nRF9160 - Why LEDs have wideband noise?

[pctj101]

In this PDF: https://infocenter.nordicsemi.com/pdf/nwp_037.pdf

On page 28 it says:

Active components on the PCB cause wideband noise coupling to the antennas. Verify that during reception, especially GPS, active components on the PCB do not radiate noise to the GPS antenna. For example, DC converters and LEDs are typical sources of radiating wideband noise. 

My question is in regards to LEDs being a wideband noise radiator.  Why is the case?

  Is wideband noise intrinsic to the generation of light? 

  Or is it the PWM signal edges?

  Or is it noisy power being supplied to the LEDs through long traces that radiate?

Thanks for any clarification... I myself see strange EMI issues sometimes and I'd like to know what you know :)

[helsing]

In general, HW design EMI must be always considered. I checked with a few team members that regularly deal with or review HW issues, and they did not report noise from LEDs as a common problem in customer designs. My impression after these discussions is that PWM control issues(and PCB traces) with their harmonics is of more of importance.

When It comes to the noise related to the operation of the LEDs themselves, I have not bee able to conclude on anything yet.

[pctj101]

Ah okay. Yeah if PWM then sure the edges and harmonics are understandable sources of concern. Just wanted to make sure a “solid red” LED wasn’t an issue :)

[helsing]

Looking into how different types of diodes generate noise is an interesting read, touching on some interesting topics. However, a solid red LED should normally not be of any concern. As always, please ask us for a free HW review on the NordicSemi related part of your design when you are ready.