Will It Make Any Difference?
Will It Make Any Difference?
Six years of building, swapping and measuring receive antennas at Lombardsijde—from a failed Wellbrook and a first OctaLoop prototype to third-generation designs backed by millions of real reception reports.
Six years ago, my trusted Wellbrook receive loop died. I still respect the electronics and the thinking behind it; the part that ended its life was mechanical. It had not been designed with the longevity I wanted from an antenna that lives outdoors. That failure started a simple question: could we build a receive antenna that was not merely different, but measurably better as a complete system?
The Failure That Started the Project
My son Jonas began designing the mechanical parts. I started on the first electronics. The private goal was deliberately ambitious: beat the Wellbrook. But “beat” needed a better definition than a louder S-meter reading.
Connectors, shield gaps, supports, fasteners, strain relief, water paths and corrosion control had to preserve symmetry for years—not only on the workbench.
The design had to remain balanced, linear and stable in the presence of strong broadcast signals and the common-mode noise carried by a real station.
A new version would count as better only if it produced a repeatable advantage under changing propagation, weather and local noise.
The first OctaLoop release passed its practical test. Early adopters sent encouraging feedback, and the antenna worked. Yet I kept seeing possible improvements. That created the harder question: how do you compare two already-good receive antennas closely enough to know whether a revision is real?
Why a Far-Field Test Was Not Enough
Far-field measurements remain useful. They can reveal frequency response, directional pattern, null depth, sensitivity and imbalance under a controlled plane-wave stimulus. We performed that kind of testing. It told us whether the antenna behaved as an antenna.
It did not fully answer the question that mattered most to us. RF.Guru designs are built around common-mode rejection, and common-mode performance is a system property. Local interference can arrive on the outside of the coax, through a mast, along a power lead, through unequal capacitance to nearby objects or through small mechanical asymmetries. A clean far-field signal does not recreate all those paths.
Far-field testing asks: how well does the antenna respond to the wanted electromagnetic field?
CMRR-first field testing also asks: how much unwanted signal does the complete installation prevent from becoming differential signal at the receiver?
A high CMRR number in an amplifier datasheet is not enough. The effective result depends on the element, matching, protection components, PCB symmetry, enclosure, connectors, feedline, mast, grounding and the impedances seen by both balanced legs. Noise that has already been converted into differential mode before it reaches the amplifier will pass through like any wanted signal.
CMRR is defined as the ratio of differential gain to common-mode gain:
CMRR(dB) = 20 × log10(|Ad / Ac|)
A 50 dB ratio is about 316:1 in voltage. That is powerful—but only for interference that truly appears as common mode at the balanced input.
The Database Behind spots.rf.guru
I left the comparison problem alone for a while, but it kept returning. Eventually I built the database behind spots.rf.guru. The public site shows the latest operational versions of the antennas. Behind that view, long-term reception records give us the raw material needed to compare designs over days and weeks instead of selecting one attractive screenshot.
The maps are based on stations the receiving system actually decoded and reported to PSKReporter. They are not modelled radiation patterns and they are not propagation forecasts. Every plotted path begins with a real reception report stored in our own archive.
Our Long-Term A/B Test
For a version comparison, we install the new and previous designs at the same site, separated by approximately six metres, and run them together for a month. They experience the same broad propagation, solar conditions, weather and station activity. Because the observations are simultaneous, a band opening benefits both candidates instead of being mistaken for an antenna improvement.
Both versions collect spots during the same minutes, band openings and local-noise events.
A long window includes weekdays, weekends, day/night cycles, quiet periods and disturbed conditions.
The antennas trade places, and the test continues. A real design advantage should follow the antenna—not the patch of ground.
Why separate them by six metres?
We want enough physical separation that the structures, immediate cable routes and local coupling are not identical, while keeping both antennas inside essentially the same propagation and noise environment. Six metres is not “electrical infinity” on the low HF bands, and the test is not an anechoic chamber. It is a practical compromise: the antennas are not mounted on top of one another, but neither is one at a different receiving site.
Why swap their positions?
