TSL2561 luminosity sensor

[espruino-discuss2]

Posted at 2014-01-30 by tom.gidden

Hey,

I've got a basic module for the TSL2561 luminosity sensor as sold by Adafruit done. It's pretty much a port of the Arduino library. It's not complete, and it's not fully working (which may or may not be a problem with my code or my specific sensor) but it's certainly sensing changes of light for me.

Anyway, thought I'd mention it as it shows a reasonably neat abstraction of an I2C-based sensor with an asynchronous read callback...

https://github.com/tomgidden/espruino-misc/blob/master/tsl2561.js

Tom

[espruino-discuss2]

Posted at 2014-01-30 by #kai

Cool! I have one lying around here and will soon try your code out with it and hopefully find some time to give feedback here!

[gfwilliams]

Posted at 2014-01-31 by @gfwilliams

Looks great! Sadly I don't have one to test with, but when it's working could we add it to the standard Espruino modules?

[espruino-discuss2]

Posted at 2014-01-31 by tom.gidden

Yeah, sure... I'll do a pull request if #kai or someone else can help test it with another sensor. I've also got a barometric sensor and a humidity/temperature sensor to do the same way.

It'd be nice to get the usual popular Adafruit etc. breakout sensors ported, especially if done to a reasonably consistent API.

Tom

[gfwilliams]

Posted at 2014-01-31 by @gfwilliams

Yes, that'd be awesome. There's some code on here somewhere from someone who got the humidity/temp working - but no module for it.

I'm going to try and come up with a template module soon though, because if you can put all the constants in a variable that's private to the module then when the module is minified it gets a lot faster and uses up an awful lot less RAM (while still being as readable as yours is :) ...

[espruino-discuss2]

Posted at 2014-01-31 by tom.gidden

Of course there needs to be a delineation between public constants needed for configuration (eg. tsl.C.address.FLOAT) and private constants used internally.

Alternatively, to use a better way of configuring than using a bunch of pseudo-enums.

BTW, one thing I did notice (but didn't have time to investigate last night) was the problem of chaining calls, eg.

obj
    .init(foo, bar)
    .setFoo(1)
    .getValue(function (val) {...});

rather than individual statements. The interpreter (or IDE... can't remember which) seemed to bork on the syntax.

[gfwilliams]

Posted at 2014-01-31 by @gfwilliams

Hmm. Could you try and come up with a really simple bit of code that fails, and stick it on https://github.com/espruino/Espruino/issues ?

This sort of thing does work though:

>function a() { return { a : function() { return 42; } } }
=function () { return { a : function() { return 42; } } }
>a().a()
=42
>a().a(function (val) { }) 
=42

Good point about the public constants, I'll make sure I think about that.

[espruino-discuss2]

Posted at 2014-01-31 by tom.gidden

Will do. I meant to investigate it and file an issue last night, but the yawning started.

[espruino-discuss2]

Posted at 2014-02-01 by tom.gidden

Following up on this thread... there's an issue at espruino/Espruino#211

Incidentally, I've refactored the TSL2561 code somewhat and fixed a few bugs. It should use memory more efficiently now. https://github.com/tomgidden/espruino-misc/blob/master/tsl2561.js

[gfwilliams]

Posted at 2014-02-03 by @gfwilliams

Thanks!

Could you do a pull-request on http://www.github.com/espruino/EspruinoDocs with it as a module (and maybe a very quick markdown file showing its use)? Either that or I could do it...

[espruino-discuss2]

Posted at 2014-02-03 by tom.gidden

Will do, but I'm not sure it's ready yet. I'd like to get the Lux calculation done, plus verify that the main code's working correctly as I'm still not getting a reading off the infrared channel, even after I double-checked the code against the original datasheet.

EDIT: Well, I guess it doesn't need the Lux calculation, but I'm still concerned about the infrared readings. Incomplete is one thing; broken is another matter entirely.

Tom