Does Reflected Power Really Flow Back? Fields, Waves and Net Power
Does Reflected Power Really Flow Back? Fields, Waves and Net Power
Yes: a mismatch creates a measurable reverse-travelling electromagnetic wave. But “power comes back” is directional shorthand, not a complete explanation of loss, standing waves, transmitter stress or radiation.
A directional meter can separate forward and reverse components, and a time-domain instrument can observe an echo after the line delay. The reverse wave is therefore not imaginary bookkeeping. The trap is treating RF power as a fluid parcel that an antenna accepts or rejects. The physical system is an electromagnetic boundary-value problem: fields, voltage and current must satisfy the load impedance.
The power-flow principle: a mismatch creates a real backward-propagating mode. Forward and reverse fields combine into the actual line voltage and current; for a lossless single-mode line with real Z0, their directional average powers subtract to give net power flow.
Power Travels in the Electromagnetic Field
In coax operating in its ordinary TEM mode, the electric field extends mainly between centre conductor and shield, while the magnetic field circles the current. The conductors impose boundary conditions and carry surface current, but the useful RF energy flow is described by the field in the dielectric.
The instantaneous energy-flux density is the Poynting vector:
S(t) = E(t) × H(t)
For sinusoidal phasors, the time-average flow density is:
〈S〉 = ½ Re{E × H*}
Integrating that vector over the coax cross-section gives the average modal power. NIST’s work on characteristic impedance and power makes this field-to-circuit normalization explicit. Saying that reflected power “flows back” means the reverse modal field has an average Poynting component toward the source. It does not mean electrons carrying labelled watts reverse course from the antenna to the final transistor.
Why a Mismatch Creates a Reverse Wave
A source launches a forward wave on a uniform line. For a lossless line, its voltage and current ratio is Z0. If the load equals Z0, that wave alone satisfies the load boundary. If the load impedance differs, the total terminal voltage and current must instead satisfy V/I = ZL. A reverse wave supplies the necessary additional voltage and current.
At the load plane:
ΓL = V−/V+ = (ZL − Z0)/(ZL + Z0)
ΓL is complex. Its magnitude sets the reflected-voltage fraction; its phase sets the spatial position of voltage and current maxima. With a passive load and real positive Z0, the reflected-power fraction is |ΓL|2, and:
SWR = (1 + |Γ|)/(1 − |Γ|)
The antenna does not inspect incoming watts and decide which ones to keep. Charge and current at the discontinuity establish fields that satisfy Maxwell’s equations and the terminal impedance. In a pulse experiment, that response launches an echo after the incident edge reaches the load. In continuous operation, the same physics appears as a steady forward-plus-reverse solution.
One Voltage, One Current, Two Useful Components
At any position there is one measurable total voltage and one measurable total current. For a uniform line using a common coordinate convention:
V(z) = V+e−γz + V−e+γz
I(z) = [V+e−γz − V−e+γz]/Z0
The reverse-current term carries a minus sign because its reference direction is opposite. The two wave components superpose linearly. They do not collide like material objects, but their coherent cross-terms shape local voltage, current, electric energy and magnetic energy.
On a lossless line, relative phase changes with distance. This creates fixed voltage and current envelopes in sinusoidal steady state: the standing-wave pattern. The words “standing wave” describe the envelope, not frozen electromagnetic energy.
Why Average Net Power Is Forward Minus Reverse
Average circuit power at a plane is obtained from voltage and current, with the numerical factor depending on whether peak or RMS phasors are used:
Pavg = ½ Re{VI*} for peak phasors
Insert the forward and reverse waves. For a lossless single-mode line with real Z0, the coherent cross-terms are purely reactive in this average-real-power calculation. They cancel from the real part, leaving:
Pnet = P+ − P−
The cross-terms have not become physically irrelevant. They create the spatial voltage and current pattern and the alternating storage of electric and magnetic energy. They simply do not add a third independent term to net average real power on this idealized line.
