(Solved) Arduino code on ESP8266 keep crashing on submitting CPM to Radmon

'overcomplicated and hard to read by all the if statements'; do you mean all the debug parts? If that, then yes I know. Lots of debug makes it messy and difficult sometimes, but a necessary evil. The debug strings offer commenting also. I generally like to keep as much debugging as I can as this can really help other people if they use the code and are struggling with things. Also with regards to your comment about putting the code on github.... Nah... I really can't be bothered in maintaining something like that for small code. If I was writing code for a nuclear power station, then maybe github would work as a collaboration tool. It takes only a few clicks to copy and paste the code into the Arduino IDE, and then viola! Syntax highlighting! 

It is all my own custom code, with the exception of a few small parts of code originally written by BroHogan (DIY Geiger - I bought one of his GK Radmon units) as the code I posted earlier on this thread is custom code for the GK Radmon. It is only really the PWM HV generation parts of code that I have kept as they are required for that particular hardware. Everything else I wrote from scratch. It isn't that BroHogan's code didn't work (with one exception that I will come to) but it didn't operate in a way that I wanted. The GK Radmon code is designed for very low power consumption and it will go to sleep, wake after 4 minutes and 30 seconds (if I recall) then counts for 30 seconds, submits to Radmon and Thinkspeak and then goes to sleep again. This is great and works fine if that is what you want. I didn't. I wanted a power-hungry always on, always connected, always up-to-date counter where I could display on my mobile phone using Blynk. So I took the functions needed to run the HV on the hardware, and also the Radmon submission part and wrote my own code around those small functions. However, for some weird reason the HTTP code would crash the ESP8266 after an arbitrary length of time. What was happening was the code was freezing at some point in the Radmon submission function and then the watchdog timer would kick in and force a reboot. I never narrowed it down to exactly what, but other people were suffering similar issues with other projects. So I found a small example of different code, tried it and it worked, so re-wrote the Radmon submission function by 'get'ting instead of 'post'ing. It worked. Has worked since so I kept it. I do need to do another post on this at some point.

The new code that I wrote this morning is the same code running on my GK Radmon (my custom code) but about 90% thinner. I took out every function except for the absolute bare essentials for counting and submitting. It may made a nice foundation for other people to write their own code upon, or just use as is.

@4L7RU157: Are you using an ESP32? If so, did the Radmon submission code just work without having to change anything?

Also, looking at your code it looks like you are using a small OLED display. It looks like you are updating the display every time the loop runs. That's generally bad mojo. You will be tying up precious CPU cycles, and it may even be using an interrupt for the SPI/I2C bus which may cause other things to hang up whilst the display is updating. It would be better to put the display code in a function and then call that function say every 1000ms or 5000ms, or whatever. Obviously this will still tie up the CPU a little, but doing it every 1000ms is way better than every loop.

Another thing I learned about the small OLED displays is they wear out, and faster than you might think. A thermostat I made about 2 years ago has the OLED always on, and that lasted less than 12 months before pixels started to fade, and now at 2 years the display is pretty much ruined. Not a biggie as they are cheap, but if you want it to last then maybe think about turning it off after say 30 seconds or so. If you look at the bare-bones arduino code I wrote you can see that I have done just that for both calculating the CPM and sending to radmon. They are both contained within their own function and the loop counts the millis. When it reaches 1000ms it will call the CPM function, same for sending to radmon. For larger programs, like the custom FW I wrote for the GK Radmon, I use the 'simple timer' library. It simply allows you to setup multiple timers and they fire automagically and will simply call the function that is defined. You can also stop and start the timers dynamically which is really handy - saves many if statements!