Maxwell’s Equations — The RF Foundations You Forgot
Maxwell’s Equations — The RF Foundations You Forgot
Antennas, feedlines, transformers and unwanted RF paths are not separate tricks. They are different solutions of the same field laws under different sources, materials and boundaries.
If you have ever asked why an antenna radiates, how power moves along coax, or why RF appears on a conductor that was not invited into the circuit, Maxwell’s equations are the deeper answer. The equations do not replace impedance, SWR, transmission-line or transformer models. They tell us when those shortcuts are legitimate—and when geometry has escaped the schematic.
Joeri’s short version: no single Maxwell equation explains an antenna. Charge and current continuity, both curl laws, both divergence laws, material response and boundary conditions produce one coupled solution. Radiation is not a field “detaching” at a point; it is the outward-power part of that complete solution.
Choose the Form and Convention First
The cleanest practical starting point is the macroscopic SI form. It separates free charge and free current from material polarization and magnetization by using D and H. The table assumes stationary boundaries, a fixed contour C, an oriented spanning surface S, and the right-hand relationship between line and surface directions.
| Law | Differential form | Integral form |
|---|---|---|
| Gauss, electric | ∇ · D = ρf | ∯S D · da = Qf,encl |
| Gauss, magnetic | ∇ · B = 0 | ∯S B · da = 0 |
| Faraday | ∇ × E = −∂B/∂t | ∮C E · dl = −d/dt ∫S B · da |
| Ampère–Maxwell | ∇ × H = Jf + ∂D/∂t | ∮C H · dl = If,encl + d/dt ∫S D · da |
| Symbol | Quantity | SI unit |
|---|---|---|
| E | Electric-field strength | volt per metre, V m−1 |
| D | Electric displacement / electric flux density | coulomb per square metre, C m−2 |
| H | Magnetic-field strength | ampere per metre, A m−1 |
| B | Magnetic flux density | tesla, T = Wb m−2 |
| ρf | Free volume-charge density | coulomb per cubic metre, C m−3 |
| Jf | Free current density | ampere per square metre, A m−2 |
| t | Time | second, s |
The familiar vacuum equations follow from D = ε0E and B = μ0H. That gives ∇ · E = ρ/ε0 and ∇ × B = μ0J + μ0ε0∂E/∂t. Those are vacuum special cases, not the best form for reasoning directly inside an arbitrary dielectric or ferrite.
Materials Need Constitutive Relations
Maxwell’s equations alone do not say how a particular material responds. For an ideal linear, homogeneous, isotropic, local and time-invariant medium, the familiar constitutive relations are:
D = εE
B = μH
Jc = σE
Permittivity ε has unit F m−1, permeability μ has unit H m−1, and conductivity σ has unit S m−1. Real cable dielectrics and ferrites can be lossy and dispersive; ε and μ may be complex and frequency-dependent. Anisotropic media require tensors. Ferrites can also be nonlinear, temperature-dependent and hysteretic. A single “permeability” number cannot describe every RF transformer operating point.
In a homogeneous lossless medium the wave speed and intrinsic impedance are:
v = 1/√(με)
η = √(μ/ε)
In vacuum, v = c = 1/√(μ0ε0). Since the 2019 SI revision, c remains exact while μ0 and ε0 are experimentally determined quantities linked through that exact relation. In a conductor or dispersive lossy medium, propagation is described by a complex propagation constant; the two simple real-valued formulas are not sufficient.
What Each Law Contributes
Electric Charge and Flux
Gauss’s electric law relates free charge to the divergence and closed-surface flux of D. At RF, charge separation appears between dipole arms, capacitor plates, coax conductors and an antenna and its environment. Voltage is therefore not a freestanding circuit label. It represents a line integral of electric field along a declared path between declared points. In a time-varying magnetic field that integral can be path-dependent, so port geometry and reference conductors matter.
At a material boundary, normal D changes according to free surface charge. Tangential E is continuous across a stationary ordinary interface unless an impressed magnetic surface current is introduced as a mathematical source. These boundary conditions help decide where electric stress and capacitance appear.
Magnetic Flux Has No Measured Monopole Source
Gauss’s magnetic law says the net B-flux through every closed surface is zero. Magnetic flux does not begin or end at an isolated magnetic charge in classical electromagnetism. That does not mean all transformer flux stays in a core. Leakage flux closes through surrounding space, and air gaps, winding placement and nearby conductors can materially change the field.
Normal B is continuous across an interface. Tangential H changes by the free surface-current density. A conducting surface and a ferrite boundary therefore shape the magnetic field differently even though both obey the same four laws.
Faraday Gives the Sign of Induction
Faraday’s minus sign is not decoration. With the stated contour orientation, increasing magnetic flux produces a circulating electric field whose induced electromotive force opposes that change. The fixed-contour integral form is the basis of transformer voltage. A physically moving conductor adds the motional v × B contribution and must not be squeezed into the stationary-loop formula without saying so.
Faraday’s law is necessary for a transformer explanation, but not sufficient. Ampère–Maxwell relates current and electric flux to H; Gauss’s laws and the boundary conditions account for charge, capacitance and leakage; the material relations set flux, loss and nonlinearity.
