Reflected Power, SWR and Tuners: What the Finals Actually See
Reflected Power, SWR and Tuners: What the Finals Actually See
A reverse wave is real, but the REV needle is not a damage meter. Reference plane, mismatch phase, output network, tuner limits and protection logic decide the stress.
“Reflected power blows the finals” and “the tuner sends every returned watt back until the antenna uses it” are mirror-image myths. A useful explanation must keep wave decomposition, net power, dissipative loss, impedance transformation and active-device stress separate.
RF-power warning: never disconnect, move or inspect a feedline, wattmeter, tuner or load while transmitting. High-power RF can cause severe burns, electric shock and arcing. Disable the transmitter and prevent accidental keying before changing the RF path.
Start with One Plane and One Reference Impedance
On a uniform single-mode line referenced to a real characteristic impedance Z0, voltage and current can be decomposed into forward and reverse travelling waves. At the load plane:
Γ = V− / V+ = (ZL − Z0) / (ZL + Z0)
P− / P+ = |Γ|²
SWR = (1 + |Γ|) / (1 − |Γ|)
The reverse wave has voltage, current, field energy and directional average power. Calling it “real” is appropriate. The next conclusion must be more careful: reverse power is not automatically dissipated, permanently lost, absorbed by the PA or available as free additional transmitter output.
Under the same real-Z0, sinusoidal steady-state assumptions, net average real power crossing that plane toward the load is:
Pnet = P+ − P−
If a calibrated directional meter reads 100 W forward and 11.1 W reverse at a 2:1 SWR, approximately 88.9 W net crosses that meter plane toward the downstream network. At 3:1, |Γ| = 0.5, so 25% of the forward component is reverse and 75% is net.
Do not rename net power as radiated power. Forward minus reverse at the meter is net power entering whatever lies downstream. Feedline, tuner, transformer, connector, antenna-conductor, ground and common-mode losses may still dissipate or redirect part of it.
What a Directional Wattmeter Knows—and What It Does Not
A directional coupler combines electric and magnetic sampling so one detector responds mainly to the forward wave and another mainly to the reverse wave. The official Bird Model 43 operation manual develops that travelling-wave model and uses forward minus reverse for load power in its defined 50 Ω measurement system.
A scalar FWD/REV meter can provide wave-power magnitudes and SWR. It does not provide the phase of Γ, and therefore cannot determine the complex impedance by itself. The manual explicitly notes that its result omits reflection-coefficient phase.
| Displayed quantity | Defensible interpretation | Not established |
|---|---|---|
| FWD | Calibrated forward-wave power at the coupler plane for the supported waveform and frequency. | DC input, PEP on every modulation, net load power or radiation. |
| REV | Calibrated reverse-wave power at that plane. | Heat in the final transistor or total system loss. |
| FWD − REV | Net real power toward the downstream network under the stated line assumptions. | Antenna radiation after downstream losses. |
| SWR | Reflection magnitude relative to the instrument reference impedance. | Reflection phase, efficiency, gain, pattern or PA-safe load region. |
Real meter limits matter
- Directivity: imperfect cancellation leaks part of the large forward sample into the reverse channel. The Bird 43 manual specifies directivity greater than 25 dB; that is finite, not infinite.
- Range: Bird specifies ±5% of full scale for CW, so choosing a grossly oversized element makes a small reverse reading uncertain. Its low-reflection procedure uses a second element rated at one tenth of the forward element.
- Frequency and calibration: the element, coupler, connectors and detector must cover the actual frequency and power.
- Waveform: detector response determines whether the display represents carrier, average or peak-envelope power. Bird states that the ordinary Model 43 is not suitable for accurate modern digitally modulated-power measurement.
- Harmonics and unwanted signals: a broadband detector can respond to energy outside the intended carrier unless filtering or frequency-selective measurement defines the result.
The Rohde & Schwarz directional-coupler analysis shows why directivity error is phase-sensitive: leakage from the forward sample combines vectorially with the true reverse sample, so a reflected-power indication can be either high or low. A calibrated VNA adds complex magnitude and phase at low power; it does not replace a suitably rated operating-power sensor.
Reference Plane: The Missing Label on Most SWR Claims
A metre of line rotates reflection phase. Loss also reduces the reverse wave before it reaches the source. For a uniform line of length l and propagation constant γ = α + jβ:
Γin = ΓLe−2γl
The e−2αl term reduces magnitude after the round trip; e−j2βl rotates phase. A lossy coax run can therefore make the transmitter-end SWR look better while less power reaches the antenna. On a low-loss line, SWR magnitude changes little but impedance can rotate around the constant-SWR circle.
