How does an RF oscillator work?
An oscillator converts DC power into a periodic RF signal. In the feedback view, oscillation starts when the loop gain exceeds unity at a frequency where loop phase is a multiple of 360° (the Barkhausen criterion), then nonlinearity reduces the gain to exactly one at steady state. In the negative-resistance view, an active device presents Re{Zin} < 0 that cancels resonator loss.[1][2]
What are the classic topologies?
- Colpitts: capacitive divider feedback; very common.
- Hartley: tapped-inductor feedback.
- Clapp: Colpitts with series capacitor, improving stability.
- Pierce: the standard crystal oscillator in digital ICs.
- Cross-coupled LC: differential negative-resistance pair; dominant in integrated VCOs.
- Ring oscillators: no resonator; wide tuning, poor phase noise.
- Dielectric-resonator and YIG oscillators at microwave frequencies.
How is startup ensured?
Design small-signal loop gain to 2–3× unity (or |negative resistance| comfortably above loss resistance) across process, temperature and frequency, so that oscillation starts reliably. Amplitude is then set by device limiting or automatic level control.[1]
What specifications matter?
Frequency, tuning range, phase noise, output power, harmonics, frequency stability with temperature and supply (pushing), sensitivity to load (pulling), and startup time.
Frequently asked questions
Why does an oscillator need nonlinearity?
To limit amplitude so loop gain settles at exactly one.
What is frequency pulling?
Frequency change caused by load impedance variation; buffers reduce it.
What is pushing?
Frequency change with supply voltage, in Hz/V.
See also
References
Online sources accessed October 21, 2026; printed works cited by edition. See all sources.
- “RF Microelectronics, 2nd ed.”. B. Razavi, Prentice Hall. 2012.
- “Microwave Engineering, 4th ed.”. D. M. Pozar, Wiley. 2012.