Ampère–Maxwell Keeps Current and Charge Consistent
The term ∂D/∂t has the same unit as current density. It is called displacement-current density, but it is not ordinary conduction of electrons through the dielectric of an ideal capacitor. Writing D = ε0E + P shows a vacuum-field contribution and a material-polarization contribution. Neither licenses us to say that copper conduction current crosses an ideal insulating gap.
Taking the divergence of Ampère–Maxwell and using Gauss’s electric law gives charge continuity:
∇ · Jf + ∂ρf/∂t = 0
∯S Jf · da = −dQf,encl/dt
That is the disciplined version of “current needs a return path.” Free charge cannot accumulate without the current divergence recording the change. The complete RF closure can include conduction on several surfaces plus electric-field coupling and displacement current through space or dielectric. It need not be one obvious wire, and protective earth should never be assumed to be the intended RF return merely because it is present.
The Capacitor Sign Check
Adopt peak phasors with an ejωt convention. Then ∂/∂t becomes jω, Faraday becomes ∇ × E = −jωB, and Ampère–Maxwell becomes ∇ × H = Jf + jωD.
For an ideal capacitor:
I = jωCV
ZC = 1/(jωC) = −j/(ωC)
XC = −1/(2πfC) Ω
The magnitude |XC| is 1/(2πfC), but capacitive reactance itself is negative under this convention. Circuit current reaches and leaves the plates through conductors; the changing D-field accounts for the Ampère–Maxwell circulation through a surface drawn between the plates.
Poynting’s Theorem Tracks the Power
The instantaneous Poynting vector is:
S = E × H [W m−2]
For a simple linear nondispersive medium, electromagnetic energy conservation can be written:
∂u/∂t + ∇ · S = −Jf · E
u = ½(E · D + B · H) [J m−3]
The right side is the rate at which the field does work on free current per unit volume. With an outward surface normal, integrating over a volume says that outward electromagnetic power plus increasing stored field energy plus work on matter balances the sources. For peak phasors, the time-average power-flux density is ½ Re{E × H*}.
In dispersive, lossy or nonlinear material, the compact energy-density expression needs additional care. Core hysteresis, dielectric loss and conduction cannot be recovered honestly by quoting ½(E · D + B · H) without the material model.
The Poynting vector is not always a simple “ray arrow.” In the far field of a locally plane travelling wave, its time average points with net power propagation. In reactive near fields, standing waves and interference regions, instantaneous or local power flow can reverse, circulate or have a zero average even while electric and magnetic energy is stored.
Feedlines Are Guided Field Solutions
A transmission line is not explained by current alone. Its conductors and dielectric support a guided electromagnetic mode with linked voltage and current waves. The characteristic impedance is their modal ratio for a travelling wave; terminations and discontinuities set forward and reflected components.
In an ideal coaxial TEM mode, the electric and magnetic fields occupy the dielectric between centre conductor and shield. The forward electromagnetic power crosses a cable cross-section through S; it is not transported solely “inside the copper.” Surface current and charge on the conductors impose the boundary conditions that guide that field. Finite conductivity and dielectric loss then remove part of the power as heat.
The intended centre-conductor and inner-shield currents are equal and opposite for the coax mode. Current on the shield exterior belongs to another mode whose return is in the wider installation. A match to 50 Ω does not prove that exterior current is absent. A choke changes the impedance of that exterior-current path; it does not block every RF field or substitute for identifying the complete parallel paths.
Transformers Need All Four Laws Too
Winding current establishes H; the core and air geometry produce B; changing flux gives the line-integrated electric field associated with induced voltage. Charges on windings also produce electric fields and interwinding capacitance. Leakage inductance, winding resistance, dielectric loss, self-resonance and common-mode coupling are therefore part of the same device, not afterthoughts added to an ideal transformer.
In a transmission-line transformer or current choke, equal-and-opposite wanted-mode currents may largely cancel core excitation when the geometry is symmetric. An exterior or common-mode current can create net core flux and see a large complex impedance. “Wanted current cancels, common mode does not” is a useful first model—not a guarantee across frequency, power, winding capacitance, imbalance and core nonlinearity.
Antennas Radiate From a Distribution, Not a Detachment Point
A time-varying charge and current distribution generates retarded electromagnetic fields subject to the surrounding boundaries. The far field is a coherent vector sum over that entire source distribution. Geometry, phase and current direction decide which contributions reinforce and which cancel.
A balanced line can carry large currents yet have small external radiation because nearby opposite-current contributions largely cancel. Opening and shaping conductors into an antenna changes the distribution and the cancellation. There is no unique point where an electromagnetic field tears loose from a wire. For a bounded antenna in a homogeneous lossless exterior, radiation is identified by the field terms that persist as outward power at large distance:
Prad = limr→∞ ∯Sr ⟨S⟩ · da
Near the antenna, reactive and radiating fields coexist and their phase relationships can be complicated. In the far field of a bounded source in a uniform lossless medium, the radiating fields are transverse, fall approximately as 1/r, have an E/H ratio set by the medium, and carry finite power through an ever-larger sphere. The near-field, radiating-near-field and far-field boundaries depend on wavelength and the maximum source dimension; “a few metres away” is not a universal category.
