What is the main difference between coaxial cable and waveguide?

Published on 7/31/2026 • Updated on 8/1/2026

Every RF signal that travels more than a few millimeters needs a transmission line, and for most of the last century the industry has had two serious answers: coaxial cable and waveguide. They solve the same problem (guide electromagnetic energy from point A to point B with minimum loss and leakage) using two fundamentally different physical mechanisms, and that difference in mechanism drives every practical trade-off between them.

The short version that every RF engineer carries around: coax is broadband, compact, flexible, and cheap; waveguide is low-loss, high-power, and physically large. Coax wins by default in most systems. Waveguide wins decisively in the specific places where its advantages are worth its bulk: high-power radar transmitters, satellite ground station feeds, and millimeter-wave systems where coax loss becomes unbearable.

This article explains why, at a level useful both to engineering students meeting the topic for the first time and to working engineers who want the practical numbers in one place. We'll cover the physics of how each line guides energy, the cutoff frequency concept that defines waveguide behavior, real loss and power-handling numbers, the WR sizing system, and a decision framework you can actually use.

How Coaxial Cable Guides Energy: The TEM Mode

A coaxial cable is two conductors: a center conductor and a surrounding outer conductor, separated by a dielectric. The RF energy travels in the electromagnetic field between the two conductors, not "inside the wire" as intuition suggests. The electric field points radially from center conductor to shield; the magnetic field circles the center conductor; and both are entirely transverse (perpendicular) to the direction of travel.

This field configuration is called the TEM mode (Transverse ElectroMagnetic), and it has a property that makes coax uniquely useful: it works at every frequency from DC upward. There is no minimum frequency. The same piece of coax that carries a DC bias voltage also carries a 6 GHz signal, simultaneously if you want. This is why bias-tees, DC-coupled instrumentation, and broadband systems are all built on coax.

Coax does have an upper frequency limit, and understanding it sets up the whole comparison. When the operating wavelength becomes comparable to the cable's circumference, the cable can start supporting additional field configurations (higher-order modes) beyond TEM. Multiple modes traveling at different velocities corrupt the signal. The practical consequence: higher frequencies demand physically smaller coax. That is why precision 1.85 mm connectors reach 67 GHz while big 7/16 DIN connectors stop around 7.5 GHz. And smaller coax means a smaller center conductor, which means more conductor loss. That spiral (higher frequency, smaller cable, more loss) is exactly the problem waveguide exists to escape.

How Waveguide Guides Energy: TE and TM Modes

A rectangular waveguide is, structurally, almost nothing: a hollow metal pipe, usually rectangular in cross-section, filled with air. No center conductor. No dielectric. The RF energy propagates as an electromagnetic wave bouncing down the inside of the pipe, confined by the conducting walls.

Without two conductors, the TEM mode cannot exist. Instead, waveguide supports TE modes (Transverse Electric: the electric field is transverse, but the magnetic field has a component along the direction of travel) and TM modes (Transverse Magnetic: the reverse). The workhorse is the TE₁₀ mode, the fundamental mode of rectangular waveguide, where the electric field spans the narrow dimension of the guide and varies as a half sine across the broad dimension.

Removing the center conductor is where all of waveguide's famous advantages come from:

• Dramatically lower loss. The dominant loss mechanism in coax (skin-effect resistance in the small center conductor) is simply gone. Current flows only in the large-area inner walls of the guide.
• No dielectric loss. The guide is air-filled. There is no polyethylene or PTFE absorbing RF energy.
• Enormous power handling. Breakdown is limited by the air gap across the full guide dimension, not the sub-millimeter gap in a coax dielectric.
• Perfect shielding. The energy is enclosed by a continuous metal pipe. Nothing leaks.

The price is equally fundamental: a hollow pipe only guides waves whose wavelength fits inside it. That constraint is the cutoff frequency, and it deserves its own section.

Cutoff Frequency: The Number That Defines Waveguide

A waveguide behaves as a high-pass structure. Below a specific frequency, the cutoff frequency, the wave physically cannot propagate down the guide; it decays exponentially within a short distance instead. Above cutoff, it propagates with very low loss. This is the single biggest conceptual difference from coax, which passes everything from DC up.

For the TE₁₀ mode in a rectangular waveguide, the cutoff depends only on the broad interior dimension of the guide:

Two practical consequences follow immediately. First, lower frequencies need bigger pipes: a waveguide for 1 GHz would need a broad dimension of about 15 cm, which is why nobody runs waveguide at VHF. Second, each waveguide size has a usable band, conventionally from about 1.25 × f_c (where loss and dispersion near cutoff have settled down) to about 1.9 × f_c (below the point where the next mode, TE₂₀, can start propagating and corrupt the signal). Roughly one octave of clean single-mode bandwidth per guide size.

Worked Example: WR-90 Cutoff and Operating Band

WR-90 is the standard X-band waveguide, and the "90" tells you its broad dimension: 0.90 inches (22.86 mm). Let's compute its behavior.

And indeed, the published operating band of WR-90 is 8.2 to 12.4 GHz, which is the X-band used by weather radar, marine radar, and many military systems. The formula is not an approximation of the datasheet; the datasheet is an application of the formula. Every WR size works the same way, which is why an RF engineer can estimate any waveguide's band from its number alone.

The WR Designation System

Rectangular waveguide sizes are standardized under the EIA "WR" (Waveguide, Rectangular) system, where the number is the broad interior dimension in hundredths of an inch. The sizes an engineer most often meets:

Notice the trend: as frequency rises, the guide shrinks. WR-284 is a substantial metal duct nearly 3 inches wide; WR-10 is a precision channel one-tenth of an inch across. At the low end, waveguide is impractically bulky. At the very high end (above roughly 60 GHz), waveguide becomes not just competitive with coax but often the only sensible choice, because coax loss at those frequencies is punishing.

