Gain scaling

[cmsavage]

After reading light levels from multiple APDS-9301 boards using this library, I have found that the values returned by APDS9301::readLuxLevel() are at least an order of magnitude lower than expected and as measured from a phone app (though such measurements are not expected to be terribly accurate). Still, I found this library differs from how the TSL2561 library treats the empirical formula for lux found in the APDS-9301 data sheet. This library assumes that formula applies to the low gain readings, while the latter assumes it applies to the high gain readings (see here); the data sheet itself is ambiguous as to the gain to be used in that formula.

The end result is that the two libraries give lux values different by a factor of 16. Given my comparison with other sensors, I suspect the scaling in this library is incorrect. A fix would be to replace:

  if (getGain() == HIGH_GAIN)
  {
    ch0 /= 16;
    ch1 /= 16;
  }

with:

  if (getGain() == LOW_GAIN) 
  {
    ch0 *= 16;
    ch1 *= 16;
  }

Can anyone confirm a scaling issue and, if so, apply the fix?

[cmsavage]

I got my hands on a lux meter and the APDS-9301 gives results accurate within 10-20% if the readLuxLevel() value is multiplied by 16. This library does seem to be applying the empirical formula at the wrong gain.

[mhord]

Please give it a try now. I've updated the library with the change you suggested.

[cmsavage]

The lux values are now consistent with a lux meter. Thanks!

[mhord]

Thanks for the feedback! We appreciate it!