VSWR and return loss describe reflection at a specified RF port, frequency and reference impedance. A lower VSWR or higher positive return loss means less reflection at that reference plane. It does not, by itself, establish antenna gain, radiation efficiency, polarization quality or GNSS positioning performance. A lossy cable can even improve the return loss measured at its near end while attenuating the wanted signal.
For a GNSS integration, first establish what was measured: a passive antenna feed, an active antenna's amplifier output, or an entire cable-and-antenna assembly. Only then compare the numbers.
One reflection, several ways to describe it
An RF wave encountering a load that differs from the line's reference impedance produces a reflected wave. At one frequency, the voltage reflection coefficient is a complex quantity: it has magnitude and phase. For a real, positive reference impedance Z₀ and load impedance ZL:
Γ = (ZL − Z₀) / (ZL + Z₀)
Let ρ = |Γ|. For the passive loads considered here, 0 ≤ ρ ≤ 1. Two scalar descriptions follow:
VSWR = (1 + ρ) / (1 − ρ)
Return loss, RL = −20 log₁₀(ρ) dB
VSWR is the ratio of maximum to minimum voltage along the standing-wave pattern. It is not a power gain. Return loss expresses how far the reflected power lies below the incident power. Thus smaller VSWR is better matching; larger positive return loss is better matching, at the same frequency and reference plane. Rohde & Schwarz explains these reflection metrics.
Watch the display convention. A network analyzer's logarithmic S11 magnitude of −20 dB corresponds to positive return loss of 20 dB. The signs differ, not the measured reflection. Keysight's reflection tutorial distinguishes scalar return loss and VSWR from the magnitude-and-phase information in the reflection coefficient.
Convert the number before interpreting it
For the real reference impedance above, the reflected fraction of incident power is ρ². The fraction accepted by the load is 1 − ρ². Mismatch loss is −10 log₁₀(1 − ρ²) dB. These quantities refer to the chosen port; they do not include every loss elsewhere in the receive chain.
The following are calculated examples, not measurements of any ResiNav product.
| Stated match at one reference plane | Equivalent reflection | What the conversion means |
|---|---|---|
| VSWR 2:1 | Return loss ≈ 9.54 dB; reflected power ≈ 11.11% | About 88.89% is accepted by the load. Mismatch loss is about 0.512 dB. |
| Return loss 20 dB | VSWR ≈ 1.22:1; reflected power = 1% | About 99% is accepted. Mismatch loss is about 0.044 dB. |
Neither row says that the accepted power is all radiated by an antenna. Conductive and dielectric losses can dissipate power after it enters the antenna. Radiation efficiency and mismatch are separate factors, as explained in the original Antenna-Theory efficiency tutorial. In an ideal thought experiment, a matched 50 Ω resistor accepts power but is not a useful GNSS receiving antenna. A good match is therefore evidence about an interface, not a complete antenna qualification.
Scalar matching data also cannot reconstruct the load impedance. Hypothetical example: on a 50 Ω line, a purely resistive 25 Ω load gives Γ = −1/3, while a 100 Ω load gives Γ = +1/3. Both have VSWR 2:1 and return loss about 9.54 dB. Their reflection phases differ. To diagnose whether an impedance needs a particular matching adjustment, magnitude alone is insufficient.
Why a longer cable can make return loss look better
The test signal travels down the cable, reflects at the load, then travels back. Cable attenuation reduces the returned wave twice. Anritsu explicitly describes this effect in Cable Loss Effect on System Return Loss: a better system return-loss reading need not mean a better antenna.
Hypothetical calculation: take the 2:1 load above and a cable with 3 dB one-way insertion loss at the test frequency. Assume a matched cable with no additional connector reflections, and measure at the cable's near end without removing its loss from the result.
RLseen = RLload + 2 × cable loss
RLseen = 9.54 + 2 × 3 = 15.54 dB
The apparent near-end VSWR is about 1.40:1, even though the load itself remains 2:1. The cable has not repaired the load. Its one-way 3 dB loss also leaves only about half the traveling power reaching the far end before the load reflection is considered.
This simple result isolates attenuation. Real mismatched cables, adapters and connectors create additional reflections that can combine in phase or out of phase. Do not subtract a single cable-loss number and assume that every real assembly has been fully characterized.
The practical comparison is antenna-plane match versus assembly-plane match, with cable insertion loss recorded separately. Calibration at the intended connection plane, or an appropriate characterized correction, makes that distinction explicit. Keysight's cable and antenna measurement documentation treats return loss, cable loss and two-port insertion loss as distinct measurements. Keep that distinction in the report rather than reducing all three to “good VSWR.”
An active GNSS antenna adds another reference-plane question
A connector on an active antenna can lead into an amplifier and filtering assembly rather than directly into the passive radiating element. A matching specification at the amplifier output describes that output interface. It does not establish the passive element's feed-point impedance or radiation pattern.
For a concrete documentation example, SBG Systems' recommended GNSS accessories labels the TW8889 matching entry as VSWR at the LNA output, separately from LNA gain, noise figure and antenna gain on a stated ground plane. That is an SBG/Tallysman example, not a ResiNav specification or compatibility recommendation.
Before interpreting an active-assembly measurement, retain the stated port, operating state, bias arrangement and frequency range. Follow the equipment and antenna manufacturer's permitted RF power and DC connection arrangements. A passive-antenna sweep setup must not be assumed suitable for an unknown powered assembly.
Read the match alongside the GNSS antenna evidence
A nominal 50 Ω label identifies the intended interface impedance; it is not a measured VSWR limit over every operating band. Likewise, a return-loss value at one frequency does not establish the match across all GNSS bands used by a project.
u-blox's GNSS antenna guidance treats impedance matching alongside radiation pattern, gain, polarization and the rest of the RF path. Those are different questions. For installation-dependent behavior, retain the mounting and ground-plane conditions rather than assuming a bench value transfers unchanged to a vehicle.
Use the RF-path documentation guide to keep the cable assembly, reference planes and insertion-loss record together. Where weak-signal reception is the question, the separate noise-figure and cable-loss explanation addresses noise added along that path. Improving a reflection metric does not answer that noise question, or prove interference suppression, tracking continuity or position accuracy.
The useful conclusion from a matching result is precise: this port reflects this fraction under these conditions. That can support a documented interface comparison. Complete GNSS reception still needs the antenna and receiver evidence appropriate to the intended installation.
NEXT STEP
Turn available platform information into an engineering review.
Use the Technology centre to frame the discussion, review relevant application scenarios, then send the available platform and receiver details for confirmation.