diondokter[m]: > <@diondokter:matrix.org> Haha yep it's hard.
> 
> And actually in our intended design I wouldn't have run into this either.
> We were planning on fully shutting off the MCU, like cut off the power.
> But it turns out that the nRF when the voltage drops, starts misbehaving sometimes on its gpio pins where a pin can suddenly shoot up.
> Sometimes that would mean it would turn off the power for our RTC and accelerometer leaving nothing to turn on the MCU again.
> But the client had already produced 10k+ of the devices and so we needed to make a workaround. So that's why we're now using system off

it's always the case :D