SWR, Reflected Power and Real Feed-Line Loss
SWR, Reflected Power and Real Feed-Line Loss
Reflected power is not automatically power destroyed, yet mismatch is not imaginary. The useful engineering question is where the waves are measured, what the source and tuner do with them, and where real power is finally accepted or dissipated.
Here is the story that deserves to be challenged: “The meter shows 50 watts reflected, therefore 50 watts have vanished as heat.” That is not valid power accounting. Forward and reflected readings describe travelling-wave components at a stated reference plane. To find the real loss, we must follow both waves through the line, tuner, source network and antenna.
A Classroom Generator Model Is Not a Transmitter Specification
A voltage source with a fixed 50-ohm series resistance is a valid Thevenin model. It demonstrates maximum-power transfer cleanly: a conjugately matched load receives the maximum available power, while a different load receives less. The mathematics is not the problem. The mistake is assuming that every HF transmitter must behave like that linear source under every load.
A radio labelled for a 50-ohm load is telling us the intended external load condition. It does not, by itself, state that a literal 50-ohm resistor sits in series with the RF output or that the small-signal output impedance is 50 ohms. A practical transmitter may include an output matching network, filter, feedback, drive control and reflected-power protection. Its voltage, current, efficiency and generated forward wave can all change as the load changes.
Do not replace one simplification with another. A transmitter is not automatically an ideal zero-ohm voltage source either. Its response to mismatch belongs to its circuit, control system, operating class and protection limits. The manual or a controlled load-pull test—not the connector label—settles that question.
Forward and Reflected Waves Keep the Books
For a uniform line with a real characteristic impedance Z0, the load reflection coefficient is:
ΓL = (ZL − Z0) / (ZL + Z0)
SWR = (1 + |ΓL|) / (1 − |ΓL|)
Preflected / Pincident = |ΓL|²
At that same load reference plane, and with the usual real-impedance power-wave normalization, the net power accepted by a passive load is:
Paccepted = Pincident − Preflected = Pincident(1 − |ΓL|²)
This equation says where power crosses the antenna terminals. It does not say that the reflected part has been dissipated. Nor does accepted power mean radiated power: conductor, dielectric, loading, ground and matching loss can consume part of what the antenna accepts.
Take a 6:1 SWR at the load. Then |Γ| = 5/7 and |Γ|² is about 0.51. On the first encounter, about 51% of the incident power is reflected and about 49% is accepted at that plane. The corresponding load-only mismatch loss is:
Lmismatch = −10 log10(1 − |Γ|²) ≈ 3.10 dB
That 3.10 dB is a real, defined mismatch quantity. It is not proof that 51% became heat at the antenna, and it is not automatically the station’s final loss. The source-end reflection coefficient, matching network, line attenuation and steady-state wave interaction decide what happens next.
A Reflection Is Not a Sack of Lost Watts
The reflected wave travels back toward the source. At the source end it may be absorbed, partly reflected toward the load again, transformed by a tuner or output network, or accompanied by a change in newly generated forward power. In steady state, the forward and reverse components coexist; the useful picture is a superposition of waves, not isolated packets waiting their turn.
Kurokawa’s power-wave treatment makes the source/load boundary explicit. Power delivery with mismatches depends on the reference impedances and the complex source and load reflection coefficients. Magnitude alone is insufficient when re-reflections occur because their phase changes the result.
Joeri’s objection is the right one: reflected power and dissipated power are different physical quantities. The correction is not to deny mismatch loss; it is to stop calling every reflected watt a permanently lost watt.
A directional wattmeter separates sampled forward and reverse wave components with finite directivity and calibration error. The transmitter’s output devices experience the total terminal voltage and current created by those waves and by the output network. “The reflected power goes straight into the final transistor” is therefore no more generally correct than “the reflected power disappears.” Device stress must be evaluated from the actual voltage, current, load trajectory, temperature and protection behaviour.
