When the operating RF frequency of signals is greater than 100 MHz, we can use the technical name “RF PCBs.” RF signals at frequencies more than 2 GHz can be found in microwave PCBs, which fall within this category. Radar, sensors, cellphones, wireless transmission systems, and security systems are just a few of the applications that employ RF PCBs. RF PCBs have a more sophisticated design than regular PCBs because of the importance of signal integrity, noise immunity, electromagnetic interference, and stringent impedance tolerances.
The difficulties to be overcome
Since RF signals are particularly susceptible to noise, there is a potential that they might ripple or reflect. The need to adjust the impedance value along each of the circuit’s traces makes impedance a particularly important factor in this type of circuits. Additionally, it’s important to limit the power losses brought on by signal reflections by designing appropriate channels for return current, which, as frequency rises, prefers to use the routes with the lowest inductance. Crosstalk, or the transfer of energy between neighbouring traces brought on by inductive or capacitive couplings, becomes increasingly important as the performance and density of the components rise.
Material characteristics like the dissipation factor and dielectric constant must be taken into account while building RF circuits. When compared to materials designed for high frequencies, such as Rogers laminates, FR-4 has a larger dissipation factor, resulting in considerable insertion losses that rise with frequency. Additionally, when the frequency rises, the FR-4’s dielectric constant might rise by up to 10%, causing impedance fluctuations throughout the PCB traces.
1 – Material choice
However, FR-4 (flame retardant level 4) and other materials frequently used in PCB manufacturing are not the best options for high frequency RF applications due to the non-uniformity of the dielectric constant and a poorer tangent angle. FR-4 is also relatively inexpensive. Specific materials, such as FEP, PTFE, ceramic, hydrocarbons, and several kinds of glass fibre, are employed for RF PCBs. The fluoropolymer family’s PFE and PTFE materials enhance the base material’s chemical resistance and feature anti-adhesion, smoothness, and exceptional heat resistance (they may survive temperatures as high as 200°C). The ideal option is PTFE with fibreglass, eventually woven glass fibre, if money is not an issue and quality is more essential than price. The PTFE with ceramic coating is used because it is less expensive and the production process is simpler. A major manufacturer of dielectrics, laminates, and pre-pregs for high frequency RF applications is Rogers Advanced Connectivity Solutions (ACS), which supplies materials to several printed circuit board makers. Although more expensive, Rogers materials enable power losses to be reduced by up to 50%, ensuring great performance even over 20GHz and a low dielectric constant value that is stable and reproducible as the frequency fluctuates. The most normal strategy is to employ several materials that satisfy the requirements for electrical performance, thermal characteristics, and cost because RF PCBs are frequently multi-layered. For instance, it is possible to employ less expensive epoxy glass laminates in the interior layers and high performance Rogers laminates in the outside layers.
2 – transmission lines
Transmission lines (such as microstrip, stripline, coplanar waveguide, or others) are needed for RF PCBs in order to prevent power losses and guarantee signal integrity. The characteristic impedance, which typically has values between 50 and 75, is determined by the width of the trace, the layer thickness, and the kind of dielectric in microstrip transmission lines (Figure 1). Striplines are utilised on the inner layers, and microstrips are used on the outer layers. On the other hand, coplanar waveguides (grounded) offer the highest level of isolation, particularly when RF signals intersect relatively near traces.
You can use one of the many online calculators to determine the characteristic impedance (and subsequently the width of the trace), but you must be aware of the exact dielectric constant R values for each layer (for instance, an external layer of pre-preg laminate could have a R = 3.8 while an internal layer of FR-4 has a R = 4.2).
3 – Impedance and inductance
A frequent strategy employed by designers is to pick a standard impedance value (usually 50 ohms), so limiting their selection to RF components (filters, antennas, and amplifiers) with this particular impedance. The 50 ohm value has the benefit of being fairly common and makes impedance matching easier, enabling each PCB trace to be given the proper width.
On the other hand, inductance should be maintained as low as possible because it can significantly affect the design of an RF PCB. This is accomplished by employing several through holes, ground planes that are appropriately large and clear of gaps or discontinuities, and suitable ground connections to each RF component. High frequency components and traces need to be located close to ground planes.
4 – Routing Rule
Curvature and Angles A trace’s curvature and angles are taken into consideration. It is ideal to generate an arc with a radius of curvature that is at least three times the trace width when a transmission line needs to change direction due to routing requirements. As a result, the characteristic impedance is guaranteed to stay constant throughout the curved segment. If this isn’t feasible, draw an angle while keeping in mind that right angles must be replaced by two 45° angles.
In order to reduce the resulting variance in inductance, it is advised to install at least two via holes for each crossing when a transmission line must pass through two or more layers. In fact, utilising the greatest diameter value feasible with the trace width for the holes, a via pair may cut the inductance fluctuation by 50%. Particularly if they are intersected by sensitive signals, the traces connecting the RF components must be maintained as short, appropriately spaced apart, and organised orthogonally on the neighbouring layers.
The multilayer structure with four layers is the greatest option for the stackup. The results are far better and are simple to duplicate, even though the cost is more than a double layer method. High frequencies cannot sustain discontinuities in the ground planes, hence continuous ground planes must be introduced beneath the RF signal traces.
5 – Insulation
It’s important to pay close attention to prevent harmful couplings between the signals. The RF transmission lines shouldn’t run parallel for extended periods of time and should be maintained as far apart from other traces as feasible (particularly if they are crossed by high-speed signals like HDMI, Ethernet, USB, clock, differential signals, etc.). In reality, as the gap between them shrinks and the distance travelled in the parallel direction grows, the coupling between parallel microstrips grows. Traces carrying high-power signals have to be segregated from other circuit components in a similar manner. The grounded coplanar waveguides can be used to get a superb insulation value.
In order to prevent coupling problems, high-speed signal traces should be routed on a distinct layer than RF signal traces. Additionally, the power supply lines should be routed on certain layers with the proper decoupling and bypass capacitors.
6 – Ground planes
It is standard procedure to place uninterrupted ground planes next to every layer that has components or RF transmission lines. For striplines, it is necessary to have dedicated ground planes both above and below the central conductor. To lessen the impacts of parasitic inductances created by current-back-to-ground pathways, via holes can be inserted on RF traces and next to RF components. The coupling between RF lines and other signals travelling through the PCB is further lessened by the use of through holes.
7 – Capacitor bypass
The power pins should be in close proximity to the appropriate value bypass capacitors, which can be arranged either singly or in a star pattern. A decoupling capacitor with a greater capacity (a few tens of micro Farads) is positioned in the middle of the star arrangement, which is particularly helpful for components with many power pins. Other capacitors with a smaller capacity are positioned close to each branch. Long return pathways to the ground are avoided by the star arrangement, which lowers parasitic inductances and prevents the emergence of unintended feedback loops. Given that the capacitor’s self-resonance frequency (SRF) rises over a certain point and takes on inductive properties, the decoupling effect is negated, it is important to pay close attention to this number.
8 – Ground plane components
On the component layer (top or bottom of the PCB), which is positioned directly underneath the component, the majority of integrated circuits need for a continuous ground plane. This plane’s job is to direct the CC and RF signal return currents in the direction of the designated ground plane. The so-called “ground paddles” (Figure 2) also serve the additional purpose of dispersing extra heat, thus appropriate through holes must to be included. To increase the dissipation effect, these vias should be through holes that span many PCB layers, be internally plated, and filled with thermally conductive paste.
