An LC oscillator is an electronic circuit that uses an inductor (L) and a capacitor (C) wired together in a resonant tank, combined with an active amplifying component, to generate a continuous AC sine wave from a DC power supply. In a real circuit, it changes a static DC voltage rail into a precise, continuous alternating carrier signal, acting as the fundamental heartbeat for radio frequency (RF) transmitters, local oscillators in receivers, and induction heating systems. Hobbyists and students frequently confuse LC oscillators with simple LC filters (which lack the active positive feedback loop required to sustain oscillation) or RC oscillators (which use resistors and capacitors and are generally limited to much lower audio frequencies).
The Physics of the Tank Circuit (and the Math That Matters)
At the core of every LC oscillator is the "tank circuit." Think of it like a mechanical pendulum: the capacitor stores potential energy (like lifting the pendulum bob), and the inductor stores kinetic energy (like the bob swinging through the lowest point). When connected, the energy sloshes back and forth between the electric field of the capacitor and the magnetic field of the inductor. Without an active transistor or op-amp to inject a tiny bit of energy on each swing to overcome resistive losses, this oscillation would quickly dampen and die. The active component provides exactly enough positive feedback to maintain a steady amplitude.
The resonant frequency ($f_r$) where this energy exchange happens most efficiently is dictated by the Thomson formula:
The Resonance Formula:
$$f_r = \frac{1}{2\pi\sqrt{LC}}$$
Where $f_r$ is in Hertz, $L$ is in Henries, and $C$ is in Farads.
Worked Numeric Example: Targeting 10 MHz
Let’s say you are designing the local oscillator for a shortwave receiver and need a precise 10.0 MHz signal. You have a high-Q, air-core inductor measured at 2.53 µH ($2.53 \times 10^{-6}$ H) on your LCR meter. What capacitor value do you need?
Rearranging the formula to solve for C:
$$C = \frac{1}{(2\pi f_r)^2 \times L}$$
$$C = \frac{1}{(2 \times \pi \times 10,000,000)^2 \times 2.53 \times 10^{-6}}$$
$$C = \frac{1}{(6.283 \times 10^7)^2 \times 2.53 \times 10^{-6}}$$
$$C = \frac{1}{3.947 \times 10^{15} \times 2.53 \times 10^{-6}} \approx 100 \text{ pF}$$
You need exactly 100 pF of tank capacitance. If you are building a Colpitts oscillator, which uses two capacitors in series to create a voltage divider for feedback, you would use two 200 pF capacitors (since series capacitance $C_{eq} = \frac{C1 \times C2}{C1 + C2}$, yielding 100 pF).
Where You Meet This in Practice
While digital clocks and microcontrollers rely on quartz crystals, LC oscillators dominate in applications where the frequency needs to be tunable, or where the operating frequency is too high for standard RC networks.
- FM Transmitters (88 - 108 MHz): The carrier wave is generated by an LC tank, often using a varactor diode to allow voltage-controlled tuning.
- Superheterodyne Receivers: The "local oscillator" that mixes with the incoming RF signal to create an intermediate frequency (IF) is frequently an LC design in older or cost-sensitive radios.
- Metal Detectors (10 kHz - 100 kHz): Beat-frequency oscillators (BFO) use two LC oscillators; when metal enters the search coil (acting as the inductor), it shifts the frequency, creating an audible beat note.
- Induction Heaters (20 kHz - 500 kHz): High-power LC oscillators (often using IGBTs or MOSFETs) drive the work coil to generate intense localized heat.
Bench Walkthrough: Building a 10 MHz Colpitts Oscillator
Theory is clean; the workbench is messy. Here is a real-world scenario of building the 10 MHz Colpitts oscillator we calculated above, using a standard 2N3904 NPN bipolar junction transistor (BJT).
1. The Setup
We assemble the circuit on a standard solderless breadboard. We use a 2.53 µH off-the-shelf molded radial inductor, two 200 pF capacitors for the tank divider, and bias the 2N3904 with standard 10kΩ/47kΩ base resistors and a 470Ω emitter resistor. We power it with a clean 9V bench supply and connect a 10x oscilloscope probe to the emitter.
