# LED luminosity decreasing over time

I'm measuring an LED with a light sensor but I'm running into a problem where the LED's luminosity (lux) is decreasing over time. Confusing since this is a simple setup, I don't know if the LED's I'm using are defective or something else is failing.

I'm running an Arduino Uno and other than the light sensor, the LED (https://www.adafruit.com/product/754) is just a simple circuit connected with an 250 ohm resistor. I've done the math and it checks out, and using a multimeter I can confirm that the current going through the LED is less than the maximum continuous current of 20mA. I even tried almost doubling the resistance, at 440 ohms, and it still had the same problem. It doesn't seem like the LED is significantly overheated, so I'm not sure what the cause of this is. Could there be something wrong with the wires or could the light sensor be the problem?

This is happening over a short period of time, 5-10 minutes.

• This is a thing LED's do. Commented Sep 15, 2021 at 19:29
• What light sensor are you using? Commented Sep 15, 2021 at 19:51
• @Finbarr adafruit.com/product/1980 Commented Sep 15, 2021 at 20:03
• What makes you think that the problem is the LED and not the Arduino or light sensor? Commented Sep 15, 2021 at 20:11
• "the LED's luminosity (lux) is decreasing over time... over a span of less than 5 minutes" - by how much does it decrease? What happens if you leave it running for a long time (eg. 1 hour)? What does the LED current (or voltage across the resistor) do during this time? Commented Sep 15, 2021 at 21:21

It's probably not the LED. It's probably the transistor that is powering the LED. To confirm, measure the voltage across the combination of the LED and the current limiting resistor. I think you will find that it decreases over time.

• Can you enlighten us with your reasoning here? Commented Sep 15, 2021 at 23:18
• What transistor? Commented Sep 15, 2021 at 23:20
• OP said "I'm running an Arduino Uno...". That tells us that OP is turning the LED on and off. And implies that OP is doing so through a transistor. Somewhere in there there is a transistor doing the turning on and off. OP could be using an actual, discrete transistor. OP could be using an IC, in which case I mean the transistor in that IC. Or, OP could be driving the LED directly from the Uno, in which case the transistor is in the microcontroller in the Uno Commented Sep 16, 2021 at 1:03
• > Can you enlighten us with your reasoning here? <BR> The reasoning is that low power LEDs do not change brightness over a few minutes all on their own. The reasoning is that therefore, something else (something that OP did not describe - a transistor controlling the LED) is what is changing over time. Commented Sep 16, 2021 at 1:07

You don't provide all the relevant information to give you absolute answers. The LED you have is a Phosphor conversion LED; it's actually a Blue LED with a phosphor disk on top and the phosphors are notorious for PCE droop (phosphor conversion efficiency droop).

You have three effects going on:

1. LED Vf is dropping due to thermal process (no matter how small) which increases the LED current.
2. Blue LED optical flux increasing as effective current increases
3. PCE droops as flux increases.

The effects do somewhat offset each other, but I seem to remember that even moderate changes in Vf resulted in 4-10% reduction in White light output. You could test this out by changing to a Yellow or Red LED which don't suffer this PCE drop. To provide better results with the LED you have you could:

1. Increase the current to very close to the rating (20mA) as the PCE curve flattens at higher flux density.
2. Drive the LED with a CC source.

The modern way to substantially eliminate PCE droop (particularly in White LEDs used for Automotive) is to use a Laser Diode PCLED.