When Boat HF RFI Refuses to Behave: Jan’s Long Road to a Real Fix

Marine HF advice often sounds like a one-evening repair: add a strap, move the tuner, fit a ferrite, problem solved. Jan, ON2ZTT, had a rather longer experience aboard SY Goudurix. From July 2025 into early 2026, he kept testing after promising improvements failed to hold. That persistence is the point of this story.

HF transmission was upsetting the boat’s solar-control chain. The possible routes included the radiator, tuner return, coax exterior, DC and control wiring, nearby structures and the controllers themselves. This was a real maritime-mobile installation on a fibreglass sailboat, with deck routing and equipment positions that could not simply be rearranged like a bench experiment. The system was not “antenna plus tuner.” It was the boat.

This was not a weekend fix

• July 2025: our discussion began with the familiar marine-HF questions about counterpoises, tuner placement and what “RF ground” could actually mean on a fibreglass boat.
• September 2025: Jan installed Schaffner filters and tried several return-path and counterpoise arrangements. There were encouraging results, but not yet a dependable answer.
• October 2025: photo-documented trials compared the SGC-230 tuner under deck and above deck, with different counterpoise combinations and choke positions. Some layouts still caused RFI on 30 m; others also struggled on 20 m or 17 m.
• November and December 2025: further trials showed why a good result on one occasion was not enough. A configuration could behave well and then misbehave later under different conditions.
• Early 2026: after replacing the solar controllers and keeping the Schaffners installed, Jan reported that the solar-related RFI was gone. Attention then shifted to matching a shorter, carbon-supported vertical on 15 m and 17 m.

That is a much more useful account than a photograph of the final wiring with “works perfectly” underneath it. The unsuccessful steps tell us why the successful combination mattered.

What Jan actually changed

The trials were substantial. The tuner moved above and below deck. Counterpoise wires of roughly 2.5 m, 3.5 m, 3.8 m, 7.5 m and 9 m appeared in different combinations. Some were shortened or removed; some were rolled up during experiments. A short copper route connected the tuner area to the boat’s existing underwater bronze plate. FT240-31 and FT240-43 ferrite arrangements were tried at different feedline and cable-bundle positions, alongside Schaffner line filters in the solar path.

The radiator changed too: a 9.2 m sloper was compared with a shorter vertical described as about 7.2 m physically and approximately 7.68 m in total radiator length. Tests took place in the marina, ashore during maintenance, and at sea or at anchor. These are the dimensions and conditions in Jan’s account, not a transferable cutting list or a filter recipe for every vessel.

One detail deserves care: rolling up a wire changes its inductance, capacitance and coupling; it does not guarantee that the wire becomes electrically insignificant. A removed conductor and a coiled conductor are different RF configurations. Likewise, naming a ferrite size and mix does not specify the impedance, winding capacitance or thermal performance of the finished choke.

The filters helped—but the apparent cure did not always last

Jan reported immediate improvement on some bands after fitting the Schaffner filters. At one stage, everything behaved except 30 m. Later, with the tuner back under deck, several ferrite chokes fitted and the experimental counterpoises taken out of active use, he reported no RFI across the tested bands, including successful operation on 17 m.

It would have been easy to stop there. Yet later trials brought trouble back on some combinations, particularly around 20 m, 17 m and 30 m. That did not make the earlier improvement imaginary. It showed that the margin depended on the complete installation and operating conditions.

A controller’s switching state and load can change its behaviour. Moving the boat, changing shore connections or moving a conductor can change the RF network. The practical lesson is not “nothing can be known.” It is that a remedy which improves the failure threshold is useful evidence, but a one-off quiet test is not the same as a lasting fix.

Why adding more metal was not a reliable answer

Extra return wires did not reliably calm this installation. Some configurations became less predictable, and removing experimental conductors sometimes helped. That is entirely plausible: the return conductor is part of the antenna system, not a place where RF politely stops existing.

