ESP8266 Community Forum

- Sun Apr 23, 2017 10:33 am #65167 Great, thank you guys for your help. Definitely looks like it is my long wiring runs. Going work on that next. It is a 12e, with the the latest nodemcu firmware.

So my regulator should be close to the esp? I thought it should be farther away to avoid interference?

Also, since I've seen a lot of these issues around, it makes me nervous about this as a production board. So if I get the power figured out, this should be able to run indefinitely without errors?

- Sun Apr 23, 2017 8:54 pm #65181 Mine seem to be reliable since I fixed the power issue. But if you are looking to buy more I would go for the ESP-12S as a first choice. Then the ESP-12F.

The regulator can only sense the voltage at it's output terminal. It sees less of a change the farther it is from the disturbance.

- Mon Apr 24, 2017 1:09 am #65183 I wouldn't trust any µC system for any long or indefinite run except when programmed in assembler where I have control over stack, heap, etc.
If indefinite or long run are needed, stability is by design. Such systems don't generaly need to be active all the time but just have to react on external events; in the time inbetween they can sleep. So the most stable design would just restart the µC "clean" every time the external event occurs.

- Mon Apr 24, 2017 9:30 am #65210

I wouldn't trust any µC system for any long or indefinite run except when programmed in assembler where I have control over stack, heap, etc.
If indefinite or long run are needed, stability is by design. Such systems don't generaly need to be active all the time but just have to react on external events; in the time inbetween they can sleep. So the most stable design would just restart the µC "clean" every time the external event occurs.

I read the first line and I wanted to argue. The company I work for makes control electronics and it needs to work continuously as long as there is power. And while I like the control of assembly language programming it really isn't a practical thing anymore.

"If indefinite or long run are needed, stability is by design."

I totally agree with that. And the problem with a lot of programmers is that they don't spend the time to take care of all the real world situations that can occur, because it is hard to do. The one place where I see the biggest problem is on power loss during the operation of the program. If you can't detect a power loss situation, and react to it properly, then you have a potential corruption problem.

"If indefinite or long run are needed, stability is by design."

Worth repeating again.