Option for duration

[naaitsab]

I experimented a bit with the "type 1" collar. It seems to support longer/continuous vibration and shocks. Tested to 7 seconds.

I've hacked together a very crude method to let it trigger longer. If you call the function in a loop. With the loop count double the amount of seconds it seems to work about right with a delay of 250 in between.

  int timestofire = sec * 2;
  for (int i=0; i<timestofire; i++) {
    process_message(commandC);
    delay(250);
   }

But after a few seconds the ESP restarts so I guess it overloads the system doing it like this. Any idea how to implement this cleanly inside of the collarTX libary?

[CrashOverride85]

I'm not too familiar with ESPs, but i think if the main loop takes too long to run, the wdt resets them? (been a few years since I've played with one).

So I'd suggest trying something like this that avoids calling delay():
https://gist.github.com/CrashOverride85/531d95e505f302c8b664b95a71ac3926

Basically it's the ColTxSerial.ino example modified to transmit for 7 seconds (set on line 161)

[naaitsab]

Thanks will give it a try later on. Not sure what the max duration of the collar itself is. Did you experiment with it yourself?

[CrashOverride85]

I honestly can't remember - I probably didn't experiment much with max duration. But I knew they could be on for a few seconds, but I can't remember if there's a limit or what it is

[naaitsab]

Got it working, even with WiFi enabled and a webpage (remote control FTW) it stays stable until 10 seconds :) Maybe even more but I think 10 is a nice max.

Made a few modifications to implement the timer in the commands

Added int sec = atoi(&input_message[5]); under power = atoi(&input_message[2]);
Made the command accept the 'sec' as input instead of the static 7 of the example.
atoi keeps 'calculating' so we need to terminate it after the power ends and the sec(onds) start with a char. So I added a : in the *input_message

So the message should look something like this 1V15:05 for collar 1 vibrate at 15% for 5 seconds.

Only thing I can't really figure out is the timing. It seems to be time + 1 second roughly. So 10 is 11 and 1 is 2.

Also want to implement a pavlov mode. So vibrate first then shock. Guess it needs the modification inside the void loop at the check "if (_transmit_end_ms > 0)" so it will set to vibrate for 1-2 seconds (or 1/3 of the time or something? and then change the "_msg" so it will shock instead of vibrate.

[CrashOverride85]

I'm not sure why it's triggering for an extra second - it might be just the response time of collar, or something like that? If it's always out by 1 second, i guess the easy approach is just to take 1 sec off seconds on transmit() 😂

Getting to vibrate first the shock is probably a bit more awkward. I guess you'd need to do something like add a _transmit_start_ms time, and a pavlov_mode flag.
So in the loop, when pavlov_mode=true and msg.mode=VIBE, if the current millis() is more than 1/3 of the way between _transmit_start_ms and _transmit_end_ms, change msg.mode to SHOCK.
There might be easier ways, though... not sure..

[naaitsab]

Do you by chance have the remote for it to test something? Perhaps the second is just the minimum trigger amount and every "loop" adds another second to it. So trigger + 1 loop is 2 seconds instead of 1 because the trigger already adds 1 second.

[CrashOverride85]

Just found it and tried it... yeah, from pressing the button to the shock collar stopping, it does seem to be around a second