Why a Hexbeam Can Sound Louder Without a Choke

Here is the uncomfortable observation: a Hexbeam can sound louder after you remove its common-mode choke. A wanted station rises. The background rises. The S-meter looks more enthusiastic. It is tempting to declare that the choke was “killing gain”—or, from the other side, that the louder result must be bad because the coax has become an antenna. Neither conclusion follows from loudness alone.

Louder Is an Observation, Not a Performance Metric

A receiver level can rise for several very different reasons:

• the installed antenna pattern has more response toward the wanted station;
• the coax exterior or station wiring has added effective aperture;
• local electrical noise couples more strongly into that external current path;
• the feed-point or receiver-plane mismatch has changed;
• the choke itself has differential loss, parasitic coupling or a damaged connection;
• the receiver has changed AGC state, preamplifier state, attenuation, bandwidth or overload behaviour; or
• propagation or the transmitting station changed while the hardware was being reconfigured.

Some of those changes can raise both signal and noise. Some raise one more than the other. A louder wanted signal with an even larger noise increase is worse for reception. A louder signal with the same noise can be better in that direction. A quieter result with better SNR can also be better. The figure of merit has to match the operating goal.

Removing the Choke Changes the System Boundary

In the intended coaxial mode, current on the centre conductor returns on the inner surface of the shield. A different current can flow on the shield exterior, closing through the antenna, mast, station chassis, protective bonding, other cables, earth and distributed capacitance.

A choke around the complete coax adds complex impedance to that exterior-current path while ideally disturbing the internal differential mode only slightly. Remove it and the coax exterior may carry more current. The feed line can then receive and radiate as part of the installed structure, and the fields from that path combine vectorially with the fields from the Hexbeam elements.

That does not guarantee a “wrecked” pattern. It guarantees only that the pattern belongs to a different structure. In one direction the added field may reinforce; in another it may cancel or fill a null. Front-to-back ratio, polarization, elevation response and azimuth response may all change. One louder station cannot tell you which of those happened.

Roy Lewallen, W7EL, separates these currents explicitly in “Baluns: What They Do and How They Do It”: the intended current returns on the inside of the coax shield, while significant outside-shield current makes the feed line part of the radiating or receiving structure. He also makes the crucial boundary clear—if the imbalance current is already small, a balun or choke may produce little useful change.

A Hexbeam Is Not Automatically Capacitive

A Hexbeam uses bent wire elements and mutual coupling to fit a directional array onto a compact support. That geometry does not force every finished antenna to present negative reactance. The feed-point impedance depends on the complete element lengths and spacing, conductor diameter, frequency, feed arrangement, nearby objects, mast, feed line and construction tolerances.

Measure the complex impedance R + jX at a declared reference plane. If the analyzer is at the shack end, the feed line transforms the antenna impedance and adds loss. Moving, adding or removing a choke can change the external current path and therefore the measured impedance. Calling one trace the “true Hexbeam impedance” is meaningful only after the system boundary and common-mode condition are fixed.

On transmit, a different mismatch can change accepted power, tuner loss and transmitter foldback. Higher indicated transmitter output is not automatically higher radiated power in the wanted direction. On receive, the same passive linear structure is reciprocal, but receiver noise, overload and local coupling still determine whether a pattern change improves usable SNR.

Use Source–Coupling-Path–Victim Diagnosis

I do not begin with “the choke is good” or “the choke is bad.” I begin with three boxes:

The current on the coax exterior is a candidate coupling path, not an automatic diagnosis. Noise can also arrive through the intended antenna mode, through finite shield transfer, or on power, USB, Ethernet, control and audio cables. Conversely, a measurable exterior current may exist without being the dominant source of the received noise.

ITU-R P.372-17 distinguishes noise received through the reference antenna and feeder from noise entering through other conductors, inadequate screening or feeder imbalance. That boundary is exactly why a single S-meter movement cannot identify the path.

