drawing in KiCad for PCB production, anyway. As soon as I have the design drawing ready I'll post it here. I guess it will be much clearer then.
One thing I can think of, I killed the led (removed it by desoldering), maybe it now short circuits and this is the current that flows through the resistor (still there). I'll have a look again, with the magnifiying glass....
I will also check the flash CS. It that remains high, the flash won't sleep either, indeed.