Question about EMI and lighting?

[rwgast_lowlevellogicdesin]

Earlier today I was running some tests on a competing audio product and which uses a really cheap wall wart, the orignal idea was just to plug it in to  a kill-a-watt and see how much power it consumed and then look at the psu noise with a scope. Anyways I had a phone pickup laying around (for those of you to young, there an inductor with a suction cup which you would stick to the phone and plug in to a tape recorder) out of curiousity I connected the scope to the pick up turning the cheap gear on and off watching it induce noise inches away.....

Anyways I seem to remember people complaining that LED lights were noisy, I have few lamps/lights around my bench, so I picked a floor standing lamp with three bulbs. Two are CFL's and one is a 5k 110 lumen LED bulb. Well I stuck the pick up on both CFL's and could see 55khz sine waves accross my scope, I then connected the pick up to the the LED shell and there was no visible noise induced above the scopes noise floor at 2mV per Div. The CFL's induced almost 15mv ptp sine waves.

So are LED bulbs noisy usually? Am I remembering this wrong or do I just have an LED light with an awesome converter in it? This experiment has made me think twice about having CFL's on near any noise sensitive or very low voltage work.

[madalf71]

This is a really good video done by another amateur on the topic. Hope it helps.
https://youtu.be/o8zMhjXcmoA

[Rerouter]

Any switching design will be noisy, even some transformer wall warts can have a resonance from the long leads to the plug, the same applies for most lighting systems, an incandescent globe by its inductive design will broadcast primarily 60 HZ (or 50Hz) but at quite low amplitude, as the current levels are fairly low once the globe has heated up,

CCFL's you can think of like a current limited spark gap, once the ark is started, it produces light, but as the ark moves about on either electrode, or impurities in the gas mix change the voltage drop, and each time it re strikes it has a current spike, and can end up generating some very broad noise,

LED's generally sit somewhere in the middle, There switchmode converters can be made fairly quiet, but as its a switching design there will always be noise, however every led bulb i have so far come across in stores under $50 has had a rubbish AC-DC power supply, which does spew forth all sorts of noise,

The work around for LED lighting is you can find LED AC-DC current sources that sit in a roof like a down light transformer, these tend to be made fairly well and either keep the noise in one specific band, or reduce most of it away, again never 100% gone, but you can filter out a lot of it with good design,

[T3sl4co1l]

You're just probing at the wrong point.   The converter (if present..) will give off most switching noise.  Magnetic fields drop off quickly so this isn't a problem.

CFLs have the added bonus, not just of having a large resonant inductor close to the surface in the base, but the load current flows through the tube as well, so some magnetic field will be sensible there, too.

Tim