State the assumptions. With appreciably lossy lines, complex characteristic impedances, multiple modes or generalized microwave reference impedances, the simple watt interpretation needs more careful power-wave normalization. Kurokawa’s power-wave formulation and NIST’s modal power treatment address those cases.
Standing Waves Still Exchange Energy
An open-circuit termination illustrates the difference between large fields and zero average load power. For an ideal lossless open:
|Γ| = 1, P+ = P−, Pnet = 0
The line can have high voltage at the open and strong current elsewhere, yet the open absorbs no average power. During each RF cycle, energy moves between electric and magnetic storage and instantaneous power at a plane can change sign. Averaged over the cycle, equal forward and reverse powers cancel.
A short circuit likewise reflects completely in the ideal model but with the opposite voltage-reflection phase. Its terminal voltage is zero and terminal current can be large. “No absorbed average power” does not mean “no field energy” or “no voltage/current stress.”
What a Directional Wattmeter Actually Separates
For real Z0, the wave voltages can be recovered from total voltage and current:
V+ = (V + Z0I)/2
V− = (V − Z0I)/2
A directional coupler approximates this separation using coupled electric and magnetic samples. One output is arranged to reinforce the forward sample while cancelling the reverse sample; another does the opposite. Keysight’s coupler documentation describes incident and reflected ports in exactly these directional terms.
Real couplers are imperfect. Directivity limits how well a large forward wave is rejected from the reverse channel. Coupling flatness, detector calibration, connector mismatch, harmonic content and load impedance add error. A small REV reading beside a very large FWD reading may be dominated by finite directivity.
| Meter quantity | What it means at that plane | What it does not prove |
|---|---|---|
| Forward power | Power associated with the calibrated forward wave component. | Transmitter DC input, antenna radiation or field strength. |
| Reverse power | Power associated with the calibrated reverse wave component. | Power dissipated in the PA or power permanently lost. |
| Forward minus reverse | Net average power toward the load for the stated assumptions. | Antenna radiated power after downstream line, tuner and ground losses. |
| SWR | Reflection magnitude relative to the chosen reference impedance. | Efficiency, gain, common-mode current or radiation pattern. |
The Reference Plane Changes the Reading
A meter at the transmitter, after a tuner and at the antenna feedpoint can report different forward and reverse values without contradiction. A lossy line attenuates the wave in both directions. A tuner transforms impedance. Connectors and filters introduce additional discontinuities. The reflection coefficient also rotates with electrical length even when its magnitude is nearly constant on a low-loss line.
For example, a shack tuner may present 1:1 SWR to the transmitter while the tuner-to-antenna line still has a high SWR. The tuner changed the boundary at its input; it did not remove the antenna-side forward and reverse waves. This is why every measurement needs a reference plane.
Does the Reverse Wave Reach the Transmitter?
It propagates toward the source end, but its eventual destination is not encoded by the word “reverse.” It encounters the complete network there: tuner, low-pass filter, transformer, coupler, relay, PA output network, protection circuit or perhaps a circulator and load.
It can be partly transmitted into that network, partly dissipated and partly re-reflected. A circulator can route it into a termination. The PA’s driven nonlinear behavior may depend on reverse-wave phase and amplitude. “The reflected watts hit the finals” is therefore poor failure analysis. The real concerns are device voltage and current, heating, load-line excursion, stability, foldback and arcing.
The companion article Reflected Power and Re-Reflection: Source, Load, Phase and Loss develops that boundary and multiple-reflection question in detail.
Reflection Is Not Loss
An ideal discontinuity can reflect without dissipating. Loss converts electromagnetic energy into heat. Real coax adds conductor and dielectric loss; tuners, filters, transformers, coils, ferrites, connectors, antenna conductors and ground systems add more.
High SWR raises peak voltage and current and makes power travel extra distance in a real line. Consequently it increases line loss for a given net delivered power compared with matched operation. The additional dissipation is caused by finite conductor and dielectric loss acting on the higher travelling-wave amplitudes—not by a mythical absorber labelled “reflection.”
