HF vs GHz Circular Polarisation: Myths, Limits and Benefits
HF vs GHz Circular Polarisation: Myths, Limits and Benefits
Most questions I receive concern HF, so a circular-polarisation question that crosses from 80 and 40 m into microwave is refreshing. The physics is continuous; the propagation paths and practical hardware are not.
Circular polarisation is not a frequency-specific trick. It is a defined field condition. Its benefit depends on the transmitted field, the propagation channel, the receive antenna and the metric being improved.
What Circular Polarisation Actually Requires
Ideal circular polarisation consists of two orthogonal field components with equal amplitude and a 90-degree phase difference. The electric-field vector rotates with constant magnitude. Real antennas produce elliptical polarisation; axial ratio describes how close the field is to circular over a stated direction and bandwidth.
Sense matters. An ideal circularly polarised receive antenna rejects the opposite sense in the same arrival direction. An ideal circular-to-linear pairing has a 3 dB polarisation mismatch because the linear antenna captures one component. Installation errors, multipath and imperfect axial ratio change those textbook values.
HF Ionospheric Propagation Is Magneto-Ionic
The ionosphere in Earth's magnetic field supports two characteristic waves, commonly called ordinary and extraordinary. For near-vertical HF propagation they can arrive with opposite circular or elliptical senses, different group delay and different fading. A linearly polarised wave launched upward does not guarantee a linearly polarised wave coming back down.
Faraday rotation and characteristic-wave splitting are path-dependent. Frequency, total electron content, geomagnetic-field direction, launch angle and time all matter. “The ionosphere randomises polarisation” is therefore too crude, just as “one circular sense always wins” is too crude.
What Witvliet and Co-Authors Measured
Ben A. Witvliet, Erik van Maanen, George Petersen, Albert Westenberg, Mark Bentum, Cornelis Slump and Roel Schiphorst built experiments that received the two characteristic waves separately. On a 7 MHz NVIS path of about 105–110 km in the Netherlands, they measured isolation exceeding 25 dB during the opening or closing interval in which only the extraordinary wave returned. A related diversity experiment reported an 8–11 dB reduction in fading margin compared with a single circularly polarised receive channel.
That is a genuine benefit for 40 m NVIS under the measured conditions: two complementary circular receive channels can supply less-correlated signal streams for selection, combining or MIMO processing. It is not evidence that a single fixed circular sense always adds the same number of decibels.
The same research programme and the literature reviewed in Witvliet's thesis report near-circular NVIS reception as low as 3.5 MHz. That makes 80 m an important practical band for circular or dual-polarisation investigation, especially where 80 m supports the near-vertical path. The strongest numerical isolation and fading-margin results cited here, however, came from the defined 7 MHz experiments. We should not silently transfer those figures to every 80 m path.
HF DX Is Not One Polarisation Case
Longer oblique paths can include several ionospheric modes, ground reflections, scatter and multiple hops. The received polarisation can change rapidly with time and frequency. Circular transmit or receive antennas may still be useful in a defined diversity experiment, but no universal advantage follows merely from writing “DX” beside “CP.”
Sporadic-E, auroral scatter and trans-equatorial paths also need their own channel evidence. The correct conclusion is not that polarisation never matters; it is that the path must be characterised before a benefit is claimed.
Why Microwave Practice Looks Different
At VHF, UHF and microwave frequencies, circular polarisation is common in satellite links, rotating spacecraft, radar and links where orientation or reflected-path behaviour makes it useful. It does not universally “fix multipath.” A reflection can change amplitude, phase and polarisation sense, and the final result depends on geometry and the receiving system.
Short wavelengths make dimensional and phase tolerances tight. Dual-fed patches, helices, crossed elements, branch-line or Lange couplers and sequentially rotated arrays can generate CP, but every implementation has finite bandwidth, insertion loss, amplitude imbalance and phase error. HF hardware can be physically larger and use transmission-line phasing, yet it is not automatically broadband or lossless.
Choose the Architecture from the Link
- Define the path. NVIS, satellite, terrestrial line of sight and scatter are different channels.
- Define the metric. Link margin, fade depth, outage probability, capacity and orientation tolerance are not interchangeable.
- Measure axial ratio and sense. State direction, frequency and bandwidth.
- Measure correlation for diversity. Two ports are useful only when their signal streams provide worthwhile independence after losses.
- Include the complete feed network. Couplers, cables, switches and combiners contribute loss and imbalance.
- Test over time. An ionospheric result based on one interval cannot establish long-term reliability.
Circular polarisation can be powerful on 80/40 m NVIS and at microwave—but for different reasons and under different constraints. The engineering win comes from matching the antenna and receiver architecture to the measured channel.
Primary and authoritative references
- Witvliet et al. — Measuring the isolation of the circularly polarized characteristic waves in NVIS propagation
- Witvliet et al. — Characteristic Wave Diversity in NVIS propagation
- Witvliet — Near Vertical Incidence Skywave: Interaction of Antenna and Propagation Mechanism
- ITU-R P.531-13 — ionospheric propagation effects and Faraday rotation
- IEEE 145-2025 — definitions of terms for antennas
Mini-FAQ
- Did Witvliet's team measure a 40 m NVIS benefit? Yes. Their defined 7 MHz experiments separated the characteristic waves and reported strong isolation and reduced fading margin with dual circular reception.
- Does that prove the same gain on every 80 m path? No. Near-circular behaviour has been observed at 3.5 MHz, but the quoted numerical results belong to specific 7 MHz paths and equipment.
- Is one circular sense always best? No. Sense, mode availability and fading change with path and time; diversity uses both channels.
- Does circular polarisation eliminate microwave multipath? No. It can help a designed link, but reflections and antenna errors still affect amplitude, phase and polarisation.