Even a small move can change the result. Soil, buried conductors, fences, buildings, lamps, mast hardware, cable direction and a local RFI source can favour one position. If antenna A wins before and after the swap, the advantage is more likely to belong to antenna A. If the advantage stays at the same position, we have measured the site instead.
| Possible source of bias | How the method reduces it | What still needs discipline |
|---|---|---|
| Propagation and station activity | Simultaneous reception exposes both versions to the same opening. | Compare matched time windows and exclude downtime. |
| Position and local RFI gradient | The position swap makes each antenna occupy both test locations. | Keep height, orientation and nearby hardware consistent. |
| Receiver-channel gain or filtering | Fixed, calibrated channels prevent a receiver difference becoming an antenna claim. | Track gain, filters, firmware, clocks and reporting rules. |
| Feedline and connector differences | Controlled cable types and routes reduce loss and common-mode variation. | Inspect weather sealing, connectors and bias feeds throughout the run. |
| One exceptional path | A month of reports replaces anecdotes with distributions. | Judge median SNR, decode counts, bands, unique stations and repeatability together. |
We have used this approach while developing the OctaLoop, VerticalVortex, SkyTracer and TerraBooster families. It does not make field data perfect. It makes the largest sources of self-deception visible.
So, Did Six Years of Revisions Make a Difference?
For most designs, the answer is yes—but not by a dramatic amount. Most current antennas are now around version 3. Compared with version 1, the improvement is typically about 1.5 dB, give or take. That is real, repeatable progress, but it is not a miracle.
That is roughly 1.41 times the received power, or 1.19 times the voltage into the same impedance.
Version 3 changed topology, not merely component values. Three decibels is approximately twice the power, or 1.41 times the voltage.
Using the common—but not universal—6 dB-per-S-unit convention, 1.5 dB is only about one quarter of an S-unit and 3 dB is about half an S-unit. Many operators will not look at an S-meter and say, “That is a completely different antenna.”
Yet a small shift can matter close to a threshold. A signal that moves from just below a decoder or readability threshold to just above it changes the outcome from no report to a report. Repeated over many stations, bands and fading cycles, a modest SNR advantage can produce more successful decodes and more robust copy without ever looking spectacular in a single comparison.
A dB result needs context. A higher spot count alone is not proof of higher gain. Uptime, reporting software, receiver settings, band occupancy and propagation all affect the count. We look for an advantage that persists in matched time windows, follows the antenna through the position swap and appears in SNR distributions—not only in the most flattering total.
Why TerraBooster Changed More
TerraBooster is the design that shifted most clearly. Between versions 2 and 3, we changed the circuit topology. On some bands, the improvement approaches 3 dB. Just as important, the new design behaves more linearly across the covered bands, bringing its overall response closer to the consistency of our other receive antennas.
This distinction matters. Refining protection capacitance, matching, layout or mechanics usually produces incremental gains. Changing the topology can move the operating point of the entire system. The larger result is therefore not evidence that every new revision should gain 3 dB; it is evidence that the previous architecture still contained a more fundamental limitation.
CMRR First: The Principle Behind the Results
We receive emails asking why the Lombardsijde station hears New Zealand, Australia and Japan so regularly. The honest answer begins with propagation: when the path is open, it is open. No receive antenna can manufacture ionospheric propagation.
What the antenna can do is avoid throwing away the available SNR. Lombardsijde is a demanding receiving site: an urban park, approximately two metres below sea level, with LED floodlight interference and restrictions on mast height. In that environment, the battle is often not against Johnson noise inside the first transistor. It is against local interference, common-mode current, overload and unwanted signals converted into differential mode by asymmetry.
That is why we design CMRR first. The two signal paths must remain electrically and mechanically symmetrical. The shield, loop gap, protection network, transformers, connectors and PCB are not separate details; together they decide whether common-mode pickup cancels or becomes part of the received signal.
It is also why noise figure cannot be viewed alone. At HF, atmospheric, galactic and man-made external noise is often far above the 290 K thermal reference. Once the first stage is quiet enough, another fraction of a decibel of NF may contribute less to real SNR than better linearity, filtering, stability and feedline-current control. At a very quiet site, toward the upper end of HF, or with a very low-output element, NF can become important again. There is no single specification that replaces system design.
