Kelvin discrepancy at higher values

[aaron severtson]

#47074

I have been using the sensor for a little over a month now, and updated the firmware to the newest version per the Android app. I have done many readings of natural daylight outdoors, and tested various lighting fixtures with the sensor. In questioning the reading accuracy, I was able to compare the color readings between this sensor and one made by AsenseTek. Based on price and industry reputation, the AsenseTek would be expected to be an accurate sensor to read light with.

Using a variable color fixture, I tested the output of the fixture with both the Illuminati and AsenseTek sensors located one foot from the light source, side by side. Starting at a 2500K fixture setting, I took readings every 500K up to 8000K per the fixture settings. This was done in a dark room. Below is a link to the data taken, and a graph showing the results. What is of concern to me is the delta between the two sensors as the Kelvin value increases. Has anyone else seen this discrepancy grow at cooler temps? Is there a firmware update that can resolve this issue? How can the Illuminati sensor be verified to know that it is working correctly? If neither can be verified, how do I at least verify the Illuminati? Thanks!

https://drive.google.com/file/d/1FiK2f2mdPBOdorZlynd5GNxMcMiKkOqU/view?usp=sharing

[Michael Okincha]

#49135

Hi Aaron.

Thanks the excellent measurements. The short answer is yes, your readings are what we see with the IM100 meter.

I assume you are using a color-temp adjustable LED light source. All white LEDs are made by putting a phosphor coating on top of a blue LED. The blue LED’s light causes the phosphor to emit yellow light, and the combination produces white light. The blue LED’s light spectrum is very narrow – less then 2% of the range of colors visible to humans.

Despite our best efforts, that blue light occurs right where our blue sensor’s spectral response is a little inaccurate. The sensor’s error is only a few percent from ideal, but the blue LED’s power is all concentrated in a narrow range. So it’s natural that as the LED’s light becomes more blue, the color temp error increases.

We plan to release an updated version of the meter by the end of the year. Among other improvements, we have increased the color accuracy for cool white LEDs.

Regards,
Mike

[Author]

Posts