Reflection and Common Mode Are Different Modes
In intended coax differential mode, current on the outside surface of the centre conductor is accompanied by equal-and-opposite current on the inside surface of the shield. The fields are mainly between those surfaces. A load mismatch reflects that differential mode. No current on the outside surface of the shield is required.
Common-mode current is net longitudinal current on the feedline exterior relative to the wider environment. Antenna imbalance, coupling, asymmetric routing and an uncontrolled return structure can convert differential energy into that external mode. Ordinary differential reflection does not, by itself, require such conversion.
| Observation | What it establishes | What it does not establish |
|---|---|---|
| High SWR | A differential impedance mismatch exists at the stated plane. | That common-mode current exists. |
| 1:1 SWR | The input impedance is matched to the chosen reference at that plane. | That feedline-exterior current is zero. |
| Clamp-measured exterior current | An external/common-mode path participates. | That differential SWR must be high. |
| Impedance changes after adding a choke | The choke altered an exterior current path or coupling. | That a choke’s intended function is to absorb reflected differential power. |
A balanced, well-choked system can have high SWR with negligible feedline-exterior current. A matched asymmetric antenna can have serious common mode. Use a directional line measurement for differential reflection and a clamp-current measurement for net exterior current; one does not replace the other.
How to Demonstrate the Reverse Wave
- Use a calibrated VNA. Measure complex S11 at a defined plane. Magnitude gives reflection; phase locates the impedance transformation along the line.
- Use time-domain gating or a TDR. A step or pulse shows an echo after propagation delay and can locate discontinuities.
- Use a dual-directional coupler. Confirm that its directivity and power rating are adequate, then compare incident and reflected ports.
- Move the reference plane mathematically or physically. Reflection phase rotates with line length; loss changes magnitude. The inferred load should remain consistent after proper de-embedding.
- Test common mode separately. Place a current probe around the entire coax so differential currents cancel in the probe and net exterior current remains.
These measurements observe different aspects of the same installation. Agreement between complex S11, time-domain location and directional power is much stronger evidence than a single SWR number.
A Practical Model of Reflected Power
Yes, reflected RF really propagates from the impedance discontinuity toward the source as a reverse-travelling electromagnetic mode. It can be isolated by a directional coupler and observed as a delayed echo. The language is physically meaningful.
It is still only the beginning of the explanation. Forward and reverse fields superpose into one voltage and current at each point. Their interference creates standing-wave envelopes and reactive-energy exchange. Under the usual lossless real-Z0 assumptions, their average directional powers subtract to give net power flow.
Reverse travel is real; “returned watts” are not a fluid. Reflection is a boundary-condition response, loss is dissipation, radiation is an antenna property, and common mode is a separate current path.
Primary technical references
- NIST — Characteristic Impedance, Power and Causality
- K. Kurokawa — Power Waves and the Scattering Matrix
- Keysight — S-Parameters and Two-Port Measurements
- Keysight — Directional Couplers and Measurement Limits
- Keysight — Time-Domain Analysis Using a Network Analyzer
- Tektronix — Understanding and Applying Time-Domain Reflectometry
Mini-FAQ
- Does reflected power really travel toward the transmitter? Yes. It is a measurable reverse-travelling wave component with field energy and directional average power.
- Is reflected power lost? No. Reflection redirects a wave; finite conductor, dielectric and component resistance dissipate energy.
- Can forward and reverse power be subtracted? At one defined plane on a single-mode line with real reference impedance, their difference is net average power toward the load.
- Does that difference equal radiated power? Not automatically. Downstream feedline, matching, conductor, ground and structural losses still matter.
- Does high SWR prove common mode? No. Differential reflection and exterior/common-mode current are distinct modal quantities.
- Why can meter readings change along the coax? Loss changes wave magnitudes, phase rotates with distance, and tuners or discontinuities change boundary conditions.
- What is the clearest proof of a reverse wave? A calibrated complex reflection measurement, a time-delayed TDR echo or a directional-coupler measurement with adequate directivity.