Light Meters, Fiber Optics and Optogenetics – Part 1 – Light Meter Blues

Prizmatix LEDs are extremely powerful and our customers are generally very satisfied with their performance. 

Sometime, however, I get complaints about the power output being lower than expected as per the published specification. Very frequently it turns out that the light source is fine, and the problem lies within the tools and methods used to measure the power.

In one specific instance, the power output of an LED light source was 16mW at the tip of the fiber when the expected power was 40mW.

After hearing about the set up, I noted 3 factors that could be causing the lower power output.

1) A flat detector designed for laser light was being used instead of an integrating sphere. Our experience shows that a flat detector will read 30% lower than an integrating sphere.

2) The laser power meter had a sliding neutral density filter to increase the power range by 100x. This works by allowing only 1/100 of the light to hit the detector. Problem is that the ND filter is designed for laser light. LED light which hits at an angle will not penetrate the filter, and the reduction factor could be larger than expected. I asked the customer to calibrate the filter and determine the correct factor to use.

3) The user had the meter set to 473nm but the LED was closer to 450nm. The lower the wavelength, the higher the energy per photon, so having the WL set too high will cause a lower than actual power reading.

After correcting these three factors, the numbers actually matched!

Here are some quotes from our email correspondence:

1) The wavelength of the LED is around 450-455nm. With the meter set to 473nm you will get lower than actual power readings. See if you can adjust the wavelength to the peak wavelength of the LED, which should be seen on the test report that was shipped with the LED. 

Yes! I forgot the correct wavelength should be 450 nm -- correcting that increased measured power to 28 mW. (from 16mW)

2) The flat detector is designed to measure laser light. To accurately measure LED light a detector with integrating sphere would be needed. In our lab the meter with integrating sphere shows 30% higher power compared to when measured with a plane flat detector.

Got it, makes sense. 28 mW * 1.3 = 36.4 mW

3) The detector has a range of up to 5mW or 500mW. The 500mW range is activated when you slide the neutral density filter into place, which should block 99% of the light from hitting the detector. That increases the power range 100x. Problem is that filters will work only when the light is collimated, like with a laser. With an LED, the light hits at an angle and does not penetrate the filter like laser light. So the power meter is expecting 1/100 of the light, and multiplies the measurement by 100x. But in fact it should multiply by a higher factor. That is why I suggested last night (in my last email) to calibrate the ND filter to determine by how much it reduces the light.

Also yes -- adding the 500mW filter decreased the reading (after the meter automatically compensated for the 100x change) by 11.5%. So 36.4 mW /0.885 = 41.1 mW.
So that agrees well with your light cube measurements.