What Makes an Oscillator
Every radio transmitter begins with an oscillator — a circuit that generates a steady, repeating electrical signal from nothing more than a DC power supply. Press the transmit button on your transceiver and an oscillator somewhere inside is producing the carrier frequency that gets amplified, modulated and sent to your antenna. Understanding what makes a circuit oscillate is the foundation of understanding every signal source in radio.
Amplifier vs Oscillator — What Is the Difference?
An amplifier takes a signal at its input and produces a larger version of that signal at its output. Remove the input signal and the output goes quiet. An oscillator has no signal input — it generates its own output continuously, powered only by its DC supply. In that sense an oscillator is a converter: it converts DC power from the supply into AC signal power at a specific frequency.
This difference seems mysterious. Where does the signal come from if nothing is feeding it? The answer is feedback. An oscillator is built from an amplifier and a feedback network connected in a loop. Part of the output is routed back to the input in such a way that the circuit reinforces itself — the signal grows, stabilizes and continues indefinitely without any external driving signal.
Think of a microphone placed in front of a loudspeaker connected to an amplifier. If you turn up the volume enough, the amplifier picks up sound from the speaker, amplifies it, feeds it back to the speaker, and the system howls with a sustained tone — even when no one is speaking into the microphone. That is acoustic feedback oscillation. An RF oscillator works by the same principle, but with carefully chosen electronic components instead of microphone and speaker, and at frequencies from a few kilohertz to several gigahertz.
The fundamental oscillator model: an amplifier (gain A) with a portion of its output (factor β) fed back to its own input. When the loop gain Aβ equals 1 and the total phase shift is 0°, the circuit sustains oscillation.
View LargerThe Feedback Loop
To build an oscillator you need two things working together: an amplifying device and a frequency-selective feedback network. The amplifying device provides gain — it makes the signal bigger. The feedback network does two jobs: it selects the frequency at which oscillation occurs, and it determines how much of the output is fed back to the input.
The feedback can be either positive (regenerative) or negative (degenerative). Negative feedback — where the fed-back signal opposes the input — makes amplifiers stable and predictable. It is intentionally used in op-amp circuits to control gain precisely. Positive feedback — where the fed-back signal reinforces the input — is what makes a circuit oscillate. The fed-back signal adds to whatever is already at the input, driving the amplifier harder, producing more output, which feeds back even more strongly, and so on.
In a real oscillator, the feedback network is not just a wire. It is a resonant circuit — typically a combination of inductors and capacitors (an LC tank circuit), or a quartz crystal — that is selective about frequency. The feedback network passes signals near its resonant frequency efficiently while attenuating signals at other frequencies. This selectivity is what makes the oscillator produce a single, clean frequency rather than broadband noise.
The Barkhausen Stability Criterion
In 1921, German physicist Heinrich Barkhausen identified the exact conditions required for sustained oscillation. These conditions are now called the Barkhausen stability criterion, and every oscillator — regardless of its topology, frequency, or technology — must satisfy them to function.
There are two conditions, both of which must be true simultaneously at the frequency of oscillation:
- Loop gain condition: The product of the amplifier gain (A) and the feedback fraction (β) must equal exactly 1. Written mathematically: |Aβ| = 1. This means the signal must return to the input at exactly the same amplitude it started with — no bigger, no smaller.
- Loop phase condition: The total phase shift around the entire loop must equal exactly 0° (or an integer multiple of 360°). The signal must arrive back at the input in phase with itself — reinforcing, not cancelling.
Let us examine each condition in turn.
The Gain Condition: |Aβ| = 1
If the loop gain Aβ is less than 1, the signal loses energy each time it goes around the loop. It grows smaller on every cycle and eventually dies away to nothing. This is a stable but silent condition — the circuit does not oscillate.
If Aβ is greater than 1, the signal grows larger on each cycle. But this cannot continue forever — practical amplifiers have limits. Eventually the transistor or other device saturates or cuts off, effectively reducing its gain until Aβ drops back to 1. In a well-designed oscillator this happens smoothly, and the circuit settles into steady oscillation with Aβ held at 1 by the amplifier's own nonlinearity.
If Aβ equals exactly 1 from the moment of switch-on — theoretically — the oscillator maintains a constant amplitude. In practice, you always start with Aβ slightly greater than 1 to guarantee startup, then rely on natural limiting to bring it back to 1.
The Phase Condition: Total Phase = 0°
Phase shift matters because a signal arriving 180° out of phase with itself will cancel the original, not reinforce it. You would have negative feedback and a stable amplifier, not an oscillator. For oscillation, the signal must arrive back at the amplifier input exactly in phase — completing a full 360° journey around the loop.
