RF Circuit PCB Design: Printed Elements, Materials, and Stackup

Radio frequency design returns to PCB engineering whenever a product needs a wireless link, a radar front end, or a transmission line with a defined impedance. Engineers trained on digital systems usually recognise the components but not the reasoning: an RF board is built from copper shapes that behave as circuit elements, and the layout is part of the electrical design rather than a drawing of it. This guide covers the concepts that a RF circuit PCB design depends on, from the printed elements themselves to material and stackup selection.

Much of the difficulty comes from the fact that RF schematics do not look like conventional circuits. Elements are distributed rather than lumped, impedance matters more than voltage, and a trace that would be a harmless connection at DC becomes a reactive component at gigahertz frequencies.

Printed Elements Instead of Discrete Parts

An RF board does not always use an off-the-shelf inductor or capacitor. Instead, copper traces are shaped to form the required element: a length of line that presents a specific impedance, a stub that transforms impedance, a spiral that behaves as a printed inductor, or a gap that couples energy between two conductors like a capacitor.

These structures work because of wave propagation. At the frequencies involved, the dimensions of the copper are comparable to the wavelength of the signal, so voltage and current vary along the trace instead of being uniform. A designer who treats such a trace as a simple connection will find that the circuit behaves in ways the schematic does not predict.

Three properties define how printed elements behave. They are passive unless an active device is added. They are linear, so current and voltage are related by a linear function until a semiconductor introduces nonlinearity. And they operate through propagation, which means that the input impedance of each section, not its DC resistance, determines how it interfaces with the rest of the circuit.

Designing the Stackup Before the Circuit

The stackup has to be settled before the RF circuit is drawn, because the substrate determines the electrical properties of every printed element. Dielectric thickness sets the trace width required for a given impedance, and the dielectric constant determines the propagation velocity and therefore the physical length of every resonant element.

The frequency of operation drives the rest. A design at a few gigahertz can use a simple two or three layer construction with standard laminate thicknesses, while a design at millimetre wave frequencies requires thin dielectric layers, tight dimensional control, and a material whose properties change little with temperature. Deciding the layer count, the dielectric thickness, and the reference plane arrangement at the start prevents the situation where a working circuit cannot be reproduced on the board.

<img src="https://www.gopcba.com/wp-content/uploads/2024/09/pcb-design1.jpg" alt="Microstrip and stripline structures on a high frequency PCB” />

The stackup also determines whether the transmission lines are microstrip or stripline. A microstrip on the outer layer is easier to tune and to probe, while a stripline buried between planes is shielded and radiates less. Most RF boards use microstrip for the elements that must be adjusted and stripline or coplanar structures for the distribution network.

Material Selection and the Limits of FR-4

Standard FR-4 is usable for RF transmission lines and interconnections up to roughly the Wi-Fi bands, around 6 GHz, where its losses remain acceptable for short runs. Beyond that, the dielectric loss and the variation of the dielectric constant with frequency and temperature become difficult to compensate.

A PTFE laminate addresses both problems. These materials use a fluoropolymer base, often with a ceramic filler, which lowers the loss tangent substantially compared with FR-4 and makes the dielectric constant stable over frequency. The practical result is that a signal propagates further before it must be amplified, and the physical dimensions of the circuit stay closer to the calculated values across the operating range.

One caution applies to all woven laminates. Glass-reinforced materials have a periodic variation in dielectric constant caused by the weave, and a trace that runs parallel to the weave sees a fluctuating impedance along its length. At high frequency the effect appears as a change in phase and impedance rather than as a simple loss, and it can be reduced by choosing a spread or flattened weave, or by routing at an angle to the weave direction.

Hybrid Stackups

Building an entire multilayer board from high-frequency material is possible but rarely necessary, and it is expensive. A more economical approach is a hybrid stackup: the RF laminates occupy the top layers where the transmission lines and printed elements sit, while the inner layers use conventional material for ground planes, power distribution, and digital routing.

The construction has to respect the process. Materials with different dielectric constants and different thermal expansion behaviour must be laminated together in a way that survives the press and later thermal cycling, and the adhesive layers between them contribute their own electrical properties. Panel sizes and drilling parameters differ between materials, so the fabricator should be consulted before the stackup is fixed rather than after.

Calculating Trace Impedance

Once the stackup is known, the trace width that produces the target impedance, usually 50 ohms for RF work, must be calculated. Closed-form expressions derived from conformal mapping relate geometry to impedance, but they cannot be inverted analytically, so a numerical solution or a field solver is used to find the width.

For anything more complex than a simple surface microstrip, a field solver is the better tool. It can account for copper roughness, the taper produced by etching, differential routing, and the position of the trace between reference planes. The calculated characteristic impedance is then used to design the matching networks that connect the printed elements to the active devices.

Impedance matching is not optional in an RF design. A mismatch causes reflection, which reduces the power delivered to the load and can degrade the performance of a sensitive receiver. Matching networks are designed using the input impedance of each stage, which comes from simulation or from the device datasheet.

Common Printed Structures

The catalogue of printed RF structures is well established. Filters of both passive and active types, attenuators, couplers and power dividers, resonators, waveguide cavities, and antennas are all implemented with copper geometry rather than with purchased components. Each has a known set of dimensions for a given substrate and frequency, which is why the stackup must be defined first.

Antennas deserve particular attention because they are intended to radiate, which means the layout around them must not obstruct or detune the structure. Keep-out areas, ground plane boundaries, and the position of the antenna relative to the enclosure all affect performance, and they interact with the mechanical design.

Verification and Simulation

Simulation belongs before the layout, not after it. The schematic should be simulated with the printed elements modelled as transmission lines, so that the electromagnetic behaviour of the board is included in the circuit response. Electromagnetic field simulation then verifies the passive structures, and a full board simulation confirms that the assembled layout behaves as the circuit model predicted.

Where the design includes digital circuitry alongside the RF section, the two domains should be simulated with their interaction in mind, because the noise environment of a mixed board is what most often degrades a receiver that works perfectly in isolation. The routing practices that keep such a board quiet are covered in PCB routing with microstrip and stripline and in high-frequency trace and data bus routing.

FAQ

Can standard FR-4 be used for a 77 GHz radar design? Not for the RF section. At that frequency the dielectric loss and the variation in dielectric constant are too large for the circuit to work reliably. A low-loss material with a stable dielectric constant is required, and the copper geometry must be controlled tightly enough to hold the impedance.

Why must the stackup be decided before the RF circuit is designed? Because the substrate sets the impedance of every trace and the physical length of every resonant structure. Changing the dielectric thickness after the circuit is designed changes the width required for the same impedance and shifts the frequency of every distributed element.

Is a hybrid stackup with RF material on the surface reliable? It is common practice and works well when the construction is planned with the fabricator. The critical points are the bond between dissimilar materials, the thermal expansion mismatch, and the drilling parameters, all of which have to be qualified rather than assumed. Layer stackup options are described in layer stackup from one to eight layers.