Radar System PCB: A Beginner Guide to Design

Radar is one of the few applications where the printed circuit board is part of the radio itself. At the frequencies involved, the wavelength is comparable to the dimensions of the board, the traces behave as transmission lines with measurable loss, and the antenna is frequently patterned directly onto the laminate. A radar system PCB is therefore designed with microwave rules rather than digital ones, and the choices that matter are the laminate, the geometry and the thermal path.

What the Board Has to Do

The signal chain begins at the antenna and ends at the digital output. On the transmit side a source is amplified, filtered and routed to the antenna, and the power amplifier is usually the single largest heat source on the board. On the receive side the signal arrives at levels that can be a fraction of a picowatt, passes through a low noise amplifier, and is then down-converted and digitised.

The same board frequently carries the digital processing that turns the received signal into range and angle information. That means the microwave section and a fast digital section share a laminate, and keeping the clock harmonics of the processor out of a receiver that is trying to detect a signal a hundred and twenty decibels below the transmit power is the central layout problem.

Laminate Selection for Microwave Work

A low loss laminate is the foundation. Standard FR-4 has a dissipation factor around 0.02 and its dielectric constant varies with the resin distribution and with absorbed moisture, which is tolerable at low frequency and fatal at twenty four gigahertz where a small phase error destroys the coherence between channels. Materials with a dissipation factor under 0.002 and a dielectric constant stable with temperature are used instead, and the dielectric constant determines the trace width required for fifty ohm.

The choice between them is a trade between loss and mechanical behaviour. PTFE based laminates give the lowest loss but are soft, difficult to process and expand a great deal with temperature. Ceramic-filled hydrocarbon laminates sit between the two, with loss low enough for most radar work and mechanical behaviour that lets them be laminated in the same press as FR-4, which makes mixed stack-ups practical. Our high frequency laminate article describes those materials in more detail.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/Automatic-Pick-Machine.webp" alt="radar system PCB with phased array antenna elements” />

Antennas and Phased Array Feeds

At these frequencies an antenna is a geometry on the board rather than a component, and a phased array is a repeating pattern of those geometries with a defined spacing. The element spacing is set by the wavelength, and it is usually close to half a wavelength to avoid grating lobes, which fixes the board dimensions before the layout begins.

The feed network is where most of the design effort goes. Every element must receive the same amplitude and the same phase, which means the paths from the distribution point to each element must be equal in electrical length, and the loss must be equal as well. A corporate feed, in which the paths split symmetrically, is the standard answer, and the board carries the transitions between the feed network, the transmit receive modules and the digital section. Grounded coplanar waveguide is common because it allows a wider trace for the same impedance and reduces conductor loss.

Transmit Receive Modules and Thermal Management

Each element or small group of elements has a transmit receive module containing a power amplifier, a low noise amplifier, a switch and a phase shifter. Those modules are the practical limit on how much power the array can radiate, because each amplifier dissipates a fraction of its output as heat and there may be dozens or hundreds of them on a single board.

Removing that heat without disturbing the electrical performance is the hard part. A metal core or a metal backed board carries heat out of the amplifiers and into a cold plate, and it also provides a solid reference for the microwave ground. The mechanical design and the electrical design are therefore the same design, and the thermal path has to be planned in the stack rather than added afterwards. Our thermal management article describes how the copper areas and interfaces are sized.

Layout Rules for the Microwave Section

Continuity is the rule that covers most of it. The reference plane under a transmission line must be unbroken for the whole length of the trace, because a plane split forces the return current to detour and produces a discontinuity that reflects energy and radiates. Vias that connect ground planes have to be placed close to any transition, and where a signal via is used its return path must be provided by ground vias around it.

Coupling between adjacent structures is the second concern. At these wavelengths, two traces that are close together form a directional coupler whether the designer intended it or not, so isolation is achieved by distance, by grounded copper between the traces and, where necessary, by a via fence. The transmit path and the receive path in particular must be kept apart, because the transmit power is many orders of magnitude greater than the received signal. Our component tolerance and reliability notes describe how the assembly is assessed for thermal stress.

Test and Calibration

A radar board cannot be tested by continuity alone. Each channel must be measured for output power, gain, noise figure and phase, and the phase measurement is what makes the array work, because the beam is steered by the relative phase between elements. Calibration structures, such as through lines and reflect standards, are often built into the panel so that the test equipment can be referenced without an external standard.

Temperature is the variable that a radar board must be proven against. A phase shift that changes with temperature will steer the beam slightly off target, and the effect appears only when the board is warm. Qualification therefore includes measurement of the channel to channel phase and amplitude at the temperature extremes, not only at room temperature, and the design must be arranged so that the temperature difference between the elements is small. Our design release checklist places those measurements in the review sequence.

Material and Mechanical Challenges

Microwave laminates are supplied in a narrower range of thicknesses and copper weights than FR-4, and their panel sizes are often smaller, which reduces the number of boards per panel and raises the unit cost. They also behave differently in fabrication: PTFE materials are soft and require specialised drilling, a plasma treatment before plating and careful handling to avoid contamination, while ceramic-filled hydrocarbon materials are closer to FR-4 but still need their own lamination profiles. Checking availability before the layout is frozen, and being prepared to adjust a dielectric thickness slightly, avoids a redesign driven purely by material supply.

Mechanically, the board is usually bolted to a metal frame that is also the heat sink and the ground reference, so flatness and hole tolerance both matter more than they would on a digital board. A board that does not sit flat against the cold plate has an air gap that no interface material will fill, and the resulting temperature difference between modules shifts their phase relative to each other. Specifying a symmetric stack-up and a flatness requirement on the drawing is what prevents that from becoming a production problem.

FAQ

Can radar boards use ordinary FR-4? Below a few gigahertz, with a carefully controlled stack-up, FR-4 can work. Above that the loss and the dielectric constant variation become the limiting factors and a low loss laminate is used instead.

Why is phase control so important in a phased array? Because the beam direction is set by the relative phase between elements. A phase error of a few degrees across the array shifts the beam and reduces the gain in the intended direction.

What limits the power of a phased array? Almost always heat. Each transmit module dissipates a fraction of its output, so the array power is limited by how much heat can be carried out of the board and into the cooling structure.