A Real Feed Line Turns Some of Both Waves into Heat
An ideal lossless line dissipates no power. It transforms impedance with position and supports a standing-wave pattern, but a mismatch alone does not heat that ideal line. A real line has conductor and dielectric attenuation. The forward wave loses power on its way to the load, and the reflected wave loses more on its return.
With propagation constant γ = α + jβ and line length l:
Γin = ΓLe−2γl
The round-trip attenuation reduces the reflection magnitude seen at the input. This is why a long or lossy cable can make the shack-end SWR look better while delivering less power to the antenna. The meter has not repaired the load; the return wave has simply been attenuated before it reaches the meter.
Under mismatch, voltage and current vary along the line. Conductor heating follows local current and dielectric loss follows local electric field, so additional loss depends on frequency, line type, length, matched attenuation, load impedance and the complete source-end termination. A claim such as “20 metres of good coax at 6:1 SWR loses about one decibel” is not an engineering result until the cable, frequency, temperature, connectors, load phase and reference planes are stated.
What the Tuner Fixes—and What It Leaves Alone
A tuner at the transmitter can transform the impedance at the feed-line input so the radio sees an acceptable load. That can prevent power foldback and place the transmitter inside its intended voltage/current region. It does not change the antenna impedance, and it does not remove the standing-wave ratio on the line between the tuner and antenna.
The matched condition at the radio is produced by the tuner, line and antenna together. If the tuner is low loss and properly rated, it can re-reflect returned energy toward the load rather than simply absorb it. Yet the high-SWR line still has its real attenuation and voltage/current maxima, and the tuner has finite conductor, capacitor, dielectric and magnetic loss.
Putting the matching network at the antenna end changes the problem. The transformation occurs before the long line, so the line can operate nearer its characteristic impedance. That can reduce line loss and stress, but the remote network must survive the actual impedance, RF voltage/current, duty cycle, weather and common-mode environment.
| Location | What a successful match establishes | What still needs evidence |
|---|---|---|
| Tuner at the radio | An acceptable impedance at the transmitter-side reference plane | Feed-line SWR and loss, tuner loss, line voltage/current peaks, antenna accepted power and radiation |
| Tuner at the antenna | A transformed load before the main feed-line run | Remote-tuner loss and stress, antenna loss, common mode, weather stability and radiation |
| No tuner | Only the naturally transformed impedance presented through the line | Transmitter response, line loss, accepted power, stress and radiation |
Reactance Does Not Dissipate Average Power, but It Can Raise Loss
An ideal reactance stores energy during part of the RF cycle and returns it during another. Its average real-power dissipation is zero. In a real station, however, a strongly reactive impedance can demand large circulating current or voltage from the tuner, feed line and transformer.
- High current increases conductor, contact and winding loss and can heat ferrite through the resulting magnetic excitation.
- High voltage increases dielectric stress and loss and can cause corona or arcing.
- Large standing-wave peaks can exceed component ratings even when average transmitter power looks modest.
- A tuning network can achieve 1:1 at its input while dissipating appreciable power internally.
The reactive part is therefore not itself a heater, but it changes the current and voltage that expose every non-ideal resistance and dielectric loss in the system.
SWR Cannot Rank Antennas
SWR measures reflection magnitude relative to a chosen line impedance and at a chosen plane. It does not identify where the accepted power goes or where the antenna sends the radiated power.
- A dummy load can have excellent SWR and intentionally radiate almost nothing.
- Loss resistance can broaden an SWR curve while reducing radiation efficiency.
- An efficient antenna can present a difficult impedance before matching.
- Two antennas with equal accepted power can have very different patterns and field strength in a chosen direction.
- Feed-line exterior current can change the measured impedance, loss and radiation pattern without announcing itself as a separate SWR number.
