issue with I2C on pico (initialization)

[espruino-discuss2]

Posted at 2018-03-29 by MikeD

Hello,
i'm writing a simple timer using an Adafruit 7 segment display which uses I2C.
I have made a working prototype using the original kickstarted Espruino and wanted to transfer the design to the pico.
The display requires initialization which I do in an E.on('init') as well as inline:

function ledInit() {
  I2C1.setup({scl:B6, sda:B7});
  initAddr = (addr) => {
    I2C1.writeTo(addr, 0x21); // clock on
    I2C1.writeTo(addr, 0x81); // flash
    I2C1.writeTo(addr, 0xE0 | 3); // brightness 15
  };

  initAddr(0x70);
}

E.on('init', function() { //doesn't work on pico
  USB.setConsole();
  digitalPulse(LED2,1,200);
  ledInit();
}
ledInit(); //doesn't work on pico

For the pico, i changed the pins to B10 and B3
The problem is that while i'm on USB, the init works just fine. However, when I save the code and run it with a battery, the LED init fails. I later tried to call the init on BTN1 and that works fine. Something about initial power that's not working. I tried setTimeout to call it and that fails as well.
I also tried the USB.console trick and that didn't help.

setTimeout(ledInit,2000); // doesn't work
setWatch(ledInit, BTN1, { repeat: true, debounce: 50 }); // works but is less than ideal

Any ideas?
I'm running latest firmware available

[gfwilliams]

Posted at 2018-03-29 by @gfwilliams

Just checking - but you also changed I2C1 to I2C2?

That is a strange one. .. and you're calling setTimeout(ledInit,2000); from E.on('init')?

Could you add a digitalPulse(LED1 to ledInit so you can make sure that it does get called at the time you think it should be called?

[espruino-discuss2]

Posted at 2018-03-29 by MikeD

Yes, i changed it to I2C2 and no, the timeout was just inline. I will try it inside the init. As it is, I'm still a bit unsure of execution. It seems when I send the code to the device, it gets executed immediately. that does not, however, happen when plugging it in? I thought the code gets parse every time it's powered on.

[gfwilliams]

Posted at 2018-03-29 by @gfwilliams

the timeout was just inline

Yes, that could be your problem.

Check out http://www.espruino.com/Saving - it should hopefully clear up what happens. By default code is executed as it is uploaded, and save() saves an image of the current intepreter state.

You can change how it works, but then you miss out on the ability to tweak things on the fly using the left-hand side of the IDE.

[espruino-discuss2]

Posted at 2018-03-30 by MikeD

tried putting timeout in E.on('init') - nothing. it almost seems as if the init function does not get called when the device is plugged into power.
if i plug it in, connect it to the ide, and then run load(), then it works as intended, but otherwise it doesnt look like it's working. any way to debug?

in fact, try running this and let me know if it works:

function ledInit() {
  digitalPulse(LED1,1,200);
  digitalPulse(LED2,1,400);
}

E.on('init',function() {
  USB.setConsole();
  console.log('ini');

  ledInit();
  setTimeout(ledInit,2000);
});

for me, it only works when i run load() from command line.. on powerup, nothing happens

[espruino-discuss2]

Posted at 2018-04-02 by MikeD

anyone?

[espruino-discuss2]

Posted at 2018-04-02 by Frida

Try with an extra timeout

function ledInit() {
  digitalPulse(LED1,1,200);
  digitalPulse(LED2,1,400);
}
//function onInit() {
E.on('init',function() {
  setTimeout(function() { // extra timeout
    USB.setConsole();
    console.log('ini');
    ledInit();
    setTimeout(ledInit,2000);
  },1500); // extra timeout (try'd 1000 but did'n work)
}); // end E.on()
//} // end onInit()

// onInit(); remove before save

[espruino-discuss2]

Posted at 2018-04-02 by MikeD

that worked! now can someone explain why? I also noticed that taking the 1500 delay down to 1000 fails as well - does the board inherently have some issue with 'settling' for 1.5 seconds?

[espruino-discuss2]

Posted at 2018-04-03 by @allObjects

I do not exactly know, but there something run after your E.on('init',yourAnonymousFunction(){}) that needs the time of 1500+ ms. I usually do not use the E.on('init',...), because it does not full control over what is initialized when, even though in your case it is just one registered (user) function. I use the 'old' onInit() {...} that was always there... and I would be very surprised that that would not run anymore... (@frida has it in the code... commented: @frida, did it work for you with old onInit()?)

function ledInit() {
  digitalPulse(LED1,1,200);
  digitalPulse(LED2,1,400);
}

function onInit() {
    USB.setConsole();
    console.log('ini');
    ledInit();
    setTimeout(ledInit,2000);
} // end onInit()
// onInit();  remove or comment before uploading last time  before save()

If it is not working, the console meddling could could have something to do with it...

Important (for my use is), that I never save() dynamic initialized values or running timers..., not even watches, even though @gordon built it the way that it - saving the state of the interpreter, configurations, timers,... are saved and restored on power up - works (...for simple things where no timing issues loom...). Sometimes I take even the port mode settings out of the upload code and put it into the initialization code...

[espruino-discuss2]

Posted at 2018-04-03 by MikeD

i suppose i2c initialization could take longer? any way to make it asynchronous?

[espruino-discuss2]

Posted at 2018-04-03 by Frida

@allObjects, I allways use onInit(), but yes it needs the time too.

[gfwilliams]

Posted at 2018-04-04 by @gfwilliams

E.on('init',...) should run just before onInit, but I didn't think there was anything particularly special about it otherwise.

What if you remove the USB.setConsole? It could be that trying to set the console device to USB before USB has initialised causes problems?

[espruino-discuss2]

Posted at 2018-04-04 by MikeD

I tried taking out setConsole.. still seems to need a delay. would probably be a good idea to have some sort of an event for 'ready' rather than 'init' if it's an issue of internal processing time to get things going..

[gfwilliams]

Posted at 2018-04-05 by @gfwilliams

So it does work with the USB.setConsole removed?

There's this issue on boards with USB where often people want stuff running as soon as possible, but then that's before USB has initialised properly. It used to be that I'd wait until USB had initialised but then people complained - but good call on the event. It would be a nice way around it.