In a common-emitter transistor amplifier, the amplifier itself introduces 180° of phase shift between input and output (the output is inverted relative to the input). The feedback network must therefore contribute another 180° of phase shift to make the total 360° — which is the same as 0°. Different oscillator topologies achieve this 180° from the feedback network in different ways: the Colpitts oscillator uses a capacitive voltage divider, the Hartley uses a tapped inductor, and the phase-shift oscillator uses a chain of RC sections.
The three cases of loop gain: if Aβ < 1 the oscillation dies, if Aβ > 1 it grows until limiting occurs, and if Aβ = 1 with correct phase, stable oscillation is maintained.
View LargerYou are designing an LC oscillator using a common-emitter transistor amplifier with a voltage gain of 50 (×50). The feedback network is a capacitive voltage divider that returns 1/50th of the output back to the input, and contributes 180° of phase shift.
Step 1 — Check the gain condition:
A = 50, β = 1/50 = 0.02
Aβ = 50 × 0.02 = 1.0 ✓
Step 2 — Check the phase condition:
Amplifier phase shift: 180° (common-emitter inversion)
Feedback network phase shift: 180°
Total loop phase shift: 180° + 180° = 360° = 0° ✓
Result: Both conditions are satisfied. The circuit will oscillate. In practice you would design the amplifier for a gain slightly above 50 (say 60–70×) to guarantee startup, relying on transistor saturation to reduce the effective gain to 1 once the oscillation is established.
How an Oscillator Starts from Noise
A common puzzle for beginners is this: if an oscillator needs a signal at its input to start oscillating, and there is no external signal source, how does it ever begin? The answer is thermal noise.
Every electronic component at a temperature above absolute zero generates tiny random voltage and current fluctuations called thermal noise (also called Johnson noise). This noise exists across a broad range of frequencies simultaneously. When power is first applied to an oscillator, this noise is present at the input of the amplifier. The amplifier amplifies all frequencies, but the feedback network selects only those components near its resonant frequency and returns them to the input.
These noise components near the resonant frequency are reinforced on each pass around the loop. Because the initial Aβ is slightly greater than 1, the signal grows exponentially. Within microseconds to milliseconds, depending on the Q factor of the resonant circuit, the signal has grown from the noise floor to full amplitude. This growth phase is called the startup transient, and you can observe it on an oscilloscope as the amplitude builds up from zero to steady state right after power is applied.
The startup time depends on how much the loop gain exceeds 1 (a higher excess gain means faster startup) and on the Q factor of the resonant circuit (a higher Q means the energy builds up faster but the bandwidth is narrower). Most practical RF oscillators are fully started within a few microseconds to a few milliseconds of power-on.
Amplitude Limiting — Why the Signal Does Not Grow Forever
If loop gain greater than 1 causes exponential growth, what stops the oscillation from growing without bound? In a real amplifier, gain is not constant — it depends on the signal level. At small signal levels a transistor behaves linearly and provides its full rated gain. As the signal level rises, the transistor begins to enter saturation or cutoff on peaks, clipping the waveform and reducing the effective gain.
This automatic gain reduction is the natural limiting mechanism of most oscillators. As the amplitude grows, the effective gain A falls until Aβ is exactly 1, at which point growth stops and the amplitude stabilises. The oscillator has found its own equilibrium.
The disadvantage of hard limiting (clipping) is distortion — the output waveform is no longer a pure sine wave. For a clean sine wave output, designers use one of several more gentle limiting techniques:
- Diode limiters: Back-to-back diodes across part of the feedback network conduct when the amplitude exceeds a threshold, gently reducing gain.
- Automatic gain control (AGC): A separate AGC circuit detects the output amplitude and adjusts the transistor bias to hold gain at exactly 1.
- FET as variable resistor: A JFET or MOSFET operating in its ohmic region acts as a voltage-controlled resistor in the feedback network, adjusting the feedback fraction to control amplitude precisely.
The Wien bridge oscillator — common in audio oscillators and test equipment — uses an incandescent lamp or thermistor as an AGC element. As amplitude rises, the lamp heats up and its resistance increases, reducing the feedback and stabilising the amplitude. This produces a very low-distortion sine wave.
