Pwm on time after calibration: ~ 200
I tried to adjust this value to see the effects on the problem: it is not related.
In my opinion, the problem probably comes from the low voltage circuit, maybe my induction or the LM2595 (Accustomed to 7805, I discover this new circuit).
I did continuity tests, the card does not seem to have cut circuits.
I would do a test, to measure the current drawn on the power supply.
Quick question: I bought this card to be able to test the detection by PIR. Not using the LEDs on the clock, could I have implemented this function on the Rev4?