Loss: The Comparison That Decides Most Systems

Here is the comparison in numbers, at 10 GHz, for representative transmission lines:

The gap is roughly an order of magnitude, and it comes straight from the physics: waveguide has no center conductor concentrating current into a thin, high-resistance surface, and no dielectric absorbing energy. Wall currents flow over the guide's large interior surface area at low current density.

The gap widens with frequency. At 77 GHz, even the best small coax runs several dB per meter, while WR-10 waveguide runs a fraction of that. This is why mmWave test benches, automotive radar test systems, and radio astronomy receivers are plumbed in waveguide even for short interconnects: at those frequencies, every centimeter of coax costs real signal.

On a receive path, this loss difference translates directly to system noise figure. A satellite ground station with a 10 m feed run cannot afford 8 dB of coax loss between the antenna and the LNA. Waveguide (or moving the LNA to the antenna) is not optional there; it is the design.

Power Handling

Two mechanisms limit how much power a transmission line carries: average power (heat: the line's losses turn signal into temperature, and materials have limits) and peak power (voltage breakdown: arc-over across the dielectric or air gap).

Waveguide dominates both. Its lower loss means less self-heating per watt carried, its all-metal construction sheds heat well, and its breakdown gap is the full guide dimension of air rather than a millimeter or two of solid dielectric. Representative numbers at X-band:

This is why every high-power radar transmitter, broadcast klystron, industrial microwave heating system, and particle accelerator RF chain in the world routes its power through waveguide. There is no coax alternative at megawatt peak powers. Pressurizing the guide with dry air or SF6 raises the breakdown threshold further, a standard trick in high-power radar.

Bandwidth and Dispersion

Here coax strikes back, on two counts.

Bandwidth. A single coax line covers DC to tens of GHz continuously. A single waveguide size covers roughly one octave. A broadband system spanning 2 to 18 GHz runs on one coax cable, or on four different waveguide sizes with transitions between them. For instrumentation, EW receivers, and any wideband application, that is decisive.

Dispersion. In coax's TEM mode, all frequencies travel at (nearly) the same velocity, so wideband signals hold their shape. In waveguide, propagation velocity depends on how far above cutoff you are: frequencies near cutoff travel slower than frequencies well above it. A wideband pulse sent down a long waveguide spreads and distorts. Radar system designers know this and account for it; it is manageable, but it is a real effect that TEM lines simply do not have.

Size, Weight, Flexibility, and Cost

The practical, unglamorous factors decide as many designs as the electrical ones.

• Routing. Coax bends around corners, coils for storage, and threads through cable trays. Rectangular waveguide changes direction only through machined E-bends, H-bends, and twists, each a purchased precision component. Flexible waveguide exists (corrugated wall construction) but is expensive, lossy compared to rigid guide, and used sparingly for final alignment rather than general routing.
• Weight. A WR-284 run is heavy metal ductwork requiring hangers and support structure. Coax hangs from cable ties.
• Installation skill. Coax terminates with connectors installed in minutes. Waveguide joins with flanges (choke or cover type) that must be flat, clean, correctly gasketed, and evenly torqued; a poor flange joint leaks RF and creates reflections. Pressurized systems add leak testing.
• Cost. Coax assemblies are inexpensive to moderate. Waveguide components (straight sections, bends, twists, transitions, flexible sections) are machined precision hardware, each priced accordingly. A waveguide run costs an order of magnitude more per meter than a coax run, before installation labor.

Head-to-Head Comparison Table

Who Uses What, and Why

Hybrid Systems: Where Coax and Waveguide Meet

Real systems rarely choose one or the other exclusively. A typical satellite ground station illustrates the standard pattern: the high-power amplifier output travels through waveguide to the antenna feed (power and loss dictate it), while the receive-side LNA output, the up/down converter connections, the reference signals, and the entire equipment room run on coax (bandwidth, flexibility, and cost dictate that).

The component that stitches the two domains together is the coax-to-waveguide transition (also called a waveguide launcher or adapter). The most common design inserts the coax center conductor as a small probe through the guide's broad wall, a quarter guide-wavelength from a shorted back wall; the probe couples the coax's TEM field into the guide's TE₁₀ field. A well-made transition achieves better than 20 dB return loss across the guide's band with a few tenths of a dB of insertion loss. Every waveguide test setup you will ever see has these on the instrument ends, because the VNA's ports are coax.

So the working engineer's relationship with this topic is rarely "choose one." It is "know where the boundary belongs in this system, and spec a good transition at the boundary."

The Decision Framework

Four questions, asked in order, settle nearly every case:

Frequently Asked Questions

Bottom Line

Coax and waveguide are not competitors so much as specialists. Coax's two-conductor TEM structure gives it DC-to-daylight bandwidth, flexibility, compactness, and low cost, which is why it carries the overwhelming majority of RF signals on Earth. Waveguide's hollow-pipe TE₁₀ structure gives it an order of magnitude lower loss and several orders of magnitude more power capacity, which is why it owns high-power radar, satellite ground station feeds, and the millimeter-wave bands where coax physics runs out of road.

The engineering skill is not memorizing which is "better." It is running the four questions (power, loss budget, bandwidth and DC, practical factors) against your actual system, placing the coax-to-waveguide boundary where the physics puts it, and specifying quality hardware on both sides of that boundary. Do that, and the transmission lines disappear into the system, which is exactly what a transmission line is supposed to do.