An LED's intrinsic capacitance works in a 650-mV LRC circuit
The LED and a discrete inductor form an LC tank circuit.
Sajjad Haidar, University of British Columbia, Vancouver, BC, Canada; Edited by Paul Rako and Fran Granville -- EDN, September 8, 2011
You can use the inherent capacitance of an LED to make a series resonant boost circuit that can create a voltage large enough to light the LED. Depending on the color of the LED, you need a voltage higher than 1.6V to turn it on. The threshold, or knee, voltage rises higher as the LED wavelength becomes shorter. All PN-junction diodes, including LEDs, have capacitance due to depletion and diffusion profiles.
You can light an LED using its capacitance
in a series LRC (inductance/resistance/capacitance) resonant circuit. In
such a circuit, the Q factor determines
the multiple of the generator voltage that
appears across LC. If you fashion a circuit
with a high enough Q factor, you boost
the generator voltage enough to light the
LED. The Q factor of the resonant circuit
is a function of the resistance, inductance,
and capacitance, as the following equation
shows: 
You can verify this calculation with a
simple circuit using a blue LED in series
with an inductor (Figure 1). The knee
voltage of the LED is 2.45V, and the signal
generator has an internal resistance of
50Ω. An inductance of 100 μH and the
50-pF capacitance of a typical LED yield
a Q of 28. The amplitude of the sinusoidal
signal generator is set at 650 mV p-p.
You can then vary the generator’s output
frequency until you see the circuit’s resonant
point. As the circuit approaches the resonant frequency, the voltage across the
LED starts to increase. The resonant point
manifests itself as a small jump in voltage,
rather than a smooth progression, due
to a positive feedback at resonance. The
positive feedback happens because the
capacitance of any PN-junction device
is not linear (Figure 2). As the circuit
approaches the resonant frequency, the
LED voltage increases, which also increases
the LED capacitance, resulting in lower
resonant frequency.
For a blue LED, the voltage waveform
as the circuit approaches resonance is
1.55 MHz. The circuit settles at 1.69 MHz
(Figure 3). The forward-biased LED is
thus emitting light, clipping the positive
parts of the boosted waveform. Using the
same 650-mV-p-p generator amplitude
on other colors of LEDs produces different
resonant frequencies. You can see a
similar effect with a square-wave generator
because it also contains the fundamental
components of the resonant
frequency.
Talkback
-
It looks like a classic parametric oscillator that could work with any diode that changes a capacitance depedently from a voltage level. The best results were achieved with varactors.
Vladimir Doubovis - 2011-4-11 10:32:04 PDT -
do you think is possible "reverse" circuit?
L=inductor(coupled-transformer)
C=diode junction from SolarPanel
R=impedance controller(to be always resonant)
From trasformer I should get "sine"-wave output?
Karm - 2011-25-10 07:55:16 PDT -
Did you measure the diode capacitance as seen in Fig 2 with 10 data points? If so, can you describe the measurement technique?
Al Daniel - 2011-25-9 07:59:17 PDT -
That's pretty cool!! I'm gonna have to try it.
Much like the "Rubik's Cube", this is a very clever idea that, once understood, is extremely simple. You know, in a kind of "Why didn't I think of that??" way. Well done Sajjid, and thanks for sharing!!!! :)
Kyle B. - 2011-12-9 10:57:11 PDT
