VSWR, Return Loss, S11 and S21: What Each Number Actually Means

I still hear that VSWR is old-fashioned and dB is the modern, more accurate answer. That argument confuses notation with information. Converting a calibrated reflection magnitude from VSWR to return loss does not improve the measurement. Keeping the complex S11 value does preserve phase, and measuring S21 adds a different transmission result. Those distinctions matter.

One-Port Reflection Has Several Valid Displays

For a load impedance ZL referred to a real reference impedance Z0, the complex reflection coefficient at that plane is:

The VSWR and return-loss equations use only |Γ|. They discard the reflection angle. A VNA can retain both magnitude and phase as complex S11, then display the same measurement as a Smith chart, impedance, VSWR, linear reflection magnitude or logarithmic magnitude.

Check the sign convention before comparing plots. Some instruments label 20 log|S11| as “return loss,” producing a negative trace. Others report positive return loss as −20 log|S11|. The physics has not changed; the label has.

S11 Is More Than a Different Unit

S11 is the complex ratio of the wave leaving port 1 to the wave incident on port 1, with the other ports terminated in their reference impedances. Its magnitude converts directly to VSWR or return loss. Its angle tells you where the reflected wave sits in phase at the chosen reference plane.

That phase is what lets a Smith chart distinguish, for example, a resistive 100 Ω load from reactive loads having the same VSWR. Move the reference plane through a transmission line and the phase rotates; line loss also changes the magnitude. A shack-end S11 trace therefore describes the antenna, feed line, connectors, chokes and matching hardware seen from the shack—not the bare feedpoint by decree.

S21 Answers a Two-Port Transmission Question

S21 is the complex forward transmission coefficient b2/a1 with port 2 terminated in its reference impedance. It belongs to a two-port model. That makes it useful for cables, filters, attenuators, transformers and properly defined test fixtures.

A logarithmic |S21| trace can show insertion gain or loss under the VNA’s reference conditions, but “how much power gets through” needs a complete definition. Arbitrary source and load mismatch create re-reflections. A transformer can change impedance as well as dissipate energy. An active device may have gain. For those cases, use the full S-matrix and the appropriate available, operating or transducer-power gain—not S21 as a universal efficiency meter.

An antenna also has no convenient coaxial output port representing all radiated power. An over-the-air S21 measurement between two antennas includes both antennas, path loss, distance, polarization, alignment, reflections, cables and the calibrated reference planes. It can support a controlled comparison, but it is not automatically the efficiency of one antenna.

Return Losses Do Not Add Through a Cascade

dB arithmetic is convenient only when the underlying quantities combine by multiplication under the stated model. Matched attenuator losses can often be added in dB. Return loss is different: reflections from several interfaces return with phase, experience transmission loss and reflect again. They can reinforce or cancel across frequency.

Two components that each show 20 dB return loss do not automatically produce a 40 dB system return loss. Cascade their two-port networks with the correct reference impedances, or measure the connected assembly. If only scalar return-loss magnitudes are available, the missing phase prevents a unique answer.

Reflection Is Not the Same as Dissipation

For a load connected to a matched source through an otherwise lossless reference system, the mismatch-loss term associated with reflection is:

A 2:1 VSWR gives about 0.51 dB by that specific expression. It does not say where accepted power goes. A well-matched dummy load intentionally converts nearly all accepted RF power to heat. A well-matched antenna may radiate most of it. A lossy feed line can make the shack-end VSWR look better while wasting power. Match, dissipative loss, radiation efficiency and directional gain need separate evidence.

Put the Reference Plane Where the Claim Lives

A useful VNA record starts with the question, frequency span, reference impedance and physical plane. Then:

• calibrate at the end of the test cable with suitable standards for the connector and frequency range;
• verify the calibration with a known load or independent check standard;
• fix cable position and connector torque, then repeat after reconnecting;
• choose test power and IF bandwidth that keep the device linear and the trace above the noise floor;
• use port extension only for a known delay, and de-embed only with a defensible fixture or line model;
• save complex data rather than only a screenshot of VSWR; and
• state uncertainty, drift and the terminations present at every unused port.

Keysight’s FieldFox guidance places calibration standards at the intended reference plane and recommends verifying the calibration and jumper. Rohde & Schwarz likewise treats S11, VSWR, return loss and impedance as representations of a calibrated reflection measurement, while its VNA manual defines S21 with the output reference-matched.

Choose the Quantity That Matches the Decision

So no, VSWR is not obsolete. It is a compact scalar display. Return loss is not inherently more accurate. Complex S11 carries more information than either scalar conversion, and S21 belongs to a different measurement. Use the representation your decision needs—and keep the underlying calibrated data.

Primary measurement references

• Rohde & Schwarz IoT Design Guide—S11, VSWR, return-loss and impedance relationships for antenna measurements.
• R&S ZVL Operating Manual—formal S-parameter definitions and matched-port measurement conditions.
• Keysight FieldFox Calibration Guide—reference-plane calibration, jumper integrity and verification.
• NIST/NBS Special Publication 300, Volume 4—RF mismatch error and the role of reflection magnitude and phase.

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