Why does reducing conduction angle improve efficiency?

Efficiency loss in a linear amplifier comes from current flowing through the transistor when the drain-source voltage is not zero. In Class A, current flows continuously (360 degrees), and during half of each cycle the transistor dissipates significant power as heat because voltage and current are both non-zero simultaneously. In Class B, the transistor conducts for only 180 degrees, turning off during the other half cycle. During the off period, no DC power is consumed. The output waveform is a half-sine, and a tuned output network reconstructs the full sine wave. The theoretical efficiency improves from 50% (Class A) to 78.5% (Class B). Class C pushes further: conduction less than 180 degrees means even less overlap between voltage and current, approaching 100% efficiency as the conduction angle approaches zero, but the output power also approaches zero.

How do switching-mode amplifiers achieve near-100% efficiency?

Switching-mode amplifiers (Class D, E, F) treat the transistor as an ideal switch rather than a current source. When the switch is on, the voltage across it is zero (no power dissipation). When the switch is off, the current through it is zero (no power dissipation). All power is delivered to the load network with zero loss in the transistor. Class E achieves this with a single transistor and a carefully designed output network that shapes the voltage waveform to reach zero and have zero slope at the moment the transistor turns on (zero-voltage switching). Class F uses harmonic tuning to shape the drain voltage into a square wave and the drain current into a half-sine, minimizing their overlap. In practice, finite switching speed, on-resistance, and parasitic capacitance limit efficiency to 80 to 95 percent.

Which class should I use for my application?

Class A: when linearity is paramount and efficiency is secondary. Lab instruments, small-signal gain stages, receive-chain amplifiers. Class AB: the workaround for most RF power amplifiers. Biased slightly above cutoff for a good balance of linearity (suitable for amplitude-modulated signals with DPD) and efficiency (typically 30 to 45% with modern waveforms). Class C: only for constant-envelope signals like FM, FSK, or CW radar where amplitude linearity is not needed. Class E and F: for high-efficiency constant-envelope transmitters (IoT, RFID readers, radar pulsed PAs) and as the final stage in outphasing or Doherty architectures where the linearity is restored by the combining network. Class D: primarily used at frequencies below 100 MHz due to switching speed limitations.

Active Components

Amplifier Classes

A cellular base station PA must amplify a 256QAM OFDM signal with 10 dB PAPR while maintaining −45 dBc ACLR. Class A would deliver perfect linearity but only 25% efficiency at the average power level, turning 75% of the DC input into heat. Class AB with DPD achieves the same linearity at 35 to 45% efficiency. A Doherty PA combining two Class AB stages reaches 50%. The amplifier class defines the fundamental operating mode: how the transistor is biased, what fraction of each RF cycle it conducts, and the resulting trade-off between linearity and DC-to-RF conversion efficiency. Every PA design begins with this choice.

Category: Active Components

Fundamental Trade-off: Linearity vs. efficiency

Determined By: Bias point & conduction angle

The Complete Amplifier Class Spectrum

Class	Conduction	Max η	Linearity	Mode	Application
A	360°	50%	Excellent	Current source	Small-signal, instruments
AB	180 to 360°	50 to 78.5%	Good (with DPD)	Current source	Most RF PAs, cellular BS
B	180°	78.5%	Moderate	Current source	Push-pull audio, some RF
C	<180°	Up to 90%	Poor	Current source	FM, CW radar, RFID
D	Switch (50%)	100% (ideal)	None (switching)	Switch	<100 MHz, DC-DC converters
E	Switch (ZVS)	100% (ideal)	None (switching)	Switch	IoT, RFID, radar pulse
F	Harmonic tuned	100% (ideal)	None (switching)	Switch	High-eff. narrowband PA

Class A efficiency:
η_max = 50% (at full output swing)
At back-off BO: η = 50% / 10^BO/10
At 8 dB back-off: η = 50/6.3 = 7.9%

Class B efficiency:
η_max = π/4 = 78.5%
At back-off: η = (π/4) × V_out/V_max

Class C efficiency (theoretical):
η → 100% as conduction angle → 0°
But output power → 0 simultaneously

Common Questions

Frequently Asked Questions

Why does less conduction improve efficiency?

Class A: current flows 360°, transistor dissipates power when V and I overlap. Class B: 180° conduction, off half the time, η jumps to 78.5%. Class C: <180°, even less overlap, approaching 100% as angle → 0 (but power → 0 too).

How do switching modes reach 100%?

Transistor acts as switch: ON = zero voltage drop, OFF = zero current. No V×I overlap = no dissipation. Class E: ZVS shaping. Class F: harmonic tuning for square voltage. In practice: 80 to 95% due to finite switching speed and parasitics.

Which class for my design?

Class A: max linearity (instruments). AB: most RF PAs (with DPD for OFDM). C: constant-envelope only (FM, FSK). E/F: high-efficiency narrowband (IoT, RFID, radar). D: below 100 MHz. Doherty combines AB stages for best wideband efficiency.

Related Terms

← Amplifier Design Antenna Gain →

PA Architecture

Amplifier Class Selector

Enter signal type (CW, FM, OFDM), required linearity (ACLR/EVM), and efficiency target. Get the recommended amplifier class and architecture for your application.

Select Your Class