Log-Periodic Antennas Over Metal Roofs: Broadband Is Not Immunity
Log-Periodic Antennas Over Metal Roofs: Broadband Is Not Immunity
An LPDA can preserve a useful match across a wide band, yet a large conductive roof can still reshape its pattern, alter gain, couple to the feedline and move the useful active region.
A log-periodic dipole array may be a practical choice over a metal building, but not because broadband antennas ignore nearby conductors. The defensible advantage is usable bandwidth; roof tolerance remains an installation-specific result.
Do not let RF performance authorise unsafe roof work. Mast loading, fall protection, weather sealing, lightning protection, electrical bonding, RF exposure and access by roof workers are separate design requirements.
What Broadband Operation Establishes
An LPDA's broadband impedance behaviour, directional radiation pattern and distributed active region do not establish immunity to a large metallic roof. They describe the antenna in a stated environment; the installed interaction with a particular roof remains a separate electromagnetic problem.
Broad impedance bandwidth can mean that roof coupling does not immediately create an unusable SWR. That is useful. It does not mean that realised gain, elevation angle, front-to-back ratio, polarisation or common-mode current stayed acceptable.
Key distinction: a broad match is not the same as a stable installed pattern. An LPDA can remain well matched while the roof redirects energy into different lobes.
What the LPDA Active Region Really Means
A conventional LPDA uses geometrically scaled elements and a transposed feed. At a given frequency, a group of elements near resonance carries most of the radiating current. As frequency changes, this active region moves along the array.
The U.S. National Telecommunications and Information Administration's LPDA theory treatment describes the array as an approximation to a frequency-independent antenna and explains this moving active region.
That mechanism supports wide bandwidth, but it does not make the antenna electromagnetically self-contained. The roof can couple most strongly to whichever elements are active at that frequency. As the active region moves, the roof interaction can also change. The result may be smooth—or it may contain frequency-dependent pattern changes and nulls.
A Finite Roof Is an Additional Radiating Structure
A conductive roof carries induced current and produces a reflected field. Its finite size, pitch, seams, gutters, edges, openings, solar frames and HVAC hardware determine how that current flows. It is not equivalent to the infinite perfectly conducting plane used in a textbook image calculation.
The direct field from the LPDA and the roof-reflected field combine with phase determined by frequency, antenna height, observation angle and polarisation. Raising the antenna changes that phase. It can reduce strong near coupling, but it can also move nulls or create additional elevation lobes.
| Height above roof | What it does—and does not imply |
|---|---|
| Electrically very small | Strong coupling is likely; impedance and current distribution can differ greatly from free space. |
| A moderate fraction of λ | Direct and reflected fields form a height-dependent elevation pattern; there is no universal optimum fraction. |
| Several wavelengths | Near coupling may be smaller, but the roof remains a reflector and the pattern can contain many lobes. |
Use wavelength at every operating frequency. A fixed one-metre mast is electrically modest at lower VHF and more than a wavelength at much of UHF.
Directionality Does Not Mean the Roof Is “Behind” the Antenna
An LPDA's boom direction establishes its main forward and rearward directions, usually discussed in azimuth. The roof is below the array, in its elevation plane. A forward-pointing antenna can still illuminate the roof strongly.
Aiming at the wanted transmitter can reject signals arriving from other azimuths, but it does not automatically reject a locally reflected copy of that same signal. Both direct and roof-reflected components can arrive through the antenna's useful angular response.
For receiving, what is often called “multipath from the roof” also includes the complete propagation environment: terrain, buildings and other reflectors. Do not infer the cause of a fading notch from the roof alone without controlled movement, modelling or field measurements.
“Balanced Elements” Do Not Finish the Feed System
The dipole elements and boom transmission line can support a balanced differential mode, but the installation includes coax, connector transition, mast and mounting hardware. If the transition permits net current on the coax exterior, that current adds another receiving or transmitting structure.
Common-mode current can change the apparent pattern, pick up building noise and make results depend on cable routing. A choke may help when its complex common-mode impedance is appropriate, but it does not repair incorrect differential impedance or an unsuitable feed transition.
The NTIA's 2008 antenna-coupling study explicitly modelled an LPDA over an infinite perfectly conducting ground plane and noted that balun effects were not included. That is a useful warning: a clean antenna model is not yet a complete rooftop installation.
