When a Better Choke Makes the SWR Look Worse
When a Better Choke Makes the SWR Look Worse
PA0LJD’s 10-metre OCFD tests show why SWR, common-mode suppression and radiation efficiency are three different questions—and why 50 cm of coax can change the whole answer.
Jaap, PA0LJD, compared several common-mode arrangements on an 80/20 off-centre-fed dipole using a 1:4 UNUN. One arrangement produced a more attractive 10-metre SWR; another placed the choking boundary closer to the feedpoint and changed the match. The result is an installation study, not a product ranking.
It is tempting to call the lower-SWR device the better choke. That conclusion does not follow from the measurement. SWR tells us about the impedance presented to the measurement point. It does not separately reveal outside-shield current, ferrite loss or radiated power.
The lesson in one line: an effective choke can remove an accidental feedline contribution that was helping the match. The intended antenna can become less dependent on cable routing while the displayed SWR becomes less flattering.
The Case in One Minute
An intentionally asymmetric, multiband structure whose feedpoint impedance and external current paths vary substantially by band.
At roughly 28 MHz, short pigtails, connector spacing and nearby conductors become a larger electrical fraction of a wavelength.
A useful observation—but not yet a measurement of common-mode current, efficiency or component loss.
One Analyzer Trace, Three Different Questions
Measure reflection coefficient, return loss or SWR at a stated reference plane.
Measure net RF current around the complete coax at several positions.
Estimate radiation, conductor, dielectric, transformer, ferrite and environmental losses separately.
Engineer’s corner: the quantities are not interchangeable
Γ = (Zin − Z0) / (Zin + Z0)
SWR = (1 + |Γ|) / (1 − |Γ|)
ηrad = Pradiated / Paccepted
SWR comes from |Γ|; radiation efficiency comes from power accounting. Neither equation contains common-mode current as a directly measured variable.
The Five Configurations
| Configuration | Observed 10 m behaviour | What the observation can tell us | What it cannot prove |
|---|---|---|---|
| Choke close to the UNUN with a short connection | Less attractive SWR in this installation. | The boundary condition near the UNUN differs from the other arrangements. | That the choke has poorer common-mode impedance. |
| Older long-pigtail unit with six clip-ons | A different 10 m curve. | Pigtail length and added ferrite affect the installed system. | Which change—length, ferrite, geometry or coupling—dominates. |
| Home-built bifilar assembly | The most attractive SWR of the initial comparison. | Its complete differential/common-mode behaviour transforms the impedance favourably. | Lower outside current, lower loss or higher radiation efficiency. |
| Short-pigtail choke plus 50 cm coax | The 10 m result worsened. | Moving the choke and exposing more coax exterior materially changes the RF system. | By itself, the exact mechanism. |
| Same 50 cm section plus ten clamp-ons | The result improved again. | Strong evidence that the external coax path participates and responds to added common-mode impedance. | The final radiation efficiency or ferrite dissipation. |
The Three Initial Measurements
Why 50 cm Matters on 10 Metres
At 28 MHz, the free-space wavelength is about 10.7 m, so 50 cm corresponds to roughly 17° of free-space phase. The internal TEM wave accumulates about 17°/VF: approximately 20–26° for coax velocity factors from 0.85 to 0.66. The outside-shield mode has a different effective velocity because its field occupies the surrounding structure and dielectric. That is not a quarter-wave stub—but it is not an invisible jumper either.
More importantly, placing 50 cm between the UNUN and the choke does two things at once:
- it adds length to the intended internal coaxial transmission path; and
- it moves the common-mode boundary 50 cm away, leaving more outside shield available to carry external current and couple to the antenna environment.
For an ideal lossless 50 Ω line observed with the same 50 Ω reference impedance, adding line length rotates impedance around a constant-SWR circle; it does not change the SWR magnitude. A substantial change after moving the choke therefore points beyond a simple uniform differential-line extension, provided the analyzer calibration and reference plane were unchanged. The antenna boundary condition, outside-shield current, loss, coupling—or some combination—has changed.
Do not call every short cable a stub. A stub has a defined termination and transmission-line role. Here the more careful description is an electrically active section of the external feedline structure, coupled to the feedpoint and its surroundings.
What the Clamp-On Test Adds
Jaap then repeated the 50 cm test with ten clamp-on ferrites placed around that section. The mechanical length remained, but its external current path received additional frequency-dependent impedance.
The 10-metre impedance response changes markedly.
The response changes back toward the useful direction.
The coax exterior and local electromagnetic environment participate in the measured load.
This is much stronger evidence than the first SWR comparison alone. Clamp-ons placed around the complete coax mainly add impedance to non-cancelling current. Their visible effect therefore supports the hypothesis that the outside-shield path was electrically active.
Evidence, not exclusivity: the experiment does not isolate every parasitic effect or quantify current directly. Clamp-on capacitance, placement and nearby geometry remain part of the structure. “Strongly supports common-mode participation” is justified; “proves the whole mechanism” would be too strong.
Why Moving the Choking Boundary Can Change SWR
A choke placed close to the UNUN removes more of the coax exterior from the feedpoint structure. In many installations that is desirable: it makes the antenna current distribution less dependent on cable routing, station wiring and nearby objects.
But if the old exposed section had been supplying useful capacitance, inductance, radiation resistance or loss, removing it changes the feedpoint impedance. The transmitter may see a worse match even though the feedline exterior carries less current.
After effective choking, the antenna may need deliberate retuning. That is not a failure of the choke; it is the difference between tuning the intended antenna and relying on a feedline contribution that changes whenever the coax moves.