Where Circuit Models Stop
Lumped circuits work when the structure is electrically small enough and its distributed fields can be represented by defined elements. Transmission-line models work when a known guided mode and reference planes describe the structure. Antenna and full-wave models are needed when distributed phase, radiation, coupling or several modes materially affect the result.
| RF shortcut | Useful engineering version |
|---|---|
| “RF takes the shortest path.” | The installed fields and impedances distribute current among every available conductive and displacement-current path. |
| “Current reaches the end and radiates.” | Ends and discontinuities alter the complete time-varying charge/current distribution; radiation comes from the resulting far-field sum. |
| “A capacitor passes RF.” | Conduction current charges the plates while the changing electric displacement provides field continuity across the dielectric. |
| “Power travels in the wire.” | Conductor charge and current guide fields; the Poynting flux describes electromagnetic power flow through the surrounding cross-section. |
| “A choke blocks RF.” | A real choke adds frequency- and power-dependent complex impedance to a particular mode at a particular position. |
| “A transformer is only mutual inductance.” | Magnetic coupling, electric coupling, leakage, loss, material dispersion, boundaries and heat all belong to the complete device. |
A Maxwell-Led Diagnostic Habit
- Draw the physical system. Include conductor surfaces, dielectric, ferrite, shield exterior, chassis, mast, nearby wiring and ground—not only schematic nodes.
- Name the mode and reference plane. Separate differential, common/exterior and radiated fields; state where voltage, current, SWR or impedance is defined.
- Declare the material model. State ε, μ and σ with frequency, temperature, linearity and loss boundaries where they matter.
- Enforce continuity and boundaries. Ask where charge accumulates, how current closes, which tangential and normal fields can change, and whether a parallel path remains.
- Close the power budget. Separate incident, reflected, accepted, stored, dissipated and radiated power using consistent time and phasor conventions.
- Measure more than one observable. Combine impedance with current, field, temperature, loss or pattern data. One low-SWR trace cannot validate the electromagnetic model.
The Historical Thread
James Clerk Maxwell’s 1865 paper A Dynamical Theory of the Electromagnetic Field connected electricity, magnetism and light in a field theory. The compact four-equation vector presentation used here is not a facsimile of that paper. Oliver Heaviside was a major contributor to the later vector-calculus form, alongside other late-nineteenth-century developments. The existing IEEE historical account of the long road to Maxwell’s equations is useful precisely because the modern notation has its own engineering history.
John Henry Poynting then made the energy-flow statement operational: electromagnetic energy can be tracked through the field, not merely imagined as cargo moving inside a conductor. That is the bridge from nineteenth-century field theory to a modern VNA trace, a feedline cross-section, a transformer loss budget and an antenna measurement.
Primary and Authoritative References
- Maxwell — A Dynamical Theory of the Electromagnetic Field, Smithsonian Libraries digitization of the Royal Society paper
- Poynting — On the Transfer of Energy in the Electromagnetic Field
- BIPM — The International System of Units, current SI Brochure
- NIST/CODATA — current recommended electromagnetic constants and units
- NASA — Fundamentals of Electromagnetics, Maxwell’s Equations and Wave Propagation
- NBS Technical Note 1157 — retarded fields, sources and boundary values derived from Maxwell’s equations
- Lawrence Livermore — NEC-5 validation, electric-field integral equations and physical-model boundaries
- IEEE 145-2025 — IEEE Standard for Definitions of Terms for Antennas
- IEEE 149-2021 — Recommended Practice for Antenna Measurements
Joeri’s Bottom Line
Maxwell’s equations are not four slogans to paste beside four radio components. They are one coupled framework. Geometry, material, sources, continuity and boundaries select the fields and currents that actually exist; Poynting’s theorem tells us where their energy goes.
When the station behaves strangely, ask the question that survives every antenna myth:
Mini-FAQ
- Do I need to solve Maxwell’s equations every time I design an antenna? No. Circuit, transmission-line and antenna tools provide useful solutions under declared assumptions. Maxwell’s equations show which geometry, material and boundary effects those tools omit.
- Are the four modern vector equations exactly what Maxwell wrote? No. Maxwell developed the field theory in a larger nineteenth-century formulation. Heaviside and others helped create the compact vector-calculus form engineers now use.
- Is displacement current ordinary current through an insulator? No. It is the time derivative of electric displacement, with units of current density. It includes vacuum-field and polarization contributions, not copper-like conduction across an ideal dielectric.
- Where does electromagnetic power flow in coax? In the intended mode, conductor charge and current guide electric and magnetic fields in the dielectric. Their Poynting flux carries power along the cable; finite conductor and dielectric loss convert some to heat.
- At what point does a field detach from an antenna? There is no single detachment point. The complete time-varying charge and current distribution produces near and far fields; radiation is the component that carries net outward power at large distance.
- Why can a good SWR coexist with poor radiation? SWR describes modal reflection at a stated reference plane. It does not separately measure conductor, dielectric, ground or matching loss, common-mode current, radiation efficiency or pattern.