Recently while working on my interim Christmas tree power supply solution, I couldn’t help but ponder whether or not modern LED Christmas tree lights are an EMI consideration. I mean, they’re powered from a high frequency switching supply, with software driven pattern sequencing. On top of this, a big long string of lights is going to make quite a good antenna. Potentially a problem?
I buried a whip antenna in the middle of the tree, which is wrapped with 1,500 LED lights, and set my Spectrum Analyser at it to see what, if anything I can pick up.
Well well. With the factory power supply, we do indeed have some EMI. I thought it’d be easiest to show with a waterfall diagram (ignore the FFT above as the lights were off at the point of taking the screen shot). The section between the “ON” and “OFF” arrows is when the lights are powered on. Quite detectable indeed, peaking at -50 dBm with a noise floor of around -80 dBm. It peters out after about 7 MHz, realistically there are no devices likely to be used around the house in this frequency range so probably not a biggie.
Might cause problems for an AM radio, but it’ll likely already be blighted by the cacophony of other rubbish on the airwaves of a modern home. Definitely something I will have to keep an eye on in my own build. Hopefully I can do a bit better than this.