Inferensys

Glossary

Phase Alignment

The critical calibration of electrical path lengths at the input and output of the carrier and peaking branches to ensure constructive in-phase power combining at the Doherty combiner output.
AI evaluator reviewing output quality on laptop, comparison metrics visible, casual evaluation session.
DOHERTY AMPLIFIER CALIBRATION

What is Phase Alignment?

Phase alignment is the critical calibration of electrical path lengths in the carrier and peaking branches of a Doherty power amplifier to ensure constructive, in-phase power combining at the output combiner.

Phase alignment is the precise synchronization of the electrical delay experienced by RF signals traversing the carrier amplifier path and the peaking amplifier path within a Doherty architecture. This calibration ensures that when the peaking amplifier activates during high signal envelope peaks, its output current arrives at the Doherty combiner in perfect phase coherence with the carrier signal, enabling constructive interference and optimal load modulation.

Misalignment, caused by differences in transistor parasitics, matching networks, or transmission line lengths, results in vector cancellation at the combiner. This degrades power-added efficiency (PAE), reduces output power, and introduces severe AM-PM distortion that cannot be fully corrected by digital predistortion (DPD) alone, making phase alignment a fundamental prerequisite for achieving specified back-off efficiency and linearity.

DOherty Combiner Synchronization

Critical Aspects of Phase Alignment

Phase alignment is the foundational calibration step ensuring the carrier and peaking amplifier outputs combine constructively at the Doherty combiner. Misalignment directly degrades power-added efficiency, output power, and linearity.

01

Electrical Path Length Matching

The fundamental requirement that the carrier and peaking amplifier branches present identical phase delay from the input splitter to the output combiner reference plane. This ensures the peaking amplifier's current injection arrives in-phase with the carrier's output at the Doherty combiner node. Mismatch causes destructive interference, reducing combined output power and forcing the carrier amplifier to dissipate excess energy as heat. Precision is typically maintained within ±5 degrees of phase error across the operating bandwidth.

±5°
Typical Phase Tolerance
02

Input Splitter Phase Balance

The input network, often a Wilkinson divider or hybrid coupler, must deliver signals to the carrier and peaking paths with a precise phase offset. In a symmetric Doherty, this is typically a 90-degree offset to compensate for the impedance inverter in the output path. Any deviation from this nominal phase relationship at the input directly translates to a phase error at the combiner, undermining the active load-pull mechanism. Active phase adjustment circuits are often integrated to calibrate out splitter imbalances.

90°
Nominal Input Offset
03

Phase Coherence vs. Frequency

Phase alignment is inherently frequency-dependent due to the dispersive nature of quarter-wave transformers and device parasitics. While alignment can be perfect at a single center frequency, group delay variations across the modulation bandwidth cause phase dispersion. This limits the instantaneous bandwidth of the Doherty amplifier. Broadband designs employ Klopfenstein tapers or multi-section matching networks to flatten the phase response and maintain constructive combining over wider channels.

< 1 ns
Group Delay Variation Target
04

AM-PM Distortion Interaction

Phase alignment is not a static condition; it interacts dynamically with the amplifier's inherent AM-PM distortion. As the peaking amplifier turns on and its input impedance changes, it can pull the phase of the preceding stages. This dynamic phase modulation adds to the static misalignment, creating a power-dependent combining error. Advanced digital predistortion models must capture this interaction to linearize the composite output effectively.

10-30°
Typical AM-PM Range
05

Output Combiner Symmetry

The physical layout of the Doherty combiner network must enforce geometric symmetry between the two paths. Asymmetric microstrip or stripline routing introduces unequal phase lengths. Even a 1 mm difference in a 30 GHz mmWave design can represent a significant fraction of a wavelength, causing severe phase error. Electromagnetic simulation of the entire output network is mandatory to verify phase balance before fabrication.

λ/100
Layout Precision Requirement
06

Calibration with Vector Network Analysis

Phase alignment is verified and calibrated using a Vector Network Analyzer (VNA) to measure the transmission phase (S21) of each branch independently. The difference in unwrapped phase between the carrier and peaking paths is computed. Phase shifter components or transmission line lengths are then trimmed to null this difference. This calibration must be performed at the amplifier's nominal operating power level to account for hot S22 effects.

S21
Key Measurement Parameter
PHASE ALIGNMENT

Frequently Asked Questions

Critical questions regarding the calibration of electrical path lengths in Doherty amplifier architectures to ensure constructive power combining and optimal efficiency.

Phase alignment is the precise calibration of the electrical path lengths at the input and output of the carrier and peaking amplifier branches to ensure that their output signals combine constructively in-phase at the Doherty combiner output. Without proper alignment, the fundamental currents from the two amplifier paths arrive at the combining node with a phase offset, resulting in destructive interference, reduced output power, degraded power-added efficiency (PAE), and severe AM-AM and AM-PM distortion. The alignment must compensate for the inherent phase shift introduced by the impedance inverter (typically a 90-degree quarter-wave transformer) in the output network, as well as any phase discrepancies in the input splitting network and the transistors' own insertion phases.

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.