Topband: Home Depot LED bulb interference.

[Bill Wichers]

It's probably worth noting that while the LED will run on DC *at full
intensity*, many/most of the controllers "dim" the light using PWM so in
that case there is a square wave of varying duty cycle driving the LED.
Human eyeballs integrate the pulses and perceive the light level as
being brighter or dimmer based on the on-time of the pulses in the PWM
output.

The switchmode power supply in the fixture is normally operating as a
current regulator since the LEDs need a constant current -- not a
constant voltage -- to operate properly. This is similar to how a
ballast works for a fluorescent or HID light. The current regulator may
run at several hundred kHz up into the very low MHz range depending on
the controller.

The dimming feature in many of the control chips will usually work by
PWM-modulating the current regulator with a low-frequency drive signal
in the 10(ish) kHz range (which can vary). The result is a drive
waveform with the current regulator maintaining a constant current
through the LED during the on-time of the PWM waveform, and the current
regulator is effectively "off" when the PWM waveform is "off". The
result will be a high frequency ripple (switching current regulator
operation) riding on top of the relatively low-frequency PWM waveform
while the light is being dimmed to something less than 100% intensity.

I could see the possibility of an LED light being RF-quiet at full
power, but noisy when dimmed due to the PWM system. That might be
something to think about if you're trying to track down RFI you suspect
is coming from LED-based light fixtures.

   -Bill KB8WYP



> LED lights work on low voltage DC, and the trash from products like
this
> is usually from the switching power supply that converts 120VAC to
that
> low voltage DC.  We know about switching power supply noise -- it can
be
> quite frequency-dependent, it is generally drifty, it usually appears
as
> a hump of noise in a spectrum display, and the relative strength in
any
> frequency range will depend both on the waveshape of the noise in the
> device and the effectiveness of both internal and external wiring as
an
> antenna.