Hey mike, I just saw this on a build come through production. It looks like your vias are 23 mil, and so the DRC has opened up the mask on all of them. Other than this minor problem the design looks fine.

Unless you'd like to take this quick rev, I think this could be handled by Patrick.

Let us know and thanks!

Currently saturation condition in the code is the following:
if ((CH0 == 0xFFFF) || (CH1 == 0xFFFF))
but according to chip's datasheet this is valid only for integration time longer then 178ms.

For tint = 101 ms saturated value is 37177 and for tint = 13.7ms it's 5047.

From product photos the silk is pretty broken. Use 0.02"(?) drill hits to get the soldermask to cover over vias.

Looking at the TSL2561 dat sheet on page 26 it says

// scale if gain is NOT 16X
if (!iGain) chScale = chScale << 4;

Looking @ line 215 SFE_TSL2561.cpp it says
// Normalize for gain
if (gain)
{
d0 /= 16;
d1 /= 16;
}

I changed that bit into

// Normalize for gain
if (!gain)
{
    d0 *= 16;
    d1 *= 16;
}

And now the integer math and float math of the sparkfun example return pretty much the same value.

gain: 1, broadband: 26693, ir: 7678, lux: 33029 -- 33260.40
gain: 1, broadband: 26718, ir: 7683, lux: 33068 -- 33298.91
gain: 1, broadband: 26720, ir: 7687, lux: 33058 -- 33289.90

Does it make sense to use a single jumper (two pad) for the I2C resistor pull up with a cuttable trace? Might cut back on solder paste. Might make layout a little easier. Not a big issue.

I used this repo a while ago for my TSL2561 and it worked fine, but now that I updated it seems the names of the source files have been changed from SFE_TSL2561.xxx to SparkFunTSL2561ArduinoLibrary.xxx. However, the cpp file still tries to include SFE_TSL2561.h, so now it doesn't compile. (Personally, I prefer the old names...)

I'm also confused why there are identical files in src/ and Libraries/Arduino/src/.