LPDA, Yagi and Dipole: No Universal Tolerance Ranking
| Architecture | Real advantage | Roof caveat |
|---|---|---|
| LPDA | Predictable operation over a designed frequency ratio. | The active region and reflected-field phase change across the band. |
| Yagi-Uda | Can provide strong performance over its design band with a chosen boom length. | Roof coupling can alter element interaction, but a well-designed wideband Yagi is not automatically more fragile. |
| Single dipole | Simple, efficient reference structure with fewer variables. | Its height over a conductor strongly shapes impedance and elevation pattern; simplicity can also make it easier to model. |
“LPDA has lower gain than a similar Yagi” is not a stand-alone engineering rule. Gain depends on frequency range, boom length, element count, design parameters, loss and what “similar” means. Compare measured or modelled antennas over the required band, not category names.
Polarisation and Roof Orientation Matter
The image-current relationship differs for electric fields parallel and perpendicular to a conducting plane. With a finite pitched roof, simple horizontal-versus-vertical slogans are insufficient because edges and vertical metalwork contribute.
Model the actual element orientation, roof slope and boom heading. At receiving sites, also match the wanted service polarisation: cross-polarisation loss can dominate any small improvement obtained from roof spacing.
Model the Installed Structure, Not an Antenna Floating in Space
Lawrence Livermore National Laboratory's NEC-5 can combine wires, conducting surfaces, transmission lines, loads and real-ground models. Its validation manual explains both numerical error and physical-model error, including limitations of older wire-to-surface junction treatments.
A credible model should include:
- the complete LPDA geometry, element diameters and transposed boom feed;
- finite roof size, pitch and the important seams and edges;
- antenna height, polarisation, boom heading and mast;
- coax exterior and a realistic choke or transition model when common mode matters;
- large nearby conductors such as gutters, PV frames, railings and HVAC equipment;
- real earth beneath the structure; and
- mesh or segmentation convergence checks.
Sweep more than input SWR. Save realised gain, front-to-back ratio, selected elevation cuts, surface currents and feedline current across the full band.
Measure the Result in Layers
Calibrated S11 across the intended band.
Net coax current at several cable positions.
Repeatable field or link measurements in several directions.
- Record geometry. Photograph roof edges, mast, boom heading, cable route and nearby metal.
- Measure S11 at a stated calibration plane. A good match is necessary for some systems, but it is not the verdict.
- Clamp-probe the feedline. Measure several points because common-mode current can form a standing wave.
- Change one variable. Compare height, polarisation, cable route or choke condition one at a time.
- Collect directional evidence. Use a controlled source or several stable paths; AGC bars and one repeater report are weak evidence.
- Repeat across frequency. The active region, roof phase and cable electrical length all move.
Rooftop Safety Is Part of the Antenna System
For transmitting installations, assess accessible fields using actual power, duty cycle, antenna gain, pattern and occupied locations. The ICNIRP 2020 RF guidelines cover human exposure from 100 kHz to 300 GHz; applicable national rules remain controlling.
Lightning and surge protection require a complete design. IEC 62305-3:2024 addresses physical damage and life hazard, while IEC 62305-4:2024 addresses surge-protection measures for electrical and electronic systems. An RF choke or convenient mast bond is not a lightning-protection system.
Have roof penetrations, wind load, corrosion compatibility, access control, lightning conductors, cable entry and surge protection approved for the building and local requirements.
Practical Design Conclusion
An LPDA is often a sensible rooftop choice when wide frequency coverage is required. Its broad designed operating range may preserve a usable match despite moderate environmental perturbation.
That is the advantage worth claiming. “Less affected by metal” is not inherent. A large roof remains a frequency-dependent reflector and current-carrying structure, and the LPDA's moving active region can make the interaction change across the band. Model it, control feedline current and verify pattern as well as SWR.
Mini-FAQ
- Is an LPDA immune to a metal roof? No. Broadband input behaviour does not prevent pattern, gain or common-mode changes.
- Does the moving active region make it roof-tolerant? Not by itself. It explains wideband operation; the roof can couple differently as that region moves.
- Does pointing the boom outward avoid the roof? No. Azimuth directionality does not eliminate the downward part of the elevation pattern.
- Is more height always better? More clearance often reduces near coupling, but height also changes reflected-field phase and elevation lobes.
- Is SWR enough to choose the installation? No. Also check feedline current, realised gain, pattern, noise and safety.