A wire’s RF impedance includes inductance, capacitance to its surroundings, loss and resonance. An added conductor can provide a useful return on one band and carry substantial current or move a resonance on another. It can also change the field at a nearby controller or cable. Jan’s observations fit that mechanism; they do not by themselves map the current on every wire.

This is why I distrust “more RF ground” as a diagnosis. The engineering task is to give the intended antenna current a workable return while limiting unwanted paths into the equipment. It is not to win a copper-strap competition. Nor does this make all counterpoises harmful: their geometry and the resulting current distribution decide whether they help.

The controller swap changed the direction of the investigation

The decisive turn came when Jan stopped assuming that the antenna side had to carry all the blame. He temporarily substituted a Victron MPPT for one Mastervolt solar controller, on the side of the boat closest to the antenna/controller area. He reported no RFI on the HF bands he tested with that substitution.

That was a strong reason to investigate the affected controller’s immunity in this installation. Chokes, tuner position and return conductors still mattered, but the receiving end of the interference path was no longer being treated as an innocent bystander. After replacing the controllers and retaining the Schaffner filters, Jan reported that the solar-system RFI had gone.

The useful conclusion is direct: on this boat, controller replacement belonged to the successful system fix. It was not enough to keep modifying the antenna indefinitely while refusing to question the equipment it was disturbing. The result does not show that every Mastervolt controller will fail, that every Victron unit is immune, or that filtering can be discarded. Those would be different claims requiring specified models, ports, loads, RF levels and repeatable tests.

For another owner facing the same dead end, a compatible controller substitution can therefore be a meaningful diagnostic step—not a brand-shopping rule. Keep the effective filtering and cable arrangement, test the same frequencies and operating conditions, and involve the equipment manufacturer with the findings. Battery, solar-input and protection compatibility still have to be right.

The bronze plate was not a magic RF ground

The short copper connection to the existing underwater bronze plate formed part of a workable return arrangement. That does not prove that the plate alone cured the RFI. The extra conductors, filtering, choking and controller selection were all part of the story.

Water and underwater metal can influence the RF network, but that connection also lives inside a real marine electrical system. Protective conductors, DC negative, galvanic isolation, cathodic protection and lightning provisions cannot be casually merged or rerouted to chase a radio symptom. ISO 13297 covers AC and DC electrical installations on small craft; changes must respect the craft’s design and applicable requirements. Use a qualified marine electrician or corrosion specialist where those systems are involved.

Controlled return paths are useful engineering. A universal “magic ground” is not.

The solar RFI was gone; the matching question remained

With the solar controllers behaving, Jan still had difficulty with the shorter vertical on 15 m and 17 m. The radiator was around 7.68 m in total length and used a carbon support. That brought us back to ordinary antenna questions: the complex feedpoint impedance, the tuner’s matching range, the return path, and coupling to the support and rigging.

Carbon composites can conduct and couple capacitively as well as introduce loss. Their effect depends on construction and geometry, so the presence of a carbon pole does not quantify the detuning. Nor should we pronounce a particular half-wave condition from the physical wire length alone: end effects, the support and the rest of the boat change the installed system.

A modest geometry change or a properly designed loading/matching element might improve the tuner’s working range, but neither is a free efficiency gain. First establish the impedance at a declared measurement plane, then consider tuner loss, component voltage and heating. A VNA or antenna analyser can help with that impedance work; neither becomes an antenna-efficiency meter simply because it displays a good SWR.

This distinction matters to the ending. Jan had made real progress: the solar-side fault was no longer dominating the job. An unresolved match on two bands did not erase that result, and a better match would not, by itself, prove controller immunity.

What the boat teaches us about coupling

The reported fault was transmit susceptibility: wanted HF energy disturbed the solar electronics. That is different from switching electronics emitting noise into a receiver. One device may behave well in one direction and poorly in the other.

