Power, Linearity, and Distortion Practical Power Topics Informational

How do I design a reactive combiner for a high power narrowband transmitter?

Q: When should I use reactive vs. Wilkinson?

Use reactive when: bandwidth is less than 20%, all PAs are identical and well-matched, PA failure is unlikely or acceptable (no need for graceful degradation), and maximum efficiency is the priority (broadcast transmitters, base stations with narrowband signals). Use Wilkinson when: bandwidth is greater than 20%, PA failures must be handled gracefully (military, satellite), PA modules have significant unit-to-unit variation, and system stability is more important than 0.1-0.3 dB better efficiency.

Q: What about hybrid combiners?

Hybrid combiners (quadrature hybrids, Lange couplers) provide a middle ground: moderate isolation (6-10 dB between ports for a quadrature hybrid), moderate bandwidth (20-50% for a Lange coupler), low combining loss (0.1-0.3 dB), and partial graceful degradation (one PA failure reduces output by 3 dB for 2-way, the surviving PA's reflected power is absorbed by the hybrid's terminated port). Hybrid combiners are widely used in: wideband military transmitters, multi-carrier base stations, and instrumentation.

Q: How do I handle the high peak power?

High-power reactive combiners must handle: the peak voltage (V_peak = sqrt(2 × P_peak × Z₀)) which determines the voltage breakdown requirement, and the peak current (I_peak = V_peak / Z₀) which determines the conductor sizing. For 1 kW CW into 50 ohms: V_peak = 316 V, I_peak = 6.3 A. For 1 MW pulsed: V_peak = 10 kV! At these power levels: use low-loss, high-voltage-rated transmission lines (air-dielectric coax or waveguide), ensure adequate spacing for voltage breakdown prevention, and use connectors rated for the peak power.

Designing a reactive combiner for a high power narrowband transmitter uses lossless reactive elements (transmission line transformers, quarter-wave lines, and resonant circuits) to combine the output of multiple power amplifiers without the resistive isolation (and associated losses) of a Wilkinson combiner. The advantages of reactive combining are: zero combining loss in the ideal case (no power is dissipated in isolation resistors because there are none), and the full power from all PAs is delivered to the load. The disadvantages: no isolation between the PA ports (if one PA fails or changes impedance, it directly affects all other PAs), and a narrow bandwidth (the reactive elements are frequency-dependent, limiting the combiner to approximately 5-15% bandwidth). Reactive combiner types: quarter-wave transformer combiner (N PAs feed into a common node through quarter-wave transmission lines of impedance Z0 × sqrt(N); each PA sees 50 ohms at the common node; bandwidth approximately 10-20% for VSWR less than 1.5:1), resonant cavity combiner (a high-Q resonant cavity with N coupling probes; the cavity mode distributes power from each PA to the output with minimal loss; bandwidth approximately 1-5%, limited by the cavity Q; used for: very high power CW transmitters (FM broadcast, industrial heating)), and hybrid-mode combiners (a circular waveguide combiner where each PA couples to a waveguide mode; the modes combine at the output; very low loss and high power handling; used for: satellite downlink transmitters and high-power radar).

Category: Power, Linearity, and Distortion

Updated: April 2026

Product Tie-In: Power Amplifiers, Combiners, Loads

Reactive Power Combiner Design

Reactive combiners are preferred when: the PA efficiency is paramount (every fraction of a dB of combiner loss is wasted power), the bandwidth requirement is narrow enough to accommodate the reactive elements, and the operating conditions are well-controlled (no PA failures expected).

Design Considerations

Load pulling: Without isolation between ports: a change in one PA's output impedance (due to gain change, failure, or load mismatch) directly changes the impedance seen by all other PAs. This can cause: oscillation, gain change, and failure cascade. Mitigation: use closed-loop power control on each PA to stabilize its output impedance
Narrow bandwidth: The quarter-wave transformer combiner's bandwidth is limited by the transformer's electrical length: it is exactly a quarter-wave at only one frequency. At other frequencies: the impedance match degrades. For wider bandwidth: use stepped impedance transformers or multiple-section matching

Reactive Combiner Parameters

Quarter-wave combiner arm impedance: Z_arm = Z₀ × √N
For N=4, Z₀=50Ω: Z_arm = 100Ω
Bandwidth (VSWR < 1.5): BW ≈ 30-40% for single-section
Combiner loss (ideal): 0 dB (no resistive elements)
Practical loss: 0.03-0.1 dB (conductor and dielectric losses only)

Common Questions

Frequently Asked Questions

When should I use reactive vs. Wilkinson?

Use reactive when: bandwidth is less than 20%, all PAs are identical and well-matched, PA failure is unlikely or acceptable (no need for graceful degradation), and maximum efficiency is the priority (broadcast transmitters, base stations with narrowband signals). Use Wilkinson when: bandwidth is greater than 20%, PA failures must be handled gracefully (military, satellite), PA modules have significant unit-to-unit variation, and system stability is more important than 0.1-0.3 dB better efficiency.

What about hybrid combiners?

Hybrid combiners (quadrature hybrids, Lange couplers) provide a middle ground: moderate isolation (6-10 dB between ports for a quadrature hybrid), moderate bandwidth (20-50% for a Lange coupler), low combining loss (0.1-0.3 dB), and partial graceful degradation (one PA failure reduces output by 3 dB for 2-way, the surviving PA's reflected power is absorbed by the hybrid's terminated port). Hybrid combiners are widely used in: wideband military transmitters, multi-carrier base stations, and instrumentation.

How do I handle the high peak power?

High-power reactive combiners must handle: the peak voltage (V_peak = sqrt(2 × P_peak × Z₀)) which determines the voltage breakdown requirement, and the peak current (I_peak = V_peak / Z₀) which determines the conductor sizing. For 1 kW CW into 50 ohms: V_peak = 316 V, I_peak = 6.3 A. For 1 MW pulsed: V_peak = 10 kV! At these power levels: use low-loss, high-voltage-rated transmission lines (air-dielectric coax or waveguide), ensure adequate spacing for voltage breakdown prevention, and use connectors rated for the peak power.

Keep Reading