This is why the same antenna can present very different resistance and reactance to a PA after the line length changes, even when a scalar SWR meter reports nearly the same ratio.
What the Tuner Actually Changes
An antenna tuner is an impedance-transforming network. When adjusted within its matching and component ratings, it can present the transmitter with the intended load—usually close to 50 + j0 Ω—even while the tuner-to-antenna line remains mismatched.
The Yaesu FTDX10 manual states the boundary unusually clearly: its internal ATU presents a 50 Ω load to the final stage, but because the ATU is inside the radio it adjusts only the impedance at the transceiver end of the coax and does not tune SWR at the antenna feedpoint.
| Tuner position | What a successful match can do | What remains |
|---|---|---|
| Inside radio or at amplifier output | Present an acceptable load to that transmitter or amplifier port. | Antenna-side feedline SWR and its extra voltage, current and loss. |
| At the feedline input | Match the source to the complete line-plus-load input impedance. | Standing waves between tuner and mismatched antenna. |
| At the antenna feedpoint | Match the feedline to the antenna system at that boundary. | Tuner loss and voltage/current stress at the remote network. |
A lossless tuner does not need to absorb the reverse-wave power to make its input matched. It changes boundary conditions and stores reactive energy. On an antenna-side line with 3:1 SWR and 100 W of net power, the steady travelling-wave decomposition can be 133.3 W forward and 33.3 W reverse; their difference is 100 W. Those are components of one steady solution, not 33.3 free watts generated by the tuner.
A real tuner has finite inductor Q, capacitor and dielectric loss, contact resistance and perhaps transformer or ferrite loss. Thus:
Pnet,out = Pnet,in − Ptuner loss
The Elecraft KAT500 manual demonstrates why “the tuner handled it” is incomplete: matching range and power are specified together, the full-power ICAS condition has a time limit, tuning uses reduced power, and out-of-normal-range matches may require reduced operating power.
A tuner is not automatically a protector. It can give the PA an acceptable load only after a valid match is found and while relays, capacitors, inductors, connectors and control timing remain within rating. During a search it may traverse worse impedances. Follow the exact tune-power, keying and amplifier-inhibit procedure.
What the Finals Actually See
A radio designed for a nominal 50 Ω external load is not necessarily a passive 50 Ω resistor looking back into its antenna socket. The reverse wave encounters the entire source-side network: tuner, filter, relay, coupler, transformer, combiner, output matching network, active device and protection system.
The reflection coefficient is complex. Two loads can have the same SWR and reflected-power fraction while having different phases. After transformation through the line and output network, they can produce different drain, collector, anode, grid, transformer, filter or relay voltage and current.
SWR gives the radius, not the position, on a Smith chart. Final-stage stress requires the complex load at the PA reference plane plus frequency, drive, supply, bias, waveform, duty cycle, temperature and protection dynamics.
NXP's AN1938 output-load study demonstrates that even a small output-load reflection with unchanged magnitude can change compressed power and efficiency as phase changes. At severe mismatch, manufacturers sweep all phases because a single open, short or reactive load does not identify the worst case.
The MRF1K50H data sheet provides a useful boundary. It reports survival above 65:1 VSWR at all phase angles with 3 dB overdrive—but specifically in NXP's 50 Ω production fixture at 230 MHz, 50 V, 100 µs pulses and 20% duty cycle. That result does not rate every HF amplifier, broadband transformer, low-pass filter, relay or tuner built around the transistor.
Valve and Solid-State Labels Do Not Set the Safe SWR
Some manually or automatically tuned valve amplifiers can transform a stated range of antenna impedances with their output tank. Some solid-state amplifiers use broadband matching and fast foldback; some devices are exceptionally mismatch-rugged. Neither technology label supplies a universal safe SWR.
The current ACOM 1003 valve-amplifier manual, for example, specifies an output-tank matching capability up to 3:1 SWR while imposing explicit reflected-power warning and protection thresholds. It even notes that a higher-SWR load may sometimes be matched but is not warranted and may require reduced drive.
The solid-state Elecraft KPA1500 manual monitors PA current, calculated dissipation, forward power, reflected power, extreme SWR, temperature and multiple supply rails as separate fault conditions. This is the correct mental model: SWR protection is one layer in a wider safe-operating envelope.
