NVIS Antenna Height: Design the Pattern for the Path
NVIS Antenna Height: Design the Pattern for the Path
Near Vertical Incidence Skywave can carry HF signals over the terrain between stations, but no mast height guarantees that link. Height shapes the installed elevation pattern; the ionosphere decides whether the frequency returns at high incidence, and the link budget decides whether the returned signal is usable.
The usual question is: “How high should I put the centre of my 80, 60 or 40 m dipole?” My answer starts with wavelength, but it does not end there. I want useful radiation across the elevation angles that reach the coverage area, and I want that result after ground loss, matching loss, noise, absorption and fading are included.
My practical rule: use a low horizontal antenna or inverted-V as a starting geometry, sweep several heights in a model, then verify the installed system over real regional paths. A narrow 0.18–0.22 λ range can be a useful test region; it is not a universal optimum, safety limit or promise of reliable NVIS.
What NVIS Actually Asks the Antenna to Do
NVIS uses ionospheric skywave at steep elevation angles to cover short and medium regional paths that would otherwise sit inside the skip zone of a lower-angle HF mode. The required angle is not always exactly 90°. It follows path length, the effective ionospheric reflection or refraction height, layer structure and the mode that is supported at that time.
A rough single-hop geometric model can relate path distance d, virtual reflection height h and elevation angle α:
Simplified flat-Earth geometry: α ≈ arctan(2h / d)
This is useful for intuition, not prediction. Earth curvature, the ionospheric electron-density profile, magneto-ionic modes and real ray bending require a propagation model such as ITU-R P.533. For a very short path, the useful departure angles may span a broad region near overhead rather than one narrow angle.
Terrain can block ground wave and VHF/UHF line of sight, while a supported skywave path can bridge it. NVIS is therefore valuable for regional nets and field communication. It is not immune to ionospheric absorption, disturbances, deep fading or a frequency that penetrates the ionosphere instead of returning.
Height Must Be Expressed in Wavelengths
A ten-metre mast is about 0.12 λ at 3.6 MHz, 0.18 λ at 5.35 MHz and 0.24 λ at 7.1 MHz. The same support therefore produces three different ground-reflection relationships and elevation patterns.
Free-space wavelength: λ = c / f
For quick HF estimates with f in MHz: λ ≈ 299.8 / f metres.
The table is a scale for modelling candidate heights, not a recommendation. Use the actual operating frequency rather than a band name.
| Reference frequency | Approximate wavelength | 0.10 λ | 0.20 λ | 0.25 λ |
|---|---|---|---|---|
| 1.9 MHz | 157.8 m | 15.8 m | 31.6 m | 39.5 m |
| 3.6 MHz | 83.3 m | 8.3 m | 16.7 m | 20.8 m |
| 5.35 MHz | 56.0 m | 5.6 m | 11.2 m | 14.0 m |
| 7.1 MHz | 42.2 m | 4.2 m | 8.4 m | 10.6 m |
| 10.1 MHz | 29.7 m | 3.0 m | 5.9 m | 7.4 m |
| 14.1 MHz | 21.3 m | 2.1 m | 4.3 m | 5.3 m |
A model sweep might include 0.05, 0.10, 0.15, 0.20, 0.25 and 0.30 λ. That reveals how broad the high-angle response is and whether a small height change materially affects the required sector. Selecting only 0.18 and 0.22 λ can hide a flatter, safer or easier solution that performs nearly as well.
Mast Height Is Not the Whole Antenna Height
For a straight horizontal dipole, centre and end heights may be similar. For an inverted-V, only the apex reaches the mast height. Much of the current-carrying conductor is lower, and the ends may approach people, vegetation and lossy ground.
Model the complete geometry: apex, end heights, included angle, sag, bends, conductor diameter, feedline route, support conductors and nearby structures. The current is not uniform along a dipole, so a simple arithmetic average height is also incomplete. The installed pattern comes from the current distribution over the entire structure.
Lowering the ends of an inverted-V can change feedpoint impedance, horizontal and vertical field components, azimuth pattern, ground loss and accessible RF voltage. An apex at the “right” fraction of a wavelength does not preserve the result if the legs are steeply folded around a site.
Ground Changes Pattern, Loss and Impedance Together
A horizontal antenna close to ground interacts strongly with its image current. Height changes the phase relationship between direct and ground-reflected fields, while ground conductivity and permittivity affect reflection and dissipation. Real ground can be layered, wet, rocky, sloped or crossed by buried and overhead conductors.
Very low placement can still provide useful high-angle radiation, especially when the installation constraint is severe. It may also increase loss, detune the antenna, reduce radiation resistance or place high electric fields near lossy material. The boundary is not a universal 0.15 λ cutoff.
Likewise, raising the antenna above 0.25 or 0.30 λ does not suddenly turn NVIS off. The pattern develops and moves lobes and nulls as height increases. Whether that hurts depends on the elevation-angle sector required by the path. Examine gain or realised gain across the whole useful sector, not only the single zenith value.