Measure Signal and Noise in the Same Reference Bandwidth

For a receive comparison, freeze every receiver setting you can: frequency, mode, resolution or audio bandwidth, preamplifier, attenuation, RF gain, AGC, detector, averaging and display reference. Record the I/Q stream when possible so both configurations can be processed identically.

Use a stable wanted source. A laboratory generator feeding a remote test antenna is better than a fading DX signal. A stable ground-wave beacon can be useful when its level and path are repeatable. If the on-air source varies, switch rapidly and repeat enough cycles to estimate the drift.

Measure the noise in a nearby signal-free interval using the same equivalent noise bandwidth. If the reported “signal” bin contains signal plus noise, subtract powers in linear units before calculating SNR:

Do not subtract dB readings directly to remove noise from a combined signal-plus-noise value. A calibrated receiver or spectrum analyzer is preferable; an S-meter can still support a relative experiment only after its steps, AGC behaviour and repeatability have been characterized.

Clamp Current Is a Separate Record

A characterized current probe around the complete coax measures the net signed RF current through its aperture at that cross-section. Because the intended centre-conductor and inner-shield currents substantially cancel, the reading is normally a useful estimate of exterior-shield current—provided the complete cable is enclosed, the probe is calibrated and no other return conductor shares the aperture.

Record the probe transfer impedance at the test frequency, receiver termination, bandwidth, detector, cable corrections, noise floor and uncertainty. Mark several positions along the coax without rerouting it. Exterior current can form a standing-wave distribution; one convenient point may be a current minimum.

A lower clamp reading after installing the choke shows that net cable current fell at that plane. It does not by itself prove less total radiation, better pattern, lower local-noise pickup or higher SNR. Those are correlated measurements, not synonyms.

Run a Real A/B/A Test

Repeat the cycle and, when practical, randomize the order. Photograph cable and choke placement. For transmit checks, use safe controlled power, log forward and reflected quantities at a declared plane, record tuner state and transmitter foldback, and de-energize before touching the feed system.

To test a pattern claim, one station is not enough. Use a controlled antenna range, a model validated against installed current and geometry, or multiple stable sources over azimuth and elevation. IEEE 149-2021 treats antenna pattern, gain, impedance, instrumentation, site errors and uncertainty as a complete measurement problem—not as an S-meter anecdote.

Interpret the Combination, Not One Number

A choke can be valuable because it makes the installed system more predictable, reduces an unwanted coupling path or restores the intended pattern. It can also be unnecessary at one frequency, poorly placed, lossy, resonant in the wrong place or bypassed by another cable. “Always fit one” and “it is louder without one” are both incomplete engineering specifications.

The Defensible Conclusion

If the Hexbeam is louder without a choke, take the result seriously—but do not let it choose its own explanation. The coax and station may have joined the receive antenna. The mismatch may have moved. The local-noise path may have changed. The pattern may favour that one station. The receiver may be behaving differently.

Better means the declared objective improved: SNR in the wanted directions, controlled pattern, acceptable mismatch and loss, adequate receiver headroom, lower unwanted cable current, or reduced EMC risk. Show the before/after records and their uncertainty. Then the choke becomes a measured system component instead of a ritual object.

If you have a documented A/B/A record—signal, noise, SNR, cable current and exact geometry—send it to RF.Guru. A reproducible surprise is much more useful than a louder S-meter.

Primary and authoritative technical sources

• Roy Lewallen, W7EL: “Baluns: What They Do and How They Do It”—coax inner/outer currents, imbalance-current paths and pattern effects.
• ARRL QST: “Common-Mode Chokes”—whole-coax clamp measurement, position dependence and installed before/after verification.
• ITU-R P.372-17: Radio noise—external noise sources and the boundary between antenna/feeder reception and other cable or screening paths.
• ITU-T K.37 (01/2024)—current EMC guidance for source verification, cabling, screening, filtering, bonding and interference mitigation.
• IEEE 149-2021—antenna pattern, gain, impedance, instrumentation, site-error and uncertainty measurement practice.
• BIPM/JCGM Guides in Metrology—measurement models, repeatability, correlation and uncertainty reporting.

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