The Doherty Power Amplifier: Achieving High Efficiency in Modern RF Systems

Wiki Article

Following up on the basics of RF power amplifier classes, one architecture stands out for its practical impact on real-world wireless systems: the Doherty Power Amplifier (DPA). Invented by William H. Doherty in 1936 for vacuum tube transmitters, this design has experienced a renaissance in the era of 4G, 5G, and beyond. It addresses a key challenge—maintaining high efficiency not just at peak power, but across a wide range of output levels.

In mobile networks, signals rarely operate at maximum power. Modern modulation schemes (like OFDM in LTE/5G) have high peak-to-average power ratios (PAPR), often 6–12 dB. This means traditional linear amplifiers (Class AB) run most of the time at "backed-off" power, where efficiency plummets. The Doherty architecture cleverly solves this.

How the Doherty Amplifier Works

The classic Doherty configuration uses two amplifiers:

• Main (Carrier) Amplifier: Typically biased in Class AB for good linearity, operates alone at low to medium power levels.
• Auxiliary (Peaking) Amplifier: Biased in Class C (deeper cutoff), remains off at low power and turns on only when needed for higher output.

These are combined via a specialized power combiner network, usually involving quarter-wave transmission lines (or equivalents at lower frequencies).

Key Phases of Operation:

1. Low-Power Region (up to ~6 dB back-off):
   • Only the main amplifier is active.
   • The impedance seen by the main amp is high (thanks to the quarter-wave line transforming the load from the off-state peaking amp).
   • This "load modulation" allows the main amp to reach saturation earlier, boosting efficiency.
2. Transition Region:
   • As input drive increases, the peaking amp begins conducting.
   • Its contribution gradually modulates the load impedance downward.
3. High-Power Region (near peak):
   • Both amplifiers are fully on and contribute equally.
   • Load impedance normalizes to 50 Ω (or design value), delivering maximum output.

This dynamic load modulation enables efficiency to remain high (often 40–50% average) even with high-PAPR signals, compared to 10–20% for standalone Class AB.

Advantages and Trade-Offs

• Efficiency: Average efficiency can be 2–3x higher than Class AB for typical cellular signals.
• Linearity: Good baseline, but often requires digital predistortion (DPD) for stringent spectral masks.
• Bandwidth: Traditional designs are narrowband due to quarter-wave lines; modern variants use broadband techniques.
• Complexity: More components and careful phasing required compared to single-amp designs.

Variations and Modern Implementations

• Symmetric Doherty: Main and peaking amps identical (simpler design).
• Asymmetric Doherty: Peaking amp larger for higher PAPR signals (common in base stations).
• Three-Way Doherty: Adds a second peaking amp for even better back-off efficiency (up to 12 dB).
• Integrated Doherty: Many MMICs/RFICs now include Doherty internals.
• GaN Advantage: Gallium Nitride transistors enable high-power, high-frequency Dohertys with excellent efficiency.

Real-World Applications

Doherty amplifiers dominate macro base stations for cellular networks (e.g., Nokia, Huawei, Ericsson equipment). They're also found in:

• Broadcast transmitters
• Satellite uplinks
• Radar systems
• Some high-efficiency handset PAs (combined with envelope tracking)

As 5G evolves to higher frequencies and massive MIMO (more antennas = more PAs), efficiency remains critical for thermal management, power consumption, and operational costs. The Doherty architecture, often enhanced with GaN devices and digital linearization, continues to be a workhorse, customized products like those offered by ZR Hi-Tech are available..

While switching classes (E/F) push theoretical efficiency higher, Doherty offers a practical balance of efficiency, linearity, and bandwidth for amplitude-modulated signals. It's a prime example of how clever topology can outperform raw device improvements.

If you're working in RF power amplifier, experimenting with a Doherty prototype (even in simulation tools like Keysight ADS or AWR) is a great way to see load modulation in action. The principles haven't changed much since 1936—but the performance certainly has!