Useful comparison: plot realised gain versus elevation angle for each candidate height over the actual ground model. Integrate the decision across the required angular sector, then include feeder/matching loss and common-mode current. A pretty zenith lobe alone is not the link budget.
The Ionosphere Sets the Frequency Window
Near vertical incidence demands a frequency low enough for the ionosphere to support the steep path. At vertical incidence, that boundary is closely related to the relevant layer critical frequency. At less steep incidence, the usable upper frequency can be higher. The electron density changes with location, local time, season, solar conditions and disturbances.
The lower end of the usable window is not set by reflection alone. D-region absorption, atmospheric and man-made noise, transmitter power, antenna efficiency, receiver bandwidth and required signal-to-noise ratio establish a practical lower usable frequency. A lower band may return reliably but arrive below the required SNR.
This is why 80, 60 and 40 m are common NVIS choices without being guaranteed choices. Near-vertical propagation on 30 or 20 m is possible only when the ionosphere supports those frequencies; at other times they pass through. Conversely, 160 m may be limited by absorption, noise and the difficulty of producing useful high-angle field from a practical antenna.
Use current ionosonde observations and a validated prediction model, but keep frequency agility. ITU-R P.1239 provides reference ionospheric characteristics; P.533 predicts HF circuit performance; P.842 turns signal and noise statistics into reliability estimates. A single “best band” cannot serve every hour and season.
Transmit Pattern and Receive SNR Are Different Decisions
A linear passive antenna is reciprocal: its transmit and receive directional properties are linked. An antenna does not acquire a fundamentally different optimum height simply because it is receiving.
What can differ is the objective. On transmit, the question may be field strength across a regional area. On receive, the question is wanted-signal SNR in a noise field that arrives from several directions and through conducted or common-mode paths. A lower antenna may sound quieter because it attenuates both signal and noise, or because its pattern rejects a local source. That is not automatically better SNR.
ITU-R P.372 treats atmospheric, galactic and man-made radio noise statistically and shows why frequency, time, location and antenna direction matter. Compare wanted signal and noise in the same bandwidth with fixed receiver settings; do not optimise from the S-meter noise level alone.
Polarisation Does Not Stay Simple Through the Ionosphere
A low horizontal dipole commonly provides the strong high-angle pattern wanted for NVIS. Calling NVIS simply “horizontal polarisation” goes too far. The ionosphere is magnetised and supports characteristic propagation modes; Faraday rotation, mode coupling and multipath can change the polarisation arriving at the receiver.
An inverted-V adds more vertical field as the legs become steeper. A vertical antenna is not forbidden from radiating high-angle energy, but its installed high-angle realised gain may be poor when the return system and ground favour lower angles or dissipate power. Compare the complete three-dimensional pattern and efficiency rather than the antenna label.
Polarisation diversity can reduce some fading on a particular path. It does not replace frequency selection, path prediction or sufficient link margin.
Control the Feedline Without Assuming One Choke Position
Coax exterior current can add an unintended radiator, disturb the pattern and bring station noise into the receive system. The intended differential transmission inside the coax and the unwanted exterior current are different modes.
A suitable choke can raise impedance in the exterior-current path, but the feedpoint is not a universal placement answer. If a defined section of coax exterior forms part of the intended return, the current boundary belongs elsewhere. If the antenna is intended to be balanced, a feedpoint boundary may be appropriate—but it still requires measurement.
Measure current around the complete coax at several marked positions and on every operating band. Check the matching network’s separate differential transfer, loss and voltage/current stress. A low SWR does not prove that the feedline stayed out of the NVIS pattern.
Build the Regional Link Budget
The useful transmit quantity is not accepted transmitter power by itself. It is the power radiated into the elevation and azimuth sector supported by the path, after feedline, matching, conductor, ground and common-mode losses.
A practical budget includes:
- transmitter power and duty cycle;
- feedline and matching-network loss;
- installed realised gain across the required high-angle sector;
- ionospheric mode availability, absorption and path loss;
- receive-antenna pattern, feeder loss and receiver noise;
- atmospheric, man-made and interference noise in the operating bandwidth;
- required SNR for the mode and information rate; and
- margin for short-term fading and day-to-day variability.
“Reliable” needs a number: required availability over named endpoints, hours, seasons and operating modes. An antenna height that performs well on a Sunday net is not automatically a year-round emergency-communications design.
A Practical Height-Selection Sequence
- Define the coverage task. Map the stations, regional area, required times, seasons, mode, bandwidth, SNR and availability.
- Estimate the useful elevation sector. Use path prediction for the endpoint distances and ionospheric modes rather than targeting only 90°.
- Choose candidate frequencies. Compare current ionosonde observations and ITU-R-based predictions with absorption, noise and frequency-allocation constraints.