A 60 cm shielded loop with a balanced input, symmetrical push-pull gain paths, controlled power-off grounding and feedline-current suppression. Typical published figures include CMRR above 50 dB, approximately +22 dB gain, +41 dBm OIP3 and +19 dBm P1dB.
A high-impedance active E-field probe with an isolated coax interface, common-mode suppression, FM broadcast rejection and a whip-dependent range extending from LF/HF monitoring into VHF, UHF and L-band.
These are different sensors. One primarily samples the magnetic field; the other samples the electric field. Their geometry, installation and use cases differ. The shared philosophy is that the front end, common-mode path, strong-signal behaviour, protection and mechanical construction must be engineered as one receive system.
From Digital Spots to CW-Equivalent Paths
The full map includes digital decodes such as FT8, which can be decoded below the audible noise floor. To make the data more intuitive for operators, the map service also publishes filtered views using signal-to-noise thresholds.
The CW-equivalent plate keeps reports at SNR ≥ −18 dB. In this particular seven-day dataset, the receivers did not report any spot below −16 dB. The CW filter therefore removes nothing, and the CW-equivalent map is identical to the complete digital map. That is a result—not a rendering error.
The More Demanding SSB-Equivalent View
The SSB-equivalent plate is deliberately stricter. It keeps only reports at SNR ≥ 0 dB. From the same seven-day archive, that reduces the map from 413,621 digital spots to 4,982 stronger reports.
Reference figures above: 17–24 August 2026. The live SSB-equivalent plate continues to refresh.
What the Maps Prove—and What They Do Not
| The maps do show | The maps do not show |
|---|---|
| Stations the Lombardsijde receiving systems actually decoded and reported during the stated window. | A modelled antenna pattern or a promise that the same paths will exist at another location. |
| Coverage across bands, continents, DXCC entities and grid squares without PSKReporter’s map-display cap. | That the antenna caused the propagation or that every path was open continuously. |
| A live operational view of the latest antenna versions at the site. | A controlled version-to-version gain measurement; that requires the separate A/B and position-swap analysis. |
| Digital reports that meet the selected CW- or SSB-equivalent SNR threshold. | Actual CW or SSB reception reports, QSOs, intelligibility scores or operator copy. |
Will It Make Any Difference?
Yes. But the honest engineering answer is more interesting than a sales slogan.
For most RF.Guru antenna families, six years of revisions have produced about 1.5 dB of measurable improvement from version 1 to version 3. The TerraBooster’s topology change produced close to 3 dB on some bands and a more linear response. Those gains can rescue marginal signals, but they do not replace propagation and they do not turn every band opening into an easy contact.
The larger difference is harder to reduce to one number: better mechanical longevity, more consistent balance, controlled common-mode paths, stronger overload behaviour, repeatable production and a test process that makes us prove an improvement before calling it one.
That is why New Zealand, Australia and Japan appear so regularly in our logs when propagation supports the path—even from an urban site. The antenna does not create the opening. A CMRR-first system simply gives the opening a better chance of surviving the local environment.
Mini-FAQ
- Is far-field testing useless for a CMRR-first antenna? No. It remains valuable for pattern, response, sensitivity and null measurements. It simply does not reproduce every common-mode and local-noise path of a real installation.
- Why test two antennas simultaneously? Simultaneous operation exposes both designs to the same band openings, weather and station activity, reducing the chance that propagation is mistaken for an antenna advantage.
- Why swap their positions? The swap helps separate antenna performance from site bias. A real advantage should move with the antenna.
- Is 1.5 dB audible? Often it is subtle, roughly one quarter of an S-unit under the common 6 dB convention. Near a decode or readability threshold, however, a small repeatable SNR shift can change the outcome.
- Are the CW and SSB maps made from CW and SSB reports? No. They filter digital reports by SNR thresholds of −18 dB and 0 dB. They are useful equivalents, not mode-specific QSO logs.
- Why is the CW-equivalent map identical to the digital map? In this seven-day sample, every stored report already exceeded −18 dB, so the CW threshold removed no paths.
Questions or experiences to share? Contact RF.Guru or join our feedback group.