OS5Z “Dreamer” Antennas: Match, Pattern and Evidence
OS5Z “Dreamer” Antennas: Match, Pattern and Evidence
The OS5Z experiment is interesting precisely because it makes strong, testable claims. The fair response is to preserve the published design, then ask which measurements establish match, efficiency, current path and gain.
La Gigazette n°179 introduced Marc, ON5CAZ’s OS5Z antenna as a compact, switchable HF experiment. The publication includes drawings, dimensions and a fourteen-question interview. It reports a double-lozenge top structure, several loading coils and a separate conductor connected to the coax shield, switched to approximately a quarter wavelength for the selected band.
That is a useful starting point. It gives us a geometry to discuss and claims that can be tested. It does not, by itself, supply the current distribution, radiation efficiency or three-dimensional far-field pattern needed to turn an on-air impression into a gain figure.
The source matters. This is a technical response to “Antenne OS5Z” in La Gigazette n°179, credited to ON5CAZ, with images and diagrams by ON5CAZ and adaptation, text and interview by ON6WG. The experiment and its context remain part of the story.
The published antenna system
The drawing presents about 42.856 metres of wire in the upper structure, with coils and a double-lozenge outline. The centre conductor of the coax feeds that structure. A second conductor leaves the shield side, drops vertically and continues near the ground. Switch positions select different lengths of this conductor for different bands.
The interview reports operation without an external tuner, balun or resistor, with an impedance described as very close to 50 Ω across selected frequencies from roughly 1.6 to 28.5 MHz. It also reports 1:1 SWR, a narrow low-SWR region, and a height-sensitive near-ground conductor above about 9 MHz. These are the publication’s claims, not values independently measured by RF.Guru.
Electrically, the most defensible first model is an unbalanced, coupled antenna system: an upper loaded radiator, a deliberately selected shield-side return conductor, the feedline and their common surroundings. That is a hypothesis to test with current measurements and a complete model—not a claim that every ampere follows one simple sketch.
A switched return conductor is part of the antenna. Calling it a dissipation wire, ground wire or counterpoise does not change the RF question: what current flows on it, what current continues onto the outside of the coax, and how do those currents change with band, height, soil and cable routing?
A good match answers only one question
SWR describes the impedance relationship at a stated reference plane. It does not directly reveal radiation efficiency or pattern. For a simplified feedpoint-resistance discussion:
Rin = Rradiation + Rloss
This identity is useful only when the terms and reference plane are defined. A 50 Ω resistive input may contain radiation resistance, conductor and coil loss, ground loss, dielectric loss and other coupled paths. The match alone cannot tell us their proportions.
The published geometry provides plausible mechanisms for both radiation and loss. The coils alter electrical length and current distribution. The near-ground conductor couples to soil and nearby objects. The outside of the feedline may carry common-mode current. Any of these may change the impedance presented to the transmitter. None may be assigned a numerical share without measurement or a sufficiently detailed validated model.
Geometry is not a performance certificate
“Fractal,” “lozenge” and “rhombic” describe aspects of shape. Shape matters because it changes mutual coupling, current distribution, impedance and pattern. The label does not import the gain of a classical travelling-wave rhombic, whose dimensions, termination and electrical aperture are integral to its behaviour.
A folded or space-filling wire can fit more conductor into a compact area and can create additional resonances. Loading can make a structure electrically longer than its straight-line dimensions. Those are legitimate design tools. They do not remove the need to account for bandwidth, loss, current phase, height and the surrounding ground.
The pattern and gain claims need the same reference
The interview describes the pattern as quasi-omnidirectional or resembling a four-leaf clover, inferred from contacts and Reverse Beacon Network observations. It also gives gain claims in the region of 9 to 14 dBd and compares the antenna with a rhombic.
Those propositions cannot be evaluated from one azimuth label. Gain combines efficiency and directivity, and must be stated against a reference antenna, direction, polarisation, frequency and installation. A four-lobed pattern may have useful coverage, but it also contains maxima and minima. A high peak-gain claim implies that radiation is concentrated in angle; it cannot simultaneously describe uniform coverage in all directions.
The way out of the apparent contradiction is measurement: publish the complete azimuth and elevation pattern, the peak direction, the reference used for dBd, the feed loss and the accepted power. IEEE 149 treats radiation-pattern measurement as a fundamental antenna property, while ITU-R BS.705 separates pattern, directivity, gain, ground and environmental effects for HF antennas.
Establishes the input condition at a declared reference plane. It does not establish far-field gain or radiation efficiency.
Show that signals were heard through a changing propagation path and a non-uniform receiver network. They are useful operational evidence, not a calibrated three-dimensional pattern.
Can predict current and pattern for a fully specified geometry, feed, conductor, ground and environment. Validation against measurement remains important.
