Forward/Reflected Power and SWR: What the Meter Can Prove

“Forward” is not a synonym for radiated power, and “reverse” is not a synonym for lost power. A forward/reverse meter samples waves travelling in opposite directions at its own reference plane. SWR and net power flow can be derived from those samples, but only within the meter's frequency, range, waveform, directivity and calibration limits.

Start With One Reference Plane

The meter separates a forward-travelling wave from a reverse-travelling wave. Its two displayed powers belong to the physical plane where its directional line section is inserted—not simultaneously to the transmitter socket and antenna terminals.

The meter itself is an inserted network. Its through loss, source/load match, connectors and adapters move the effective plane and can perturb a sensitive mismatch. “At the feedpoint” therefore means a stated port boundary, with the meter and interconnections included in the uncertainty.

The Four Quantities and Their Conditions

For a single-frequency measurement on a uniform line referenced to a real Z0, let PFWD and PREV be calibrated wave powers at the same plane and in the same detector mode. Then:

The equations do not rescue invalid readings. Forward power must be above the specified measurement threshold; reverse power must be distinguishable from leakage and noise; both channels need the correct frequency correction and compatible time response. A display that rounds reverse power to zero does not prove an infinite return loss or exactly 1.000:1 SWR.

Ordinary amateur instruments assume a real 50 Ω reference. General power-wave definitions become more careful when the reference impedance is complex, so these simple square-root relations should not be exported uncritically to an arbitrary complex measurement system.

A checked 2:1 example

At one valid plane, PFWD = 100 W and PREV = 11.11 W give |Γ| = 1/3, SWR = 2:1, return loss ≈ 9.54 dB and net downstream power ≈ 88.89 W. The 11.11 W is 11.11% of the incident wave power at that plane. None of those numbers alone establishes antenna radiation efficiency.

What Forward, Reverse and Net Power Mean

Forward power belongs to the wave travelling toward the instrument's load-labelled side. It can subsequently be dissipated in line and matching losses, accepted by the antenna system or reflected by one or more downstream discontinuities. It is not automatically the transmitter's generated power, because a mismatched network can re-reflect returning energy and contribute to the steady-state forward wave.

Reverse power belongs to the wave travelling toward the source-labelled side. A load, connector, filter, cable fault, tuner output or combination of discontinuities may contribute. The meter measures their vector result at its plane; it does not locate the cause.

Net power is the difference of valid forward and reverse wave powers at that plane. In a passive steady-state downstream network, it is the net time-average real power crossing the plane toward that network. It is not a universal “delivered to the antenna” number and certainly not a radiated-power measurement.

The source-side termination matters to the steady-state wave pattern. Returning energy may be absorbed, dissipated in protection or isolation hardware, or re-reflected according to the transmitter/tuner network's complex source reflection coefficient. A generic meter cannot determine which. Describing the result as repeated wave interactions is more accurate than claiming that a particular reflected watt is always “reused” or “burned in the finals.”

Why a Tuner Can Produce Large Wave Readings

A tuner changes the impedance presented at its input; it does not require the output line to have low SWR. A meter before the tuner may therefore show a near match while a meter after it shows substantial forward and reverse components.

For an ideal lossless output line terminated in a 3:1 SWR load, |Γ| = 0.5 and the reflected fraction is 0.25. If the steady-state net power delivered past a plane is 100 W:

The arithmetic is not energy creation. Those are counter-propagating steady-state wave powers whose difference is 100 W. A real tuner also has loss, voltage/current limits and a source match; its transmitter-side input power must cover the net output plus tuner dissipation.

Feed-Line Loss Hides the Load Reflection

Let L be the line's one-way power transmission factor between the input and load planes, with 0 < L ≤ 1. For a uniform line:

If the one-way attenuation is A dB, then L = 10-A/10. The input return loss improves by approximately 2A dB because the reflected sample experiences the line in both directions relative to the input forward sample. This is a prettier input SWR created by attenuation, not a better load.

A checked 3 dB example

With 100 W forward at the line input, exactly 3.0 dB one-way attenuation gives L ≈ 0.5012. A 2:1 load has a reflected-power fraction of 1/9:

• forward power reaching the load: ≈ 50.12 W;
• reverse power launched at the load: ≈ 5.57 W;
• reverse power returning to the input: ≈ 2.79 W;
• input indicated SWR: ≈ 1.40:1;
• power accepted at the load: ≈ 44.55 W.

The input difference is ≈ 97.21 W, not 44.55 W. The missing ≈ 52.66 W is the combined forward- and reverse-trip line dissipation under these stated conditions. In practice, matched-line attenuation itself varies with frequency, cable, temperature and construction, and mismatch raises loss beyond the simple matched-line value.

Directivity Sets a Low-Reflection Boundary

A directional coupler is not perfectly blind to the unwanted direction. Some incident signal leaks into the reverse channel. Keysight's measurement guidance identifies this leakage as directivity error and treats reflection error as a complex quantity. Leakage can therefore add to or subtract from the true reflected wave; it is not a fixed positive number that can always be removed by subtracting watts.

An illustrative 30 dB raw directivity corresponds to an unwanted voltage-wave magnitude of 10-30/20 ≈ 0.0316, or a power ratio of 0.001. With an ideal load, that scale alone corresponds to an apparent SWR of about 1.065:1 before tracking, source match, detector, connector and noise errors. This does not say every 30 dB coupler will read exactly 1.065:1; leakage phase changes the vector result.

Adapters and imperfect connectors can degrade effective directivity and source match. Calibration should therefore establish the desired reference plane using the same connector path used for the measurement. A precision load is useful only when its reflection, power, frequency and connector condition are suitable and known.