2. The Numbers and Outcome
We expect a crisp 10.0 MHz sine wave at roughly 1V peak-to-peak. Instead, the oscilloscope shows a slightly distorted, asymmetrical waveform oscillating at 9.42 MHz, with an amplitude of only 400 mV.
3. What Went Wrong?
This is a classic bench failure driven by three invisible parasitic factors:
- Breadboard Stray Capacitance: A standard solderless breadboard introduces roughly 2 pF to 5 pF of parasitic capacitance between adjacent rows. This adds directly in parallel with our tank capacitor, lowering the resonant frequency.
- Transistor Junction Capacitance ($C_{ob}$): The 2N3904 has an output capacitance of about 4 pF. In a common-base Colpitts configuration, this appears in parallel with the tank.
- Capacitor Dielectric Choice: If you grabbed standard X7R or Y5V ceramic capacitors from your bin, their capacitance drops significantly under DC bias voltage and shifts with temperature. This causes both the frequency drop and the waveform distortion (acting as accidental frequency modulation).
The Fix: To hit 10.0 MHz accurately, you must use C0G/NP0 dielectric capacitors (which are voltage- and temperature-stable). You must also account for ~8 pF of stray/parasitic capacitance in your initial math, targeting a physical tank capacitance of 92 pF instead of 100 pF. Finally, for RF work above 5 MHz, abandon the breadboard and build "dead-bug" style on a copper-clad board to eliminate stray row capacitance.
LC vs. RC vs. Crystal: Choosing the Right Heartbeat
When designing a system that requires an AC signal, choosing the wrong oscillator topology leads to poor phase noise, frequency drift, or outright failure. Here is how they compare on the bench.
| Criterion | LC Oscillator | RC Oscillator (e.g., Wien Bridge) | Crystal Oscillator (XO) |
|---|---|---|---|
| Frequency Range | 100 kHz to >1 GHz | 1 Hz to ~1 MHz | 10 kHz to >200 MHz |
| Tunability | Excellent (via variable cap or varactor) | Good (via potentiometer) | Very Poor (fixed, or narrow VCXO pull range) |
| Frequency Stability | Moderate (drifts with temp/parasitics) | Low to Moderate | Extremely High (ppm level) |
| Phase Noise | Low (if high-Q inductor is used) | Higher | Very Low |
| Best Use Case | RF transmitters, tunable local oscillators | Audio test equipment, low-frequency clocks | Microcontroller clocks, digital comms, GPS |
For deeper theoretical foundations on how these topologies sustain oscillation via the Barkhausen criterion, the Electronics Tutorials guide on LC Oscillators provides excellent AC equivalent circuit breakdowns. Additionally, All About Circuits offers a fantastic primer on the transition from linear amplifiers to non-linear oscillators.
FAQ: Troubleshooting LC Oscillator Failures
Why won't my LC oscillator start up when I apply power?
Oscillation requires a loop gain strictly greater than 1 at startup. If your circuit is dead, check your biasing. The transistor must be biased in its active (linear) region. If the base voltage is too low, the transistor is in cutoff; if too high, it's saturated. Furthermore, ensure your feedback network (the capacitive divider in a Colpitts, or the tapped inductor in a Hartley) is providing enough signal back to the input. A 10:1 impedance ratio is a safe starting point for BJT Colpitts designs.
Why is the output frequency drifting as the circuit warms up?
Thermal drift is the enemy of LC tanks. First, verify you are using C0G/NP0 capacitors; X7R capacitors will shift value as they self-heat or as ambient temperature changes. Second, check your inductor. Ferrite-core inductors change permeability with temperature. For high-stability bench designs, use an air-core inductor or a temperature-compensated ceramic resonator. Finally, ensure the active device isn't dissipating excessive power; a 2N3904 running hot will shift its internal junction capacitances, pulling the frequency.
Can I use an op-amp instead of a transistor for an LC oscillator?
Yes, but only at lower frequencies. Standard op-amps (like the TL072 or NE5532) have a Gain-Bandwidth Product (GBWP) that limits their useful open-loop gain. If you try to build a 20 MHz LC oscillator with a 10 MHz GBWP op-amp, the loop gain will drop below 1, and it will fail to oscillate. For audio-frequency LC or active-filter-tuned oscillators, op-amps are fine. For RF, stick to discrete BJTs, JFETs, or dedicated RF ICs.






