181123-0929 EST
LEDs are a moderately constant voltage device. This means that, if supplied from a constant voltage source, then some means of current limiting is required between the constant voltage source, and the constant voltage load, or if the voltages are not exactly equal you will burn out the LED. Three components exist that can limit current. These are resistors, inductors, and capacitors. Resistors dissipate energy, to a major extent inductors and capacitors don't.
Under steady state DC conditions an inductor becomes a resistor, and a capacitor an open circuit.
Under transient conditions with no initial stored energy, and at the instant of applying a voltage an inductor is an infinite impeder, and a capacitor a zero impeder.
Under steady state AC conditions an inductor is an impedance of some value, and a capacitor is an impedance of some value.
In a normal AC distribution systems we are dealing with an approximately constant voltage source. Thus, an impedance is needed between the source and the LED. This impedance is usually mostly an inductor to reduce power loss and avoid current peaking.
Current adjustment to the LED for dimming purposes is usually done by time controlled input to the inductor by a controlled switch. The switch is likely an FET, and not a Triac. For screw in bulbs dimming is usually controlled by an external phase shift dimmer (a chopped input sine wave). The phase angle of turn on is sensed inside the bulb and its internal circuitry uses that information to internally control dimming. Or average input voltage is used as the signal to control dimming. In all cases there will be an internal high frequency oscillator to reduce component size.
Even when dimming is not part of the objective it is mostly likely that high frequency switching is used as part of the current limiting function to reduce component size and cost.
Unless AC is converted to DC and then back to AC there will be higher than line frequency components going back into the AC supply.
Since each individual bulb has its own internal oscillator that runs independent of other similar bulbs on the same circuit we can not expect synchronization between the separate oscillators. This means it is hard to predict what the actual net current on the neutral will be when we try to balance two circuits with the same neutral.
Certainly there can be both conducted and radiated noise from the oscillators in the bulbs. I have some Feit bulbs that over about 110 V input will oscillate at about 1 MHz. This noise is conducted throughout my home, and on my power company, DTE, secondary lines. These secondary lines also feed two street lights. From the power company secondary lines there is radiated noise. I don't know how much of this 1 MHz power goes thru the 50 kVA pole transformer to its primary lines. Tried to do a test today, but there is a lot of RFI from something else in the 900 to 1000 kHz range today.
.
