Start a new topic

Live and neutral crossed

Just an FYI -

I've just installed my first 5 Sonoff POW's.  3 worked wonderfully and connected.  The other 2 connected intermittently, and were often in the "double-blue-flash" state.

Solution: in these 2 circuits, the neutral was actually the live and the live was actually the neutral.  When I switch the connections on the POW, the suddenly connected stably!

The symptoms were that the POW was connecting not connecting to the Wifi (single-blue-flash), or if it connected, it wasn't managing to connect to the Sonoff server (double-blue-flash).  I don't know how or why or what in the crossed live/neutral caused this.


This has my thinking on alternating currents completely confused. I thought the only difference between L and N was a phase shift. Maybe the PCB is using the phase difference to time the clock or something I haven’t though of. I’ve realised I don’t actually know how AC to DC converters work. I had mentally designed just two paths of one way diodes between the two inputs with a capacitor to reduce fluctuations & a transformer to convert down the voltage, but maybe it is more complicated than that…

Completely agree with everything you wrote, which is why this surprised me enough to post it.  My guess is that it has something to do with AC making it into the circuit board.  There was another phenomenon too ... the multi-led floodlights that I connected were kind of glowing all the time until I switched the live and neutral, so there was definitely some kind of leakage that there should not have been.

I just looked up AC to DC conversion. I now realise I did learn about bridge rectifiers at school many years ago. I’d be really surprised if they used something else and can’t see how reversing the input could make any difference. That only leaves one thing: the circuit for measuring current flow. Maybe they just soldered the current measuring device upside down and that was confusing the ESP chip so it kept crashing. The current measuring PCB I have though allow both forward and backward current, giving 2.5V for 0 current, more for +, less for -. I don’t think the ESP has analogue read ability, so I guess there is an ADC. Maybe the analogue V exceeds the capabilities of the ADC or something weird…or the firmware on the ESP can’t handle unexpected values. But then, it sounds like you were getting problems when nothing was connected to them? Now I’m really confused. It’s a pity the warranty is voided by opening things or we could open them and check nothing is soldered upside down when we get weird stuff. There was a batch of RF Bridge where a Switch was off instead of on…there is also a PCB with mislabelled pins ? 4Ch
Login or Signup to post a comment