- Model the installed antenna. Sweep height in wavelengths using the real dipole or inverted-V geometry, ground model, feedline and nearby conductors.
- Include losses and stress. Calculate or measure feeder/matching loss and verify voltage, current and thermal limits at the intended power and duty cycle.
- Map common-mode current. Measure along the coax and connected conductors before and after any choke change.
- Select safe support options. Check mast capacity, guy geometry, anchor loads, wire tension, wind, ice, falling radius, access and local permissions.
- Run height trials. Compare at least three practical heights with unchanged power, receiver bandwidth and antenna geometry where possible.
- Use A/B/A evidence. Restore the first height or configuration to expose propagation drift; use several remote receivers and paths.
- Repeat across conditions. Test day/night, seasons and disturbed periods, then choose the height that gives the required margin most often.
If one mast must serve 80, 60 and 40 m, do not call one physical height optimal for all three. Find the safest candidate that gives acceptable angular coverage and loss on every required band, then keep more than one operating frequency available.
Mechanical, Electrical and RF Safety Set Hard Limits
A taller support is not an RF improvement if it exceeds the mast, guy, anchor, soil or attachment rating. Wind pressure, gust response, ice, wire tension, mast buckling, guy preload and foundation capacity belong in the design. Use manufacturer data and the structural rules applicable at the site; no generic antenna article can provide a universal guy spacing or wind rating.
Place the antenna so the mast, wire, ladder, tools and complete falling envelope cannot contact or approach overhead electrical lines. Electricity can flash over without physical contact. Identify lines before work, assume they are live unless their owner confirms otherwise, and obtain the safe clearance and work plan from the network operator.
Keep wire ends, feedpoints, tuners and current hot spots inaccessible during transmission. Assess RF exposure under the applicable national rules using frequency, power, duty cycle, installed pattern and simultaneous transmitters. ICNIRP’s 2020 RF guidelines provide a health-protection framework; local regulation determines the compliance procedure.
Plan lowering, disconnection, bonding, surge protection and lightning risk as one station system. A common-mode choke and a low mast are not lightning protection.
Primary Propagation, Antenna and Safety References
- Recommendation ITU-R P.533-14—current in-force method for predicting HF circuit modes, elevation angles, field strength, usable frequencies and SNR.
- Recommendation ITU-R P.1239-4—reference ionospheric characteristics, including foF2 and its spatial, diurnal, seasonal and solar dependence.
- Recommendation ITU-R P.842-5—HF circuit, reception and service reliability from signal, noise and variability.
- Recommendation ITU-R P.372-17—current atmospheric, galactic and man-made radio-noise data and variability.
- Lawrence Livermore National Laboratory, Numerical Electromagnetics Code—Method of Moments—wire-current, ground-interaction and installed-pattern modelling.
- ICNIRP Guidelines for Limiting Exposure to Electromagnetic Fields—RF exposure framework from 100 kHz to 300 GHz.
- UK Health and Safety Executive GS6, Avoiding Danger From Overhead Power Lines—planning, line-owner coordination, flashover and exclusion principles.
- European Commission JRC, Eurocode 1: Actions on Structures—wind, snow and other actions that structural design must consider.
The Height Answer
Start with what you can build safely. Convert that height to wavelengths on every candidate band, model the complete antenna over realistic ground, and inspect realised gain across the elevation sector needed by the regional paths.
Then ask whether the ionosphere supports those steep paths and whether the link retains enough SNR after absorption, noise and fading. If 0.2 λ is practical, it is a useful candidate—not a magic number. If 0.1 λ is the safe limit, measure what it delivers before assuming it cannot work. If raising the antenna develops lower-angle lobes, decide whether they actually reduce the required high-angle coverage.
For NVIS, height shapes one end of the circuit. Frequency choice, ionospheric state, pattern, loss, noise and safety complete the answer.
Mini-FAQ
- Is 0.18–0.22 λ the universal best NVIS height? No. It is a useful modelling and trial region. The installed geometry, ground, required elevation sector, losses, ionospheric support and safety constraints decide the useful height.
- Does an inverted-V use only its apex height? No. End height, included angle, sag, bends and current distribution all affect impedance, loss, polarisation and the three-dimensional pattern.
- Are 80, 60 and 40 m always the NVIS bands? They are common choices, but support changes with ionospheric critical frequency, time, season and disturbance. Absorption, noise and link margin set the lower usable boundary.
- Can 20 or 30 m support NVIS? Yes, when the near-vertical ionosphere supports the operating frequency. At other times the signal penetrates rather than returning, so current observations and prediction matter.
- Must the common-mode choke be at the feedpoint? Not universally. Place it at the intended exterior-current boundary, then verify current along the complete coax on every band.
- How should I verify a candidate mast height? Model realised gain over the required elevation sector, measure losses and current paths, then compare practical heights with equal power, fixed receiver settings, several regional paths and restored-baseline A/B/A tests.