Oscillator Families Used in Radio
There are many different oscillator topologies, each suited to particular frequency ranges, stability requirements, and applications. The table below summarises the main families used in radio work:
| Oscillator Type | Frequency Element | Typical Range | Stability | Radio Application |
|---|---|---|---|---|
| LC (Colpitts, Hartley) | Inductor + Capacitor | 100 kHz – 1 GHz+ | Moderate | VFO, BFO, LO in receivers/transmitters |
| Crystal (XTAL) | Quartz crystal | 1 kHz – 300 MHz | Very high | Fixed-frequency reference, channel oscillators |
| RC Phase Shift | Resistor + Capacitor chain | 1 Hz – 1 MHz | Low | Audio tones, low-frequency test signals |
| Wien Bridge | RC bridge | 1 Hz – 1 MHz | Low-moderate | Low-distortion audio oscillators, test equipment |
| Relaxation (astable multivibrator) | RC charging/discharging | 0.1 Hz – 1 MHz | Low | Square wave clock sources, timing circuits |
| Voltage-Controlled (VCO) | Varactor diode + LC | 1 MHz – 10 GHz | Moderate (used inside PLL) | PLL synthesizers, FM demodulation |
| Phase-Locked Loop (PLL) | Crystal reference + VCO | 1 kHz – 10 GHz | High (crystal-derived) | Modern synthesized transceivers |
| Direct Digital Synthesis (DDS) | Digital numerics + DAC | DC – ~1 GHz | Very high (crystal-derived) | SDR, modern computer-controlled transceivers |
Why Oscillators Matter to Radio Operators
A radio operator who understands oscillators can diagnose many common faults and make better equipment choices. Here are the most important places oscillators appear in your shack:
The transmitter carrier oscillator sets the frequency you transmit on. In older radios this was a crystal that fixed you to a specific channel. In modern transceivers it is a synthesizer that can be set to any frequency in the amateur bands with 10 Hz or better resolution. The stability of this oscillator determines how accurately your signal is on the frequency you intend.
The beat frequency oscillator (BFO) is a low-level oscillator in SSB and CW receivers that beats against the IF signal to produce audio. Its frequency offset from the IF determines whether you hear the upper or lower sideband, and its stability directly affects the pitch of received audio. A drifting BFO makes CW sound wobbly and SSB voices sound as if they are talking through a fan.
The local oscillator (LO) in a superhet receiver mixes with the incoming signal to produce the intermediate frequency (IF). If the LO drifts, the station you are listening to drifts out of the IF passband and disappears. LO phase noise — rapid random frequency fluctuations — directly limits the receiver's ability to separate nearby signals (called reciprocal mixing).
The reference oscillator in GPS receivers, frequency counters, spectrum analyzers and precision test equipment determines the absolute frequency accuracy of the instrument. High-end equipment uses temperature-controlled crystal oscillators (OCXOs) that are accurate to parts per billion.
The signal generator in your shack is essentially a precision oscillator with a calibrated output level control. When you align a receiver or test a filter, the accuracy of your results depends entirely on the frequency accuracy and stability of the signal generator's oscillator.
Frequently Asked Questions
If Aβ must equal exactly 1, how do you achieve that precision in practice?
You do not design for exactly 1 — you design for slightly greater than 1 (typically 1.5 to 3) and rely on the amplifier's natural nonlinearity to reduce the effective gain to 1 once the amplitude builds up. The amplitude stabilises automatically at the level where the nonlinear gain reduction brings Aβ to 1. For very low-distortion applications you use an active AGC circuit to control the gain more precisely.
Why is phase shift so critical? What happens if it is slightly off?
If the total loop phase is not exactly 0°, the oscillation frequency shifts slightly from the resonant frequency of the tank circuit until the phase shift of the tank adds the missing correction. This is called frequency pulling. A reactive load connected to the oscillator output can also pull the frequency by changing the phase in the loop. This is why oscillator buffers (isolation amplifiers) are important — they prevent the load from pulling the frequency.
What is the difference between a free-running oscillator and a VCO?
A free-running oscillator runs at a frequency determined entirely by fixed LC components or a crystal. A voltage-controlled oscillator (VCO) includes a varactor diode — a variable-capacitance diode — whose capacitance changes with the applied reverse-bias voltage. By varying the control voltage you tune the oscillator frequency. VCOs are the tunable elements inside phase-locked loops and FM modulators.
Can I use an oscillator as a transmitter directly?
A direct-conversion transmitter using a crystal-controlled oscillator directly modulated and amplified is technically possible and has been used in simple CW transmitters. However, the output power of the oscillator itself is typically very low (milliwatts), so one or more power amplifier stages are always needed to reach useful power levels. Also, keying an oscillator directly can cause chirp — frequency change at the instant of keying — which is why most transmitters key a later amplifier stage rather than the oscillator.
What is phase noise and why does it matter?
Phase noise is rapid, random fluctuation of the oscillator's instantaneous frequency around its nominal frequency. On a spectrum analyzer it appears as a skirt of noise that broadens the signal on either side of the carrier. In a receiver, the local oscillator's phase noise is mixed with the IF signal — a strong nearby station can mix with the phase noise skirt of the LO and appear as noise on your desired frequency. This limits how close you can receive a weak signal to a strong one. High-Q resonators (crystals, cavity resonators) and well-designed VCOs have lower phase noise.
Test Your Knowledge
Answer the questions below to check your understanding. Every answer can be found in the lesson above.