Foldback or a trip reduces damage probability; it is not a licence to transmit continuously into a fault. Sensor bandwidth, detector accuracy, control latency and stored energy matter, and a severe transient can precede a slow display. Repeatedly resetting without correcting the load, switching or tuning problem can convert a recoverable warning into permanent damage.
Where Reverse Energy Can End Up
There is no single destination encoded by the word “reflected.” Depending on the actual boundary and operating state, reverse-wave energy can:
- be partly re-reflected toward the load;
- be dissipated in line, tuner, filter, transformer, connector or antenna-system loss;
- be routed by a circulator into a rated termination;
- alter active-device voltage, current, dissipation, gain or stability;
- cause power foldback, transmit inhibit or a latched fault; or
- combine with source-generated waves into a new steady-state solution.
A circulator does not destroy reflected energy; it redirects it into a dump load that must survive the power, waveform and duty. A tuner does not normally function as that dump load. A PA is not a bucket that must swallow the REV reading.
A Safer Measurement and Commissioning Workflow
- Read the exact manuals. Record the permitted load/SWR region, reflected-power limit, tuner range, tune power, duty cycle, keying sequence and fault-reset procedure.
- Name every reference plane. Radio output, amplifier output, tuner input, tuner output, feedline input and antenna feedpoint are not interchangeable.
- Verify into a rated 50 Ω load. Start at low drive and confirm forward power, reverse null, gain, current and temperature before introducing the antenna system.
- Measure complex impedance at low power. Calibrate a VNA at the relevant connector or de-embed the known line. Scalar SWR cannot show the phase-dependent load.
- Characterise the feedline. Measure matched attenuation and account for the two-way reduction of the reflected wave. Do not mistake a lossy cable's improved shack SWR for a better antenna.
- Tune only by the prescribed sequence. Inhibit the amplifier where required, use the specified low tune power and never hot-switch relays or tuner elements.
- Raise power in steps. Watch input/forward/reverse power, supply current, gain, tuner or component temperature and fault logs. Stop on instability, unexpected heat, smell, noise or arcing.
- Investigate every repeatable trip. Check antenna switching, connectors, water ingress, cable damage, tuner state, band data and load changes before resetting.
The Practical Verdict
Reflected power is a real reverse travelling-wave component. At one suitable reference plane, calibrated forward minus reverse power is net real power toward the downstream network. Neither number, by itself, reports radiation efficiency or watts dissipated in the final stage.
A tuner changes the impedance presented at its input; its location decides whether antenna-side feedline SWR remains. The finals respond to the complex load transformed through the complete output path—not to SWR magnitude alone. Use the exact amplifier and tuner envelope, complex low-power measurement, properly ranged operating-power sensors and conservative fault handling.
Primary technical references checked
- Bird Model 43 RF Directional Thruline Wattmeter Operation Manual
- Rohde & Schwarz — influence of directional-coupler parameters on forward/reverse measurements
- Yaesu FTDX10 Operation Manual — internal tuner boundary and matching range
- Elecraft KAT500 Owner's Manual — matching, tune power and reduced-power limits
- Elecraft KPA1500 Owner's Manual — layered PA and mismatch protection
- NXP AN1938 — sensitivity of RF power transistors to source and output loads
- NXP MRF1K50H Data Sheet — conditional all-phase mismatch qualification
- ACOM 1003 User's Manual — output-tank matching and reflected-power limits
Mini-FAQ
- Is reflected power real power? Yes. It is a measurable reverse travelling-wave component. It is not automatically dissipated, lost or absorbed by the final transistor.
- Can reflected power be subtracted from forward power? At one calibrated plane on a single-mode line with real reference impedance, the difference is net average real power toward the downstream network. It is not automatically radiated power.
- Does an antenna tuner absorb reflected power? Not as its intended function. It transforms impedance and stores reactive energy; a real tuner also dissipates some power because its components have finite loss.
- Does 1:1 SWR at the transmitter remove SWR from the feedline? Not when a tuner in the shack matches the line-plus-load input. The tuner-to-antenna line can still carry forward and reverse waves; a feedpoint tuner changes that boundary.
- What actually threatens the finals under mismatch? The complex load at the PA plane can create excessive device or network voltage, current, dissipation, arcing or instability. Reflection phase, frequency, drive, supply, duty, temperature and protection all matter.