That is why a low reading cannot prove that one wire architecture, transformer ratio or feed arrangement outperforms another. Compare accepted power at the antenna plane, matching and feed-line loss, radiation efficiency, common-mode current and the installed pattern. Otherwise the SWR meter is ranking the easiest quantity to see, not the result you wanted.
Measure at Declared Reference Planes
A defensible station measurement keeps every quantity tied to a plane and condition:
- Calibrate or de-embed. Move the VNA reference plane to the connector or network being discussed, or include the intervening fixture and cable uncertainty.
- Keep complex data. Record resistance, reactance and complex reflection coefficient across the band; scalar SWR discards phase.
- Characterise the feed line. Use manufacturer data only within its conditions, or measure the actual assembly’s S-parameters and temperature.
- Measure the tuner as a loaded network. Insertion loss at an easy 50-ohm termination does not describe every high-current or high-voltage transformation.
- Check the directional coupler. Directivity, calibration, harmonic content and placement limit forward/reflected-power accuracy.
- Observe the transmitter. Record DC input, generated forward power, foldback, device current and temperature rather than assuming one source model.
- Separate accepted from radiated power. Use field, gain, efficiency or comparison measurements suitable for the antenna and environment.
- Repeat the starting state. An A/B/A sequence helps expose drift in power, temperature and propagation when comparing station configurations.
A single SWR value can be useful for checking whether a transmitter is likely to operate happily. It cannot close the station power balance.
Primary and Authoritative Technical Sources
- K. Kurokawa, “Power Waves and the Scattering Matrix,” IEEE Transactions on Microwave Theory and Techniques—the primary power-wave treatment of source/load exchange and reflection.
- Keysight, Fundamentals of RF and Microwave Power Measurements—mismatch loss, source/load re-reflection and mismatch uncertainty.
- Rohde & Schwarz, dB Calculator application note—the relationships among reflection coefficient, reflected-power fraction, return loss, mismatch loss and VSWR.
- Rohde & Schwarz, Accurate Test Fixture Characterization and De-embedding—calibration and movement of the VNA reference plane.
- IEEE Std 145-2025, IEEE Standard for Definitions of Terms for Antennas—the current standards framework for separating antenna quantities.
- ARRL, More About Antenna Tuners—the practical distinction between the match seen by the transmitter and the unchanged SWR on the line beyond a shack tuner.
Joeri’s Bottom Line
Do not subtract the reverse meter reading from the transmitter label and call the difference “radiated power.” Do not declare mismatch harmless either. Keep the quantities separate: reflection redirects a wave; attenuation dissipates guided power; a tuner transforms impedance and has loss; the antenna accepts power and divides it between radiation and internal loss.
Once the reference planes and the source, line, tuner and antenna are treated as one network, the apparent contradiction disappears. SWR matters because it changes power transfer, line stress and loss. It is simply not a magic meter for antenna efficiency—and reflected watts are not automatically burnt watts.
Mini-FAQ
- Is reflected power the same as power dissipated? No. It is a travelling-wave component returning from a discontinuity; dissipation must be located in the line, tuner, source network, antenna or another real loss mechanism.
- Is mismatch loss a myth? No. Mismatch loss is a defined reduction in accepted or delivered power under stated source, load and reference-plane conditions; it should not be confused with heat at the reflection point.
- Does a shack tuner remove the SWR on the feed line? No. It can present a suitable impedance to the transmitter, but the SWR between the tuner and antenna remains set by that line and load.
- Why can lossy coax make the radio-end SWR look better? The reflected wave is attenuated on its return trip, so the input reflection is smaller even while the cable is dissipating more power.
- Can SWR prove antenna efficiency? No. SWR describes reflection magnitude; radiation efficiency, matching loss, feed-line loss, common-mode current and pattern require separate evidence.
- What should I record when investigating SWR-related loss? Record frequency, complex impedance, reference planes, line data, tuner loss, forward and reverse waves, transmitter response, temperature and a separate radiation or field comparison.