class f amplifier
A Class F amplifier pushes efficiency past what a single-tuned Class C stage can reach by shaping the output network to control not just the fundamental frequency but a whole set of harmonics, forcing the drain or collector voltage toward a square wave while the current stays a half-sine — the two waveforms overlap as little as possible, and since instantaneous power dissipation is voltage times current, minimizing that overlap is what pushes theoretical efficiency toward 100% without requiring the hard on/off switching that Class D and E rely on.
Shaping the voltage waveform, not switching the device, is the mechanism — which is the key difference from Class D and E. Class D and Class E both force the active device fully on or fully off to avoid ever having significant voltage and current present simultaneously; Class F instead keeps the device operating as a conventional biased amplifier and uses a multi-resonant output network to reshape the voltage waveform the device sees into something closer to a square wave, so the overlap reduction comes from waveform engineering in the matching network rather than from the device's switching behavior, which is why Class F can be implemented with the same transistor technology as a Class B or C design, just with a more elaborate output network.
Odd and even harmonics get opposite treatment, and that split is the whole design recipe. A true Class F output network presents a high impedance at odd harmonics of the carrier (letting them contribute to flattening the voltage peak into a square-ish shape) and a near-short-circuit impedance at even harmonics (suppressing their contribution so the current waveform stays close to an ideal half-sine) — practical designs typically control just the second and third harmonics with dedicated resonant stubs or lumped LC traps, since each additional harmonic controlled adds real efficiency but also real network complexity and bandwidth cost.
It inherits Class C's narrowband problem and makes it worse, not better. Because the output network needs multiple precisely-tuned resonances instead of Class C's single tank, Class F amplifiers are even more sensitive to component tolerance, load impedance, and frequency drift than a Class C stage, and the harmonic-shaping effect collapses outside a narrow design bandwidth — this tradeoff is accepted in applications like cellular base station power amplifiers and dedicated ISM-band transmitters where the operating frequency is fixed and known at design time, never in general-purpose or wideband RF amplification.
GaN device technology made Class F genuinely practical rather than a mostly-theoretical curiosity. Gallium-nitride RF power transistors from vendors like Wolfspeed and Qorvo combine the high breakdown voltage and low parasitic capacitance that harmonic-tuned output networks need to actually deliver their theoretical gains at microwave frequencies, which is why Class F and its inverse variant (Class F-1, which swaps the roles of voltage and current shaping) have become mainstream techniques in high-efficiency cellular infrastructure power amplifiers over the last two decades rather than staying a textbook topic.
Read a Class F amplifier through a *harmonic-engineering* lens rather than a *switching* lens: nothing about the active device changes from a conventional biased amplifier, but the output network is deliberately built to control multiple harmonics instead of just the fundamental, reshaping the voltage waveform the device experiences so that voltage and current spend less time being large at the same moment — recovering efficiency through waveform geometry rather than through hard switching.