There are several possible paths. Differential-mode disturbances appear between the conductors of a circuit. Common-mode cable current uses a wider return network, and a nearby field can also couple directly into a cabinet, circuit or sensing input. Those mechanisms can coexist. Jan’s sequence supports treating the boat as a coupled system; it does not establish which cable or internal circuit carried every disturbance.

IEC 61000-4-6 provides controlled conducted-RF immunity methods for 150 kHz to 80 MHz, while IEC 61000-4-3 addresses radiated-field immunity. A close HF antenna and a boat’s wiring do not automatically reproduce those laboratory conditions. IEC 60533 explicitly concerns ships with metallic hulls; it should not be treated as a blanket standard for Jan’s fibreglass vessel.

For common-mode suppression on a DC pair, both conductors normally pass through the same core so their intended DC magnetic effects largely cancel. A filter only across the pair primarily treats differential mode unless it provides another defined path. Neither “fit a ferrite” nor “fit a line filter” is a complete specification: the port, coupling mode, frequency response, current, fault rating, environment and thermal behaviour matter. Keep unfiltered leads short at the susceptible equipment and do not introduce an unsafe path through filter capacitors or grounding connections.

Borrow the discipline, not the exact wiring

The point of telling Jan’s story is not to promise that his final arrangement will cure every boat. It is to show where to look when the easy answers have run out:

• Keep a repeatable baseline. Record frequency, modulation, power, duty cycle, tuner state, controller load, battery and solar conditions, location and shore-power state. Photograph cable and choke positions. Repeat successful tests later and under realistic load.
• Separate the branches. A suitably rated dummy load at a declared reference plane can help distinguish antenna-related coupling from station-wiring paths. Its enclosure and coax may still carry common-mode current, so it is a clue, not an absolute isolation test.
• Follow the cable current. A characterised RF current probe can compare coax-exterior and controller-cable current at fixed positions. A near-field probe can locate suspect areas. Absolute claims require calibration, bandwidth, detector, geometry and uncertainty records; a probe indication is not a compliance certificate.
• Change one element when practical. A cable route, choke position, non-safety branch or compatible controller can reveal more than a pile of simultaneous remedies. Where safe, restore the previous arrangement to see whether the symptom follows the change.
• Do not stop at the first good band. Retest the full operating set. An improvement on 20 m can leave 17 m or 30 m troublesome. Keep the effective parts of the configuration while testing the next suspect.

For interference to radio reception, CISPR TR 16-4-6 offers field-verification and root-cause guidance. Its receiver-interference scope should not be confused with a certification test for the solar-controller malfunction in this case.

No RFI cure is worth defeating the boat’s safety systems

Begin transmit tests at low power and increase cautiously. Stop if there is heating, arcing, a protective-device trip, accessible RF voltage, or disturbance to navigation, alarms or other essential equipment. Do not interrupt protective earth, defeat overcurrent protection or galvanic isolation, or alter cathodic and lightning provisions to obtain a quiet radio test.

Filters, ferrites, cables, tuners and terminations need appropriate electrical, thermal and marine environmental ratings. Do not modify equipment internally without manufacturer or qualified responsible-engineer approval. Verify human RF exposure and contact-current risks for accessible locations and foreseeable crew movement; ITU-T K.52 and ICNIRP’s RF guidance provide the assessment framework, not a universal safe distance for every boat.

A dependable outcome means the intended bands, power levels and controller loads work repeatedly without overheating or upsetting essential systems, with all safety functions intact. It must survive more than the particular afternoon when the last ferrite was fitted.

Why Jan’s persistence mattered

Jan did not get there by finding a more convincing slogan. He kept going when a partial fix looked tempting, tested the controller as well as the antenna, and separated the fault that had gone from the matching problem that remained.

That is why I wanted to tell this story. Filters helped. Return-current discipline helped. The controller substitution changed the outcome. Real RF work can be slow, iterative and slightly annoying—but that is how a hopeful explanation becomes a configuration you can actually use aboard a boat.

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