Are you sure that "seeing the extra outputs" eliminated by just cutting the DC is enough of an answer?

There were a lot of ideas thrown out regarding why this may be an issue when cutting the AC power. Wouldn't it seem prudent to investigate and know exactly that was going on there? I'm sure that many readers that have been following this thread would be curious as to the root cause as well. So my take is that is premature to close this thread.

I do not know the acual nature of your current project, whether it be a hobby project, school project or actual commercial product development, but a rigorous exercising of the engineering discipline really does direct that you should investigate to root cause. Certainly there are design decisions that would have a big impact on the future nature of your effort that would be based upon data you've collected during this investigation. Even of this is a hobby/school project it may seem like now you can move on to some other part of the work but this is still the wrong approach. If your approach was to decide to simply switch the DC power to solve the current symtom it would not be good enough for a robust product design. Here are two reasons:

1) If you build a product where you just switch the DC on and off to the load the whole AC section of the power supply stays active as long as the unit is plugged into the wall power. There are getting to be more and more "green requirements" that demand that electronic devices not consume standby power as long as the device is plugged in. Your product could even be violating a legal requirement in some regions.

2) Even if you intentionally switch the DC power of the product to get the firmware to behave at power off time it is still possible to lose the AC unexpectedly. This could make the product firmware "act up" and do its strange behaviors such as were being discussed in this thread. Extend that to some product where some big dangerous machine is controlled by your MCU. If that machine had some unexpected loss of AC the repeated code could be a lot more than a just a simple UART output message that gets repeated. A machine could drop a blade repeatedly and cut off an operators arm. A earth mover machine could unexpecetelly dump its load of dirt out over the top of some workers causing them to be maimed or killed. A production like robot could unexpectedly drive ahead multiple times and crush line workers against the wall.

So do you agree with me that this story is not yet really closed and resolved???

Michael Karas