Can quantify repeatable differences when switching is fast, source power is monitored and both antennas share comparable height, environment and feed losses.
What Reverse Beacon data can and cannot say
The Reverse Beacon Network records which volunteer receivers heard a transmission, when, on which frequency and at what reported level. Its own comparison tool can produce quantitative comparisons when stations are heard by the same beacon on the same band and at closely related times.
That makes RBN valuable for disciplined A/B work. A map of all spots is not a radiation pattern, however. Receiver locations, beacon availability, propagation, noise, transmit timing and calling behaviour are all mixed into the result. The data become stronger when an experiment alternates two antennas rapidly, holds transmitter power constant and compares paired reports from the same receivers over many cycles.
Soil and feedline routing belong in the test
The source says the shield-side conductor must run close to the ground on parts of HF, yet reports little difference between wet and dry soil. That is worth testing, not dismissing. ITU-R’s HF antenna guidance explicitly treats ground conductivity, topography and surrounding structures as influences on practical radiation patterns.
Measure the installed system in more than one soil condition and repeat the test with the feedline routed differently. Add a high-impedance common-mode choke at a declared position and measure the currents on both sides. If the impedance, field strength or pattern changes, that is useful evidence about the current path—not a failure of the experiment.
The Sixpoles and Octopus discussion
The related Sixpoles and Octopus ideas use several arms from a common region. These structures deserve their own full-system analysis. The arms are mutually coupled; their lengths cannot simply be added as though they formed one series conductor. Depending on geometry and feed, a multi-arm structure may provide several resonances, change impedance bandwidth or reshape azimuth coverage.
Crossed and multi-arm arrangements can reduce a deep null in one direction while reducing the maximum somewhere else. With controlled phase they can form an array; without it, current magnitude and phase emerge from the coupled structure. “More wire” is therefore neither proof of more gain nor proof of no benefit.
A fair measurement programme
Record every wire dimension, coil construction, switch state, feedline length and route, choke position, height, ground condition and nearby structure.
Record complex impedance across each operating band at the antenna feed, then separately account for feedline loss and the transmitter-end match.
Use a calibrated clamp or repeatable current probe along the feedline and return conductor. Repeat after rerouting the coax and adding a known choke.
Switch between OS5Z and a declared reference antenna while monitoring forward power. Use paired RBN reports or a stable local field-strength receiver over many cycles.
Use an appropriate field, pattern, calorimetric or equivalent method. Do not infer efficiency from feedpoint resistance alone.
Show azimuth and elevation cuts, frequency, polarisation, peak direction, reference and uncertainty rather than one gain number detached from its conditions.
The engineering conclusion
The OS5Z antenna is not “nothing.” It is a documented amateur experiment with a compact loaded top structure and a switched return conductor. It reportedly makes contacts and presents a convenient match in its installation. Those are meaningful observations.
The stronger claims remain open. A near-50 Ω input does not establish efficiency. RBN coverage does not establish a radiation pattern. A geometric resemblance does not establish rhombic gain. Claims of quasi-omnidirectional coverage and 9–14 dBd need a shared definition, a direction, a reference and measured evidence.
That distinction is not hostile to experimentation. It is what makes an experiment reusable. Publish the geometry, the reference plane, the current map, the loss method and the pattern, and other amateurs can reproduce the result rather than merely inherit the claim.
Primary and authoritative sources
- La Gigazette n°179 — “Antenne OS5Z”, source drawings and interview.
- IEEE 149-2021 — Recommended Practice for Antenna Measurements.
- Recommendation ITU-R BS.705-2 — HF transmitting and receiving antenna characteristics and diagrams.
- Reverse Beacon Network, network scope and comparison method.
Mini-FAQ
- Does a 1:1 SWR prove that the OS5Z is efficient? No. It establishes a match at a stated reference plane. Efficiency requires the radiated power to be separated from conductor, coil, ground, dielectric and feed losses.
- Is the shield-side wire merely a ground connection? It carries an RF current determined by the complete antenna, feedline and environment. Its switched length and installation make it an intentional part of the RF system.
- Can Reverse Beacon reports compare antennas? Yes, when used as paired observations from the same receivers with rapid switching and controlled power. A general spot map is not a calibrated radiation pattern.
- Does the lozenge shape make it a rhombic antenna? Not by itself. Classical rhombic performance depends on electrical aperture, travelling-wave behaviour, termination, height and a defined pattern—not visual resemblance alone.
- Can a Sixpoles or Octopus arrangement be useful? Yes. Coupled arms can create several resonances, alter bandwidth and reshape coverage. The benefit and any gain still depend on measured current, phase, loss and pattern.
- What measurement would clarify the claim fastest? Measure complex feedpoint impedance and common-mode current, then run a rapid power-controlled A/B comparison against a declared reference antenna in the same environment.