The Bifilar Assembly: Possible Explanations, Not a Verdict
A bifilar common-mode choke is not inherently bad. If its conductor spacing and dielectric create a suitable differential characteristic impedance, and its winding provides useful common-mode impedance without excessive parasitics, it can work well.
It is nevertheless a different RF component from intact 50 Ω coax wound through ferrite. At the upper end of HF, the bifilar assembly may introduce:
- a differential impedance discontinuity if the pair is not near 50 Ω;
- leakage inductance and interwire capacitance;
- frequency-dependent common-mode resistance and reactance;
- self-resonance and coupling to leads or enclosure;
- conductor and ferrite loss.
Any of those can move the SWR. The present plots do not show which contribution dominates.
What would demonstrate ferrite heating?
ZCM = RCM + jXCM
Pferrite ≈ ICM,rms2RCM
The current and ZCM must use the same common-mode convention and reference plane; mixing per-conductor and summed-aperture definitions can create a factor-of-two current error.
Loss requires both current and resistance. A lower SWR does not supply either quantity. Measure common-mode current, component temperature under controlled power, and the assembly’s differential insertion loss before claiming that it “eats power.”
What the Existing Data Support
| Hypothesis | Support from this test | Next measurement |
|---|---|---|
| The section between UNUN and choke is electrically active. | Strong. Changing its length and adding ferrite both altered the response. | Map outside-shield current along the section. |
| The short-pigtail choke suppresses more common mode. | Plausible, not yet measured. | Compare net coax current with unchanged geometry and power. |
| The bifilar unit provides accidental matching. | Plausible. It produces a favourable impedance, but its modal behaviour is unknown. | Measure differential S-parameters and common-mode impedance separately. |
| The bifilar unit wastes more power as heat. | Not established. SWR cannot prove dissipation. | Temperature-rise or loss measurement at controlled current and duty cycle. |
| The lower-SWR configuration radiates better. | Not established. | Relative field-strength/pattern test plus accepted-power accounting. |
A Better Follow-Up Measurement Plan
Use the same low power and measure before/after the choke, along the 50 cm section, farther down the feedline and at the shack entry.
Record complex impedance, not only SWR. Keep the main feedline length, routing and analyzer plane unchanged for every configuration.
Compare component temperature at equal accepted power. A repeatable far-enough field-strength point is useful, but it can confuse a pattern change with an efficiency change; use several documented directions and polarizations, or a controlled efficiency method.
- Photograph and measure every physical spacing before changing hardware.
- Sweep all relevant bands, but analyse 10 m separately from the lower bands.
- Record
R + jX, SWR and feedline-exterior current for each configuration. - Measure outside-shield current at several points; a standing-wave minimum at one point can otherwise look like success.
- Repeat at low power first, then increase power in steps while checking choke, cable and connector temperature.
- Move or reroute the 50 cm section as a geometry-control test. A large response to position is evidence of external coupling.
RF.Guru product-rating boundary: RF.Guru common-mode chokes are specified from completed-assembly impedance, loss, power and thermal qualification under declared conditions. An installed SWR trace is a system-level measurement involving the antenna, transformer, connection length, coax exterior and surroundings; it neither establishes nor overturns those product ratings. Select the choke from its measured operating limits, then verify common-mode current, temperature and station behaviour in the actual installation.
What This Means for Everyday Antenna Work
- Do not rank common-mode chokes by the SWR they happen to produce.
- Expect the antenna match to move when a feedline that was radiating becomes better isolated.
- Keep the feedpoint-to-choke connection short unless that section has a deliberate, measured role.
- Retune the intended antenna after controlling common mode.
- Treat clip-ons as real, frequency-dependent impedances—not decorative insurance.
- Do not infer ferrite loss from SWR; measure current, impedance and temperature.
- Do not infer radiation efficiency from SWR; measure field behaviour or make a defensible power budget.
- Remember that an OCFD, transformer, choke, coax exterior and nearby structure form one electromagnetic system.
The Final Interpretation
PA0LJD’s follow-up test is persuasive because the same 50 cm section behaved differently when common-mode impedance was added around it. The most defensible conclusion is that the feedpoint-to-choke region and outside of the coax were participating in the 10-metre system.
The measurements do not yet crown a universal “best choke,” prove that the bifilar assembly is lossy, or establish which arrangement radiates the most power. They do show why a nicer SWR can be the signature of accidental impedance tuning—and why a cleaner current boundary can make the analyzer look worse before the antenna is deliberately retuned.
Technical background
- Rutgers University: Transmission Lines — reflection coefficient, impedance transformation and standing-wave behaviour.
- University of New Mexico, Measurement Note 44 — a three-surface coax model for separating intended and external current.
Mini-FAQ
- Does lower SWR mean better common-mode suppression? No. SWR measures impedance mismatch, while common-mode suppression must be judged from external current or a properly defined modal measurement.
- Can a better choke make SWR worse? Yes. It can remove an accidental feedline contribution that previously helped transform the impedance.
- Can 50 cm really matter on 10 metres? Yes. It is about 17° of free-space phase at 28 MHz, roughly 20–26° inside typical coax, and moving the choke also changes the external current boundary and near-field coupling.
- Did this test prove that the bifilar assembly wastes power? No. That requires current, loss or temperature measurements; SWR alone cannot establish dissipation.
- What did the ten clamp-ons demonstrate? Their effect strongly supports the conclusion that the exposed coax exterior was participating in the RF system.
- What should be measured next? Complex impedance and outside-shield current at several positions, followed by controlled temperature and relative field-strength tests.