Accuracy Is Not Display Resolution

A defensible result identifies the measurand and an uncertainty budget. Relevant terms include forward/reverse calibration factors, linearity, range, coupler directivity, frequency response, insertion mismatch, source match, detector response, temperature, connector repeatability, load quality, noise and reading resolution.

As a current product-specific example, Bird specifies the Model 43 at ±5% of full scale and describes it as a CW wattmeter. On a 100 W full-scale element, that one specification contributes ±5 W regardless of whether the indication is 100 W or 20 W. At a 20 W indication it is already ±25% of the reading; a 4 W indication on that scale is smaller than the ±5 W full-scale contribution. A permitted lower-range element can improve deflection, but it does not remove directivity, mismatch, calibration or overload constraints.

Bird's Model 43 manual also requires the correct element frequency and full-scale range, gives a nominal 50 Ω impedance and less than 1.05:1 insertion VSWR for the listed standard models, and specifies ±5% of full scale for CW. These are instrument boundaries, not generic specifications for every crossed-needle or digital station meter.

For small independent fractional errors only, the ratio relation suggests:

That approximation does not include correlated calibration terms or complex directivity/source-match error, so it is not a complete uncertainty budget. It does show why a poorly resolved reverse channel can dominate the calculated SWR.

The Detector Must Match the Waveform

“Watts” is incomplete without a time definition. A continuous unmodulated carrier, time-average power, average burst power and peak-envelope power are different measurands. Speech processing and crest factor alter the relation between SSB average and peaks; pulse width and repetition rate alter the relation between average and pulse power.

The U.S. FCC definition of peak envelope power (PEP) is the average power supplied to the antenna transmission line during one RF cycle at the crest of the modulation envelope under normal operating conditions. That definition does not mean any analog needle captures PEP. The sensor still needs adequate envelope bandwidth, peak acquisition/hold behaviour, calibration and range.

Bird's standard Model 43 manual says its diode indication on amplitude modulation responds mainly to the carrier, with little response to the added sidebands. The separate Model 43P adds peak mode, is specified at ±8% of full scale in that mode and has stated pulse limits. Bird's current product page points users measuring modern digital signals to a different true-average architecture. Model names and modes therefore matter.

Rohde & Schwarz's current NRT2/NRT-Z directional system illustrates the same principle in a modern sensor: average power, average burst power, PEP and crest factor are distinct functions, and selectable video bandwidth affects PEP, crest-factor and burst measurements. Do not transfer those capabilities to a meter whose data sheet does not specify them.

What the Meter Cannot Establish Alone

• whether the load impedance is inductive or capacitive;
• whether an antenna is resonant;
• which of several discontinuities created the measured vector reflection;
• matched-line attenuation or total feed-line dissipation;
• tuner, balun, trap, loading-coil, conductor or ground-system loss;
• antenna radiation efficiency, pattern, gain or radiated power;
• common-mode current on the outside of the feed line;
• local voltage/current maxima or component stress elsewhere on the line;
• safe operation of the transmitter, tuner, line or antenna at the measured mismatch.

Moderate SWR on a short low-loss line need not waste much power, but it can increase voltage/current maxima, cable loss and stress in tuners, connectors, ferrites and transmitter protection. Conversely, 1:1 at the shack can describe an efficient antenna, a dummy load, a lossy system or a tuner's transformed input. SWR is a match metric at a stated plane—not a system-quality score.

A Measurement Workflow That Survives Scrutiny

1. Define the measurand. State CW, time average, burst average or PEP; forward, reverse, ratio or net; and the required uncertainty.
2. Define and label the plane. Record whether the meter is before/after the tuner or at the antenna port, including adapters and jumpers.
3. Choose the correct sensor or element. Confirm direction, frequency, full-scale range, power and mismatch limits, detector mode and waveform capability.
4. Inspect de-energized. Check connector condition, torque practice, element seating and zero. Never insert, reverse or reconnect components while transmitting unless the manufacturer explicitly permits it.
5. Verify with a suitable load. Use a load with adequate frequency, power, thermal and reflection specifications. A load check tests the setup; it does not locate every station fault.
6. Keep the signal stable. Use a specified carrier or repeatable test waveform. Do not compare a moving SSB needle with a CW rating.
7. Record raw readings. Keep forward and reverse values, range, frequency, mode, time, temperature and configuration—not only the calculated SWR.
8. Apply the uncertainty. Use manufacturer accuracy in its stated form, especially full-scale versus percent-of-reading, plus directivity and connection terms.
9. Use another measurement for another question. A VNA can locate complex match versus frequency; two-plane power measurements can estimate line loss; current probes can examine common mode; field or calorimetric methods address different efficiency questions.

Primary Sources Checked

• Bird Model 43 current product page: CW use, current frequency/power overview and ±5% full-scale accuracy.
• Bird Model 43/43P instruction manual: travelling-wave definitions, element direction/range, nominal impedance, insertion VSWR, CW/peak accuracy, modulation response and pulse limits.
• Keysight measurement-error guide: directional-coupler leakage, source/load match and reflection-tracking errors as complex systematic terms.
• Keysight directional-coupler application note: coupling, isolation and directivity definitions and measurement boundaries.
• Rohde & Schwarz NRT-Z14/44 sensor manual: distinct average, burst, peak-envelope and crest-factor functions, video bandwidth and reflection measurement limits.
• 47 CFR §97.3: the U.S. amateur-service definition of peak envelope power.
• NIST RF power metrology overview: traceability and measurement uncertainty as necessary parts of a power result.

Questions, antenna-factor records or height trials to share? Contact RF.Guru.